BULLETIN OF THE ILLINOIS STATE LABORATORY OF NATURAL HISTORY Urbana, Illinois. U.S.A. Vol. VL NOVEMBER, 1903 Article II. -; 9'r_ ,Af-i<^\<. I I- !--^y THE PLANKTON OF THE ILLINOLS RIVER, 18Q4-1899, WITH IN- TRODUCTORY NOTES UPON THE HYDROGRAPHY OF THE ILLINOIS RIVER AND ITS BASIN. PART I. QUANTITATIVE INVESTIGATIONS AND GENERAL RESULTS. BY C. A. KOFOID, PH.D. Article II.— Plankton Sfiiclies. /TV The Plankiun of the Illinoiii R/rer, 1894-1899, irlth Introdiietori/ Notes upon the Hi//irog- nij)Jni of the I/linols River and its Basin. Part I. Quantitative Invest ii/at ions and General Results. By C. A. Kofoid. Introduction. When the work of the Illinois Biological Station was be- gun in lS94r. it seemed to the Director desiralile to determine as far as possible the normal routine of aquatic life as a necessary basis for the detection of problems for investiga- tion and experiment, and as an indispensable backgi'ound for their adequate solution. Such an investigation demands not only the discovery and specific determination of the biological population, Init involves also the study of life histories, seasonal changes, and mutual dependencies of the assembled organisms by quantitative and statistical methods, together with a study of the environment and an analysis of its factors. The plank- ton presented itself as the most available and concrete assem- blage of organisms to which this method of study could be ap- plied, and it afforded, moreover, a prolilem not only of prime scientific interest, but also of some important practical relation to fishculture. A presentation of the most general results of this investigation of the free microscopic fauna and flora, or plankton, of this typical stream of the Mississippi Valley is the object of the present paper. Inasmuch as this is the first of a series of reports upon the plankton of the Illinois River system, it has seemed advisable I. The three preceding numbers of this series, all by the present writer, have been published as articles in the Bulletin of the Illinois State Laboratory of Natural History, X'olume V., as follows : — Article I. Plankton Studies. I. Methods and Apparatus in Use in Plankton In- vestigations at the Biological Experiment Station of the University of Illinois. Article V. Plankton Studies. II. On P/t-odoriiia iV/inoisensis, a New Species from the Plankton of the Illinois River. Article IX. Plankton Studies. III. On /'Aj/(v;'<''v'//(tO second-feet, only one third more than Greenleaf's ('S5) estimate of the average discharge at that point. At Copperas Creek, on the other hand, the more moderate estimates of Cooley ('91) place the average discharge at 10.500 cubic feet 138 per second, while the bank-full capacity is estimated at double this amount. In either case complete overflow stages appear more I'eadily than they do in the majority of streams. The impounding action of the bottom-lands, on the other hand, begins with every rise qf the river, for as the water rises large amounts are drawn off from the main stream by the adjacent lakes and bayous, many of which retain their connec- Bank-full Capacity of Lower Illinois River. Locality 189 the inadequacy of the stream for the prompt removal of flood waters, says: "This lack of capacity, while it explains the wide and deep overflows, liy no means implies that any large pro- portion of the volume moves down the valley for considerable distances except in the river bed. The dense timber and the vegetation in summer, the higher ground leading across from the bluffs along every tributary, the occasional approach of bluff, terrace, or ridge, the frequently returning sloughs from interior ponds and lakes, all forbid this. The bottoms are really storage grounds to impound the flood waters that arrive faster than the channel can carry them away, and they prolong the floods in some inverse ratio to the reduction of volume." Observations made in the course of field work during the floods of 1896 to 1900 at Havana, lead me to suggest that there may often be developed a fair current outside of the channel of the main stream in such localities as Thompson's, Flag, Quiver, and Phelps lakes (Plate II. ), where a considerable reach of open territory lies in the general direction of the main cur- rent. Even in the wooded districts the cun-ent may not be wholly absent, though it is often very slight. The duration of the overflow in various parts of the valley illustrates the reservoir action of the bottom-lands. In the period from 1883 to 1889 the river was out of its Ijanks at Morri.s 60 days, or 8.5 days per year; at Copperas Creek, for 444 days, or 63.5 days per year; at LaGrange, 526 daj's, or 75 days per year. I quote from Cooley ('91)the following discussion of this subject: "A better appreciation of the reservoir action, or equalizing effect, of overflows may be obtained by a consideration of the impounding area of the bottoms. An area of 704 square miles submerged to a uniform depth of four feet—this is a flood height of sixteen feet and not aii unusual occurrence—represents 1.21 inches of water running off the entire watershed, and will sup- ply the river at the rate of 110,000 cubic feet at the mouth for 8.26 days, or at half this volume, which is an approximation to the true maximum discharge, for 16.52 days. An overflow of eight feet, or a flood of twenty feet, which is an extraordinary occurrence, represents 2.42 inches of water running off the 140 entire watershed, and will supply the river at the rate of 110,- 000 cubic feet for 16.52 days, or at half the volume for 33.04 days. * * * During flood stages the valley is a great lake of, say, 700 square miles, into which flood waters from above and from tributaries are precipitated, and from the lower end of which they run out more at leisure and in reduced and equalized volume." When we remember that even in the average year over 21%, or more than 8 inches, of rainfall escapes by way of the river, that the greater portion of this run-off takes place at times of flood, and that the overflows are greatly prolonged in the lower river by the inadequacy of the channel to carry off the excess of water and by the imperfect development of the flood-plain consequent upon the past history of the valley, we realize how important, and at the same time how unique, a fac- tor is the retardation of the run-off in relation to our plankton operations. The past decade has witnessed the completion of a vast amount of surface and under-drainage throughout large areas in the watershed of the Illinois River. Extensive open ditches have been dredged through localities where the slope or other conditions do not favor the establishment and maintenance of natural channels. These have been supplemented by miles upon miles of tile drains, thus bringing under constant cultiva- tion hundreds of square miles of territory occupied in former years by pond, marsh, or meadow of the original prairie. Even in the rolling prairie the thousands of little ponds and marshes which formed the head waters of the various tributaries of the river have been blotted from the landscape by the tile drain. In addition to this the natural lines of drainage have been sup- plemented in a great many cases by under-drainage, in order to facilitate the run-off of the rainfall and the ground water, and thus bring the soil as soon as possible into condition for culti- vation. This work of drainage is to a great extent completed throughout a large part of the catchment area. The principal exception is the basin of the Kankakee River, but the drainage of even this has already been projected. 141 The outcome of this wide-spread iuterference with the estalilished eouditiou of natural drainage has given ground for the almost universal testimony that streams which in former years held a continuous flow throughout the summer no longer run in the dry season. The reservoirs at their head waters are emptied and the supply of ground water is early exhausted by the artificial drainage in their basin. There is also a consider- able concurrence of opinion that, in the smaller streams at least, the floods come more suddenly and rise to greater heights than they did in former years. The '•wash" along the banks and consequently the amount of silt carried in suspension by the flood waters are thus increased. The presence of under-drainage undoubtedly facilitates the discharge of such water as reaches the drains, but this impetus is in large part counteracted by the greatly increased power of absorption of the soil when thus drained. Heavy rains upon a soil already surcharged with moisture may lead to even a greater run-off than the same rainfall upon the same territory rendered porous and capable of absorbing and retaining, for a short time at least, a large amount of moisture. From many points of view the subject of the effect of drain- age of the catchment-basin upon the flow of streams is one of interest and importance. For its adequate discussion records of a long series of years of the stages of tril)utary streams and the river both before and after the installation of the drainage system are needed. With a view to throwing some light upon the possible effect of drainage upon the floods in the main river, and consequently upon the plankton. 1 have talmlated the fluctuations (in excess of .25 foot in 24 hours) in the river level at the lower gage at Copperas Creek for an earlier and a recent period, each of Ave years. The earliest authentic records which I have been able to secure begin with lS7y. This antedates the completion of a considerable portion of the artificial drainage of the river basin. 1 have accordingly chosen the records of 1879-1883 inclu.sive for comparison with those of 1892-1896 inclusive. This choice is unfortunate in one respect, for the earlier series lies in a period of heavy rainfall and the later in- 142 eludes two years of unusual drouth. The fact that the earlier period is in a series of wet years and the later in one of dry years will, it seems, tend to obliterate whatever contrast may exist in the rate of the fluctuations. Thus the extent of the fluctuations is much greater in the earlier period and the dura- tion of high water is longer. The river was above the six-foot level during 1,028 days in the first period, as against 709 in the second; and it was above ten feet, that is at a stage of overflow, 645 days in the earlier period, and only 297 in the later one. The dam at LaGrange, completed in 1889, raises the water 2 feet on the lower gage at Copperas Creek at the low-water stage. Its effect at the stages above cited is not, however, according to Cooley, perceptible at the upper end of the pool. The results of the tabulation do not reveal any alarming changes in the flood habits of the river. There is, however, a well-defined increase in the rate of movement in the later period as compared with the earlier. The average daily movement (above .25 foot) is in the first period .416 foot, in the second period .492 foot,—an increase of 18%. The difference is still n^ore marked when the comparison is made in the rate of rise alone. In the earlier period the rate of movement (in excess of .25 foot) was .4848 foot per day; in the later period it was .592 foot; an increase of 22%. The distribution of this increase through the year is somewhat irregular, and, owing to the in- sufficiency of the original data, is probably of slight significance. The greatest increase occurs, however, in the months of May, December, September, and March, all months in which floods prevail, or at least occur occasionally. As shown on previous pages the conformation of the valley is such as to induce a prolongation of the floods. The records and the hydrographs show that the decline from a rise is in most cases much less rapid than the approach of the flood. We find accordingly in the above table that the number of days of decline (at a rate exceeding .25 foot per day) is considerably less than those of rising waters, and that the rate of fall is also less than the rate of rise, being only .3185 and .337 foot, respec- tively, per day, in the two periods. The increase in the rate of 144 fall is thus less than 6%. So far, then, as the table goes, it indicates that there has been a moderate increase in the rapidity with which floods rise and a slight increase in the rate of their fall, as compared with the rate of their movements in a corresponding period thirteen years ago. It is impossible from data at hand to determine whether or not the general drainage of the basin has shortened the period of high water and extended that of low water with a consider- able diminution of volume at low-water stage. The decade prior to 1896 seems to have witnessed such a change, as hydrographs on Plate VII. show. This diminution is the more marked when allowance is made at low-water stages for the 2 feet which the dam at LaGrange (completed October 12, 1889) is estimated to raise the water on the lower gage at Copperas Creek. The fact that the rainfall was deficient during this decade doubtless Monthly Means of Gage-readings (in feet) below Copperas Creek Dam 1879-1899. (Basis of reference, low water of 1873 and 1879.) Year 145 Monthly Means of Gage-readings (in feet) below LaGrange Dam 1883-1899. (Basis of reference, low water of 1879.) Year 146 Fluctuations at Copperas Creek. Months 147 siderable decline. The rise of the flood may occur in any of the succeeding months up to and including April. There is a predominance of flood movement in January as compared with February, thei-e being more instances in the former month of movements exceeding 3 feet and culminations of floods exceed- ing 10 feet than there are in the latter. What is popularly known as a " January thaw " is probably the occasion of this predominance. The rise of the flood is often very rapid, as, for example, in 1881, 1883, and 1895, the rise in 1883 being over 11 feet in 8 days. These rapid rises often occur after heavy pre- cipitation in winter month.s when the conditions favor a very rapid run-off. The initial stages of the flood are usually less precipitous, as are also the final stages preceding the culmina- tion, especially at overflow stages, when the flood capacity is greatly increased by the impounding action of the flood-plain. The flood curve is rarely an even one, such as that of 1887, since fluctuations of more or less importance occur, as a rule, during both the rise and fall. A well-defined culmination is, however, present, except in a few instances, such as in 1890, when three about equal maxima appear. The maximum of the flood occurs most frequently in the latter part of March or the first part of Api'il, though it may appear as early as December, as in 1895, or as late as June, as in 1889. Of the twenty-one maxima six have been in April, five in March, three each in January and Feliruary, two in May, and one each in December and June. The highest point reached at Copperas Creek in 21 years was 19.25 feet in 1885. The rises in excess of 10 feet above low water in the 21 years have appeared nine times in April, six in June, five in March, four in January, three each in December, February, and May, and but once in November, while no flood has reached this limit in the remaining four months. The decline of the flood is much less rapid than its rise, and is often marked by secondary rises which more or less delay the return of the low-water period. The "June rise," caused by the heavy rainfall of that month, is masked by the averaging process in the mean hydrograph, 148 but appears occasionally (nine times in twenty-one years) in the annual ones as a well-defined flood. In the . majority of years, however, it appears as a slight interruption in the decline of the earlier spring flood. In a few cases these fluctuations occur late in May or early in July. The fact that this June rise is so little felt in the Illinois, while it is so prominent in other streams of the state, is explained by the fact that it often occurs within the period of overflow, when large accessions of water produce relatively slight rises of river levels. A compar- ison of the hydrographs of 1889 and 1892 will illustrate the points in question, the June rise of the latter year appearing as a slight fluctuation in the declining flood, while in the former year it stands out as a well-marked rise, owing simply to the previous low water. As compared with Leverett's normal regimen for an Illinois stream, we flnd in the case of the Illinois River that the high- water period exhibits a considerable range; that it extends over a much longer time ; and that the phenomenon of the June rise is less pronounced,—all of which deviations may be explained by the impounding action of the slightly developed flood-plain of the Illinois. Following the period of high water comes an equally pro- nounced period of low water, extending through the summer months until the late autumnal or winter rise. As shown by the averages, this low-water period (below four feet) extends from August to November, inclusive. It varies, however, with the high-water period, appearing even in May, as it did in 1895, and frequently continuing through the fall and winter till late in February, as it did in 1891, 1893, and 1894. The average time during which the water was below 4 feet for the 21 years at Copperas Creek is 134.1 days, two days more than the high- water period. This low-water stage is quite variable in its du- ration, ranging from 260 days in 1895 to 5 in 1885. The low- water period, which Leverett estimates as continuing at least ten months in normal Illinois streams, is thus much shortened in the Illinois River. The lowest levels of the year are reached 149 usually in September, the occurrence of minimum levels being distributed among the months as follows: In the records at Copperas Creek, eight occur in September, six in August, five in October, three in July, and one each in Novemlier, December, and June. The lowest level was recorded at Copperas Creek in 1879, when the low-water mark was established. It has not again been reached, owing since 1889 to the dam at LaGrauge. The low-water period is often one of marked stability as contrasted with other parts of the year, the total movement of river levels falling to 0.10 foot per month in November, 1893,. and frequently amounting to less than one foot in September, October, and November, while even this movement is probably caused to a considerable extent by the operation of the locks and by changes in the direction or force of the wind. The stage of extreme low water is followed by a gradual but very slight rise during the fall months, which cannot be attributed entirely to rainfall since, as shown in the table on page 126, these are months of lessened precipitation. This increase is well shown in the hydrographs of 1898 and 1897. It seems more probable that with the falling temperatures the loss by evaporation, both from the river and its tributaries, is sufficiently lessened to account for this slight rise in levels, amounting in most cases to about one foot. This low-water period is frequently interrupted by minor fluctuations, some of which appear at or subsequent to the autumnal equinox. These fluctuations are due to heavy sum- mer rains, and usually appear suddenly and decline with almost equal abruptness. They rarely rise to eight feet and are usually below five; they thus do not cause overflows, and affect only those bayous and lakes which maintain connections with the river at low-water levels. Their duration is short also, being but a week or ten days, rarely a fortnight. In 1896 there was a repetition of such rises of more than usual prominence and duration, giving a unique character to the hydrograph of that year. The equinoctial period, marked by the slightly increased rainfall of September, is not marked in the average hydrograph 150 by any increase. Indeed, the average level for this month at Copperas Creek is only 2.82 feet, the lowest average for the year. The rises attending this period appear but twelve times in twenty-one years, are usually insignificant, and are often less than two feet, owing doubtless to the greater capacity of the soil for absorption at this season of the year. The time not included in the high- and low-water periods as here defined amounts on the average to 96 days, a relatively short time for the transition between these two extreme condi- tions. This abruptness of the transition stages is to some ex- tent apparent in the hydrographs. On Tables I. and II. will be found tabulations of the total movement, both -f- and — , of the river levels in each month in the period covered by the records at the two dams. The monthly and yearly averages of these data are also given. The figures are to a certain extent an index of the relative stability, both monthly and annual, of the river. The averaging process has to some degree masked the differences in the several months, as will be seen on a comparison of the mean monthly movements with those for any single year, the latter exhibit- ing at some times of the year much greater contrasts than the means. Thus the greatest and least movements in 1897 are re- spectively 10.37 feet in January and 0.40 in October, while the corresponding limits in the means are 4.18 feet in July and 2.78 in November. The means indicate two periods; one of consid- erable movement, corresponding to that of high water, and one of less movement, representing the low-water stages. The greatest movements occur in February and July, indicating the rise and decline of the flood. The least movement is found in November, a period of low water and freedom from sudden and heavy rainfall. The several years exhibit quite a range in the total amount of change in levels, the extremes in Copperas Creek in twenty-one years being 68.70 feet in 1881 and 32.92 in 1894; at La Grange, in seventeen years, 85.36 feet in 1898 and 40.41 in 1887. The movement is thus somewhat greater and more variable at the lower dam. In general there is some cor- 151 relation between the average heights of the water for the year and the total movement of levels, though this correlation is by no means continuous or uniform, as will be seen, for example, on comparison of the average heights and movements for 1895 and 1896 (3.61 and 7.26 feet; and 51.89 and 53.16 feet). The ab- normally low water in 1895 was followed by an unusually early rise in December. Had this rise occurred ten days later the correlation of average heights and total movements would have been more apparent in these years. The range in movement of river levels between the highest and lowest water of the year as show^n in the tables varies at Copperas Creek between 9.00 feet in 1889 and 17.70 in 1883, and at LaGrange between 8.75 in 1894 and 20.52 in 1883, while the extreme range for the period of record at each dam is 19.27 and 22.92 feet respectively. Somewhat greater fluctuations thus occur at the lower dam. The preceding discussion of the fluctuations in river levels affords evidence for the extreme instability of the environment of the plankton with which we are to deal. Indeed the only really constant feature seems to be this very instability. The mere presentation of the statistics of the fluctuations can give but little life or color to the great modifications which these changes produce in the physical conditions environing all the aquatic life of the river. The great increase in area and vol- ume (not far from one hundredfold) which occurs at high water (Plate III.) affords an opportunity for a great increase in the total production of the plankton, especially when the flood period extends into the spring and early summer months, when the maximum development of the plankton sometimes occurs. The submerged flood-plain also affords the greatest variety of conditions of depth, current, vegetation, bottom, light and shade, temperature, and sewage, thus favoring the diversifica- tion of the plankton locally produced but carried away to some extent into the channel by the receding floods. A trip by boat across the submerged bottom-lands from the Quiver shore to the western bluff (PI. II. and III.) in the latter 152 part of May would be far more enlightening than any descrip- tion that might be given. As we leave the sandy shore of Quiver we traverse the clear, cold, and spring-fed water along the eastern bank with its rapidly growing carpet of Cerato- phijlhon, and in a few rods note the increasing turbidity, rising temperature, and richer plankton of the water which has moved down from the more or less open and slightly submerged bottom to the north (PL II.). As we cross the muddy bank of Quiver ridge and enter the main channel of the river we find rougher water, caused by the wind which usually sweeps up or down the stream with considerable force between the bordering forests. The water also appears much more turbid by reason of silt and plankton, and no trace of vegetation is to be seen save occasional masses of floating Ceratophylhim or isolated plants of Lemna, Wolffia, or SpirodeJa. Huge masses of cattle- yard refuse, veritable floating-gardens, may also at times be seen moving down the channel or stranded in some eddy along shore. As we plunge into the willow thicket on the western shore we have to pick our way through the accumulated drift lodged in the shoals or caught by the trunks of the trees or the submerged underbrush. The surface of the water is one mat of logs, brush, sticks, bark, and fragments of floating vegeta- tion, with its interstices filled with Lemnacece dotted with the black statoblasts of Plumatella. From this dark labyrinth we emerge to the muddy but quiet waters of Seeb's Lake with its treacherous bottom of soft black ooze. We next enter a wider stretch of more open territory with scattered willows and ma- ples and a rank growth of semiaquatic vegetation, principally Polygonums. The water is clearer and of a brownish tinge (from the diatoms), while mats of algae adhere to the leaves and stems of the emerging plants. A flock of startled water- fowl leave their feeding grounds as we pass into the wide ex- panse of Flag Lake. We push our way through patches of lily- pads and beds of lotus, past the submerged domes of muskrat houses built of last year's rushes, and thread our way, through devious channels, among the fresh green flags and rushes just 153 emerging from the water. Open patches of water here and there mark the areas occupied by the "moss" or Ceratophyllum, as yet at some depth below the surface. The Lemnacea' are everywhere lodged in mats and windrows, and, amidst their green, one occasionally catches sight of a bright cluster of Azolla. The water is clear and brownish save where our movements stir the treacherous and mobile bottom. We now enter a second time the partially wooded country, and cross the submerged ridge to the sandy eastern shore of Thomp- son's Lake. This ridge is covered by submerged vegetation which has as yet attained but little growth. The "breaks" of the startled flsh show that we have invaded favorite feeding grounds. The waters are evidently moving towards the river, and they bear the rich plankton of Thompson's Lake, while their turbidity is doubtless increased by the movements of the fish. Schools of young fry can be seen feeding upon the plankton in the warm and quiet waters. Thompson's Lake, the largest expanse of water in the neighborhood, is wont to be rough in windy weather, but if the day be still we can see the rich aquatic vegetation which fringes its margin and lies in scattered masses toward its southern end. Its waters seem somewhat turbid, but more fi'om plankton than from silt, though the deep soft mud which forms much of its bottom is easily stirred. The slender transparent limnetic young of the gizzard-shad may be seen swimming near the surface. There is a perceptible drift to the south in the open lake, though this current is deflected by the elevated banks of Spoon River ( PI. II. ) towards the Illinois River, crossing the lower bottom-lands above this region. If we push on through the fringing willows at the south we find a series of open places locally known as "ponds". The warm still waters are turbid in places from the movements of fish, and at times we see the compact schools of young dogfish [Aii/ix ca/va) and, if we are late enough in the season, the myriads of young black, tadpole-like catfish (Aineiurus), likewise in schools, while young cry]) (Ct/priinis carpio) are everywhere. The new vegetation is already spring- 154 ing from the decaying and matted stems of the preceding sum- mer. Turning back towards the river we pass through the heavy timber where the still brown water, cool and clear, over- lies the decaying leaves and vegetation of last season's growth, now coated with the flood deposits of the winter. Emerging again upon the river channel we may find a turbid yellow flood pouring out from Spoon River, bringing down its load of drift and earth, and marking its course down the stream as far as the eye can see. From an environment even more varied than this come the different contributions to the plankton of the river in the flood seasons. Every change in level modifies this environ- ment by connecting or cutting off backwaters, shifting or check- ing currents, disturbing vegetation and temperature in a man- ner the very complexity of which beggars description. Contrast with the extent and variety of conditions at flood the limitations placed upon the stream at low water (PI. IV. and V. ) . Instead of an unbroken expanse of four or more miles we find now a stream only 500 feet in width (at Station E), while the adjacent territory is dry land save where the sloughs, marshes, and lakes remain as reservoirs. Quiver Lake is now much reduced in width, and it may be choked with vegetation except in a narrow channel where the clear water shows little or no current. A half mile below we find the river water rush- ing in a narrow " cut-off " across the ridge of black alluvium into the lower end of the lake. The wooded banks which sep- arate the river from Quiver and Seeb's lakes are now crowded with a rank growth of weeds and vines. The latter " lake " is re- duced to a shallow stagnant arm of the river, whose warm turbid waters are foul with dead mollusks, and whose reeking mud- flats beneath the August sun shine green and red with a scum of Euglena. As we pick our way through the tangle of rank vegetation we come upon Flag Lake, now a sea of rushes. The discharge from this marsh to the river ceased in the early sum- mer, and its margins are even now dry, with gaping cracks. Beyond the marsh we pass to the shore of Thompson's Lake to 155 find its southern end choked with vegetation, though the greater part to the north is open water. The woodland and open ground to the south are now pastures and fields of waving corn The only outlet to this large body of water, now somewhat re- duced in area but warm, turbid, and rich in plankton, is a tor- tuous slough six miles to the nox-th. The discharge, however, is in any case but slight, the lake being, indeed, not infrequently the recipient of river water. Spoon River still pours a sluggish but constant stream into the river, but save for a water- bloom of livid green (EncjU'iia) its waters yield but little plank- ton. Thus, of all the wide area contributing to the plankton of the channel at high water there now remain only Thompson's and Quiver lakes and Spoon River, each much diminished in volume, but all diversified in character. Returning now to the river itself we find a gently sloping bank of black mud, baked and cracked by the sun's heat, ex tending towards the softer deposit at the water's margin. A low growth of grasses, sedges, and weeds springs up as the water recedes. The river margin does not often have much aquatic vegetation. In low-water years, such as 1894 and 1895, a considerable fringe is formed along the shore, but this is quickly cleaned out on the seining grounds, which occupy a large part of the shore, as soon as the fishing season opens in July. In years of normal high-water the vegetation rarely gets much of a foothold along the shores, even at low-water stages. Save for the few sandy banks where springs abound, such as those below Havana along the eastern bluff, there is little, at least in the LaGrange pool, to vary this monotony of mud banks and fringing willows. The backwaters have been reduced to the lakes, sloughs, bayous, and marshes (PI. II.) which abound everywhere in the bottom-lands. Many of these, as, for example, Phelps and Flag lakes, have ceased in their re- duced condition to contribute to the river. Others, like Thomp- son's Lake, maintain a connection with the river by means of a long and tortuous bayou or slough through which the cur- rent flows in or out as the relative levels of the two fluctuate. 156 This lake receives but little water from a few springs and creeks along the bluffs, and like many others in the bottom- lands serves only as a reservoir from which the water is slowly drawn off as the river falls, but when once the lower stages are reached its contributions cease. Still others, like Quiver and Matanzas, maintain direct and open connection with the river, and since they receive tributary streams they continue to feed the river, but in reduced volume. Though the number of tributary areas is thus much reduced at low-water stages, the individual peculiarities of the tributary waters in the bot- tom-lands become more pronounced. As each one loses its connection with the general flood it becomes a separate unit of environment, with its local differences in those factors which determine the character of the plankton developing in its waters. The resulting contributions may thus differ greatly in amount and component organisms, and accordingly tend to diversify the river plankton of low water to a degree even more marked than that of high water. With the confinement of the river waters to the channel goes a marked condensation of the sewage, which, under condi- tions of uninterrupted low water, leads at times to an excessive development of the plankton, or, if the river is closed by ice, to stagnation conditions. But few years, however, offer such op- portunities; for, as a rule, in most low-water periods sudden and heavy rains are wont to occur which flush the stream, wash away the sewage and plankton-laden waters, and store anew the reservoir lakes without causing any considerable overflow. After each catastrophe of this sort the decline of the flood affords a new and favorable opportunity for the develop- ment of the plankton. In this instability lies the great distinction between the river and the lake as a unit of environment—for I believe we are justified in applying this phrase to the conditions of fluvia- tile life, though it must be admitted that the "fluviatile unit" is an exceedingly complex one. As the discussion of the river fluctuations indicated, there is a seasonal routine which the 157 cycle of seasonal fluctuations more or less approximates year after year, and not a few of the important environing factors are operative in much the same way upon by far the larger part of the biological area. There is thus a common basis upon which the other less constant factors produce their effect. The best justification, however, for the use of the term lies in the results of our work, which show a biological assemblage adapted to this complex environment, and exhibiting in some of its phases at least as much uniformity as the more stable factors of its surroundings. During 1894 and the early part of 1895 readings were re- corded only at occasional intervals owing to the fact that the Station was occupied but part of the time, its work being as yet in the preliminary stage. From August to October, 1895, bi- daily readings were made by Mr. Newberry at a gage located by us at our field headquarters and based upon the government gage on the protection at the wagon-bridge at Havana. From October, 1895, to January, 1896, the readings were taken by Mr. Henipel at the government gage, and since that time bi- daily or daily readings have been taken under the direction of the city authorities of Havana. These readings are given in the tables which follow. 158 Readings of River Gage at Havana, 1895. (Plane of reference, low water of 1873.) Q 1:V.) Readings OF River Gage at Havana, iSg6. (Plane of reference, low water of 1873.) Day 160 Readings of River Gage at Havana, 1897. (Plane of reference, low water of 1873.) Day 161 Readings of River Gage at Havana, i8g8. (Plane of reference, low water of 1873.) Day 162 Readings of River Gage at Havana, iSgg. (Plane of reference, low water of 1873.) Day 168 Monthly and Yearly Averages of Readings of River Gage at Havana, 1894-1899. (Plane of reference, low water of 1873.) Year 1894. 1895. 1896. 1897. 1898. 1899. Jan. 164 rant hydrographs are found. In the discussion which follows, the average or "normal" conditions and figures pertaining thereto are based upon the twenty-one years of record at Cop- peras Creek, eighteen miles above Havana. The year 1894 (PI. VIII.) is typical in that the high- and low-water periods are normally located as to season and also in the presence of a March, June, and September rise. Both the extreme and average heights for the year, 10.4 and 4.63 feet respectively, are, however, much below the general aver- age (13.8 and 6.74 feet). The high-water period (above 8 feet) is shortened to three weeks, and the overflow stage is thus al- most eliminated. The concentration of the sewage in the nar- row limits of the channel during the early summer favors the greater development of the plankton. With the exception of the September rise the extreme low water continued without interruption for a period of eight months—till the last of Feb- ruary, 1895. These are conditions which cause the drying up of extensive backwater areas, and' also the development of a large amount of aquatic vegetation in those lakes and marshes which remain—a circumstance which reduces their plankton, and their contribution, if there be any, to the river. The auton- omy of the river plankton is thus emphasized in such a year as this, which may be briefly characterized as one of predom- inant low water and unusually stable conditions. In 1895 (Plate IX.), another low-water year, we flnd, on the other hand, little that approaches the normal. There is, to be sure, a diminutive March rise and a sharp but very brief equinoctial one, with very low water in the autumnal period. The abnormal features are the failure of overflow, the long low stages,—almost ten months, with unusually low water in Feb- ruary and June,—the July rise, and the December overflow. The extreme low water of the year is apparent in the average, 3.61 feet, the lowest on record in twenty-one years. The low water in the winter combined with ice produced a stagnation fatal to the plankton, while the June minimum favored an un- usually large development for that season. The July, Septem- ber, and December rises flushed out the river. The low water Kir) of this year, following that of the previous year and combined with the al)8enoe of overflows with rise and current sufficient to lift and carry away the vegetation, resulted in a very unus- ual growth of the aquatic flora in the lakes and even along the river mai-gins. The conditions prevailing throughout the greater part of the year thus continued to favor the autonomy of the main stream noted in the previous year. In l)rief. the year may be characterized as one of extreme low water, with some minor and unusual fluctuations. The contrast with 1S94 is best seen on comparison of the total movements of the two years, viz., 39.98 and 51.75 feet respectively. The year 1896 (PL X.) is one of still more unusual char- acter, since it presents a series of bimonthly rises culminating in step-like succession throughout the year, hi none of these. however, save the initial one, is more than a 'very moderate stage of water i-eached. This results (PI. VII. ) in a reduction of the normal March Hood, the isolation of the June flood in the hydrograph, and the submergence of the Septemlier rise between the abnormal rises of August and October. The gen- eral result of such a series of rises is to bring the average level for the year up to (5.98 feet (_7.26 at Copperas Creek), 0.71 feet above the general average, though the rainfall for the year is slightly below normal. The increased average height does not, however, in this case carry with it the usual extension of tfie flood period. The river was above ten feet for less than a month and above eight feet only three months. The overflow stage was thus slight, and in addition it occurred in the flrst months of the year, during the winter minimum of the plankton, while during the spring months, when the normal overflow occurs, the river was practically conflned to its banks. The succession of minor floods and the slight increase in the average level does, however, greatly extend the reservoir action of the permanent backwaters. The repeated floods also had the effect of clearing out the vegetation in the river and lakes where some current develops, as, for, example, in Quiver Lake. This reduction in the amount of vegetation in the reservoir 166 waters is accompanied by a considerable increase in their plankton. This fact, combined with the increase in the vol- ume of their contributions due to higher levels, augments the relative importance of their share in the formation of the river plankton, tending to increase its quantity and variety. On the other hand, the repetition of floods, no less than eight of which may be found in the hydrograph, flushes the river so often that no concentration of sewage and marked maximum of plankton occur. The unusual extent of these movements is apparent when the total movement for the year, 50.7 feet, is compared with the totals of other years having about the same average height. For example, 1S90 and 1897, with an average height of 6.9 feet, have a total movement of only 44.2 and 36.56 feet respectively. In brief, the year was one without extended overflow, with lower water than usual at the normal flood sea- son, with prolonged bank-full river and reservoir action of the permanent backwaters, and with more than the usual turmoil. In 1897 we find a hydrograph (PL XL) which approaches the mean closely in its main features, and exhibits all the ex- pected movements excepting the equinoctial rise. The average height for the year, 6.90 (6.86 at Copperas Creek), is also near the general average (6.74). The year thus approximates the normal. The high-water period is of 141 days' duration, al- most exactly the average (140), but it occurs somewhat earlier in the year and attains 16 feet—a little more than the usual height. The earlier decline renders more prominent the June rise, and gives an early start to the extreme and uninterrupted low water of the remaining five months of the year. The low- water peiiod (155 days) is normally located but is somewhat in excess of the average (147), and it is also unusual in the fact that the extreme low-water level (1.7 feet since the completion of the dam at LaGrange) continued almost unchanged from the middle of August till the first of November. This was fol- lowed by the usual slight increase in water in the closing months of the year. The total movement of the year (43.1 feet) is considerable in view of the average height (6.9 feet). 167 but this was less disastrous to the plankton than usual since it was in the main due to the spring flood and not to minor changes when the stream was within its banks. This freedom from minor interruptions during the low-water period is some- what unusual, and resulted in a concentration of sewage ap- proaching stagnation and in a marked increase in the fall plank- ton. The overflow period, in which the reservoir action of the bottom-lands a.s a whole was operative, prevailed during the first five months; the change to low water, during w^hich the reservoir action of the more jiermanent and diversified waters was in force, took place very rapidly: while the low-water stages, during which it is a minimum, were Ijoth pronounced and iirolonged. These circumstances combine to emphasize in this year l»oth the unity and the autonomy of the river. In brief, 1897 was a year of normally located but pronounced high and low water, of marked freedom from interruptions, and of unusually favorable conditions for the unity and autonomy of the plankton of the river and for the full development of its normal seasonal cycle. In 1898 (PI. *XII.) we find another year whose hj-drograph approaches the normal in its main features. There is a well- defined period of high water followed by one of much inter- rupted low stage. The spring flood is normally located, con- tinues (al)ove 8 feet) for 164 days, and culminates at 18 feet on April second. The extension of the flood period for 24 days beyond the normal is due largely to the "June" rise of unusual proportions, which culminated in the last of May at 13.8 feet, and covered a period of five or six weeks. The impounding action of the bottom-lauds as a whole is thus shifted forward into the late spring and early summer, while the concentration of the overflow into the channel occurs in the early part of May and again in June, and the conditions of rainfall, season, and overflow combine to favor the production of a relatively large amount of plankton at these times. The decline is rapid in July to low-water stage, which continues but three weeks, the lowest record l^eing 2.5 feet. This is followed l)y a series 168 of minor rises, which flush the river at short intervals during August and September, and a rise to bank height in November —fluctuations which favor the reservoir action of the perma- nent backwaters, and at the same time introduce much silt and interrupt and diversify the plankton cycle. Of all the years of our operations at Havana this was the one of highest average level—8.02 feet (8.11 at Copperas Creek)—and greatest move- ment (66.2 feet). The dilution of the sewage, the increased current and silt, and the flushings incident to such hydro- graphical conditions tend under most circumstances to de- crease the relative amount of the plankton, though doubtless they also tend to increase the total production of the stream. In brief, the year was a typical one of high water with much delayed run-off and interrupted low-water period. In 1899 (PL XIII.) we find another year conforming very closely to the normal hydrograph in its main outlines. We are concerned only with the first three months, at the close of which occurs the maximum (14 feet) of the spring flood. The greater part of the rise occurs in a brief period at the close of February, and the declining waters or more stable conditions at other times reduce considerably the flushing and silt attend- ing most winter floods, such, for example, as that of the pre- ceding year. The decline in February also afforded a good op- portunity for the reservoir action of the permanent backwaters under midwinter conditions. Our collections of 1899 thus cover a period of winter flood of more than usual stability. The wide range of hydrographical conditions during the six years of our plankton work at Havana have afforded a unique and, up to this time, unexampled opportunity to follow the effect of flood and drouth, of changing season, and of yearly fluctuations upon the life in the waters of a stream, and to give to the conclusions here reached the confirmation which repeti- tion alone can bring. TEMPERATURES. The fluctuation in the temperature of the river water con- stitutes for the plankton one of the most marked evidences of 1(U> the climatic changes of the recurring seasons. This factor in the environment of the plankton is thus an ever changing one, but at the same time it runs an annual cycle of the same gen- ei'al character j'ear after year with ever present minor varia- tions of a seasonal or local origin. The extremes of tempera- ture in bodies of water in this latitude are so divergent that they afford the basis for marked seasonal changes in lioth the constitution and the quantity of the plankton. Adaptations on the part of the organisms of the plankton to definite tempera- ture limits thus occur. Records of the temperature of the air and of the surface and the bottom water have been taken regularly at all stations where quantitative plankton collections were made. These are recorded in Talile I. The temj)eratnres were taken with a Negretti-Zambra self-recording thermometer from 1894 till ^lay 24, 1898, after which time a Hick's self-recording maximum-mini- mum thermometer w as used. Under stable conditions no appre- ciable variation was noted in the reading of the thermometers, but at the times of sudden change, as in the mingling waters of a rising flood, I'eadings would sometimes vary as much as four or five degrees at one location and level. The temperature of the river water is influenced by a vari- ety of causes in addition to the immediate action of solar heat. The most prominent of these are the access of the tributary water from streams, springs, and imi^ounding liackwaters. The temperature of trilmtary streams, such as the Spoon River (Table IV. ). is often, though not always, warmer in winter and colder in summer by several degrees than that of the main stream, as a result probably of the greater proportion of spring water and the greater nearness of the same to its subterranean source. A good illustration of this was to be seen along the eastern shore of Quiver Lake, where at low-water stages spriugs near the water's margin kept up a continuous How. The tem- perature of the water in summer was 54 \ while in winter it fell only to 51 "\ The smaller tributary waters also respond more quickly to fluctuations of temperature than does the river 170 itself. In like manner the backwaters, which are usually much shoaler than the river, are subject to greater changes, exhibit- ing in warm days greater extremes of heat, as high, for example, as 96° having been found in the margins of bottom-land ponds. On the other hand, the flood waters in the foi'ests and marshes, where the vegetation protects the water from the direct rays of the sun, remain at lower temperatures than those of more open tracts. The lakes and bayous with aquatic vegetation also re- main cooler in their deeper waters, as, for example, Thompson's Lake, where, among the CeratopJii/UiiiJi, the temperature at the surface on the fifteenth day of July was 88.'2°, while only six inches below, in the vegetation, it was 80°, the difference being due to the protection from sun and wind which the veg- etation afforded. Another factor tending to modify the temperature is the earth temperature, which in the very shallow waters of our environment becomes relatively important in both summer and winter. In the low temperature of winter this is heightened by the fact that most of the bottom of the backwaters is strewn with a mass of vegetation whose decay must produce some heat. This probably accounts for the higher bottom tempera- tures sometimes observed in winter (cf. Tables III. and VIII.) in Flag Lake, where such detritus was more abundant than in the river, where but little is found. For example, on February 26, 1897, the bottom temperature in Flag Lake was 36°, while in the river, with about the same surface temperature (32°) and greater depth, it was only 32.5°. This difference may also be due to the effect of the current in the river in mingling more quickly the surface and bottom waters and thus equalizing their temperatures more rapidly. The temperatures recorded in the Illinois River, Spoon River, and in Thompson's, Quiver, Dogfish, Flag, and Phelps lakes are to be found in Tables III.-IX. respectively, and they appear on the plates with the hydrographs and plankton data of the re- spective years and stations. The extreme range of temperature observed by us in the river and its adjacent waters at Havana 171 was 32 "-96". The highest temperature recorded in the I'iver was 89% on the afternoon of August 3, 1897, and again, at the same time of day. Julj^ 26. 1898. The diurnal range in temperature is considerable at times, depending naturally upon that of the air. On August 3-5,1898, in connection with a test of the diurnal movements of the plankton and accompanying analysis of the gases dissolved in the water, the temperatures recorded indicate in the surface waters a range of 5.5 . with a maximum of 79.5 at 5:00 p. m. and a minimum of 74° at 2:00 a. m. The bottom water (depth 2.44 meters) showed a range of but 2 ".from 74° at 8:00 a. m. to 76° at 11:00 a. m. The air temperatures on the days in que.s- tion ranged from 83' at 5:00 p. m. to 58° at 5:00 a. m. A diur- nal variation of 5.5° in surface waters and 2° in bottom waters is thus indicated at this time. Other conditions will probably show a slightly greater range. Monthly and Yearly Averages of Surface Temperatures, 1894-1899, Illinois River. Year 1894 ... 1895... 1896... 1897... 1899.... Monthly average Jan. 32-75 32-7 32.9 Feb. Mar. April 32 , 58 33-7 39-5264.54 32.2543.8 60 32-1243-3 32-6 35-2 32 . 78 32 . 73 40 . 45 60 . 46 68 . 27 53-32 May 72.7 66.3 65.8 June 80.25 80 74-7 75 „ 78.8 77-75 July 82.25 79 80.7 .02 Aug. 837" 80.51 82 80.9 Sept. 82.S780.56 81 .0381 .49 77-5 78.87 65-75 77-07 71.87 74.21 Oct. Nov. 58 I41 54.2642.5 56 44 Dec, 39 33-6 > 56.66 65-' 45-7 i33-02 54.3741-4232-9855-84 57. 55143. 00 35. 22 57.08* *Averag'e of monthly averages. The temperature records are too isolated to plot complete thermographs of the river and its backwaters, though they do give a very fair idea of the seasonal fluctuations, especially in the later years, when they were more evenly distributed, and in the midsummer, when they were more numerous. They were usually taken between 7:00 a. m. and 5:00 p. m.. and may be regarded as day temperatures. A comparison of the records in Table III., the plottings (PL VIII.-XIII.), and the above table, giving the monthly averages of surface temperatures of our records, shows the following .seasonal routine in the river: 172 During the months of January and February there is a period of minimum temperature approaching 32°, averaging 32.75°, and rarely exceeding 34°. The constancy of the tem- perature at this season is probably due in large part to the equalizing effect of the ice which normally covers the stream, and especially its backw^aters, at this season of the year. Dur- ing the early part of March the temperature rises, but the rate becomes more I'apid in the last part of the month. The up- vpard movement reaches 40° or 50°, in early springs, such as 1898, attaining the latter temperature. The average for the month rises to 40.45 ° and the fluctuations increase in extent. The rapid rise continues through April, attaining 60°-70°, and averaging 60.46°. The records for this month are somewhat meager for any comparisons. In May the season of maximum temperature is approached and occasionally reached, as in 1896, the average temperature from somewhat scanty records being then 68.27°. This month is one not only of marked rise but also of considerable fluctuation. The period of maximum tem- perature is in full swing in June, and continues through July and August and well over into September. The average rises from 77.75° in June to 81.49° in August, and falls to 74.21° in September owing to the decline which begins in the latter part of this month. This period of maximum summer heat is fairly well defined in the thermograph and continues at or near 80° approximately three months, from the middle of June to the middle or latter part of September. It is a time of consider- able fluctuation, most of the movements being within 10°, though the range for August in the five years of record was from 74.3° to 89°. These fluctuations combined with the di- urnal changes and the wind are effective in producing a con- siderable vertical circulation of the water. Following the summer maximum comes the fall decline, which begins late in September and is practically completed in November. The greater part of the change takes place in October, the average decline in that month being 16.76°, while that in November is 14.55°. In some years, as 1897, the de- 17n cline is a gradual one; in other.s, as 1S95 and 1S9S, it is subject to some irregularities. With December the winter minimum returns, but with less ])ersistence than in the months which follow, flood waters at this season bringing their higher tem- peratures. The annual temperature cycle thus falls into four periods: one of minimum and quite constant temperatures, including December, January, and February, and a varying portion of March ; one of maximum and more fluctuating temperature, approaching 80° and extending, with some interruptions, from the early part of June till about the middle of September; and, separating these, the two shorter intervals of change. The period of increase in temperature, which is also one of rapid change and increase of the plankton, includes the latter part of March and the months of April and May. The period of de- cline, which is sometimes more abrupt than the spring rise, as in 1895, 1897, and 1898, extends from the latter part of Septem- ber until the end of November. This is also a period of change and of frequent but not universal diminution in the plankton. The average temperature for the years, as expressed approxi- mately in the table, is 57.08°. This point is passed about the middle of April and again about the middle of October with considerable regularity. Since, however, these dates both lie in periods of rapid change, the average temperatures are of much less duration than the more extreme ones. The existence of these well-defined periods of maximum, minimum, increase, and decline of temperatures affords the basis for corresponding seasonal changes in the minute life of the water as fundamental and extensive as those which affect the plant and animal life of terrestrial and aei'ial environment. This subject of the rela- tion between temperature and organisms of the plankton will be fully discussed in connection with the statistical study of their seasonal distribution. A comparison of the thermographs (PI, Vlll.-Xlli.) of the different years and an inspection of the table on page 171, reveal but few significant annual differences. The spring rise 174 in temperature was somewhat delayed m 1896 and again in 1899, and the summer maximum was less pronounced in 1895 and 1897, though in compensation the summer heat was pro- longed into September in these years. The spring rise in 1896 and the autumn declines in 1895 and 1898 are rather more abrupt than usual. These annual differences extend and cur- tail the plankton periods characteristic of the seasons, or render their changes more abrupt. The difference between surface and bottom temperatures is, as a rule, but slight. It is perforce usually lacking during the period of decline in the autumn, and at other seasons varies in amount with the air temperature, the wind, and other attendant circumstances. So long as the temperature is above the point of maximum density of water, 39.2°, the surface waters are the warmer by an amount ranging from a fraction of a degree to 5°, the latter occurring on still, hot days. With air temperature falling below that of the water the surface and bottom quickly come to have the same degree of heat. Below 39.2° the colder waters are at the surface, though at this season of the year there is usually much less contrast at different levels than in the warmer months. Temperature fluctuations, following those of the season and the day, occur in the watersof this region to a degree not realized in the typical lake, whose deeper waters respond but slowly to the surface changes, and thus exercise an equalizing effect. Examples of this quick response are found in the unusually high temperature (82.3°) in both top and bottom waters of the river on May 13, 1896, while temperatures of five days later showed a drop to 71.2° in both regions. A decrease equal in suddenness and extent occurred in September, 1898. The surface layers of water, quickly affected by temperature changes, form relatively a very large part of the volume of the river and its backwaters, and thus instability of temperature becomes an important feature of the environment of the plankton of the river as contrasted with that of the lake. Changes of the ex- tent above noted must affect considerably both the movements and the multiplication of the plankton organisms. 175 The temperature conditions here described are those as- signed by Whipple ('98) to lakes of the temperate type and third order, those whose bottom temperatures are seldom very far from their surface temperatures, and in which there is con- siderable vertical circulation at all seasons when the surface is not frozen. At no place in the region examined by us has a depth been found sufficient to permit the occurrence of a stratum of cold water at the bottom unaffected by the vertical circulation and warming process in the surface regions, such, for example, as has been found by Birge ('97) in Wisconsin lakes. This absence in the river environment of the "thermo- cline" and of summer and winter periods of stagnation in lower levels, marks another point of contrast between the river and some lakes as units of environment. The temperature conditions in the bodies of water adja- cent to the river do not differ to any considerable degree from those here discussed. The limited extent, greater amount of vegetation, shallower waters, or greater access of spring water in some of these will cause slight variations from the condi- tions found in the river. The ice conditions attending the winter minimum are of profound biological significance, since they produce important alterations in the winter routine. As a result of the presence of an ice sheet on a body of water, the temperatures become more constant, the mingling of waters due to winds ceases, the usual processes of aeration are interrupted, and the propor- tions and amounts of the gases dissolved in the water may be very much altered, the degree of the change depending upon conditions such as the completeness with which the surface is sealed by the ice, the amount of sewage, the relative abundance of plant and animal life, the duration of the ice, and the exist- ence of currents. So far as our observations go at Havana, the stage of stagnation attended by the destruction of the animal life which is sometimes found in small lakes is rarely realized in this environment. Several reasons may be assigned, the principal one being being the instability of river levels in the 176 winter season, which prevents the culmination of stagnation conditions. Again many of the backwaters are rich in vegeta- tion, and some of them are spring fed at the margins which thus remain open even in the coldest weather. The river itself rarely closes over entirely, air-holes remaining where the cur- rent is rapid. Thus, below the mouth of Spoon River (PL II.) a large area was usually free from ice even when the river was closed above this point. The currents due to tributary waters, as in Quiver Lake, and to changes in level, as in all impound- ing waters, also tend to prevent stagnation conditions. In spite, however, of these favoring circumstances one catastrophe of this nature did occur in the years of our work at Havana. In the winter of 1894-95 prolonged low water and heavy ice upon the river and lakes combined to render the conditions unfavorable to life in the river, and to some extent in Quiver Lake. Conditions in other localities at this time were not observed. The practical extinction of the plankton and the death of large numbers of fish attended this period of stagnation. The duration of the ice at the various stations in the sev- eral years is indicated at the bottom of the diagrams which give the hydrographs and plankton data of the several stations by black lines of a thickness proportional to the icfe. The occurrence of ice in the different years at Havana has varied considerably. No records were made in 1894-95, but from other sources, river stages and weather reports, it seems probable that the river closed in the last days of December, and that the ice continued until the rise of February 25, a period of almost 60 days. In the winter of 1895-96 there was but little ice, the river and backwaters being partially closed only for the first fortnight in January. In 1896-97 the river did not close until after the rise in the early part of January, the ice remaining about one month, going out with the rise of Febru- ary. The lakes, on the other hand, were closed to a large ex- tent throughout December, and again, to varying extents, dur- ing January and a part of February, the current due to high water keeping portions free from ice at times. 177 In 1897-98 the lakes closed the last days of November and opened again on December 12, freezing again December 17, and not clearing entirely until February 14. Rising water continued from January lU, so that stagnation conditions did not ensue. The river also closed partially early in December, opening and closing again with the lakes. The first ice went out with the rise on January 11. The river closed again Jan- uary 27, and the ice went out February 9 and 10. Again on February 21 ice was present, and for several days following. In 1898-99 ice again formed early in December and par- tially closed the river during the month, going out about the 27th and reappearing on the 30th. This went out gradually January 17-24, and the river froze over again on the 26th and remained closed for a month. Thin ice formed March 5, re- maining only three days. The lakes closed early in December, the ice never entirely disappearing until the middle of March. Partial breaking up occurred at the times of breaking up of the river ice. These partial openings and the changes in level were sufficient to prevent a period of stagnation. OTHER METEOROLOGICAL FACTORS. As indicated in Tables III.-IX.. at each plankton collection observations of the direction and force of the wind, with its effect upon surface conditions and on the state of the sky, were recorded. The relation of these factors to the plankton may not seem intimate or apparent. They have more bearing on the subject of vertical movements of the plankton, data upon which will be found in the study of the surface and bottom col- lections made with each of the vertical collections which form the basis of the present paper. The surface w^aters affected by the intensity of the sunlight and the movements caused by the wind form i-elatively so large a part of the environment of the river plankton that these factors are much more widely opera- tive here than in the lakes, where the surface stratum thus affected is relatively small. 178 The wind conditions on the river and the lakes adjacent to it—which are generally, elongated in the direction of the main stream (PI. II.)—are somewhat peculiar. Owing, it may be, to the configuration of the river valley, or perhaps still more to the bordering forest of the contiguous bottom-lands, the pre- vailing direction of the wind is either up or down the river or lake, especially during the summer season. The effect of an up-stream wind is greatly to increase the disturbance of the surface when wind and current are thus opposed. These winds, when prolonged and violent, decidedly affect the levels of the different parts of the lakes, and, for example, in Thompson's Lake (PI. II. ) determine at low-water levels whether the lake shall discharge its waters into the river or itself receive an access of river water. Owing to the mobile condition of the abundant bottom deposits, at low stages the winds also add very much to the silt in suspension in the water, and thus hin- der the penetration of light. The effect of varying sky conditions lies primarily in their re- lation to the temperature of the water, but is due in a less degree to the influence of light upon the multiplication of chlorophyll- bearing organisms—the primal food supply of the plankton — and upon the movements of these and other plankton organisms. The abundant silt in suspension in waters of the river and most of the adjacent lakes doubtless hinders the penetration of the sunlight, but modifies to a much slighter extent its effect upon temperatures. Wind and sky conditions combine to favor or prevent the appearance of the "water-bloom." This is a char- acteristic green scum which coats the surface of the river, and occasionally of the lakes, on still, warm days in midsummer. On cloudy or windy days the minute organisms {Euglena, Chlam- ydomonus, etc. ) which form the bloom do not rise to the sur- face. The conditions of wind and sky are thus important fac- tors in the economy of limnetic life and, by reason of their rel- atively greater effectiveness in the river and its adjacent waters as contrasted with the typical lake, add to the elements of instability in the fluviatile environment. 179 TURBIDITY. Records of the turbidity were made (Tables III.-IX.) in gen- eral descriptive terms during the first two years of our work at Havana. After April 29, 1896, the turbidity was tested by means of a white plate of semi-porcelain, 10 cm. square. The depth at which this square disappeared from view was recorded in centimeters as a measurement of turbidity. Although the method is somewhat primitive and subject to some variations with the conditions of sky and daylight, it is still sufficiently accurate for the purposes of the present paper. The disc method has not as yet been correlated with the platinum-wire method, the diaphanometer method, or the silica-standard method of Whipple and Jackson ('00), and comparisons with these are consequently excluded. As might be expected in the river environment, when floods occur the turbidity is often extreme, and is exceedingly varia- ble according to the locality and the river levels. (Of. Tables III.-IX.) The extreme range of our records extends from 1.3 cm., in a Spoon River flood, to 260 cm., in Quiver Lake, under the ice. In the river (Table III.) the great majority, about two thirds, of the records lie between 20 and 50 cm., while the ex- treme range is from 2 cm., in the flood of May, 1897, to 115 cm., in the declining waters of July, 1896.. The clearer waters ap- pear, as a rule, with declining floods and stable low stages, especially under the ice. With the inception of floods the most turbid water is found, which gradually clears even while the rise continues. The river varies in clearness according to the instability of the river levels, as will be seen on comparison of the turbidity in 1896 and 1897, the latter year being more stable and having relatively fewer records of a marked turbidity. The turbidity of the river is due to both plankton and silt, the latter being as varied as the character of its tributaries, with the added contamination from the cities along its banks. In Spoon River (Table IV.) the extremes are even more marked than in the main stream, varying from 1.8 cm., in flood 180 conditions, to 165 cm. at low water under the ice. The turbidity here is almost entirely due to silt, that at flood being largely composed of earth and clay, giving a black or yellow tinge to the water. The amount of comminuted vegetable debris found in the waters is considerable. In Thompson's Lake (Table VIII.) the turbidity is not so frequently marked by the extremes seen in the other bodies of water examined, the range being from 115 cm. in the declining waters of May, 1896, to 6 cm. in invading floods, and again in the late autumn of 1897, when high winds roiled the shallow waters. As a general rule the turbidity of this lake is somewhat less than that of the river, but as great or greater than that of other backwaters. This turbidity is often due in part to the heavy planktons occurring here, and also to the floccu- lent debris loosened from vegetation and stirred up from the mobile bottom by fish and the waves. Very little silt enters the lake except at times of inundation, especially with back- water from Spoon River. Owing to its origin the silt in this locality is usually of finer, more flocculent material than that found elsewhere. In Quiver Lake (Table V.) the extremes are much more marked than in any other locality, ranging from 3.5 cm., in flood water from the river, to 260 cm. on June 5, 1896, in clear impounded water. In winter, under the ice, the bottom was vis- ible on December 3, 1896, in 260 cm. of water. A great deal of variation in turbidity occurs in this lake. In years of low water, as 1894 and 1895, when vegetation is abundant, the turbidity is very slight, the bottom being visible much of the time. In the three succeeding years the lake was free from vegetation, and the turbidity was considerably increased as a result largely of the increase in the plankton. The sources of the silt in this body of water are varied; occasional freshets in Quiver Creek, which enters the upper end of the lake (PL II.), invading floods from the river, and debris from vegetation and the bottom put in suspension by the wind, current, or movements of fish,^—all contribute their share to the pollution of the otherwise clear water of this lake. 181 In Dogfish Lake (Table VI.) the conditions are essentially those of Quiver Lake, of which it is an arm. The principal difference lies in the fact that flood water entering Quiver Lake at low stages never moves as far up as our station in Dog- fish Lake (PI. II.). Floods from Quiver Creek also merely back up the clear water in Dogfish Lake without themselves invad- ing that territory. The only flood silt entering this region is, consequently, that which comes with general inundations. In Flag Lake (Table VII.) the conditions at inundation are similar to those of other impounding backwaters. The great amount of vegetation found here adds to both the vegetable and the flocculent debris which roil the water whenever this is disturbed by waves or the movements of fish. Turbidity is but rarely caused by plankton here, with the exception of the few instances when diatoms or Osrilldria became very abun- dant. The water is thus usually clear, the bottom being commonly visible in the small spaces left free of vegetation, even at a depth of 215 cm. In Phelps Lake (Table IX.) the silt conditions are peculiar. The high level at which the lake lies and the intervening stretch of bottom-lands (PI. II.) combine to keep out all silt- laden flood-waters except those that enter by a now abandoned channel from Spoon River or from the main stream at times of their maximum floods. At other times the silt consists prin- cipally of particles of bark and dust from the adjacent forests, or of fragments of loam from the bottom, which is here unusually stable. The comparative freedom from vegetation removes a large element common in the silt of the other lakes. The turbid- ity, however, is very marked in this lake, falling in many cases below 20 cm., and in the majority of instances is lai-gely due to the very abundant plankton characteristic of its waters during the greater part of the summer. Movements of fish and water- fowl add considerably to the silt in suspension in this lake at some seasons of the year. The color of the water has not l)een made a subject of special inquiry. In general the turbidity gives it a grayish cast 182 that varies to yellowish or blackish tints with silt of clay or loam origin. When diatoms are abundant a brownish tinge is very evident, and with OsciUaria rising in quantity, as it does in some semi-stagnant waters in late summer, a blackish tint becomes pronounced. In midsummer and early fall, when water-blooms rise, we find varying tints of green according to the kind and quantity of chlorophyll-bearing organisms present. The turbidity, as above suggested, is due to a great variety of factors, one of the most important of which is the plankton itself. Indeed, under some conditions turbidity be- comes a token by which the relative abundance of the plank- ton may be estimated. The presence in our plankton of vast numbers of the most minute planktonts, such as the flagel- lates, renders this relation of plankton and turbidity more prominent in our waters than it is in waters where such organ- isms are less abundant. The turbidity otherwise is due to non-living solid matter in suspension. This is brought in by tributary streams, and is torn loose from the shores and bottom by the current of the river, the movements of fish, the wash from steamboats, and the constant sweeping of the river channel by fishermen's seines during the open season at stages when seining is possible in this place. The dust from prairies and forests brought by winds; the waste from factories, distilleries, glucose-works, and cattle-yards; and the sewage of a score of cities along the banks, —all make additions to the burden of the water. Microscopical examination of the plankton has revealed the diverse character and origin of the silt which accompanies it. Fine fragments of quartz, bits of mollusk shells, small pieces of coal or ashes, minute particles of loam or clay, and the fecal pellets of aquatic organisms—especially of mollusks and of insect larvae—constitute the heavier element of the silt. To this is added a variable but ever considerable quantity of exceedingly fine sediment of earthy or clayey origin, some of which remains long in suspension. The coarser and lighter silt consists largely of comminuted vegetation, both terrestrial 183 and aquatic, minute bits of leaves, stems, bark, and wood, with the characteristic grain refuse from distilleries and glucose- works and the offal from the cattle-yards at Pekin and Peoria. At all seasons of the year and in all waters the scales of Lepi- doptera and the pollen of coniferous trees are of common oc- currence. Mingled with this material, especially when aquatic vegetation is present, is a very light flocculent material con- sisting, in part at least, of the zoogloete of bacteria. It is in the midst of debris of this varied composition that the plankton lives, and it is in collections consisting to a greater or less ex- tent of silt material that the river plankton must be studied, its species determined, and its individuals enumerated. In collections made with the silk net the greater part of the fine silt passes through the meshes with the water. In filter-paper catches some of it adheres to the paper, and the finer flood silts will even pass through hard-pressed filter paper in small quantities. With silt of so varied a character it is practically impossible to establish and continue any standard of measurement or estimate which affoi'ds a satisfactory basis for the determination of the relative amounts of silt and plank- ton present in the collections. After considerable experience in the examination of our collections I have endeavored to estimate the amount of silt present in them as they appear in the Rafter counting-cell. The distribution of the material in the cell and the conditions of examination are such as to favor a uniform standard of estimation. On the other hand, the estimates are purely personal, without any volumet- ric check, and are thus only comparable with each other. This method seems to be the only solution at present available for this perplexing problem. These estimates are given in Tables III.-IX., together with computations, based thereon, of the amount of both plankton and silt per cubic meter. These figures form the basis of the diagrams in Plates VIII.-XIII. and XXII.-XLII. As will be .seen in the tables, the per cent, of silt varies from a mere trace to almost the entire catch, changing with the river conditions as previously stated. 184 A still more accurate determination of the total amount of solids in suspension in the river, both silt and plankton, is afforded by the catches made by the Berkefeld army filter, data concerning which will be found in Table XV. This filter removes all of the suspended solids and permits their complete removal from its surface, but adds a small portion of its own substance to the catch. After the first few catches with this filter the wear becomes somewhat uniform and is thus dis- tributed. On computing the loss from the filtering sur- face by wear, and quadrupling this volume to allow for its less compact condition, we find that it constitutes less than five per cent, of the catches washed from its surface. The true amount of solids is thus about five per cent, less than the figures cited in the tables and in the following discussion. The amount of water strained in making these catches was usually 5 liters, while the tables give the computed amount per cubic meter. The amount of solids was measured by our usual method of measuring plankton, that is, by condensation in a centrifuge. In this treatment it usually attains the consistency of soft mud. For the river the amount ranges from 148 cu. cm. (per cubic meter), in declining water under the ice in December, to 5,416 cu. cm., in the incipient stages of the winter fiood of February 28, 1899. The average amount of the weekly catches for 1898 is 592.2 cu. cm. per cubic meter, which for an average flow of 24,600 cubic feet of water per second (see page 132) means a discharge of 14.57 cu. ft. of solids per second, or, 459,794,232 cubic feet (1,301,990 cubic meters) per year, or 16,472 cu. ft. (46.64 cu. meters) per square mile of the catchment-basin of the river. The average amount, per cubic meter of water, of solids taken at fortnightly intervals in 1898 in Quiver Lake was only 378 cu. cm., a fair index of the greater clearness of its waters. In Thompson's Lake similar collections average 557 cu. cm., in- dicating waters somewhat clearer than the river. In Phelps Lake the average amount is large, 1,572 cu. cm., due in no small 185 degree to the very abundant and minute plankton organisms. In Spoon River the average of the monthly collections is 1,740 cu. cm., three times as much as the main stream carries. The heavy floods and rapid current of this tributary are responsible for this large amount of earthy solids in suspension. In this matter of silt and turbidity the river as a unit of environment stands in sharp contrast to the lake. Deposition of solids and clear water are normal to the environment of the lake, while fiolids in suspension and marked turbidity are the rule with river waters. Owing to their varied occurrezice these elements, silt and turbidity, also add to the instability of fluviatile, as contrasted with lacustrine, conditions. Silt and turbidity are usually attendant upon floods, so that their unmodified effect upon the plankton is not easily deter- mined. Some inferences and observations regarding the rela- tion of these factors to the economy of the plankton may, how- ever, be made. The silt affects the plankton indirectly by hastening the solution of nutrient substances from the organic detritus that forms a considerable portion of the unstalile de- posits which accumulate in shoal and in sheltered parts of the stream. It hinders the penetration of light, thus checking the development of the chlorophyll-beainng organisms while favor- ing the multiplication of bacteria and hastening the decay of organic matter in suspension. It also seems probable that it produces a deleterious effect upon the Entomostraca by ad- hering to the hairs which clothe their various appendages, thus hampering their movements and causing them to sink to the bottom. Accessions of flood water are frequently followed by an increase in the relative number of moribund and dead Entomostfaca, especially of the Copepoda. CHEMICAL CONDITIONS. The food supply is the most fundamental feature in the environment of the plankton. Its abundance or scarcity de- termines to a large degree the growth and reproduction of or- ganisms, and its fluctuations are important factors in deter- 186 mining the seasonal and local production of plankton. The primary source of the food of the plankton lies in the water and in the gases and inorganic salts dissolved therein, the oxy- gen, the carbon dioxid, the nitrates, and the phosphates being usually regarded as of prime importance to the growth of chlo- rophyll-bearing organisms. The phytoplankton, which utilizes these inorganic materials, then becomes itself the food for the zooplankton. These inorganic substances, the primary food supply, are thus indices of the capacity of the water for the production of plankton. With the inauguration of the work of the Biological Sta- tion at Havana arrangements were made whereby collections of water taken by the Station staff from the river and some of the adjacent lakes were sent to the Chemical Department of the University of Illinois, at Urbana, for analysis. In 1895 the Chemical Survey of the waters of the state was es- tablished at the University under the direction of Prof. A. W. Palmer, and in September of that year regular shipments for analysis from the Illinois River and from Quiver Lake were made at intervals of one week, and in January of the following year Spoon River was added to the collection points. These collections were continued throughout the period of our oper- ations at Havana. In September, 1897, collections were insti- tuted in Thompson's Lake, and from that time on the samples for chemical analysis were taken at the same time and place as the plankton and, like that, by the plankton pump. After the date above named a fortnightly interval corresponding to the plankton interval was made between collections in Quiver Lake, though the weekly interval was continued in Illi- nois and Spoon rivers. August 16, 1896, a disastrous fire in the chemistry building of the University destroyed many of the records, and this fact accounts for the absence of data of the analyses in the months of the year prior to the fire and for some other gaps in the record. Special collections were made during the last twenty months of our operations for the deter- mination of the oxygen and carbon dioxid dissolved in surface 187 and bottom waters, but determinations of these dissolved gases made by Professor Palmer immediately upon collection in the held, yielded results which throw some doubt upon the value of those made on samples which were shipped for analysis at the laboratory of the C!hemical Survey. Twenty-four to forty-eight hours elapsed between the time of collection and that of anal- ysis, and during this time changes no doul)t took place in the gases dissolved in the samples, so that the results of the analy- ses give no trustworthy basis for a statement of the amount of dissolved oxygen and carbon dioxitl in the water at the time of collection. I am indebted to Professor Palmer not only for the data of the chemical analyses which he has furnished me from the records of the Chemical Survey, but also for many other cour- tesies in connection with this subject. COMPARISON OF CHEMICAL CONDITIONS AND PLANKTON AT THE DIF- FERENT STATIONS. In Tables X.-XIII. will be found data from the chemical analyses of the waters of Illinois and Spoon rivers and Quiver and Thompson's lakes, together with plankton data of the same or contiguous dates. The most important of the determinations, those of chlorine, oxygen consumed, free and albuminoid am- monia, organic nitrogen, nitrates and nitrites, as well as the plankton, are graphically shown in Plates XLIII.-L. CHEMICAL ANALYSES OF WATER FROM PLANKTON STATIONS. PARTS PER MILLION. AVERAGE* OF ALL .ANALYSES. 188 determined for each. Since the samples were collected at in- tervals throughout the year, the averages may be regarded as presenting in succinct form the chemical characteristics of the stations examined, and they may therefore serve as a basis for a comparison of the relative fertility of the localities. The residue upon evaporuflon, which comprises the solid matters left upon evaporating the water and drying the residue, includes both organic and inorganic substances. The inorganic constituents are salts, and comprise mainly compounds of lime, magnesia, soda, potash, iron and alumina with chlorine and with carbonic, sulphuric, nitric, and silicic acids. In this residue lie both the mineral constituents of the food of the phytoplank- ton and the undecayed organic matter found in the water. Not all of the constituents of the residue are equally utilized as food by the phytoplankton, so that the quantity of the residue gives a basis only for a very rough estimate of the fertility of the different waters. Some significance, however, attaches to the marked differences shown in the table. The differences in total residue in Illinois and Spoon rivers (367.5 and 522.3) and Quiver and Thompson's lakes (268.9 and 326.4) show no particular correlation with those of the average plankton production of these waters for corresponding periods (1.91, 0.384, 1.62, and 6.68 cm.^ per m.^, as shown in Tables X.- XIII.). The amounts and relative proportions of the dissolved and suspended residue in these localities show some relation to the plankton production. The residue in suspension is not, in its present form at least, available for plant food. Its occurrence in the four localities is almost directly correlated with the rela- tive turbidity of the water Spoon River has from four to eleven times as much suspended matter (274.3) as the other localities, and this consists largely of clayey material with considerable fine quartz, neither of which contributes any considerable source of nutrition to the phytoplankton. The suspended ma- terial in the other locations at times of flood partakes of the character of that in Spoon River. At other times it contains a considerable proportion of debris of plant or animal origin !^ 1S9 including the plankton itself. The current of the river is doubt- less responsible for the e^ccess (61.4) which its waters carry above that in the lakes (25.1 and 44.6). The greater amount in Thompson's Lake (44.6) may be due to two sources, its greater dependence on the river for its water supply and the greater disturbances in its waters due to fish and to waves. The fact that the total catches of the plankton net (8.91, 1.35, 2.24, and 7.68) do not on the average more nearly approximate in their ratios to each other the ratios of the chemical resi- dues (61.4, 274.3, 25.1, and 44.6) is due to the great leakage of the finer suspended particles through the silk, especially in Spoon and Illinois river waters. The residue in solKfioii contains the available supply of mineral salts for the phytoplankton as well as some organic materials which become sources of plant food, and its distribu- tion in the four localities is correlated with the plankton pro- duction in the direction of the differences, though not in their quantity. Thus S]joon River with the least dissolved residue (167.1) has the least plankton production (.384), and Quiver Lake has likewise less residue (248.2) and less plankton (1.62) than Thompson's Lake (282.9 and 6.68). The Illinois River ex- ceeds all of the localities in its dissolved residue (304.1), which may be attributed to the fact that the water is " older," afford- ing greater time for solution, and that it is the recipient of considerable sewage and industrial wastes which add to its burden of substances in solution. The small amount in Spoon River may be attributed to the fact that it is largely uncon- taminated surface water of recent origin. The greater amounts in the two lakes (248.2 and 282.9) are due in part at least to their dependence upon the river, which in the case of Quiver Lake is slight during the summer season. In so far as the total res- idue held in solution is an index of fertility, the data indicate that the river itself carries the greatest store of food (304.1); Thompson's Lake, somewhat le.ss (282.9): Quiver Lake, still less (248.2): and Spoon River, least of all (167.1). On this basis and in the light of the production of Thompson's Lake it would 190 seem that the river water might under more favorable condi- tions develop a more abundant plankton. These favorable conditions are to be found in the quiet backwaters of river-fed lakes, where time for breeding is afforded. The loss which the residues of total solids suffer upon igni- tion (heating to redness) includes the organic matters which are burned away and such constituents of the mineral matters as are volatile or are decomposed by heat into volatile sub- stances. In stream waters the suspended portion of this mate- rial may be a rough index of the quantity of plankton and silt of organic origin, all of which on decay add to the water sub- stances available for plant food. From the data in the table it may be ascertained that the four localities yield respectively, in the order of the the table, 7.7, 17.5, 1.9, and 8.2 parts per million of such material. The excess in Spoon River (17.7) is doubtless due to silt of organic origin, while the plankton pre- sumably forms a larger proportion in Thompson's Lake and in the Illinois River. The poverty alike of plankton and of silt in Quiver Lake is reflected in the small amount (1.9) lost on igni- tion in its waters. The loss, on ignition, of substances held in solution shows no differences at all commensurate with the relative production of plankton, though the trend of the differ- ences is similar in three instances of the four. The chlorine is contained in surface waters in combination with various basic elements, but chiefly in the form of common salt. Its principal source is animal matter, sewage, or drain- age from refuse animal matter. In our river and lake waters it is largely an index of their relative contamination with sew- age from cities within the drainage basin. Since its combina- tions are not utilized by plants as food in any considerable quan- tity, at least as compai'ed with other constituents of the sew- age, such as the nitrates, the chlorine becomes the best crite- rion of the amount of sewage and thus of the principal adventi- tious fertilizer which the waters examined by us contain. The differences in the four localities are striking and significant. The average chlorine in the Illinois River (21.6) is more than 191 five times as great as that iu Spoon River (3.8), while that in Thompson's Lake is more than three times the amount in Quiver Lake. The large amount of chlorine in the Illinois and in Thompson's Lake—which draws its water supply mainly from the river—is due to contamination by the sewage of Chi- cago, Peoria, and other cities within the drainage basin. Quiver Lake receives water from the river only during flood periods, when the sewage is diluted, and at other seasons it contains more nearly the chlorine of the uncontaminated prairie stream. Its chlorine thus averages low (3.8). That of Spoon River runs higher (4.8), in part tiecause of backwater from the main stream to the point of collection. The sewage systems discharging into this stream are few and but slightly developed, and its chlorine is correspondingly low. While it is true that the chlorine is not a precise measure of the amount of sewage or of the adventitious fertilizing material received by a stream, it is nevertheless significant that ratios of chlorine and plankton production not only trend in the same direction but are quanti- tatively somewhat similar when lake is compared with lake and stream with stream. Thus in Quiver and Thompson's lakes the ratio of their chlorine content is 1 to 3.4 while that of the plankton production is 1 to 4.2. The corresponding ratios in Spoon and Illinois rivers are 1 to 5.7 and 1 to 5. An increase in chlorine due to sewage or animal wastes seems thus to be accompanied by a proportionate increase in the plankton pro- duced. It is safe to infer that it is one of the factors producing the increase, but, as shown elsewhere in this paper, other fac- tors, such as vegetation and current, are also potent in produc- ing the contrasts in plankton production above noted. The oxtii/oi cunsuiiu-(l in oxidizing the organic matters af- fords an additional index of the quantity of these substances present in the water, but since all kinds of organic matter are not oxidized in the analysis it does not yield a criterion of the total quantity of organic matter. A comparison of the oxygen consumed in the four localities yields results very sim- ilar to those obtained by a comparison of loss on ignition, ex- 192 cept in the case of Quiver Lake, where the oxygen consumed (5.9) is proportionately very much lower than the loss on igni- tion (27.5). The amount of oxygen consumed is greatest in Spoon River (14.1), and may be attributed largely to the detri- tus of organic origin which the stream carries, or to the prod- ucts of its decay held in solution. It may also be due in part to the organic material of the water-bloom (Enr/lrna) which es- capes the silk of the plankton net. There is, however, no in- crease in the oxygen consumed in the season of the water-bloom which can be considered commensurate with its development. Nitrogen is an essential constituent of protoplasm and of many of its products. It is taken up by plants in the form of nitrates and free ammonia, and there is increasing evidence that it may be utilized, especially by the lower plants, in more complex combinations, such as the amido-compounds. Since the other principal constituents of protoplasm—carbon, hydro- gen, and oxygen—are present in inexhaustible quantity in the air, water, and carbon dioxid, and since the nitrogen available for plant food is practically limited to that contained in the above-named compounds, the nitrogen in combination in any given body of water becomes par excellence, an index of its fer- tility. These compounds exist in living plants and animals, in their wastes, and in the products of their decay. They enter stream and lake waters in various ways : in the debris of veg- etable and animal origin washed into the stream, especially by flood waters ; in leachings from such matters drawn from the soil in seepage and spring waters ; and, especially (in the Illi- nois River) in the sewage and industrial wastes of Chicago, Peoria, and other cities within the drainage basin. In the lake and stream waters these nitrogenous compounds are found in solution in the water, in the sediment and debris of organic or- igin in suspension, in the zoo- and phytoplankton, and in the macroscopic aquatic plants and in the larger animals^such as fish, mollusks, insects, and crustaceans. The chemical anal- yses show only those nitrogenous compounds in solution, in silt, and in plankton, while that stored in the larger plants and 193 animals is not determined. Since the silt is undergoing decaj-, and since the individuals of the plankton are short-lived and rapidly release their nitrogenous compounds into the water by waste and decay, the determinations of nitrogen in its various forms in the analyses represent both the present fertility and that in immediate prospect. The contributions from the ma- croscopic plants and animals not included in the samples analyzed constitute an undetermined element in the sum total of the nitrogenous matter available for the sustenance of the phytoplankton. The relative amounts of nitrogen in the several stages of decomposition are shown in the determinations of total organic nitrogen, of nitrogen as albuminoid and free am- monia, and of nitrites and nitrates. The total orf/aiiic nitrogen includes all nitrogen that is in combination with carbon (together with other elements) in the tissues of living plants and animals and in many of the waste products of the latter. It is also present in organic matter in the early stages of decay, and is accordingly found in organic debris and sewage of stream and lake waters. It is accord- ingly an index of the quantity of organic matter which in its present form is not available for plant food ( with the possible exception of certain amido-compounds I l)ut is destined to be- come available by decay. It thus indicates the potential fer- tility of the water. The differences in the amount of total organic nitrogen present in the four localities are not in each case correlated with the actual plankton production. Spoon River, which contains the least plankton, has the greatest amount (1.292) of organic nitrogen. The absence of any ex- cessive contamination by sewage in this stream combined with the paucity in plankton, makes it apparent that this mat- ter is probably in the organic detritus of the silt, which is pres- ent in an unusual amount in this stream. The close resem- blance of the Illinois River and Thompson's Lake in the matter of total organic nitrogen (1.03 and and 1.05 ) is explained by the dependence of the latter upon the river for its water supply, and by the excess of sewage in the former and of plankton in 194 the latter. The small amount in Quiver Lake is attributable to its greater independence of the river, to the paucity of its plankton, and to the sandy nature of its drainage basin and consequent share of spring water in its water supply. It is noticeable that the large amount of submerged vegetation in this lake does not contribute any great amount of organic nitrogen to the water at any season of the year. The nitrogen as albiiininoid ainmonia is included in the to- tal organic nitrogen, and exhibits almost identical relative amounts in the four localities, though actual quantities are only half as great. It represents the nitrogenous materials which have not undergone decomposition. The nitrogen as "/ree" ammonia represents the ammonia contained in the water in free or saline condition. It is a prod- uct of the decomposition of organic matter in the first stages of oxidation, and its quantity is an indication of the amount of such matter present in the water in a partially decomposed state. It is abundant where sewage occurs, and together with the chlorine affords evidence of the degree of contamination. The occurrences of free ammonia in the four localities (.86, .245, .165, and .422) are not in most instances in the same ra- tios as those of the chlorine (21.6, 3.8, 4.8, and 16.8) or of the plankton (1.91, .384, 1.62, and 6.68). The excess (two to five times as much) of decaying organic matter in the river as com- pared with the other situations is apparent, and is doubtless due to the concentration of sewage in its channel and to the more recent access of the sewage there as compared with that in the reservoir backwaters, as, for example, in Thompson's Lake. The early stages of decay are in consequence more active in the river. The free ammonia is high in both the river (.86) and Thompson's Lake (.422) but lower in Spoon River (.245), where the organic material in suspension is considerable, as indicated by the loss on ignition, the albuminoid ammonia, and the organic nitrogen and oxygen consumed. The decay of this matter and the accompanying release of free ammonia has not been attained as yet in a part at least of the silt in Spoon River to the same 195 degree that it has in the older river and lake waters. Its bur- den of silt thus adds to the sources of fertility of the main stream and of the reservoir backw^aters at times of flood. The small amount of free ammonia in Quiver Lake (.165) is corre- lated with the small amounts of the substances above named in its waters and the sandy nature of its drainage basin. The differences in the two streams in the tjuantity of free ammonia (.86 and .245) have the same trend as the differences in plank- ton production (1.91 and .384), but they are not commensurate quantitatively, owing apparently to the more recent origin of the water in Spoon River. In the lakes the free ammonia (.165 and .422) and plankton (1.62 and 6.68) exhibit a similar trend and a like absence of quantitative differences in the plankton commensurate with the free ammonia available for support of the plankton. The effect of the relative food supply is thus apparent in the trend of the differences, and the operation of other factors is suggested by the quantitative contrast. The factors in Quiver Lake tending to reduce the plankton below the amount that the food supply would make possible are to be found in the passage of tributary waters through the lake and in the excessive aquatic vegetation. It is noticeable that the considerable amount of submerged vegetation in Quiver Lake does not seem to effect any appreciable increase in the free ammonia. The abundance of free ammonia in the Illinois Eiver would seem to afford a basis for a greater development of the phytoplankton than it attains under the conditions in that stream. The time for breeding which is afforded in the backwaters is one factor involved in this contrast. The nifriten constitute a second intermediate stage in the oxidation of nitrogenous substances into inorganic products. Their presence indicates organic matter in the hnal stages of decay, and that decompositions due to the vital processes of living organisms are under way. The nitrites exhibit a distri- bution in the four localities which in the trend of the differ- ences is similar to that of the free ammonia. The ratio of the free ammonia in Spoon River to that in the Illinois is 1 to 3.4, 196 while that of the nitrite content of the two streams is 1 to 3.7. The ratios in the two lakes, Quiver and Thompson's, are 1 to 2.6 and 1 to 2.1 respectively. Spoon River and Quiver Lake are thus poorer in nitrites than Illinois River and Thompson's Lake. The same contrasts are to be found in their production of plankton, though the differences in the amounts produced are greater than those in this source of fertility. The amount of nitrites (.048) in Thompson's Lake is quite, low when the large plankton production in this lake (6.68) is contrasted with the much smaller amounts (1.91, .384, and 1.62) in the other local- ities, where the nitrites are but a little less or even greater (.147, .039, and .023). Either the nitrites are an inadequate measure of the potential fertility of the water, or the other waters named might, in the environment of Thompson's Lake, support a more abundant plankton. The nitrates are the final products of the oxidation of ni- trogenous matters, in which the nitrogen returns to inorganic compounds and is once more in a form most available for util- ization as food for the phytoplankton or other aquatic plants. The quantity of these compounds is a prime index of the im- mediate fertility of the water, and becomes a basis for future growth of the phytoplankton and other aquatic plants. The amounts of nitrates present in the waters of the four localities are very different, and at first glance exhibit little correlation either with the other forms of nitrogen present in the water or with the quantity of plankton produced. It should be noted in this connection that the nitrates, more completely perhaps than any other form of nitrogen, are utilized by the chloro- phyll-bearing organisms as food, and if taken up by the phyto- plankton the nitrogen appears in the subsequent analysis as organic nitrogen. If, however, the phytoplankton or the zoo- plankton feeding upon it is utilized by some macroscopic animal, —as, for example, by Polyodon, or by the Union idee which cover the river bottom in places,—it is removed from the field of analysis, excepting only in such animal wastes as are returned to the water by the feeding organism. If it is utilized by the 197 grosser forms of submerged aquatic vegetation, it is likewase effectually removed fi'om the field of analysis until again released by the decomposition of this vegetation. The nitro- gen as ammonia in organic compounds, or as nitrites, is either entirely unavailable for plants or, with the probable exception of the free ammonia and the amido-compounds, is less availa- ble than the nitrates. These other forms consequently more fully represent the potential fertility of the water than the ni- trates do, for the latter indicate mainly the KindUlzed portion of the nitrogenous plant food immediately available. In the light of the foregoing conditions more significance attaches to the distribution of nitrates and plankton in the four localities. The excess in the river (1.58) over that in the tributary waters of Spoon River (1.01 ) and Quiver Lake (.66) may be due in part to the greater age of the waters of the main stream and the opportunity thus afforded for the completion of the processes of decomposition of organic substances delivered to the main stream by tributaries above the point of examination. When the quantity of nitrates in the river is compared with the or- ganic nitrogen, free ammonia, nitrites, and nitrates in Spoon River or Quiver Lake, it becomes apparent that the tributary waters of this stream still act as a diluent of the river water. The source of this excess in the main stream is to be found in the sewage and industrial wastes of Chicago and Peoria. The unutilized nitrates are two and a half times as great in the river (1.58) as in Thompson's Lake (,.64). In so far as the ni- trates are concerned, both Spoon River and the Illinois might support a much more abundant plankton than they now pro- duce 1^1.91 and .384) if the conditions permitted. Thompson's Lake, drawing its water from these sources, does maintain a greater production (6.68) and exhibits a great reduction in the amount of nitrates (.64), the unutilized residium being less in this lake than in any of the other localities. The increase in the amount of plankton in Thompson's Lake over that in the river (3.5 times as much) is roughly proportional to the deci'ease in nitrates in the lake as compared with the river (.4 as much). 198 The similarity of the residual nitrates in the two lakes is strik- ing (.66 and .64), and it bears no apparent relation to their plankton production (1.62 and 6.68). The excess of other forms of nitrogen in Thompson's Lake (roughly twice that in Quiver) would seem to indicate either that the decomposing nitroge- nous substances are utilized before they reach the form of ni- trates, or that they are abstracted from the water so promptly that they do not accumulate above a certain residual minimum which is apparent during the growing period of the phyto- plankton and of the coarser forms of aquatic vegetation. (See Plates XLIX. and L.) It is evident that the nitrates in the two lakes (.66 and .64) cannot adequately represent the nitrogenous resources of the two bodies of water; neither can they furnish any reliable clue to their actual productiveness in plankton. Other factors of the environment are equally or even more potent. The number of analyses and of plankton catches is so great ( 188 and 156 from Illinois River and 40 of each from Thomp- son's Lake), and they are so distributed through the year, that the inference is justified that the nitrates shown by chemical analysis in the water of a lake or stream, especially during the growing period of vegetation, afford no reliable basis for judg- ment as to its plankton production. The sewaffe received by the Illinois River bears an impor- tant relation to the chemical condition of its water and thus to the plankton which it produces. No measurements are made by boards of public works of the amount of sewage which mu- nicipal systems discharge into the various streams which unite to form the Illinois River. Two sources of information are, however, available which throw some light on the extent of sewage pollution arising from these sources. They are the population of the cities in question and the pumpage of their water-works. Municipal engineers are accustomed to estimate the sewage discharged from a city with well-established sewage and water systems as approximately equivalent to the pump- age of the latter. I have accordingly prepared a table which includes practically all of the cities provided with these works 199 in 1897, and states the pumpage in gallons per day, the popula- tion, and pertinent data concerning the systems in discussion. The population is that reported by the census of 1890, and the figures for 1897 would show a considerable increase owing to the rapid growth of the urban population in the vicinity of Chicago during the past decade. The second part of the table includes the smaller cities with water-works but without de- veloped sewage systems. These do not contribute to the stream Population and Pumpage in Cities with Sewage Systems. City 200 waters a volume of sewage equal to the pumpage, though their imperfectly developed systems of drainage, combined with the surface run-off, carry some sewage to the stream. Population and Pumpage in Cities without Sewage Systems. City 201 from the population of Chicago, as before stated, because of the fact that the drainage of certain districts did not enter the Illi- nois River. It is apparent that Chicago, with a population four times and a pumpage twenty-hve times as great as that of the remaining territory, is the principal source of sewage, over- shadowing all others by its magnitude. The sewage of Chicago during the period of our operations was mainly discharged into Chicago River, a trilnitary of Lake Michigan. An area of 50.63 square miles lying within the city limits and having in 1897, according to estimates kindly fui-- nished us by the engineering department of the Sanitary Dis- trict, a population of 250,000 to 800,000, drains directly into Lake Michigan. The water supply of Chicago is drawn directly from the lake, and to decrease its pollution by sewage, pumping works were established at Bridgeport which raised the fouled water of Chicago River into the Illinois and Michigan Canal, which emp- ties into the Illinois River at La Salle. At low-water stages the pumpage of Bridgeport prevented the discharge of a considerable amount of the sewage into the lake, reversing at times the direc- tion of the current in the river. During floods the pumps were powerless to prevent the discharge of large amounts of sewage into the lake. Under the conditions prevailing during the years of our operations a considerable portion of the sewage of Chicago thus found its way into the Illinois River. This sewage included a large amount of industrial wastes, especially from the I^nion stock-yards and slaughter-houses connected therewith. The average daily pumpage of the city water-works in 1.S97 in Chicago was 265,530,910 gallons—an amount 50 per cent, less than the pumpage at Bridgeport. The amount discharged into the Illinois and Michigan Canal thus represents a somewhat di- luted sewage as compared with that from other sources. Chemi- cal examinations of the canal water indicate (see Palmer, '97) that the maximum period of decomposition of the sewage passed before the water entered the river. The location of the crest of this wave varied with the temperature, ranging from Lock- port to Morris. Bacteriological determinations (see Jordan '00) 202 also indicate a somewhat similar wave of bacterial develop- ment, which is to be correlated with the wave of nitrification detected by the chemical analyses. By the time the sewage of Chicago entered the Illinois River at La Salle it was thus al- ready in the advanced stages of decay and available for the sup- port of the phytoplankton or other vegetation, if, indeed, it was not already used to some extent by these agencies. The progressive nitrification of the sewage in the canal is shown by the average nitrates found by Palmer ('96) at Lockport (.84), Morris (1.44), and La Salle (2.51 parts per million). The average at Havana, about one hundred miles below La Salle, in the same year, was only 2.34 with the added amount from Peoria's con- tribution. At Kampsville, about 190 miles from La Salle, the amount falls to 1.39. The sewage of Chicago under conditions prior to the open- ing of the drainage canal in 1899 thus enters largely into the sources of fertility of the river water. It reaches the maximum of decomposition before mingling with the channel waters at La Salle, and is reinforced by the sewage and wastes of Peoria. The products of decomposition ( nitrates ) continue in dimin- ishing quantity, diluted by tributary streams—as, for example, by Spoon River, where the average amount of nitrates ( 1.01 ) is somewhat less than that of the river at that point ( 1.58, for 1894-99 )—and utilized by the developing phytoplankton and other aquatic vegetation. Entering practically at the head- waters of the Illinois, it becomes one of the most potent fac- tors in the maintenance of the abundant plankton found in the river and its backwaters. The sewage of Peoria, as represented by the pumpage of the water-works, is but a small fraction of the total amount re- ceived, being less than one per cent, if industrial waters are included. For two reasons its effect upon the plankton in the river at Havana is proportionally much greater than the fig- ures indicate. The first is the proximity of Peoria, it being 55.7 miles above Havana. The maximum stages of decomposition are usually passed, even in the coldest weather, before the sew- 203 age reaches our plankton station, so that its fertilizing effect upon the water has been operative for some time. The second reason lies in the fact that large industrial plants with private water supplies—such as the distilleries and cattle-feeding j'ards connected therewith and the glucose factory—discharge im- mense amounts of organic wastes directly into the river. As many as thirty thousand head of cattle are often on hand at one time in these feeding-yards, and the refuse from the feed- ing-pens is Hushed into the stream or piled at the river's edge till a rising flood carries it away in huge floating islands. The contributions from these sources at Peoria and Pekin are con- siderable. The comminuted vegeta])le debris of the silt owes its origin to this source in some degree, and it shares also in producing a wave of bacterial development (Jordan, '00), of putrefaction (Palmer, '97), and of the rapidly developing plank- ton organisms whose crest lies lietween Peoria and Havana. The contributions of sewage from the smaller cities in the drainage basin above Havana are relatively so small, so scat- tered, and so mingled with triluitary waters in many cases be- fore they enter the river, that no localized effect upon the plankton of the stream can be traced. The direct conveyance into drainage channels of so large an amount of animal wastes as occurs in sewage diverts from the soil and adds to the water an unusual, and, owing to the narrow conflnes of our streams, a proportionately great, source of fertility. In these particulars, together with its unusual ex- tent of impounding backwaters, its low gradient, and its im- mediate access to markets, the Hlinois River offers a magnifi- cent field for the development of a scientific aquiculture. CHEMICAL CONDITIONS AND PLANKTON PRODUCTION. A summary of the chemical conditions as related to the production of plankton in the four localities. Hlinois and Spoon rivers and Quiver and Thompson's lakes, yields some evidence of correlation, and also some points of difference which indi- cate the operation of other factors than nutrition in determin- ing the production of plankton. The following table—giving 204 the sum total of the averages of the nitrogenous matters (free ammonia, organic nitrogen, nitrites, and nitrates) and also the average plankton production—sets forth in brief the relative fertility and production of the four localities. Locality. 205 (1.292), all of which are in excess in its waters. The freedom from sewage is evidenced by the low chlorine (3.8), while the considerable amounts of free ammonia (.245), nitrites (.039), and nitrates (1.01), indicate organic decomposition in progress or completed. In the absence of any considerable contamina- tion by sewage it seems probable that these substances have tlieir origin in the organic silt and the soil waters of the very fertile catchment-basin of the stream. The water of Spoon River, in so far as the nitrogenous substances (2.586) are con- cerned, could support a much more abundant plankton than it produces (.384). As in the case of the main stream, the ex- planation of the slight production lies in the recent origin of the tributary water. Impound Spoon River water in Thomp- son's Lake, and it produces an abundant plankton. In food resources Quiver Lake is the poorest locality of the four (1.456, total of nitrogenous substances), having 40 percent, of the amount of the nitrogenous substances in the Illinois, 56 per cent, of that in Spoon River, and 67 per cent, of that in Thompson's Lake. The suspended solids (268.9), the loss on ignition (27.5) , and the oxygen consumed (.5.9), are least here as a result of slight access of silt-laden waters. The chlorine is low (4.8), and would be much lower if contaminations from river water at overflow could be eliminated ; and corroborative evi- dence of the slight contamination of the waters of this lake by sewage is seen in the amounts of free (.165) and albuminoid (.251) ammonia, of organic nitrogen (.61) and nitrites (.023), all of which exhibit minimum averages in this lake. Organic mat- ter in decay is less abundant here than in the other localities, being, for example, about 50 per cent, less than in Thompson's Lake. The final products of decay, the nitrates, are greater (.66) than the amoimts of organic matter would lead us to ex- pect, and are probably due in large part to soil waters from the drainage basin. In the light of the production of Thompson's Lake (6.68) the small amount of plankton produced in Quiver Lake (1.62) hnds no adequate explanation in a reduction of 33 per cent, in the total nitrogenous substances. Flushing by tribu- 206 tary water and abundance of submerged non-rooted vegetation are the more potent factors in the failure of the plankton de- velopment in Quiver Lake. In most particulars the averages of the analyses of water from Thompson's Lake approach those of the river water, from which it draws its main supply. There are less solids in sus- pension (282.9) than in the river as a result of sedimentation, and less in solution (44.6)—probably the effect of the small amount of creek water, or of the utilization by the plankton and vegetation of substances held in solution. The loss on ig- nition (36.5), oxygen consumed (11.9), albuminoid ammonia (.546), and organic nitrogen (1.05), all run higher than in the river as a result of the greater amount of plankton. The de- creased amounts of free ammonia (.422) and of nitrites (.048) as compared with those in the river (.86 and .147) would seem to indicate less decomposition here, while the small amount of nitrates (.64)—the least of all the averages—suggests utiliza- tion of these matters by the plankton, which here reaches a greater development than in any of the other localities under present consideration. SEASONAL CHANGES IN CHEMICAL CONDITIONS AND PLANKTON. The data concerning these changes are given in Tables X- XIII., and they are presented graphically in Plates XLIIL- L. They afford evidence for the following general conclusions: There is a major seasonal movement in the chemical con- ditions which can be traced in the analyses for each year and each locality. There are, in addition to this wide-spread and recurrent cycle of changes, many abrupt and often considerable fluctuations due to floods, while others are of minor importance and apparently of local origin. The various nitrogenous sub- stances to a considerable extent fluctuate together. The quan- titative fluctuations in the plankton show no intimate and im- mediate correlation with those of any substance determined in the analyses. Certain relations of the plankton to the quantity of nitrogenous substances are however indicated, but precise quantitative correlations cannot be established. The operation 207 of other factors is evident, chemical conditions alone offering no satisfactory clue to causes of many of the fluctuations in the amount of plankton. The cycle of seasonal fluctuations in chemical conditions is best seen in years of more normal hydrograph, such as that of 1898, and it is more regular in the backwaters, such as Quiv- er and Thompson's lakes, than it is in tributaries such as Spoon River, or in the Illinois itself. In the streams the floods produce irregularities which either do not enter the reservoir backwaters or i-each them only in diminished volume. The varying degree of contamination by sewage in the different lo- calities and in different seasons in the same locality adds an- other element which diversifies the seasonal changes and makes it more difficult to detect the common features which the fluc- tuations exhibit in all the localities. The cycle of sea.sonal fluctuations (see PI. XLIII.-L. ) in the chemical conditions is, in the most general terms, an in- crease in the nitrogenous compounds during the colder months and a decrease during the warmer ones. The maximum period usually appears in October and continues until the following summer, declining in May and June to the summer minimum, which in the following October and November rises again to the winter maximum. This fluctuation is somewhat similar to that found in soil waters. This coincidence suggests the oper- ation of fundamentally similar causes liack of the common phe- nomenon. These maximum and minimum pulses in the Illinois River in 1896 (PI. XLIII.) are most evident in the nitrates and free ammonia, though traces of their influence can be detected in the curve of the albuminoid ammonia. The suppression of this spring flood and the recurrence of four minor but unusual floods during the summer and fall are probably the cause of the nonconformity of some of the substances to these pulses and of the irregularity which they all exhibit in this year. In 1897 (PI. XLIV. ) the curve of the nitrates again exhib- its these pulses, but they are not apparent elsewhere unless it 208 be in the free ammonia. The prolonged and unbroken low water from August to the end of the year, and the consequent concentration of the sewage in the river and the marked de- velopment of the water-bloom during this period, seem to have obliterated the minimum pulse in all but the nitrates. The marked rise in chlorine and free ammonia gives some idea of the unusual degree of concentration of the sewage. In 1898 and the first three months of 1899 (PI. XLV.) these pulses are much more evident, being traceable in the nitrates, albuminoid ammonia, organic nitrogen, and oxygen consumed. The marked depression of the free ammonia during the flood season in a measure modifies its conformity to these pulses. A relation of these maximum and minimum pulses to the growth of the plankton is suggested by the chronology of the chemical (especially that of nitrates) and the plankton curves. The spring maximum of plankton production, which normally occurs in the last of April and the first of May, comes toward the close of a long period of high content of nitrogenous mat- ters. It is followed by or is coincident with the decline in these substances. With the decline in plankton production in late autumn the nitrogenous substances again increase (PI. XLIII.-XLV.). During the low water of 1897, when the mid- summer minimum of nitrogenous substances was overshadowed by the concentration of the sewage, we also find a marked in- crease in plankton production as contrasted with that of cor- responding seasons of 1896 and 1898. The warm season is pre- sumably one of more rapid nitrification, the heat favoring the more rapid decomposition of the organic matter in water, but excepting instances of great sewage concentration, as in the late summer of 1 897, we do not find an increase or an accumu- lation of the products of such decay in the water during the warm season. Indeed, the opposite seems to be the tendency. The explanation of this phenomenon lies, it seems, in the rajiid utilization of the nitrogenous products of decay by the nitro- gen-consuming organisms of the water. In open water these are the chlorophyll-bearing organisms of the plankton. In 209 lakes rich in vegetation the grosser forms of aquatic vegetation draw heavily upon these resources. The accumulations of de- cay in winter and the increased products of decomposition in summer are all largely and promptly transformed again into organized matter, leaving only an unutilized residual mini- mum which represents an equilibrium of the processes of growth and decay in progress in summer waters. The seasonal distribution of floods may also enter as a determining factor in the problem. The coincidence of the spring plankton maximum and the decline of nitrogenous matters in the river water has its par- allel in the decline of nitrates in soil waters with the pulse of spring vegetation. In both cases the decline in nitrogenous matters seems to be due to utilization by growing vegetation, by chlorophyll-bearing organisms. These maximum and minimum pulses of nitrogenous mat- ters may also be traced in the analyses of samples from Spoon River. In 1896-97 (PI. XLVI.Uhe nitrates exhibit most clearly the fluctuations in question. Traces of their presence can be detected in the plottings of the organic nitrogen, albuminoid and free ammonia, and oxygen consumed, though in all these cases the effect of Hood waters is also evident and cannot be eliminated from the problem. Invasion of Illinois River water is also apparent in October of the low-water period of 1897, be- ing shown especially by the chlorine curve. In 1898 and the hrst three months of 1899 both the cold weather maxima and the warm weather minimum are more sharply dehned and appear in all the substances above enumer- ated. The plankton of Spoon River, with the exception of that of the low-water period of 1897, is too insignificant to make much of a showing even when plotted upon a scale tenfold that used for other stations (see explanation of PI. XLVI.); nevertheless we still find here the same midsummer reduction in nitrogenous substances which has just been explained as the result of the utilization of such matters by the phytoplankton. 210 In spite of this seeming contradiction, I believe the explana- tion still holds in the case of Spoon River. The minimum peri- od occui's during the time of low water, when the principal source of the flow in the stream is ground water which has already been robbed of its nitrates to some extent by terrestrial vegetation. Again, the plankton production of Spoon River, judging from the development of the water-bloom {Eiiylenu), consists largely of chlorophyll-bearing organisms, which also rob the water of its nitrogenous substances. The period of de- velopment of the water-bloom covers the months of summer and early autumn, thus coinciding with the period of depressed nitrates. It is quite certain that the collections of the silk net fail completely to represent the quantity of those minute or- ganisms which compose the water-bloom, and thus give no adequate clue to the amount of nitrogen-consuming organisms present in these or other waters. The reduction in nitrates in this stream during summer months is not, however, as great in quantity as it is in the Illinois River (cf. PI. XLV. and XLVIL). The excess of sewage in the latter creates a greater winter maximum, thus permitting a greater range in reduction to the residual minimum of midsummer, which is about the same in both streams. But little correlation between the chemical conditions of Spoon River and its plankton production can be established beyond the reduction in nitrates in the plank- ton maximum of the autumn of 1S97 at a time of abnoi'- mal low-water. Under normal conditions the plankton curve (silk-net catches) exhibits no movement correlated with or commensurate with the changes in chemical conditions. Flood and current afford here no time sufficient for the expression of the chemical factors. In Quiver Lake the maximum and minimum periods ap- pear with distinctness and affect all of the substances in ques- tion. This is partly the result of the diminished effect of floods in this reservoir area, and also of the delimitation of the lake as a separate unit of environment with the cessation of overflow. During the flood period (see PL III. and hydro- 211 graphs on PI. XLVIII. and XLIX.)the lake receives in addition to the drainage of its own catchment-basin some access of flood waters from the l)ottom-lands above and from the adja- cent river. The water along the eastern shore, even in flood conditions, is " lake " rather than river water, as a comparison of the plottings of the analyses of water from the two sources clearly demonsti'ates. Our collections of plankton and water samples were taken within or near this l)elt of lake water, in which contamination by flood waters was not usually noticea- ble. Compare in this connection the chlorine curve of the river and lake ( PI. XLV. and XLIX.). To some slight extent, then, the analyses pertain to two sources : to the waters of overflow, largely belonging to the colder months and period of the maximum of nitrogenous substances ; and to the waters of a spring-fed lake, delimited during the period of low water and of the minimum of nitrogenous substances. The data at hand do not cover low-water conditions during a "maximum" period, which might give evidence of a seasonal cycle in chemical conditions in this lake independent of the river overflow. From conditions elsewhere it seems probable that such a cycle does occur here also, though the overflow and proliable contamina- tion may serve here to heighten somewhat the contrast be- tween the maximum and minimum periods of the seasonal cycle. In the autumn months of 189(5 and 1897 covered by the analyses, the rise in nitrates only is indicated (PL XLVIII.), the summer minimum continuing through the low-water period of autumn. In 1898 and the first three months of 1899 (PI. XLIX.) the period of maximum, November to May, is well distinguished from that of the minimum, May to November, and not only in the nitrates but to some extent also in all of the other sul)stan- ces, appearing most clearly in the free and albuminoid ammonia and the organic nitrogen. As in the Illinois River, so here also the spring maximum of the plankton (PL XLIX.) comes at the close of the period 212 of maximum of nitrogenous substances in the water and is fol- lowed by a period of depression in these substances, and in this case by a much more marked fall in the amount of plankton, which does not again rise until the return of the nitrogenous substances in the autumn. The unutilized minimum of nitrates during the summer season is but a trifle less than that in the river (cf. PI. XLV. and XLIX.), but the fact that all the other forms of nitrogenous matters are not only low but are lower than in the river throws some light on the slight devel- opment of the plankton here as compared with that in the river during this period of the summer minimum of nitroge- nous substances. While the small amount of plankton seems inadequate to explain the marked reduction in the various ni- trogenous substances, it may be that the more permanent veg- etation, the submerged aquatic flora of this lake, is an import- ant factor in the reducing process. In its seasonal production the plankton of Quiver Lake shows a general correlation with the movement of the chemical changes, though all of its fluc- tuations are not commensurate with the fluctuations of the ni- trogenous materials. The operation of other factors—such as the submerged aquatic flora and replacement by tributary wa- ters—must be called in to throw light on all the plankton changes in this lake. In Thompson's Lake the seasonal cycle of periods of maxi- mum and minimum amounts of nitrogenous matters is almost as well defined as it is in Quiver Lake. The plottings of the analyses (PI. L.) from September, 1897, to March, 1899, include two periods of winter maximum and one of summer minimum, all of which are well defined, and affect not only the nitrates but also the organic nitrogen, the albuminoid and free ammonia, and the oxygen consumed. The diminished effect of floods and of unusual flushes of sewage in this reservoir backwater is evident in the greater regularity of its seasonal curves of nitrogenous substances as contrasted with those of the river- Its close dependence upon the river for its water supply is shown by the similarity of its chlorine curve to that of the 213 river. The rise in chlorine during the minimum period, July to November, indicates the entrance into this lake of sewage- laden vi^aters of the river during this period, but it brings with it no corresponding increase in the residual nitrogenous sub- stances. The depression of the nitrates, and possibly of the other forms of nitrogen, may be referred here as elsewhere to their utilization by the phytoplankton and submerged vegetation of the lake during their period of growth. As in the Illinois River and Quiver Lake, the spring maximum of the plankton appears at the close of the maximum of nitrogenous substances and is followed by their minimum period. The autumn maximum appears, at least in 1897, somewhat before any marked in- crease in the I'esidual nitrates, though in both this year and the following one it extends into the period of rising ni- trates. A general correlation thus exists between the seasonal production of plankton and the seasonal fluctuations of ni- trogenous substances. The seasonal ftuctiiaiion of the several nifrogennus sithsfnnces exhibits some interrelations with the changes in the plankton, and especially with the accession of flood waters, and some variations from the general maximum-minimum cycle above discussed which call for brief notice. The nitrates, the final products of decomposition, exhibit the maxiinum-minimum cycle most clearly, as, for example, in PI. XLV.,XLIX., and L. The fluctuations which affect the other substances appear here in diminished prominence, as may best be seen by comparing the plottings of Spoon Eiver (PI. XLVI. and XLVII.) with those of Thompson's Lake (PL L.). The close of the maximum period of nitrates is usually later than that of the free ammonia (PI. L.), and extends for a varying distance into the period of growth of vegetation. This growth in our latitude becomes marked in the last days of April and the first of May, and continues, in some plants at least, until the frosts of October. The nitrates do not reach minimum levels, how- ever, (see PI. XLV.-L.) until late in June. In like manner the close of the minimum period is frequently delayed beyond the 214 limits of the growing period of vegetation into November or even December. This seeming inertia in the seasonal move- ment of the nitrates seems to be due on the one hand to the gradual utilization of the accumulations of the winter and con- tributions of the spring floods by the spring plankton ; and on the other, to the slow accumulations of the autumn and to utilization by the autumn plankton, which, as in Thompson's Lake in 1897, often attains a considerable development. A somewhat intimate connection between the nitrates and the plankton maxima can be detected in many instances in the diagrams. When the plankton increases, the nitrates often exhibit a depression, the extent of which, however, is not always proportionate to the change in the plankton. This absence of any constant ratio between the apparent changes in these two factors indicates the operation of other factors, one doubtless due to defects in the quantitative plankton method, and another due to changes in the component organisms of the plankton. In the Illinois River in 1895-96 (PI. XLIII.) the plankton maxima of April and October are accompanied by a marked fall in nitrates ; on the other hand, those of November, June, and August appear with rising nitrates, the last two accompanying floods. The depressions in nitrates in October, December, Feb- ruary, June, July, and September are not in any case associ- ated with a rise in the plankton, though often with the initial stages of the flood. In 1897 (PI. XLIV.) the April-May, July, September, and October maxima are all associated with de- pressions of the nitrates. The February and March depressions of nitrates occur with floods, while in November and Decem- ber no correlation is apparent. In 1898-99 (PI. XLV.) the effect of the May, June, and July maxima can scarcely be detected in the nitrate curve, while those of December and March pro- duce corresponding depressions. In this diagram neither plank- ton nor nitrates show marked changes after July. In Spoon River the development of plankton is apparently so slight and the nitrates are relatively so abundant that no 215 correlation between the respective fluctuations is apparent in the data except in the fall of 1897. when an unusual minimum of nitrates appeared in conjunction with an unusual develop- ment of plankton (PI. XLVL). Decrease in nitrates often at- tends the initial stages of flood independently of plankton development, as in December. 1896 (PI. XLVI.). Some nitrate increases, as in the autumn of 1896 (PI. XLYI.). appear with the crests of floods, especially those of the gradual type. Other fluctuations in the nitrates—and they are often considerable — show no correlation with available data In Quiver Lake in 1898-99 (PL XLIX.) the plankton maxima of April-May. June, and December all occur when nitrates de- crease. The tendency of nitrates to increase and then fall again with the crest of the floud is apparent in January, March, May, November, January, and March. In Thompson's Lake in 1897-99 (PI. L.) practically all of the maxima are attended by a greater or less diminution of the nitrates. This appears in October. Xovember. December, April- May, June, July, August, and December. The effect of floods in decreasing the nitrates in their initial stages and suh.^^e- quently increasing them is slightly indicated in January, Feb- ruary, November, and February. The tiifrites exhibit a tendency in the Illinois River to ex- cess during the low-water period of midsummer iPl. XLIIL- XLV.). averaging about .3 to .4 parts per million to .1 during the remainder of the year. This excess was prolonged into Novem])er in 1897 with the low-water period of that year. It seems thus to attend the concentration of sewage in the river. No constant correlation of movement between the nitrites and plankton can be detected. In a few instances, however, plank- ton maxima coincide with marked decrease in nitrites, as. for example, in the river in September and October, 1897 (PI. XLIV.), and the spring maximum precedes the rise in nitrites in each year. The changes in the nitrites .show no constant correlation with those of other forms of nitrogen, though at 216 times they exhibit indications of a common movement with the nitrates or the free ammonia. In Spoon River (PI. XLVI. and XLVII.) the summer rise in nitrites is not apparent except in the low water of 1897. The decay of organic matter is thus less active during this season in tributary water than it is in the main stream. In contrast with the summer, the winter exhibits somewhat more nitrites, but these are not markedly different in amount from those in the main stream at that season. The only correlation between the nitrites and the plankton of this sti'eam appears in 1897 from May to December, when plankton maxima are uniformly attended by decrease in nitrites. As elsewhere, they present no constant relation to the fluctuations of other forms of nitrogen. In Quiver Lake ( PI. XLVIII. and XLIX. ) the nitrites have their maximum during the colder months and the flood period. A marked depression of nitrites appears with the May maxi- mum of the plankton in 1898 (PL XLVIII.). In Thompson's Lake ( PL L. ) the changes in the nitrites are slight, irregular, and without apparent correlation either with other nitrogenous substances or with the plankton. Like the nitrates, the nitrites are not greatly and immediately affected by the accession of flood waters, and they run lower in the reservoir backwaters than in the main stream. The albuminoid ammonia and the total organic nitrogen fluctuate together so closely (see PL XLIX.) that it seems un- necessary to distinguish between them in this discussion. The seasonal fluctuations in these substances in the Illinois River (PL XLIII.-XLV.) are not marked, as a result apparently of the somewhat uniform accession of sewage. The dilution of the sewage consequent upon overflow is to some extent offset by the large accessions of these substances, which as silt and leachings accompany flood waters. A slight increase attends concentration in the low water of 1897 (PL XLIV.), and a slight decrease comes with the period of overflow of the same year. Similar movements are less evident in the other years 217 (PI. XLIII. and XLV. ). The effect of sudden floods, presuma- bly those of tributaries but a short distance above Havana, appears in February, 1896 (PI- XLIIL), and in 1899 (PI. XLV.) as a twelve- and two-fold increase respectively, which is re- markably abrupt and is followed in both cases by a quick but somewhat more gradual return to the previous condition. Owing to the complexity alike of the substances included in these items of the analysis and of the plankton itself, no uniform correlation of these factors can be discovered. Two different and in a certain sense opposite tendencies can be de- tected in the relationship of the movements of the plankton to those of the substances under discussion. During the winter season and the pei-iod of excess of nitrates, plankton pulses are attended by hicrease in the albuminoid ammonia and organic nitrogen. This appears in the Illinois with the pulses of April, 1896 tPl. XLIIL), and December, 1898 (PI. XLV.). During the warmer months, when most of the plankton pulses occur, the opposite tendency is seen in the movement of these substances. They tend to decrease at times of plankton pulses, as may be seen in August and October, 1896 (PI. XLIII), in May, July, September, and October, 1897 (PI. XLIV.), and in June, 1898 (PI. XLV.). With the pulses of December, 1895 (PI. XLIIL), and May, 1898 (PI. XLV.), no marked effect in either direction is apparent. In Spoon River any seasonal movement of the albuminoid ammonia and organic nitrogen is quite thoroughly masked by the disturbances due to floods. In 1898 (PI. XLVII.) these substances are a trifle lower in the warmer months than in the colder, a condition which may result from the prevalence of floods in the latter season. In 1897 (PI. XLVI.) they increase during the warm season and period of low water attending a development of the plankton unusual in the water of this stream. The effect of flood upon the quantity of these substances in the water of this stream is well deflned, and seems to throw light upon the relation which flood waters bear to the plankton 218 of the main stream. Not all of the floods which flush this tributary appear with corresponding prominence in the hydro- graph of the main river, which is the one plotted upon all the diagrams pertaining to Spoon River. In many instances they coincide. All instances in the chemical diagrams (PL XLVI. and XLVII. ) of abrupt, steeple-like eminences in the curves of albuminoid ammonia and organic nitrogen (and also of oxygen consumed) are due to sudden floods, and appear most promi- nently when the date of collection of the water sample coin- cides with the initial stages of the flood. This is well shown in September, 1898 (PI. XLVII). Not all of the samples from flood waters were collected at times which afford evidence for the enriching effect of the initial stages of these tributary flushes. The relative amount of these and other forms of nitrogen which floods bring to the river is well shown in this flood of September, 1898 (Table XL and PL XLVII.). On August 30 the amounts of albuminoid ammonia (.32) and organic nitrogen (.6) are normal for that season of the year. With the flood of the flrst week of September these amounts increase more than tenfold (being 3.6 and 8.32 respectively), falling again a week later to the normal (.2 and .48). A large part of this matter is in suspension. For example, in the flood of May, 1898 (PL XLVII.), about 86 per cent, of the albuminoid ammonia (2.32) and 90 per cent, of the organic nitrogen (5.46) was in suspension. It is not plankton, neither is it to any large extent sewage, which the tributary floods of Spoon River bring to the Illinois as organic nitrogen, but largely organic debris not yet decayed. The sewage-laden river habitually carries much less of these substances than these tributary flood waters laden with this organic debris from fertile prairies. The latter thus become very important agents in maintaining the fertility of the river water. The effect of these periodic additions of nitroge- nous substances by tributary floods upon the plankton of the river will be discussed in another connection. A decrease in these nitrogenous substances attends the two 219 plankton pulses of 1897 (PI. XLVI.) in the warm months of May. and September, but the increases noted with the pulses of plankton in the winter in the Illinois River are not apparent in the case of the pulses of February and December in this stream, though no decrease appears as in the summer months. In Quiver Lake in 1898-99 (PI. XLVIII.) a seasonal move- ment in the albuminoid ammonia and the organic nitrogen is evident, though it seems to accompany the access of sewage- contaminated waters of overflow, as appears on comparison with the chlorine curve. This seasonal movement is evident as a depression of the curves during the warm and low-water months, and as an elevation during the colder months of the flood period. As in Illinois and Spoon rivei's. the plankton pulses in Quiver Lake of the warm period, in May and June, are attended by a temporary decrease in these nitrogenous sub- stances. A still more marked decrease in both albuminoid ammonia and oi'ganic nitrogen attends the winter pulse of plankton in February. 1899. while that of the preceding De- cember appears with an upward movement of the organic nitrogen and a downward one in the alliuminoid ammonia. The correlation between the movement of albuminoid am- monia and organic nitrogen and of the plankton is thus in this instance (predominantly, at least) similar to that noted else- where in the warmer months. The very slight ripples in the plankton curve in July. August, and September attend minor increases in these nitrogenous substances, a feature noted else- where in colder months. In Thompson's Lake (PI. L.) the albuminoid ammonia, the organic nitrogen, and the i^lankton are all more abundant and ex- hibit greater fluctuations than they do in Quiver Lake. These conform in a general way to the tendencies noted in other localities. The amounts present during the colder months, October to May. are a trifle greater than in the intervening warmer period. There is also a tempoi-ary decrease in those nitrogenous matters attending plankton pulses in the uarm months. This appears with the pulses of June. July, and 220 August-September, 1898. A temporary increase appears with plankton pulses in cold months, as in December-January- and February, 1899. Some exceptions to these general tendencies appear here as in Quiver Lake; such, for example, as that in the low water of the autumn of 1897, when the great plankton pulse of October-Novenaber attends an unusual wave of both the albuminoid ammonia and the total organic nitrogen. Temporary decrease in the former appears with the crest of this plankton pulse, and again in the pulse of December, along with an increase in organic nitrogen. The spring maximum of April-May, 1898, comes with a rising wave of both substances, whose crest coincides with the fall in the plankton. It is evident from the data here presented that the fluctu- ations in the volume of the plankton, as determined by the methods employed by us, show some intricate correlations with the changes in the quantity of albuminoid ammonia and organic nitrogen. The massing together of all organic matters, both living and dead, indigenous and adventitious, in the de- termination of these two substances, and the composite nature of the plankton itself, including both the synthetic phytoplank- tonts, and the analytic zooplanktonts, alike combine to conceal the relationship which exists between the succession of living forms in the plankton and the flux of nitrogenous matters in suspension and solution therein. Furthermore, the plankton is not the only assemblage of organisms concerned in this flux of matter; the bottom fauna, the flshes and other aquatic ver- tebrates, and aquatic fauna of the grosser sort, all share in effecting the changes here manifest. We lack a common unit of measurement in terms of which we can express the values alike of the chemical analyses and of the volumetric and the statistical determinations of the plankton. Precise comparisons, for example, of the changes in the organic nitrogen with the cubic centimeters of plankton and the number of diatoms cannot be made. The direction of the changes in these several elements can, however, be noted, 221 and its interpretation, ill many cases at least, becomes proI)aliie, if not, indeed, certain. In the first place it may be noted that the fluctuations of the plankton are not paralleled l»y proportionately great move- ments in the total nitrogenous substances in the water which enter largely into their composition. For example, the s])ring maximum of the plankton is accompanied by no such wave in these substances. Indeed, aslightrippleof depression seems to be the only concomitant fluctuation. Even granting a large mar- gin because of the absence of a common unit of measurement, it remains apparent that the Huctuations of the substances in question and of the plankton are not proporfioiidl. A single illustration, found in the spring maximum in Quiver Lake in 1898 (PI. XLIX.), will suffice to make this point clear. The following table, drawn from Tal)le XIII.. gives the amounts of plankton and of the several forms of nitrogen present l)efore Quiver Lake. Date 1898 222 Three causes may be assigned in explanation of the absence of proportional correlation in the flux of these nitrogenous sub- stances and of the plankton, all of which are operative, but in varying effectiveness at different times and under different conditions . In the first place, the plankton itself constitutes but a part of the total organic nitrogen; how small a one the data at hand do not determine. Barring out error arising from the death of the plankton and from the solution of the products of its decay which might take place during the interval between the collection and analysis of the sample, we find in the relative amounts of albuminoid ammonia and total organic nitrogen in solution and in suspension respectively some evi- dence as to the possible limits of the proportionate amount which the plankton and silt together form of the total nitrogenous substances. The average amount ( Table X. ) of albuminoid ammonia in solution and in suspension from July 6, 1897, to March 28, 1899, is .355 and .131 parts per million respectively. Plankton and silt together thus constitute about one third of the total albuminoid ammonia in the Illinois River. The rela- tive amounts of dissolved and suspended albuminoid ammonia at the weekly intervals of analysis fluctuate according to access of flood waters and increase in the plankton. The former is the more potent factor. Usually the amount in suspension is from one third to one half that in solution, rarely equaling or surpas- sing it, as in the flood of February, 1898, when it rose to .4 as compared with .28 in solution. The plankton pulse of April-May, 1898, accompanies a rise in total albuminoid ammonia from .4 to .6—an increase of 50 per cent. The increase lies almost entirely in the suspended form, which rises from a previous level of .04 to .08, to .08 to .20, that is, it is more than doubled. The volumetric increase in the plankton is, however, over thirty-five-fold. Thus, under the most favorable conditions, receding flood, little silt, and plankton maximum, the increase in suspended albuminoid ammonia attending a thirty-five-fold increase in the plankton constitutes but 33 per cent, of the total 223 amount present. During the long-extended periods of plank- ton minimum it is apparent that the plankton must constitute a very much smaller pai-t of the total amount of albuminoid ammonia in the water. That much of the albuminoid ammonia may be in the silt is shown especially in the table (Table XL) and diagi-ams (PI. XLVI. and XLVII.) of Spoon River at times of Hood. In such waters there is practically no plankton—as will be shown else where from the examination of our plankton collections in that stream—although the amount of albuminoid ammonia is often very gi-eat. The average amounts of total organic nitrogen (Table X.) in solution and in suspension during this same period are .69 and .34 pai-ts per million respectively. The latter (.341. which represents plankton and silt, thus constitutes about one third of the total amount ( 1.08) of organic nitrogen in the water. The proportion of this fraction which the plankton may consti- tute under the most favorable conditions may be inferred from the increase in the suspended organic nitrogen which attends the spring pulse of 1898 (PI. XL.). This rises from a previous level of .12^.16 parts per million to .24-. 64. the latter with the de- cline of the plankton, and at its maximum (^.64 I it constitutes 46 per cent, of the total amount of organic nitrogen in the water. On May 3, when the plankton is at its maximum, the suspended organic nitrogen is but .24. or .25 per cent, of the total. During the periods of plankton minimum the propor- tion which the plankton forms of the total organic nitrogen must be very much less than at times of plankton maximum, since the amount in suspension shows no decrease at all pro- portional to the fall in the amount of plankton. Here also floods are quite as potent as plankton in causing marked in- crease in the amount of total organic nitrogen in suspension, as will be seen on comparison of the curves of this matter with the hydrographs on Plates XLIII.-L. It is thus evident that the plankton does not form, even under most favorable conditions, any large part of the total 224 organic nitrogen—certainly less than 50 per cent, and on the average much less than 33 per cent., which figures represent the total organic nitrogen, both plankton and silt, in suspen- sion. The fluctuations of the organic nitrogen contained in the plankton are thus masked by the predominance of the dis- solved form, and by the undetermined quantity of nitrogen- containing silt. A second cause for the lack of proportional correlation between the movement in these nitrogenous substances and the plankton may lie in the utilization by the plankton itself of some forms of nitrogen included within the range of sub- stances reported in the analyses as albuminoid ammonia and total organic nitrogen. For example, some organisms of the phytoplankton may utilize as food such forms of organic nitro- gen in solution in the water as the amido-compounds and the humus acids. It may be that some of the animal wastes are turned into the more highly oi'ganized nitrogen of the phyto- plankton without passing through complete oxidation and a return to the inorganic nitric acid and nitrates. If this be the case the flux of nitrogenous matters may lie quite within the range of substances here discussed, and the movements in nitrogen incident to these changes will consequently produce no pulses in the common curves of these substances. When, however, the inorganic nitrogen enters largely into the ebb and flow of the nitrogen of the plankton, the possibility of a correlated movement of plankton and organic nitrogen be- comes apparent, though proportionate pulses in the two remain improbable so long as the organic but non-living nitrogen con- tributes also to the flux of matter involved in the plankton changes. That the phytoplankton, as other low forms of vegetation, may thus utilize organic nitrogen in some of its forms as food, has been rendered probable by the experimental work of Loew ('96), Bokorny ('97), Maxwell ('96), and Zumstein ('99). The work of the latter is especially in point in this connection, since his experiments deal directly with a genus, Eugleiia, which 225 furnishes a large part of our phytoplankton of midsummer and the bulk of the water-bloom. The experiments of this author have shown conclusively that this chlorophyll-l)earing organism is usually autotrophic (holophytic) in the light and in the ab- sence of abundant organic nitrogenous matters in solution, and under these conditions its chromatophores are a bright green. When organic nitrogenous matters in solution are abundant the organism becomes mixotrophic (half saprophytic) even in the light, and its chromatophores may become paler. In the dark it becomes colorless, and depends entirely (saprophytic) upon the dissolved uryanic nitrogen for its growth and multi- plication. The waters of the Illinois Eiver and its backwaters are unusually turliid, thus excluding more than the usual amount of light. The plankton of this environment is rich in species and individuals of flagellates, algte, and diatoms, many of which exhibit this tendency to become paler. This I have noticed repeatedly in the examination of the living plankton, and to some extent in material preserved in formalin-alcohol. It has occurred in the several species of Euylena, viz., vividis, saiiffidiica, descs, anis, spiroffi/ra, and (/racilis. I have noted it also to a very marked degree in Cldp^wf/r/ abound more generall}-. Ward and Grajbill ('00) find that the decrease ranges from 60 per cent, to 70 per cent, and averages slightly less than two thirds. Juday ('97) finds that the decrease amounts on the average to SO per cent. Both of these records deal only with a small number of midsummer planktons, while our records cover the whole round of sea- sonal changes. This may be an element which tends to increase the range of the decrease shown in our results. The gravity work in our experiments was done in tubes identical in pattern with those used by Ward and Clraybill ('00). and the time for settling was the same. Juday ('97) gives no account of his gravity method. The centrifuges which we have severally used are only approximately of the same pattern, and they have not been used in exactly the same way by any two of us. We may compute the specific pressure in dynes per square centimeter at a given distance from the axis of rotation accord- ing to the formula ''<" { ~^'' j : in which r> is the density of the contents of the tulie; <«. the angular velocity in ra- dians per second; r the di.stance in centimeters from the axis of rotation to the bottom of the tube; and n, the dis- tance from the axis to the top of the fluid. For density I have used the specific gravity of water, since the extractions from the plankton considerably increase the specific gravity' the alcohol in varying degrees in which the plankton was preserved. 2 ~ 11 The formula for .« is ^^ . in which » is the number of revolutions per minute. 258 On this basis the pressure in dynes in our use of the centri- fuge, where there were 1 ,000 revolutions per minute with dis- tances of 16.7, and 4.5 cm. to bottom of tube and top of liquid respectively, was 1,420,484. Professor Ward writes me that these distances r and )\ in the Bausch & Lomb machine he used were 14 and 5 cm. respectively, which with a density of 1 and 80 turns of the crank, equaling 1,840 rotations of the axis (on the manufacturer's authority that the machine is geared to give 23 rotations of the axis to one of the crank), gives 2,774,- 897 dynes. Mr. Juday ('97) states that the pressure in his cen- trifuge was 391,680 dynes, but he writes me that the density used was the difference between that of alcohol and dried plankton, amounting to only .2 and .25. Reducing his calcu- lations to our basis in so far only as the matter of density is concerned, we find the dynes to be from 1,958,400 to 1,566,720. A further difference between our methods lies in the fact that in our use the pressure was exerted two minutes, and but one minute in that of the other investigators. The practical differences may not really be so great as the figures indicate because of the asymptotic character of the curve of reducing volume as the pressure continues or is in- creased. It it desirable that some standard unit of measure- ment be agreed upon for purposes of comparison. The use of the centrifuge in the many measurements here recorded has only confirmed the views expressed by me ('97) regarding its utility, greater accuracy, and convenience for volumetric plankton work. Material which has been properly preserved has not suffered in the compression. All of the enu- meration work to be reported in the second part of this paper has been done upon plankton which has been centrifuged at least once and in some cases six times. I have not detected any mutilation or distortion of the constituent organisms unless it be of Leptodora hyaJina, an elongated and delicate cladoceran, and several other organisms of great delicacy of organization. These are often crumpled as a result either of the compres- sion or manipulation of the plankton, but the crumpling rarely 259 interferes with identification lof the species or determination of sex or breeding condition. The statement made liy me ('97, p. iU) that "this is, I be- lieve, the first application of the centrifugal machine to quan- titative plankton work" requires modification. In Kraemers account of Samoan plankton (,'97) he describes a traveler's centrifuge for use on shipboard for measurement of plankton, and I infer from his text that this machine was in use by him in 1893-95 in Samoan waters, though 1 find no explicit statement to that effect. If this be true, his use antedates ours, which did not pass the experimental stage until January, 1896. In any case the measurements he publishes ('97) were made by the centi'ifuge. The following statement made by Ward and Graybill ( '00 ), — "Juday ( "97 ) was apparently the first to publish an account of the use of the centrifuge for this pur- pose. Both Dr. Kofoid and 1 had, however, experimented indepen- dently for more than a year before that, and had written to var- ious investigators regarding the advantage of such an instru- ment,"—requires notice in this connection, since the question has been raised by these writers as to the priority in publication of the use of the centrifuge for plankton measurement, and the facts are incorrectly given. The date of Juday's paper is subse- quent to May 28, 1897.* for the volume containing the paper contains the records of the held jueeting of the Indiana Acad- emy held on that date. The date of publication of my ac- count of the centrifuge was March 10, 1897, and it was also mentioned by Professor Forbes I '96) in the biennial report of our Station operations. This report was again issued January 24, 1897, in separate form, and distributed to plankton investi- gators generally. The date of publication of Kraemer's work is "Ende Januar oder Anfang Februar 1897; geuauer konnen wir das Datum leider nicht augeben," according to inforuia- tion received by me from the publishers Lipsius and Tischer, of Kiel, Germany. The empty honor of first publication thus probably belongs to us, and certainly not to Juday as Ward *In a letter Mr. Juday informs me that liis paper was published in August, 1897 260 and Graybill state. The real credit for first centrifuging plank- ton belongs evidently to Kraemer. In view of the statement above quoted it may be well to add that our experimental use of the centrifuge was begun in the autumn of 1896. It was at once adopted, ai;d our plankton thus far collected was meas- ured by it at that time. The measurements made are those used in the present paper. Our experiments with and adop- tion of the centrifuge were independent of and without knowl- edge of similar work elsewhere. SILT ESTIMATION In Tables III.-IX. the amount of the actual catch of the plankton net will be found, and in subsequent columns the estimated percentage of silt, and computed volumes of silt, plankton, and total catch are given in cubic centimeters per cubic meter of water. In all our discussions of the plankton the amounts used are those of plankton only, unless otherwise stated, that is, of the total catch /?,« the e^itimuted anion nt of silt, and they are always quoted in terms of cubic centimeters per cubic meter of water. The determination of the amount of silt has been of ne- cessity a matter of personal estimation, and involves a soui-ce of error of uncertain extent. The estimates have been made large- ly by myself, with some aid from Mr. R. E. Richardson, and no effort has been spared to maintain a uniform standard of esti- mation so as to distribute, as far as possible, the error incident to the process. Accordingly the estimates were revised and cor- related after the qualitative and numerical analysis of the col- lections at Station E, and they consequently rest upon a com- parative basis upon the examination of the catch as it appears in the Rafter cell.* In the case of the collections from the Illinois River only, they have been controlled in some degree by the re- sults of the numerical analysis. The estimates of silt were made Closer study of the quantitative and qualitative data since the above was written leads me to suspect that in some cases the silt estimates are too high, especially where there is much light flocculent debris, which occupies considerable space in the Rafter cell but may be compressed considerably in the centrifuge. 261 without reference to. and in most eases prior to. the prepara- tion of tables and plots of seasonal distribution. Estimates made by othei's, not accustomed to judging the plankton, re- veal a wide divergence in percentages, and independent esti- mates which I have made of the same material on different oc- casions show some divergence, though, as a rule, quite within the probable error of plankton method. The silt estimation does not, I believe, essentialh' vitiate any of the conclusions drawn in this discussion of quantitative results, and in no way enters into the qualitative analysis. THE CLOGGING OF THE NET. The collections made with the silk net drawn from the bottom to the surface of the water in vertical or oblique hauls are all diuiiuished in vohame to some extent by the resistance of the silk to the rapid passage of the water. As a result of this, the net pushes aside some of the water in the column which it is supposed to traverse. As its meshes clog with the accu- mulating catch the amount pushed aside is increased progres- sively during the haul. The actual catch of the net is there- fore only a poi'tion of the total contents of the column of water, whose length is that of the haul and diameter that of the mouth of the net. The volume of plankton actually present in this column can be computed if we can determine the factor of correction. Hensen ('87 and '95) has sought by experiment with filtered water to determine the mathematical formula which will give this correction for a net of known silk and di- mensions drawn at given velocities. This factor he calls the "coefficient of the net." Keighard ('94) attempted to de- termine the coefficient of his net by using a miniature model in water in which Lobelia seeds had been placed, but ultimately adopted the formula of Hensen. It was necessary that in our final computation of the volume of plankton we should make some correction in all catches of the drawn net for this loss due to the pushing aside of the water. Since our net was constructed after the Henseu-Apsteiu model it was possible to apply the mathematical method of 262 Hensen ('95). This coefficient—computed for us by Instruc- tor W. C. Brenke, of the University of Illinois—is 1.320 or 1.303, according to the area of the silk in the bucket of the net, for the velocity of one-half meter per second, which vpe uniformly employed. Our net has an area of 81.72 square centimeters in the opening; of 1,847.5 and 22.66 square cen- timeters respectively in filtering cone and windows of the bucket; and its coefBcient is 1.32. With a disk of silk clamped to the lower end of the net,—a method used prior to the adop- tion of my detachable bucket,—the area of the silk of the buck- et falls to 15.2 and the coefficient is accordingly reduced to 1.303. A second net frame which we used, with an opening two millimeters less in diameter, has a coefficient of 1.318. Apstein's ('96) net has an opening of 92 square centimeters, a filtering cone of 1,730, and the silk of the bucket measures 62 square centimeters. The coefficient of his net is computed to be 1.39. Observations on the operation of the net in the field through the seasonal changes of the plankton led me to be- lieve that a uniform coefficient, and, moreover, one founded on the operation of the net in filtered water, would not ade- quately correct the error, since it takes no account of the sea- sonal changes in the quantity and kind of plankton, and does not recognize the effect of the progressive clogging of the net by the catch, or the change of the net with use. Upon the adoption of the pumping method of collection a number of tests were made for the comparison of the amount of plankton taken by the drawn net and that by the pump and filtering net along parallel courses of thirty meters—our usual haul. The results amply justified my belief that the coeffi- cient fluctuates with the conditions above named as well ag with the condition of the net. We therefore sought by this empirical method to determine the coefficient under a variety of conditions representative of our environment. These tests were not carried beyond this point, since we had adopted the pumping method for the later and major part of our work. ] t ] 1 \ ( i i r. Date 263 In thih method the question of coefficient is entirely eliminated. The I'esults of these tests are given in the preceding table. The variation in the coefficient with different seasons, lo- calities, and planktons is apparent upon the first glance at the table. The greater straining capacity of a new net as com- pared with one which had been used for some time may be seen in the tests made June 16. and July 9, 11, and 21, 1896. The new net—the silk in which had been shrunken by washing and pressing several times prior to use—catches at least 50 per cent, more than the old one which had done service since February 25. The rise in the coefficient as the net progressively clogs by plankton is demonstrated by the tests of October 14, 1S99, and July 27, 1897. The latter test is graphically presented in the accompanying figure. S 10 5 Dtti. Fig. B.— Catches of [jlankton made on 30-meter haul by drawn net and pump The upper, continuous line represents the pump catch; the middle, dotted line the catch of the net measured by the settling method; the lower, heavy line the same measured by the centrifuge. 264 The coefficient has been computed in each case on the ba- sis of measurements by the gravity method, by the centrifuge, and by enumeration of all the larger and quantitatively more important constituents of the catch. An examination of the table w^ill indicate that the relation and direction of the dif- ferences of the various coefficients do not materially differ by the three methods. The results by the enumeration method give the largest coefficient—probably as a result of the elimina- tion of the silt factor in some instances, and possibly by reason of the large margin of error involved in the method. It is evident from the table that an average of a number of catches, not only by the net but also with the pump, should be used if empirical coefficients are to be established w^ith ac- curacy. It is probable that the low^ coefficients seen in a few instances result from insufficient pump catches, or from some error in paralleling the catches. Since the coefficient problem was eliminated in our later work by the use of the pump, fur- ther efforts to establish empirical coefficients were abandoned for lack of time to carrv on more elaborate tests. Three alternatives were thus before us. First, to adopt the coefficient computed according to Hensen's formula, and use this one factor, 1.32, for all catches irrespective of the age of the net and of seasonal, local, quantitative, and qualitative differences in the catch. This method Apstein ('96) and other European planktologists have adopted. Reighard ( '94), Ward ('95), and Juday ('97) have also followed this plan, but in each case they were dealing only with catches taken in midsummer from the same or similar bodies of water, and the resulting error thus introduced was much less than would result from the adoption of a uniform coefficient for our varied catches. Furthermore, we had the evidence of the probable extent of this errror which the pumping method afforded. A second alternative was to ignore the coefficient question entirely ; but this involves even greater distortion of the prob- able seasonal and local fluctuations in the plankton. A third method, and the one finally adopted, was that of 265 assigning an empirical coefficient to each catch. This coeffi- cient was decided upon after analysis or inspection of the plank- ton, and in view of its quantity and constituent organisms, the amount and nature of the silt, and the age of the net, the basis of estimation in each case being the coefficient test by the pumping method whose conditions most nearly approached those of the catch in question. These coefficients were decided upon without knowledge of or reference to the effect which they might have upon the theoretical questions arising from the analysis of the quantitative results, and prior to the organ- ization and analysis of the volumetric data. Olniously this latter method involves both possible and probable error in estimation of similarities and differences in the catches examined and in maintaining throughout a uniform standard. Nevertheless, for plankton catches as varied as those with which we deal, it is probable that this method involves less distortion of volumetric results than the omission of the coeffi- cient factor or the adoption of a uniform factor for all catches irrespective of the tluctuatious in this factor as revealed by our field tests. Accordingly all of the volumes of catches by the drawn net, of plankton, silt, and total catch per cubic meter recorded prior to May 20, 1896, in Tables III.-IX. have been computed with this coefficient as one of the factors, the actual factor employed being given in the tables in each case. The results of my efforts (see Kofoid. '97a ") to find an ade- quate correction for the loss by leakage through the silk by the use of hard-pressed filter paper and the Berkefeld army filter will be discussed in another connection. The volumetric data of the plankton at the seven stations (see PI. II.) at which periodical collections were made, name- ly, the river channel (E), Spoon River (M), Quiver Lake (C), Dogfish Lake (L). Flag Lake [K], Thompson's Lake (G). and Phelps Lake (Fj will now be discussed, and the general ques- tions arising from the investigation as a whole will then be treated. The chronological series of collections at these seven 266 stations included in this discussion number in all 643, distrib- uted as follows: Illinois River 235, Spoon River 36, Quiver Lake 115, Dogfish Lake 48, Flag Lake 44, Thompson's Lake 99, and Phelps Lake 67. ILLINOIS RIVER CHANNEL, STATION E. (Table III.; PI. I., V.,VII.—XIII.) DESCRIPTION OF LOCALITY OF COLLECTION. The collections were made two and a quarter miles above the city of Havana, a short distance above the outlet of Quiver Lake (PI. XL), at a point where the river was about 500 feet in width at low water and about 600 feet from crest to crest of the banks, which are here fringed by willows {Salix nigra and S. Inngifolia) on both sides. The eastern shore is a nar- row spit, 6 to 8 feet above low water, separating the river from Quiver Lake. The western bank is higher, 8 to 10 feet, and is covered by bottom-land forest. This is also a spit or "towhead" between the river and Seeb's Lake. At low water (PI. IV. ) the eastern bank is exposed as a gentle declivity of 25 to 40 feet, while the western one is much wider—a belt, 50 to 75 feet in width, of soft black mud with gaping cracks (PI. V.). A short distance from the low-water shore-line the bank shelves some- what abruptly to the bottom, which with the exception of a slight ridge near the center of the channel extends in an un- broken level from side to side of the stream. The depth at low water for a width of over 400 feet is 8 to 9 feet. To the north- ward the river deepens slightly, while towards the mouth of Spoon River it shoals to 6 feet, and below it to less than 5 feet. The banks are of black alluvium, hardened in the upper levels by exposure at low water, but al- ways soft and treacherous near the low-water line. The bot- tom in the channel is firm, being a compact bed of heavy blu- ish mud mingled with sand and the shells of Union idee, which form in many places continuous beds of large area. A slight curve in the river above our plankton station shifts the current at that point towards the eastern shore, but at the 267 point of collection the run of driftwood in the stream exhibited no marked difference in current in the channel for an extent of fully 400 feet. Approximately uniform conditions thus pre- vail over a considerable extent of the river channel at this point. At high water (PI. III.) the banks are submerged, but aside fi-om increased rapidity and some lateral extension there is no noticeable difference in the conditions of the current. MODIFICATIONS OF .METHODS AT THIS STATION. Collections made by the oblique-haul method were always taken on the western side of the stream, across the current from a point in deep water, the surface end of the haul being completed in shoaler water. At times of high water it was necessary. I)oth on account of the streugth of the current and the depth, to shift the apparatus still more towards the shore, and, finally, in the flood of Decembei-, 1S95. to abandon the method and substitute temporarily a series of four to six vertical hauls, amounting to about 80 metei's—the usual dis- tance of the oblique haul. These were made in midstream from a floating or anchored boat, and were continued from De- cember 27. 1895. to May 20. 1896. After the adoption of the pumping method on the latter date the boat or launch was at first allowed to drift with the current while the collection was made. Owing to frequent difficulty caused by the drifting of the boat into .shallow water by the wind before the catch was completed, we finally adopted the plan of anchoring the boat or launch in or near the mid-channel while the collections were being made. During the winter season, owing to air-holes and weak places caused by the irregular melting of the ice upon the low- er surface, the ice on the river channel was rarely firm enough to permit safe transit of our plankton outfit, whose total weight was over 800 pounds. Steel runners were placed upon the bottom of the boat, and by the aid of ice hooks it was pos- sible to run over or to break one's way through thin or rotten ice to the mid-channel station, where open stretches of 268 water were not infrequently found. In a few instances, owing to roughness and rottenness of the ice, it was not possible to reach the point up-stream where the collections were usually made, and in such instances the catch was taken nearer Ha- vana but always above the mouth of Spoon River (PI. I.). Even when the ice was running at the time of break-up, it was possible by floating in rifts of the floes to secure a catch of the channel plankton. Thus in all seasons our catches at this station are typical of the channel plankton. CHRONOLOGY OF COLLECTIONS. As shown in Table III., the collections at this station cover the period from June 12, 1894, to March 28, 1899, in which time catches were made on 235 different days, 10, 50, 76, 34, 52, and 13, respectively, for the several years included. The interval between collections in 1894 (PI. VIII.) ranges from 14 to 34 days. In the first half of 1895 (PI. IX.) they were few and irregular, but four being taken, while in the second half of that year 46 were taken at intervals of one to twelve days, the interval varying with flood conditions, since an attempt was made to follow closely the effect of changing river levels upon the quantity of plankton. The December flood of this year was followed at intervals not exceeding five days until Februa- ry 10 of the following year (PI. X.). From this time till April 24 the intervals average about seven days, in no case, however, exceeding eleven. From this date till the end of August, 49 col- lections were made at intervals of one to seven days, following thus closely the fluctuations attendant upon the two recurrent floods of that season (PL X.). The field station at Havana was then closed, and until it was reopened in the following July fortnightly or monthly trips were made to Havana for collec- tions (PI. XL). From this time until the suspension of opera- tions March 28, 1899, the collections—with the exception of a few extras and two delays due to sickness—were made at regu- lar weekly intervals (PL XII., XIII.). Thus, in one or another of the years in question all months but October, November, 269 February, March, aud April have l)een covered by collections at intervals of five days or less, and from July 14, 1897, to March 28, 1899. a period of nearly twenty-one months, the series of regular weekly collections is almost unbroken. The following table gives the distribution of the collections by months in the several years. DISTRIBUTION OF COLLF.CTInNS BY MONTHS. 270 from the assumption that a collection at a single place is rep- resentative of a larger area—an assumption necessary if any wide signilicance is to attach to the analysis of plankton data. It has been a matter of observation that the quantities of plank- ton taken at different places in a body of water, or even within very narrow limits, are not equal in volume under similar methods of collection. Thus, Apstein ('96) in eighty catches in German lakes finds the average deviation from the means to be 5.52 per cent, when the plankton is computed per square meter of surface. Of the eighty catches 68 or 85 per cent, ex- hibit a departure of less than 10 per cent, from the mean, and only four have a departure in excess of 15 per cent. These de- partures are derived from averages of comparabld* collections on various dates and in several different lakes in groups of onh/ two fojice, evidently from mid-lake waters, and hardly afford suf- ficient data for an analysis of the conditions of distribution in any given lake or in atypical lake, since on account of their small number they do not throw any light on the effect of shore, tribu- tary waters, vegetation, currents, or other factors of the environ- ment. They indicate, however, the probable error of ±5.52 per cent, in mid-lake collections, and seem hardly sufficient to sub- stantiate fully the more general conclusion that "das Plankton sehr gleichmassig in einern Seebecken vertheilt ist." Reighard ('94) finds in the case of twenty-nine hauls that his results "agree very well with that of Apstein." Of his twenty-nine hauls, 26, or 90 per cent., have a "percentage of difference from the average" which is less than 20. These percentages yield, I find, an average of 9.7 per cent, to Apstein's 5.52 per cent. Reighard does not, however, compute his "percentage of dif- ference from the average" in the same manner as Apstein deter- mines the "Abweichung von Mittel," the former using the vol- ume of each catch as the basis for the determination of the per- centage of difference from the average, while the latter employs the average of the two or more catches for this base. Obviously this slightly increases one half of Reighard's percent- ages and decreases the other half, though it does not materi- 271 ally affect the average of all of them. Apstein's method of stating the ± error in terms of departure from the mean is, it seems to me, to be preferred to the "percentages of difference" which Reighard uses. In all instances but one Reighard averages but two collec- tions, made at some one of fourteen points of collection in Lake St. Clair. His percentages of difference, therefore, refer only to these individual points of collection and tiof to flic lake as a irliolr. His collections were all made within an interval of ten days, and it is p)-obable that the results can be used to deter- mine the departure from the mean in the lake as a whole. This he has not done, though he cpncludes from these percent- ages of differences of ihe jjairs of colled ions that "llw jtlaidioii is distribidi'd over Lake St. Clair iritli great imi/orutiti/." In the case of Apstein's data the sets of collections are scattered over several seasons and represent a number of lakes, and range in number from two to five in each test. It is obvious that conclusions as to tlie uniformity of distri- bution of the plankton in the lake ((s a irhole should be based upon a comparison of all catches with their average, and are best expressed in terms of departure from their mean, employ- ing the mean as a base and expressing the deviation in percent- ages whose average will constitute the mathematical expres- sion of the variation in distril)ution or the ± error of the method. For reasons al)ove stated, this method cannot lie applied to Ap- stein's data as a whole, though it is the one he uses for indi- vidual lakes or tests. Applying this method to the data of Reighard's ('94, p. 33) table, as in the accompanying tabulation, I find that the aver- age departure from the mean is ±31.8 per cent., with a range of + 111.5 to- 57. 5,—a total of 169 per cent,—on the basis of the amount of plankton per square meter of .surface ;and ±28.8 with a range of +91.3 to —55.4,—a total of 146.7 percent.,—on a basis of plankton per cubic meter of water—a deviation much greater than that expressed by Reighard's method. This deviation is much greater than that found byApstein ('96), and it re- 272 suits in large part, doubtless, from the greater number of catches averaged, and from the fact that they represent a num- ber of more widely separated points in a larger body of water, DISTRIBUTION OF PLANKTON IN LAKE ST. CLAIR. Haul No. 273 sample of the contents of the water in contiguous parts of the stream, or of the plankton present in the water passing a given point of the stream for any considerable length of time. Will not the conditions pertaining to tluviatile life cause such local variations in the plankton and such changes in it fi'om day to day that chronological series of isolated collections will reveal only erratic and meaningless fluctuations, without significance for the analysis of the factors of the environment and incapa- ble of revealing an orderly regimen of aquatic life ? In other words, is the river a unit of environment sufficiently compact to yield, by the plankton method, data of scientific value com- parable with those derived from other bodies of water, types of which we find in the sea and the lakes .? As contrasted with the lake, the river as a unit of environ- ment pi'esents a constant and excessive predominance of the longitudinal over the transverse axis. This feature, combined with the fact that in a river the relative shore development is much greater than it is in the lake, makes it necessary to discuss the longitudinal and transverse distribution of the plankton iu the stream separately. LOXGITUDINAI, DISTRIBUTION. With a view to testingthis question of the local longitudinal distribution of the plankton in the Illinois River. I made a series of ten catches in immediate succession from a boat anchored at our usual station in mid-channel on October 29. 1S9(5. This was at a time of a considerable autumnal development of .S'^- )uini and Si/nrlid'ta. and the quantity of plankton present (see Talile III.) was sufficient [P\. X.) to allow room for cousidera- lile fluctuation and to minimize the error attributable to meas- urement. In the following table the volume of thecentrifuged plank- ton per m.-"^ and the deviations from the mean in volume and in percentages of the mean are given. The similarity iu the amounts of these successive catches is shown in the fact that the average departure from the mean catch is only ±3.58 per cent, and the total range of the limits of 274 departure only 14.1 per cent. I know of no test of similar ex- tent elsewhere with which this may be compared, but its range is well within the limits of the records of repeated catches in TEN CONSECUTIVE CATCHES FROM ANCHORED BOAT. No. 275 smaller number in waters of European lakes. For example, Apstein ('96) records 23 instances of hauls on the same date from equal depths and evidently in every case within distances between catches less than that represented in the extremes of our test. The number of hauls did not, however, in any of his tests exceed four. I have compiled or computed from Apstein's table (pp. 56-57) the average and limits of departure from the mean in these 23 cases. In 12 of the 23 the average de- parture exceeds ±3.58 per cent.—the average depai'ture in our test, in which there were from two and a half to five times the number of hauls. In 10 of the 11 instances in which the departure from the mean in Apstein's records falls below ±3.58, only two hauls were averaged. The total range of the limits of departure also exceeds that found in our test in 8 of the 23 cases. In the light of Apstein's results and considering the larger number of catches averaged in our test, and also the considera- ble length of the channel that it covers, it seems beyond rea- sonable doubt that single catches of the plankton in the channel of the Illinois at our station of collection afford as trustworthy a basis for the analysis of plankton problems as do similar catches made in a lake. The margin of error thus introduced is no greater, if indeed so great, as that appearing in investiga- tions in such waters. Since these catches were made from an anchored boat, the water from which the plankton was taken was distributed over a considerable length of the stream. The test was made between 7:30 and 9:30 a. m. The river stood at 5.1 ft. above low water and was falling rapidly, so that the current was noted at the time as unusually swift, probably approaching two miles an hour in mid-channel at this point. At this rate the collections represent plankton taken at ten intervals from a body of water about three miles in length. This areal distribution is compar- able with, if it does not exceed, the limits of widest distribu- tion of catches in Apstein's tests, but it is much less than that of Reighard's, which lay within an area of about ten by thirty miles. 276 A series of ten consecutive hauls made on the afternoon of August 21, 1896, from a floating boat between the bend in the river above the plankton station and the tow-head below it (PI. II.) throws some light on the questions of local distribution and of variation in catches from a limited area. Owing to the wind it was not possible to float with the current, and the apparatus also served to impede the boat. The river stood at 7.1 ft. above low water and was falling slowly, so that the current was not so strong as when the ten were made from the anchored boat. The test occupied about eighty minutes, and the boat drifted about a mile, so that the body of water actually passing it, from which the plankton was taken, was less than half a mile in length. Considerable dislodged vegetation and some cattle- yard debris were floating at the time, causing more than the usual inequality in the distribution of the silt which these elements introduce into the plankton. The catches ranged in centrifuged volume from A to .575 cm. 3, averag- ing .48, and showing an average of divergence of ±11.2 per cent, from the mean, with limits of +19.9 and —16.6—a total of 36.5 per cent. The divergence in this test is greater than that from the anchored boat, owing in part to the floating debris, and in part, probably, to the fact that the wind drifted the boat across fully three quarters of the channel. These divergences, both in average and limits, fall within the figures of parallel catches in lake waters quoted above from Apstein ('96) and computed from Reighard ('94). The fact that the range of variation on the whole is greater than the average run of Apstein's results is doubtless due in part to the larger number of catches included in my test. These two tests thus indicate that the plankton of the main channel waters of the Illinois at the point where our col- lections are made, is distributed quite as evenly as that in lakes thus far examined from this point of view, and in consequence single collections may be utilized for the study of plankton problems with no greater error for the potamoplankton than for the limnoplankton. The divergence from the mean will upon 277 the average, in all probability, fall within ±10 per cent. Our chronological series of collections affords a few instan- ces of catches under somewhat stable conditions of river levels and temperature, and at intervals so short that they may be utilized as tests of local distribution within certain larger limits of error, since the utilization of such data introduces the er- rors resulting from changes of chemical conditions due to rot- ting of sewage, and from growth, reproduction, and destruction of the plankton in the interim between collections. The follow- ing tabulated instances (p. 278) from Table III. and Plates X. and XI. may be cited as throwing light on this question of local distribution along the length of the stream. The fourteen groups of collections were selected wdth refer- ence to stability of conditions, therefore in falling or low w^ater and in periods of relatively even temperatures. Inspection of the tables and plates above referred to will show that the selection has not Ijeen made so as to eliminate wide varia- tions, and it may therefore be regarded as fairly typical. The periods included, range from 2 to 15 days in extent, and upon estimated rates of current the several tests include plank- tons taken at intervals in reaches of channel water from 24 to 252 miles in length. The average departures from the mean, range from ±0 to ±29.8, andyieldagrand average of ±14.1. In view of greater number of catches averaged and extended time element involved, these results compare very favorably with those derived from Reighard's data and Apstein's results. The probable error resulting from variations in the longitu- dinal distribution under stable conditions seems to be less than ±15 per cent. An inspection of Table III. and Plates X.-XIII. will show that in the case of invading flood waters the departures from the mean of catches at similar intervals would be consideralily greater than the averages above computed. Also, that in case of plankton pulses in stable conditions—for example in Sep- tember and October, 1897—collections at weekly intervals may exhibit departures in excess of ±50 per cent. It is evident, how- 278 LOCAL DISTR 279 ever, that in all tests extending over many days other factors than variation in local distribution come in to modify the re- sults. LONGITUDINAL DISTRIBUTION FROM THE MOUTH TO HENNEPIN. By com-tesy of the Illinois State Fish Commi.ssion 1 made on May 18-21, 1899. a trip on their steamer "Reindeer" from the mouth of the Illinois to Hennepin, about 205 miles from the mouth, making ichthyological collections for the State Survey. Incidentally plankton collections were also taken continuously from a short distance above the mouth to Henne- pin—in all. 21 collections. Of these. 19 will be utilized in the following comparison, the tir.st being omitted because of uncer- tainty as to the distance, and one other because of loss of the collection. The catch was made by means of a H in. iron pipe carried from the guards of the boat to a depth of 18 in. below the surface of the water. The intake was reduced to 'I in. and turned toward the prow of the vessel, so that, while moving, a continuous stream of water was dischai'ged into the plankton net. immersed in a barrel on deck. In this fashion a con- tinuous stream from the level of the intake was filtered. The contents of the net were removed approximately every ten miles of transit, and its clogging to the point of resistance pre- vented by shaking it down whenever necessary, thus minimiz- ing, in part at least, this source of error. The following table gives the data concerning these catches and the measurements and silt estimations,* together with my computations of the de- partures of the total catches from their mean and of the esti- mated planktons. The distances between points of collection were not deter- mined with great accuracy, since we had no log. and maps give no clue to the not infrequently toiiuous steamboat channels. The distances are therefore approximations based on the expe- I'ience of the pilot and engineer in charge of the boat. *By the generous permission of Professor Forbes, centrifuge measurements and silt estimations have been kindly furnished tome by .Mr. R. E. Richardson, who is preparing for publication in this Bulletin ,i detailed report upon these collections. 280 No. 4 5 6 7 8 9 10 H 12 13 14 IS 17 18 19 20 Time V-pEt Vol. in cm. 3 of catch Prnkl I To- ton I Silt tal Departure from mean in per cent. Total catch 1 m. above Hardin Kampsville Dam--- 1 m. above C. A. bridg-e Florence Mauvaise Terres Cr'k Meredosia La Grange Beardstown Browning. . Holmes Landing Havana.---- Liverpool 2 m. above Copperas Creek Mackinaw Creek 6 m. above Pekin 7 m. above Peoria Chillicothe Park 1 m. below Lacon Henry Hennepin 5:45-6:45 p. m. 6:45-7:45p. m. 4:30-=-:4Sa. m. 6:00-7:05 a. m. 7:10-8:10 a. m. 8:10-9:15 a. m. 10:00.11:00 a. m. 11:00-12:00 a. m. 2:45-3:45 p. m. 3:45-4:45 p. ra. 5:15-7:30 p. m. 4:20-5:20 a. m. 5:25-6:30 a. m. 7: 10-8:20 a. m. 8:40-9:45 a. m, ) 9:50-10:10 a.m. I 12-12:40 p. m. 12:50-1:50 p. m. 2:30-3:40 p. m. 4:00-4:55 p. ra. 5:20-6:30 p. m. .03 .04 .04 .04 .04 .05 .05 .05 .04 .04 .05 .06 .08 .12 .15 .20 .20 67.5 68 67 67 .18 .23 lost .02 .05 .11 .07 .22 .11 1.80 ,53 2.12 1.32 1.47 .95 2.19 .12 .12 .26 .91 2.32 2.36 2.15 2.21 1.61 2.15 .20 2.13 2.12 5.29 3.42 8.59 3.29 .16 .22 .90 2.20 .44 1.14 2.34 2.41 2.26 2.28 1.83 2.26 2.00 2.66 4.24 6.61 4.89 9.54 5.48 1.08 1.10 1.02 2.32 - 85 - 51 - 21 - 18 - 23 - 24 - 38 - 23 - 32 - 10 + 44 -fl25 + 66 +224 + 86 - 53 - 52 - 61 - 21 Plank- ton _"^75 - 68 - 97 - 93 - 85 - 90 - 59 - 85 +153 - 25 +199 + 86 +107 + 34 +209 + 30 + 27 - 83 - 83 Average I -71 2.23 I 2.94 ±57 I ±89 An inspection of the data of this table at first gives little comfort to one desiring to establish even an approximate uni- formity in the distribution of the plankton along the length of the stream. The average departure from the mean is ±57 per cent, in the case of the total catch and ±89 per cent, in the esti- mated plankton, with ranges respectively from —85 to -|- 224, a total of 309 per cent., and from —97 to +209, a total of 306 per cent. This is greatly in excess of the figures above given from the work of Reighard and of Apstein, and as a tvhole the data are so aberrant as apparently to disqualify them for scientific use. If, however, we take into consideration the conditions un- der which the collections were made, the aberrancy of this se- ries loses its force. In addition to the errors introduced by the slight clogging of the net and the uncertainty as to the precise distance, there is an error of undetermined proportions caused by the vertical movement of the planktonts and consequent possibility of uneven distribution at the 18-in. level between 4:30 a. m. and 7:45 p. m.—the extremes of our time of collection. Furthermore, an examination of the planktographs in the river and its backwaters for 1896 and 1898 (PI. X., XII., XXVII., XXIX.,XXXI., XXXIII.) —in which years the collec- 281 tions were sufficiently frequent to trace the movement in plank- ton production—shows that this season of the j'ear is wont to be a period of rapid change in plankton content. Thus, in the river in 1S96 on Maj- 13-18 the plankton fell from 3.56 to .86. or 76 per cent., in stable hj'drographic conditions. A similar phenome- non maj- be involved in the fluctuations in plankton content found in this transit of the river. The time intervening between the first and last collections was a little over two days. To this must be added the consideration that the collections represent a strip more than 200 miles in length, since we were traveling against the cuiTcnt. and, furthermore, that we have to deal with the volumetric changes in plankton content, as it passes down stream, due to growth and decay. Allot these influences are apparently but slight in comparison with the effect of certain environmental factors which are local- ly dominant within certain sections of the river. We can distin- guish OH flic (liii/s of collfction four sections or minor units of en- vironment dominated by different factors. The nrst three col- lections made in the lower river lie in a region of comparatively clear water free from flood invasion. Unfortunately the third collection was lost, but the remaining two exhibit a departure in the case of the estimated plankton of ±12 percent, and of ±44 per cent, in the total catches. The next six collections, covering a stretch of 60 miles, from Florence to Browning, were all taken in a section of the river invaded 1iy flood water of recent origin and poor in plankton, as was evident from the increased tur- bidity, the large amount of drift floating, and the discharge from tributary streams—principally on the right bank. In such conditions the amount of plankton (estimated) is small, and its variations form proportionately large percentages of its mean, the average departure being ±51 per cent., with a range from —79 to T^127—a total of 206 per cent. If, however, simply the total catch is taken, the average departm-e is±5 per cent., witha rangeof — ISto ^8—a total of 26 per cent. In view of the extent of the river included in this section—60 miles—and the uneven distribution of the flood contributions, it is not surprising that 282 we should find such irregularity in the (estimated) plankton. We now come to the section of the river dominated by the Peoria-Pekin pulse of sewage, including 70 miles of channel — from Holmes Landing to Peoria. The flood waters are still in evidence, but in reduced volume, and there is marked increase in the plankton content. The average departure from the mean plankton is ±32 percent., with arange of —64 to +48—a total of 112 per cent. In the case of the total catches the average de- parture from the mean is ±36 per cent., with a range of —60 and +89—a total of 149 per cent. The upper section of the river, above Peoria, a stretch of 40 miles, was less disturbed by flood conditions, there being only slight local invasions. This region is within the sphere of influence of Chicago sewage, and not receiving any large tributaries, we might expect but do not find conditions some- what equalized here. The average departure from the mean plankton is ±76, with a range of —76 to +80 per cent.—a total of 156 per cent. The average departure of the total catch is ±34 per cent., with arange of—27 to +66 per cent.—a total of 93 per cent. These departures will be much reduced if we break this section into an upper and lower region of two collections each, the percentages falling from ±34 to ±2 and ±0 for plankton, and to ±1 and ±39 per cent, for the total catch for the two sections, each of which represents 20 miles approximately. The average departures from the mean plankton in the four sections are respectively ±12, ±51, ±32, and ±76 per cent., yielding a grand average of —43 per cent. ; while the corre- sponding average departures for the total catches are ±44, ±5, ±86, and ±34, with a grand average of ±29.7 per cent. These four subordinate units of environment represent longitudinal extensions of 20, 60, 70, and 40 miles. The area included in Reighard's Lake St. Clair collections has a length of 32 miles and a maximum width of 5|, and the average departure from their mean (computed by similar methods for all localities) is ± 28.8 per cent. Similar methods of computation thus yield for Lake St. Clair and these sections of the Illinois River almost an identical ± error of distribution. 283 Tn the lioflit of these volumetric data the eonclusiou is pat- ent that plankton data from fluviatile environment contain on the average a distribution error which approximates that in plankton data from limnetic areas of similar extent selected with reference to uuitj- of environment as determined by local factors. It should be noted in this connection that the conditions prevailing when this plankton traverse of the Illinois River was made, were most adverse to an equalized plankton in the fol- lowing particulars. It was at a time of rapid seasonal change in plankton during the decline of the vernal pulse, and it was at a time of intercalation of flood water of local and recent ori- gin, whose poverty in plankton is brought into contrast with the larger content of the run-off of impounded backwaters else- where. Finally, the river stage, which was 9 feet at Kamps- ville and 6.9 at La (Trange. was such that the equalizing effect of general overflow on plankton content in impounded back- waters had ceased and local differences were emphasized, while at the same time their discharge continued in considerable volume. All of these factors, the last two of which are more impor- tant in the river than in the lake, tend to diversify the plankton content in the river at this season. It is reasonable to suppose that under other conditions—such as general overflow, the more stable features which attend falling levels above or below 9-7 feet, or in prolonged low water—we should And the uni- formity of distribution of the plankton more pronounced than it was on May 18-21, 1899. barring, however, the effect caused by sewage contamination, which at all stages and seasons is the most potent factor in the environment of the plankton of the Illinois River. TRANSVERSE DISTRIBUTION AND RELATION OF SHORE TO PLANKTON. The shore is a factor of great importance in the aquatic environment. It is here that land and water come into most intimate relation; seepage and drainage waters enter here; vegetation gains its foothold, affects the gaseous contents of 284 the water, and contributes by its decay to the nutrition of aquatic organisms; rise and fall of temperature are more pro- found here in shoal surface waters; light pervades more com- pletely; and currents are less rapid. It is in many respects a less stable region than the central waters which it bounds, and it may, indeed, be regarded as a separate unit of environment, in contrast with mid-lake or channel waters. The effect of the shore-line upon the distribution of the plankton in the lake has not entered into the data referred to in the previous section, for in the investigations of both Ap- stein ('96) and Reighard ('94) along-shore collections were not made, and, moreover, the shore-line is less important relatively in the lake as compared with the stream. For example, the absolute development of the shore-line in Lake St. Clair—de- termined by the method of Seligo ('90) (=shore-line divided by square root of area) is given by Reighard ('94) as 9.23. In the Illinois River at high water, from Utica to the mouth it is ap- proximately 17.1, and at low water 78.3, omitting all the con- necting lakes and bayous, computing the area on the basis of the average of the low-water widths given on page 110, and ig- noring sinuosities exceeding that of the channel. The relative development (absolute development divided by absolute devel- opment in a circle in which r-1) in Lake St. Clair is 2.607, in the Illinois River at highest water, 4.83, and at low water, 22.1. These figures serve to show in a general way the exceed- ing importance of the shore-line in the environment of the po- tamoplankton. Owing to the great sinuosities of the shore- line as rising waters invade the bottom-land, these figures are probably very much smaller than actual measui'ement would make them. It is probable that the relative shore develop- ment in the Illinois is ten times that of Lake St. Clair, and fif- teen times that of most lakes. Added to the diversifying action and predominance of the shore-line in the river, there is the tendency of its tributary waters, especially of the smaller lateral feeders, to follow their shore for some distance. The absence of great sinuosity in the 285 Illinois as compared with other streams, as shown hy the slight ratio of development of the stream (see p. 102 ), tends to prevent the rapid mingling of channel and marginal waters, and thus gives cumulative effect to their differential charac- ters. In order to trace the quantitative effect of the shore and determine the variation in transverse distribution, I made two series of ten collections each along a transverse line, the first at our usual plankton station and the second below the mouth of Spoon Kiver (see PI. II.). The results of the first-named test, made Augu,st 26, 189G, are given in the following table, to- gether with conditions of distance from shore, depth, tempera- tures, and turbidity. The river at this time stood at 6.5 feet above low water, and had a width at the station of 150 meters. PLANKTON IN CROSS-SECTION OF RIVER AT STATION E. c the increase in the initial collection at the east shore is caused in part by the greater abundance of Wolfia drifted there by the prevailing wind. It is obvious that for comparison with lake collections these shore catches should be excluded, for the former are rarely taken so near shore. Furthermore, all our chronological series on which this paper is based were taken in mid-channel, far from the shore belt, and in excluding those marginal collections but one sixth to one third of the total width of the stream is removed from the test. After all al- lowances are made, it is obvious that quantitative differences in the plankton are much greater in a single transverse trav- erse of the stream than they were found to be in a longitudi- nal test extending over approximately thirty times the width of the stream. Indeed, it is to be expected that differences arising from the effect of the shores and of tributary waters would be carried by the current far down the stream. On the basis of volume per m." the probable error of distribution is ±27.23, with a range of —61.7 to +58.1, and a total between limits of 114.89—all within these limits of variation in Reig- hard's data from Lake St. Clair, but exceeding somewhat the more limited data of Apstein. If we omit the three inshore collections, Nos. 1, 9, and 10, the probable error of distribution falls still lower,— to ±21.9 per cent., with a range of —22.1 to + 39.5, a total of 61.6 per cent, between limits. If we take the amount of plankton under one square me- ter as the basis of comparison the results will be much more di- vergent, owing to the greater relative difference in depth in my locations and to the introduction of variation due to verti- cal distribution of the plankton . In Apstein's tests the great- est departure from the mean depth in no case exceeds 10 per cent., and with but four exceptions his 81 tests are in water from 15 to 45 meters in depth, where differences in depth are of less importance than in shoaler water. In Reighard's se- ries the greatest departure from the mean depth is 66.9 per cent., the range being from 1.17 to 5.54 meters. In my test the range is from .56 to 4.88 m., the greatest departure being 81.5 per cent., 287 and my inshore collections were all probably very much nearer the shore than any of his were made. It is therefore legiti- mate to omit these inshore collections in comparisons based on amounts under one square meter. Accordingly, if we omit Nos. 1, 8, 9, and 10, the probable error of disti'ibution becomes ±15.4 per cent., with a range of —25.2 to +34.2, a total of 59.4 per cent, between limits. This is far within the limits of error which Reighard's St. Clair data yield. Since his catches include two at depths of 1.17 m., we may include all of my catches except No. 10, in which case the probable error of dis- tribution rises to ±38 per cent., with a range of — 77.1 to + 71.7, a total of 148.8 per cent., Reighard's data yielding on this same basis of computation ±31.8 per cent., —57.5 to + 111.5, and 169 per cent. The greater average ± error of distribution in my river test when these lateral collections are included is manifestly an expression of the effect of shore—an element not so pronounced in Reighard's tests. On this basis the limits and total range still remain less in the river test than in the lake. From the data of transverse distribution in the Illinois River it is apparently demonstrated that, on the whole, the dis- tribution is no more variable than it is in Lake St. Clair; and if we eliminate marginal collections and consider only channel waters, that is the middle two-thirds beyond 20 meters from shore, the variation falls considerably within the margin of er- ror found in the lake, being in the six centrally located col- lections ±15.4 per cent, on computations per square meter of sui-face, and ±24 per cent, for the same on the basis of plank- ton per cubic meter. The variation was also tested by counting the planktonts in the catch, with the resulting error in distribution of ±7.8 per cent. /b/' r/// /f>>/ ra^7/r.s, with limits of —14.7 and +31.8— a total of 4(5.5 per cent. The cross-section made below the mouth of Spoon River September 30, 1897, contains ten collections made at equal dis- tances, about 12 meters apart, and the first and last this same 288 distance from the east and west shores respectively. As will be seen in Plate XL, this was made after nine weeks of uninter- rupted low water, when the river had been standing at 2 ft. for some time. The catches were made between 2 and 4 o'clock p. m. There was no vegetation in the river at this point in this season, though both Havana Lake and Quiver Chute, to the north (PI. XL), contained a small amount. The discharge from Quiver Creek and Lake makes its way along the eastern margin of the river, while that of Spoon River un- der these hydrographic conditions hugs the western shore. The effect is seen in the turbidity records, the clearer water be- ing on the eastern side and the more turbid on the western. The following table gives the data of collection. There was almost no silt in the catch, and the silt estimates are therefore omitted. PLANKTON IN CROSS SECTION OF ILLINOIS RIVER BELOW MOUTH OF SPOON RIVER. Num- ber of col- lec- tion. Temperature (F.) Sur- face Bot- tom Depth in m. Tur- bidity deptb (in m.; disc visible Plankton per ra.3 in cm. 3 Plankton under 1 sq. meter Vol- ume Departure from mean in percent. Vol- ume Departure from mean in per cent. Departure from mean in per cent, (omitting Nos. 1 and 2) Per m.3 Under 1 sq. m. I 2 3 4 s 6 7 8 9 10 71.5 71.S 70.5 70.5 70.3 70.5 70 5 70.5 71 71.2 70 70 70 70 70 70 70 70 70.5 71 .66 1.06 1.42 1.58 1.58 1.68 1.83 1.88 1.83 i.n .35 289 The hydrographic conditions and the location of the test in the stream are responsible for a large percentage of this va- riation. Though the low-water levels cut off and reduce the diversifying action of impounding backwaters, the slight cur- rent minimizes the equalization due to mingling by the flow of the water in the channel, and, most of all, the location of the test just below the outlets of Quiver Lake and Spoon Riv- er (PL II.) involves the full effect of the diluent action of their relatively poorer .waters. In Spoon River, on the day of the test, 3.12 cm.* of plankton per m.^ of water was found (Table IV.), while in Quiver Lake on October 1 there was only .07 cm." per m.-' (Table V.). The discharge from Quiver Lake is reinforced by the seepage from springs along the eastern shore, and these diluents are probably the cause, to some extent, of the low plankton content in the two collections nearest the eastern shore—2.4 and 8.SS cm.'' to an average of 6. for the ten collections. The effect of Spoon River is seen in the much smaller decline in the inshore collection on that side of the river. Combined with the diluent action of these plankton- poor tributaries may also be the effect of shoal water and the horizontal stratitlcation of the plankton. If we eliminate from the test the two collections made in the marginal belt of spring-fed waters, 24 meters wide along the eastern shore, the ± departures from the mean fall from ±22.3 and ±38.8 to ±12.1 and ±20.2. These latter figures more truthfully represent the variation in distribution of plankton in channel waters including four Mfths of the width of the stream—a lateral extension far beyond the range in that direction of the mid-channel collections of our chronological series which form the basis of the conclusions of the present paper. The data concerning the local distribution of the plankton in the Illinois River in longitudinal and transverse directions presented in the preceding pages may be summarized as fol- lows: The average ± departure from the mean longitudinal distribution in consecutive catches at the same point in the 290 stream is 3.58 per cent; from a floating boat, 11.2 per cent.; at intervals of 1—7 days for periods of 2 to 5 days in the more stable hydrographic conditions, 14.1 per cent.; and in the stream as a whole for 200 miles of its course, 57 (total catch) or 89 per cent, (plankton estimated). If, however, we break up the 200 miles into four sections representing sub- ordinate units of environment, each dominated by some local factor, the ± departures from the mean are 12, 51, 32, and 76 percent, respectively for estimated plankton (i. e. after silt deduction), or 44, 5, 36, and 34 per cent, for the total catches, the averages for the two methods being ± 43 and ± 29.7 per cent. The average departure from the mean catch in two trans- verse series of 10 catches each is ± 27.2 or ± 22.3 on the basis of plankton content per m.'' If we eliminate the shallow-wa- ter shore collections, the departures fall to ± 21.9 and ±12.1, or on the basis of volumes under 1 sq. m., to ± 15.4 and ± 20.2. The departure from the mean number of planktonts is only ± 7.8 for the whole cross-section. These results are in the main within the ± error of distri- bution of the plankton in lakes arrived at by similar methods of computation. The plankton method may tJierefore he applied to the quant dative investigation of the life of a stream as legitimately as to that of a lake. The laivs of the horizontal distribution of the plankton are in this respect essentially the same in both types of aquatic en oironment. Whether or not a fundamental source of error as large as this—pi'obably the greatest of all the errors in the method as we have used it—vitiates the utilization of such data for scien- tific conclusions must be to some extent a matter of opinion. The extent to which it renders conclusions tentative must de- pend upon the distribution of the error, the extent of the data, and the method of their utilization. Personally I may say that close study of the at first sight aberrant data upon which this paper is founded, has led me to attach less significance to this source of error than I was at first inclined to do. Readers 291 of the paper will, I believe, find that in the main the conclu- sions arrived at rest on a body of confirmatory data so large as to counterbalance to some extent the probability of vitiating error from this source. The distribution of the error is, more- over, continuous throughout the whole series of data, with, however, some probability of valuation with the stability of the hydrographic conditions. Finally, the conclusions to be drawn in subsequent pages rest upon data which to a large extent rise above the level of the error resulting from the irregularity of distribution. PLANKTON PRODUCTION. 1894 (Table III., PL VIII.) Ten collections were made by the oblique-haul method in this year between June 12 and December 15. The v^olumes of plankton, silt, and total catch per cubic meter average 2.49, .28, and 2.77 cm.' respectively. The maximum catch, 10.18 cm.' per m.' (plankton, 9.67; silt, .51) was taken Aug. 15, and the min- imum, .25 cm.' (plankton, .10 ; silt, .15), on Nov. 11. The series of ten catches form a somewhat regular curve, rising during July and August, and declining, most rapidly in September, to a minimum in October-December. A comparison of the record of 1894 (PI. VIII.) with that of other years (PI. IX.-XII.), as shown in the accompanying table of avei'ages (p. 292), and with the conditions of temperature and hydrograph, will serve to throw light on the significance of the plankton volumes of this first year of our collections. As shown on pages 168 and 164, this wasa year of normally located high and low water, with March, May, and iSeptember rises all so reduced as almost to eliminate overilow stages and to prolong low-water stages, resulting in the low average height of 4.63 ft. above low water. Our collections all fall in the sta- ble period, broken only by the September rise. They therefore afford no data on the spring maximum of plankton production, 292 00 Z o < Pm (•< O O o & Q o ». >- .J X H Z O t a e B a 298 revealing only a single midsummer pulse, culminating in the August maximum in a period of maximum heat and lowest water. In the light of collections of later years in this and other localities it seems probable that collections at more frequent intervals vpould have yielded a curve of greater irregularity, vfith other fluctuations than the single one apparent in the present record. It seems probable from the records of 1896 (PL X.) and 1898 (PL XII.) that the small average (0.74) in June is due to the fact that the dates of collection fall in a period of decline from an April-May pulse, hastened by the rise in May and pex'- haps reduced in volume by the relatively small contributions of impounded backwaters resulting from the depression of the spring flood. It may also be that the collection of June 29 ex- hibits the flushing, depleting effect of the rise of the preceding ten days. It will be noted that the collection of June 12 lies about four weeks after the crest of the May rise—a location which is attended in 1896, 1897, and 1898 (PL X.—XII.) by a decline to a minimum after a pulse of plankton development. The hydrographic conditions of July in 1894—decline of flood to low-water levels—are approximately realized with va- rying stages of river and rates of decline in all the other years but 1895 (PL IX.-XIl.). In 1894 they attend a tenfold in- crease in the plankton during this month. The movement of production is in the same direction approximately in July in 1896, 1897, and 1898, though its development is less in 1896 and 1898, and data are lacking for its progress in 1897. In 1894, and to a varying extent in other years, this rise attends among other factors the restriction of contributions from impounding backwaters and the differentiation of what might be called channel plankton proper. The July production in 1894 aver- ages 5.12 cm.' per m.'—the largest, with the exception of that for 1895, of any year, and a fact to be correlated with the un- usually stable conditions then prevalent. In August of this year the single collection forms the apex of the season's production, reaching 9.67 cm." per m.'—an 294 amount not surpassed for this month in any suh.sequent year. It may also be correlated with the continuance of^stable con- ditions. The nearest approach to this amount is found in 1897 (9.45 cm.', Table III.), likewise in stable conditions. It is not im- possible that there is more than one culmination in the months of June and July, collections being at too great intervals to suggest the direction of the movement in production. The flood of September attends a decline of the plankton to a minimum of .34 on the 17th in the fluctuations in level on the crest of the flood (PI. VIII. ). Similardirection of movement in production may be traced in 1895, 1896, and 1898. This decline in production attends the beginning of the autumnal de- cline in temperature, 10°-15° of which occur within this month. The hydrographic conditions during the remainder of the year are exceedingly stable, there being a gradual rise of only .5 ft. from the middle of October until the middle of December. Beyond the insignificant rise in the October catch no movement in production is evident. A comparison of these scanty data with the curves of production in these months in 1895 and 1897, both with low-water autumns, makes it evident that collections in 1894 were too infrequent to serve as a basis for any conclu- sions as to the average autumnal production in this year, and raises the query as to whether considerable fluctuations of pulse-like character might not have run their course in the in- tervals between our collections. The higher averages in Octo- ber-December in other years supports this suggestion. It is evident that the monthly interval of plankton collection is too infrequent to afford usable or significant data. The average of the ten collections in 1894 is 2.49 cm.', and that of the seven monthly averages 2.53 cm.' This is larger than the averages for a similar period in 1896 and 1898 (.99 and 1.09 respectively), both years of disturbed autumnal hydrograph. It is much less than that of the last seven months of 1895 (7.15). If, however, the exceptionally large collections of June-July be omitted in this year, its average (of monthly means) falls to 2.05. In the main, the hydrographic conditions in 1894 and 295 1895 in the last five montlis are somewhat alike, and their plankton production is somewhat similar (cf. PI. VIII. and X.). In 1897, however, the uninterrupted and prolonged low water yields a much larger production of plankton (3.56 for the last seven months). Though incomplete, the evidence in a general way indi- cates that 1894, in the period included in the collections, was a year of abundant plankton production, approximating 2.5 cm.' per m.' of water. 1895. (Table III.; PI. IX., XLllI., LI.) Of the 50 collections of this year but 4 were made in the first six months. This was particularly unfortunate, for the spring was one of exceptionally low water, and the collections are so in- frequent as to give only the faintest clue to the curve of plank- ton production in this important period. All of the collections were made by the oblique-haul or repeated vertical-haul method. Omitting the very unusual catch of June 19, the mean volumes of plankton, silt, and total catch per cubic meter are respective- ly 2.12, 1.88, and 4.01 cm." As an average, the proportion of silt in the catches is thus quite low—a fact explained by the absence of considerable floods during the period of most frequent col- lections. As is shown on pages 164 and 165, this was a year of unusu- ally low water, the mean annual stage of the river being 3.61 ft. The spring rise did not bring the river to much more than min- imum bank height, and there was no June rise. Aside from a few minor meteoric rises to less than 6 ft. in July and Septem- ber the low-water period was unbroken until the December flood culminating at 12.6 ft. at the close of the year A glance at Plate IX. will indicate that the collections suffice to trace the production during the last six months,—a low-water period with minor rises,—and to follow somewhat closely the effect of these hydrographic changes upon the volume of the plankton. The isolated collection of Feb. 23, made beneath 37 cm. of ice at the close of a period (PL IX.) of ice blockade of approx- 296 imately two months' duration, reveals an almost comijlete ex- termination of the plankton, the amount given in the table,. 01 cm.', being only an expression for an amount beyond the reach of our methods of measurement. As shovrn in Plates VIII. and IX., there was prolonged and quite stable low water from Oct. 15, 1894, till the flood of the last week of February of the following year which carried away the ice. The concentration of sewage under such conditions was shown by the stench of the water, by the departure of fish into tributary backwaters, and by the death of many not escaping. Unfortunately no chemical analyses of river water at this season are now availa- ble, and the chemical conditions can only be inferred from those in later years at times of briefer ice blockade, higher water, and presumably less contamination. For example, in January, 1898, following the low water of 1897, we And under ice of three weeks' duration (PL XL, XIL, and XLIV., XLV.) great excess of free ammonia and chlorine, and high albuminoid ammonia, organic nitrogen, and oxygen consumed—all, in- dices of contamination. The ice sheet upon a contaminated stream must also profoundly affect the equilibrium of oxygen and carbon dioxide dissolved in the water, and thus directly influence the life of all constituents of the plankton. It is there- fore not surprising that these unusual conditions should ex- terminate all but the most resistant members of the plankton. The catch consisted almost entirely of flocculent debris (zo- ogloese?) with a few minute filaments of bluish green alga of uncertain affinities, while the usual plankton was represented by only 43 individuals, representing 14 + species, as follows. Protozoa: Diffugia sp. (deformed?) 1 Epistylis sp., heads 3 Carchesium lachmanni, head 1 Ciliate, indeterminate 1 Heliozoan 1 297 Rot ifera: Brachionus dorcas, female 1 Rotifer tardus, female 9 Pnli/((rflira plufi/pfira, female 2 I'olijartlira pliifjiptcra, female with 1 egg 1 Indeterminate rotifer sp 1 Eiitoniostraca: Cyclops bicitspidatus, female, young, dead or moribund 1 Cyclops, young 2 Cyclops, nauplii 4 Canthocamptiis, nauplii 4 Cliydorus ijlohosiis 1 Miscellaneous: Rhabdoca?l 1 Dcro racjfi 2 Indeterminate 7 Total 43 In towings made at the time of the quantitative collec- tions Mr. Hempel found an individual each of Pterodiiia patina and Ndtholca aciiniitiafa. The list includes representatives of the prominent winter planktonts excluding alga? and diatoms. The effect of the sewage contamination is observed in the reduced numbers both of individuals and species, in the moribund condition of Lyifflnyia, Carcliesiioii, Epistylis, Bracliiomis, and Cyclops, and in the fact that apparently the only breeding forms, with the exception of the Cyclops, were the ubiquitous and pei-enuial Folyarflira and the muck-loving Cnntltocably the decline) is detected in catches of June 19 and July 6. Of the occurrence of this latter and larger pulse there is little doubt, though the data are not available for its location and delineation. The unusual and prolonged low water of these spring months thus seems to result in a marked increase in the plank- ton content. The causes which lead to this are not far to seek. 300 The decreased volume of water causes a relative concentration of the sewage and consequent increase in fertility of the chan- nel water over that of the usual high water of this season. Lower levels insure more rapid rise in temperature, and the slackened current affords more time for the breeding of the plankton. The occurrence of Molna micrura, a lover of foul water, is in itself an index of the character of the stream in this low-water spring. The contributions of the impounded backwaters to the stream during this April-June period (see PI. IX.), owing to the small areas submerged, are reduced in volume so that both the relative and actual share which they have in the formation of the channel plankton is probably less than in years of normal spring flood. Nevertheless, as seen in Plate XXXVI., such waters as Thompson's Lake tend by their run-off to enrich and increase the channel plankton. The month of July (PL IX.) witnesses the rapid decline of the second vernal pulse from 29.68 cm.^ on the 6th to 6.8 on the 23d and .33 on the 29th—a fall of 98 per cent, in 23 days. The last stages in this decline were hastened by the rise of 3 ft..in the third week of July, the flushing and destructive action of the flood waters continuing until the close of the month. In this and subsequent years I shall call attention—when- ever the interval between collections is brief enough to afford adequate data—to the phenomenon of recurrent pulses of plankton production. I am led to make this emphasis by ob- serving in the numerical analysis of these catches recur- rent pulses in most if not all of the more abundant species, pulses, moreover, which exhibit a degree of concurrence in many species which I believe to be expressed in the faintly traceable volumetric pulses which run like waves, erratic in amplitude but more regular in interval, through the seeming vagaries of the volumetric data. I shall therefore treat the volumetric data from this point of view, endeavoring to discov- er evidence of cyclic production wherever it exists, and seek- ing to correlate this phenomenon with the more patent fac- tors of the environment. 301 Considering, then, the data from Jul}' to the end of the year in 1895 (PL IX. ), we find that the month closes at a min- imum of .33 cm.^—the end of a pulse of uncertain limits and the beginning of the next, which culminates in the third week of August. This August pulse is followed by one of less ampli- tude and duration, culminating about three weeks later , by one of slight amplitude in October, culminating at an interval of about four weeks , by one of greater amplitude in November, after an interval of about five weeks, and by one in December, al- so of considerable amplitude, at an intei'val of about four weeks. The fluctuations of some of the component groups of organ- isms are shown in Plate LI., and considerable correspondence in the volumetric and statistical pulses will be apparent on com- parison. The AwjHst pidfic has a duration of 39 days,—from July 29 to Sept. 9,—and a maximum amplitude of 7.63 cm.' on August 24. The mean of the pulse,* that is the line upon which the center of gravity of the polygon formed by connecting the ordi- uants lies, falls upon August 22. This pulse occurs in a period of somewhat stable low water, and its decline fi-om the maxi- mum of 7.63 reaches 2.07 on Sept. 4 and occurs without the assistance of flood waters. On Sept. 5 a sudden minor flood, due to local rains, flushes the stream and completes the deple- tion of the plankton to .69 on the 7th. The August average for 1895, 4.03 cm.\ is higher than that of any other year excepting 1894, in which but a single collection was taken, which may not be representative of the whole month. Freedom from ris- ing flood waters in 1895 is doubtless one cause conducing to this high average of production. The enriching effect of the minor flood which culminated in the closing days of July may also contribute to this end. The absence of rises in May and June would also tend to increase the contributions of organic material to the stream by this July flood as compared with *The mean was computed by multiplying the volume of each catch in the pulse (ordinant) by the number of days from preceding minimum to date of collection (abscissa) and dividing the sums of the products by the sum of the catches in the pulse, the quotient being the abscissa of the mean. 302 floods which followed normal sjaring rises. The correspond- ence of the August pulse of plankton with a heat wave of 10° amplitude is well shown in Plate IX. Similar correspondences may be detected in some instances elsewhere in the plankton and temperature curves, but neither the completeness of our temperature data nor the corroborative evidence is sufficient to lend much support to a causal nexus between the phenomena. The Septeniher pttlse has a duration of 25 days,—from Sept. 7 to Oct. 2,—and a maximum amplitude of 3.25 cm.' on the 14th. Its mean falls on the 17th, 26 days after that of the August pulse. This is a month of considerable hydrographic disturb- ance, the rises of the 6th, 17th, and 27th causing almost twice as much movement (8.75 ft.) in river levels (see Table I.) as is found in other years of our operations. These accessions of flood water in each instance attend a fall in temperature of 5° to 8°, though that on the 27th is combined with normal autum- nal decline. None of the three is sufficient to cause overflow; and each is of but few days' duration. Their effect upon plank- ton production is, however, considerable. In the first place, the immediate result of the invasion of flood water is an in- stant decline in the plankton, as shown by the change from 2.07 on the 4th to .69 on the 7th, the flood in this case acceler- ating and perhaps continuing the normal decline of the August pulse. So also the little rise of the 17th checks the rising curve of production, the fall being from 3.25 cm.'' on the 14th to .89 cm.' on the 18th. The rise of the 27th evidently occurs towards the minimum of a declining plankton pulse, and the fall from 1.03 cm.' on the 25th to .37 cm.' on Oct. 2 is of less extent. The location of these floods in the pulse is such that if my conjec- tures as to their reducing effect be true they cause a shifting of the apex of the curve and of the location of the mean to the left of their probable position had not the floods occurred. In technical phraseology the mode of the curve of this pulse ex- hibits left-handed skewness. In the second place, the general effect of these recurrent rises is a reduction in total production the extent of which can only be conjectured. It seems proba- 303 ble that the rise of the 17th is responsible for the suppression of a rising pulse whose cuhnination had not yet been reached. The slight recovery in the following week is indicative of the upward tendency in production thus interrupted. That con- tinued low water in this month may attend great plankton production is seen in the records of 1897, when the monthly average (see table on page 292) is 8.(S3 cm.'' to 1.52 cm.' in 1S95. On the other hand, in 1896 and 1898 the disturbed conditions, with higher water and more current, are accompanied by much reduced production, averaging only .38 and .69 cm.'' The last week in September witnesses the hrst stages of marked decline in temperature from the well-sustained summer heat of 75°-85°. The decline reaches 68° at the end of the month. This phenomenon combined with the last flood to ac- celerate and complete the decline of the Septeml)er pulse which had already appeared prior to the last flood. The Ortoher pulse has a duration of 29 days,—from the 2d to the 30th,—and a maximum amplitude of .76 cm.' on the 11th and 15th, following a rise in nitrates and attending in- creased sewage contamination (Pl.XLIIL). Its mean falls on the 18th, 31 days after that of the preceding pulse. This is a month of stable low water approaching minimum levels, the total movement in the pulse period at Havana being only 1.03 ft. The temperature in this period falls from 61° to 45°, and this taken in connection with the fall of 11.5° in the preceding week bi'ings to bear upon the plankton production of this month the cumulative effect of a decline of 27.5° and the results of the low temperature of 45°. The consequence is that the summer planktonts are killed off or reduced in numbers, and the winter planktonts have not as yet had time or temperatui-e to reach any considerable development. The plankton produc- tion is therefore low ; so low, indeed, that its pulse-like char- acter is largely a matter of conjecture in the volumetric data (cf. statistical data on PI. LI.). Phenomena of like character are to be detected at corresponding periods of autumnal decline in temperature in September-October, 1896 ; in October-Novem- 304 ber, 1897 ; and in October, 1898. The effect of this autumnal decline of temperature may also be traced iu monthly averages of production in the table on page 292. Rapid decline in tem- perature is thus immediately followed by rapid decline in pro- duction in the channel plankton. Such correlation in decline of temperature and plankton cannot, however, be found as a general phenomenon in the bottom-land lakes (cf. PL XXX.- XLII.) and a causal nexus between the two declines must there- fore be of limited operation and at the best highly conjectural. The operation of other factors than that of direct temperature is probable. The monthly mean of production for October in 1895 is .57 cm.^ approaching that of 1894 (.61) and 1898 (.24). In 1896 and 1897 it is much higher (1.11 and 5.95 respectively), an earlier decline of temperature in 1896 (PI. X.) and a later one in 1897 seeming to shift accordingly the attending decline in plankton, so that the September (.38) and November (1.) aver- ages respectively more nearly represent the October averages of 1895. The November pulse has a duration of 35 days,—from Oct. 30 to Dec. 4,—with a maximum amplitude of 4.37 cm.' on Nov. 27. Its mean falls on the 22d, 36 days after that of the pre- ceding pulse. This is also a month of continued stable low water, with a slight rise of .75 ft., due to the checking of evap- oration and to autumnal rains. Thetotal movement is only .99 ft. at Havana. Temperatures during the first three weeks are somewhat stable for this season of the year, exhibiting a range of only 5.3°—from 48.5° to 43.2°. The last week, however, ex- hibits a fall of 10.2°, to minimum winter tempei'atures and the beginning of the ice blockade. Under these stable conditions the plankton production in November rises to a level approach- ing that of midsummer of the current year, its apex (4.37) fall- ing 44 per cent, short of the August apex (7.63), and its aver- age (3.02) 25 per cent, short of the August average (4.03). Both volumetric (PI. IX.) and statistical data (PI. LI.) demon- strate the rapid multiplication of the plankton in these stable 305 condiiions, and the result is a pulse of considerable amplitude, moreover, one not attained in any other year in channel plank- ton ; a fact whose significance is apparent when we find (PL VIII.—XII.) that no other year combines to the same extent stability of hydrographic and thermal factors. The relative production in different years (see table on p. 292) bears upon the point in question, the monthly mean (3.02) being from 150 to 3 times as great as that reported for other years. The chem- ical conditions attending this remarkable plankton production (PL XLIII.) are those following increased sewage contamina- tion, namely, a rise in nitrates, free ammonia, and chlorine. The Derciiiher piil.'^ie has a duration of 21 days,—from De- cember 4 to 25,—the limits being taken from the statistical data (PL LI.), which are based only on catches at intervals of 5 to 7 days. Its maximum amplitude (2.60) occurs on the 20th, and its mean falls on the 16th, 24 days after that of the preceding pulse. The first 18 days of the month are relatively stable, with a movement in levels of only 1.05 ft. and stable minimum temperatures under the ice sheet. During this period a slight- ly developed pulse begins its course (cf. also PL LI.), but its apex does not rise much above the level of previous produc- tion. It is noteworthy that this takes place beneath the ice sheet which covered the upper river during the fortnight pre- ceding the flood. It is in this month that the contamination noted in Novemlier reaches its maximum (PL XLIII.), at least as shown by nitrates and free ammonia. The chemical condi- tions thus favor a continuance of the productive activity of the previous month. On the 19th heavy general rains started a flood of unusual magnitude wdiich continued to rise, culmina- ting at 12.6 ft. at the end of the month. This raised the temper- ature about 9°, brought in an immense load of silt, flushed out the plankton, and increased the rate of the current so as to greatly reduce the time for breeding. The first two days of rising water did not materially change the quantity of plank- ton per m.^, indicating a considerable rise in production had not the flood occurred. By the 25th, however, the flood waters had 306 swept away all but a vestige of the rich plankton of the earlier weeks. The amount remaining was so small that its quantita- tive changes were swamped in the errors of the volumetric method and silt estimation. The large amount of silt carried in this and subsequent floods of the winter is due to the fact that bottom-lands and fields covered with a rich vegetation were now submerged for the first time in two years ( PI. VII.), and vast quantities of debris from this region and tributary streams now entered channel waters. In comparison with other years December in 1895 is, in spite of its fortnight of flood, the most productive December recorded (see table on p. 292), averaging 1.14 cm' to .76 in 1896, .56 in 1897, and .99 in 1898. It shares the large development of the preceding month, and with it presents the most marked late autumnal development in channel waters, though falling far below the production of some of the permanent backwaters in this and other years in this season. The unusually stable hydrographic conditions in the river doubtless contribute in large measure to this exceptional development. That low tem- peratures alone do not prevent the development of a large win- ter plankton is apparent from this December development of 2.6 cm.' per m.' and 11.1 cm.' per square meter at temperatures but little above 32°. The year 1895 as a whole may be summed up as one of mid- winter stagnation followed by excessive spring and early sum- mer development of the channel plankton, of midsummer and equinoctial floods, which check development at that season, of stable autumnal conditions and exceptional production in late autumn, and of catastrophic reduction by flood to a minimum. As a whole the year was one of exceptionally heavy production when expressed in terms of plankton per cubic meter. This is seen in the high average—3.22 cm.' of all catches, 5.31 cm.' of monthly averages. When total production is considered it may be that the decreased volume of water at the time of the maxi- mum in the low water of June will at least counterbalance the excess per cubic meter, and that the total production will not 307 exceed, if indeed equal, that of years of more normal hydro- eraph. 1896. (Tables III., X. ; Pi. X., LI.) There were 76 collections made in this year, of which 69 are prior to Sept. 1 and are. moreover, at intervals brief enough to enable us to trace the curve of plankton production with some degree of accuracy. In the last four months the fort- nightly interval is too great to permit more than conjecture as to the probable course of the plankton curve. The collections prior to May 22 were all the result of com- bining 4 to 9 repeated vertical hauls of the net. Subsequent to that date they were made with the plankton pump. This, as is shown on page 165. was a year marked by recurrent floods, which bring the average height for the year up to 6.9S ft. in a year of less than average rainfall. This is almost twice the average height [S.6\ ) of the preceding year. Since the flood did not in most cases reach bank height, the overflows were not extensive and did not occur during periods of large plank- ton production (PI. X.). The distribution of the collections with reference to the floods is such that we have again in this year the opportunity to test the effect of the access of flood wa- ter upon the curve of plankton production at all seasons of the year but the autumn months. In this year the vernal rise in temperature occurred abruptly in the middle of April, and the autumn decline began quite early but progressed slowly. Summer temperatures were also lowered somewhat by access of flood water. The plaitktoit of JaiiKanj, Fcbnaifij, (iiul Mfurli (PI. X.) forms so small a portion of the total catches that its quantita- tive changes are swamped by the probable error of silt estima- tion, and are apparently of such slight extent that their signifi- cance cannot be detected. The amount of silt carried is very large, doubling ox trebling in quantity on rising floods, and reaching a maximum of 14.77 cm.' per m.' on the crest of the March freshet. No recurrent pulses appear in the volumetric 308 data, though the statistical data (PI. LI.) indicate the recur- rence of three such pulses in this period. The volumetric production is very small throughout this whole period, rising above an estimated amount of .01 cm.' per m.'' in only 7 of the 18 catches, and not exceeding .13 in any of them. This results in monthly averages of .01, .02, and .07 re- spectively for the three months ( see table on p. 292). These are lower than those of any other years excepting only that afforded by the single collection of February, 1895. The cause of this slight production is, I believe, the high water and increased current resulting therefrom, which does not affoi'd to the channel plank- ton the time requisite for breeding a more abundant plankton. Some corroboration of this view may be found in the fact that the February collections in the high water of 1897 (PI. XL) like- wise yielded minute quantities of plankton (average .04 cm. ^ per m.^), while the channel waters of 1898 (PI. XIL) in January and the early part of February, and of January-March, 1899 (PI. XIIL), produced at stages below that of overflow (8 ft.) a more abundant plankton—.07 to 1.15 cm.' per m.^ of water. High water with accompanying rapid current is thus deleteri- ous to plankton production in channel water's, in midwinter at least. It is noteworthy that this minimum production occurs in the presence of nitrates in great excess, in fact in quantities larger than those recorded at any other period of our records. (cf. PI. XLIII.-XLV.). It is not therefore for lack of nitrates and other products of decay that the plankton fails to develop. The data of the collections in the latter part, of March in- dicate a rising production as levels fall and temperature rises. The direction of movement is upward, though the quantity at- tained in this month is not great. The interval of collections and the quantities of plankton obtained from March to September enable us to trace with some probability the course of the recurrent plankton pulses of this season. The April pulse has a duration of 32 days,—from March 30 to May 1,—with a maximum amplitude of 9.39 cm.' on the 24th. 309 Its mean falls on the 23d. This is the vernal pnlse. often the lar- gest of the year, this distinction being attained in 1896. It rises in 25 days from a minimum of .13 cm.' on March 30 to a max- imum of 9.39 on the 24th—an average daily increase of .37 cm.' This pulse, as elsewhere, follows immediately npon the vernal rise in temperature, which in this spring reaches 72° on the day of the maximum of the pulse and passes from 46° to 66° a week prior to it. The maximum thus lies a fortnight after the mo.st rapid vernal rise in temperature begins. It attends a sharp decline in nitrates and free ammonia, and its maximum coin- cides (PI. XLIII.) with that in the organic nitrogen. It also occurs in a period of apparently stable hydrographic condi-. tions, the total movement in April in this year being only 1.4 ft. less than in any other year. This stability is more appar- ent than real. The decline of the March flood (PL X.) was checked, and slight rises resulted from spring rains which brought large quantities of silt into the stream, so that move- ment in river levels is not in this instance a sufficient index of hydrographic stability. The result was apparently the sup- pression to some extent of the vernal pul.>^e (cf. on this point 1896 and 1898, PI. X. and Xll. ). The amplitude of this vernal pulse (9.39) is less than that of 1898 (35.68 ^l—the only other year in which our collections are frequent enough to lo- cate and delineate this pulse with sufficient accuracy. This may be due to the operation of one or more of the following factors. First, to the spring rains above referred to, at the time of the apparent maximum of the pulse, which flush it out and dilute it, and to some extent destroy the plankton. In the second place, there was no general overflow at this season, and plankton bred in the less current-swept, impounded backwaters is not entering the channel to the usual extent at this period of the year (cf. PI. X. and XII.). Again, the periods of standstills in levels and those of rise check the outflow from impounding areas or turn channel water into the bottom-lands, conditions which obtained in 19 of the 30 days in April. Lastly, there is some possibility that 310 the days of maximum production were not touched in our col- lections. The meteoric character of the vernal pulse of 1898 in channel waters (PI. XII.) is indicative of such a possibility. If a greater production than that recorded did occur, it prob- ably fell between the 17th and 24th—a period of non-interfer- ence by flood and of rising production. The location of the apparent maximum in this year is sig- nificant. This was an early spring, the average of the surface temperatures in April in 1896 (see p. 171) being from 4° to° 11° higher than that in any other April represented in our records. The temperature of 70° degrees is attained almost a full month earlier in 1896 than in 1898 (cf. PL X. and XII. ). The maximum production was recorded in 1898 on May 3; in 1896, on April 24, nine days earlier, and it may have antedated even this. Early spring thus affects the life in water much as it does that upon land. Vegetation bursts into leaf and insects mul- tiply in field and forest in proportion to vernal rise in tempera- ture; so in lakes and streams, in like response, the algae multiply with meteoric rapidity, and the animal planktonts dependent on them follow in their wake. In 1896 the early vernal rise in temperature deflects the maximum of the vernal pulse to an earlier date by virtue of this response on the part of aquatic life to the environing factor. The average production in April in 1896 (5.67) exceeds that in any other year of our records, in large part, it seems, because of the early spring and the deflection into that month of the maximum production, which in other years passed un- detected or fell in the following month in consequence of la- ter vernal rise in temperature, as in 1898. The May pulse has a duration of 31 days,—from the 1st to June 1,—with a maximum amplitude of 3.56 cm.^ on the 13th. Its mean falls on the 15th, 22 days after that of the preceding pulse. There is in this month considerable hydrographic dis- turbance—a total movement of 7.5 ft., consisting of a fall of 3.1 ft. followed by a rise of 4.4. The maximum production oc- c urs during the decline in the earlier weeks, which is practi- 311 cally the run-oif of the April rains which checked the fall of the March tlood (PI. X.). This is also a period of rising tem- perature, a rise of 12° (to 82°) attending the decline of river levels and the rising plankton production. The rising plank- ton pulse is. however, flushed out by the entrance of flood waters in the closing fortnight of the month. The plankton falls at once from 3.56 cm.^ on the 13th to .86 on the 18th with the first stages of the flood, and the fluctuations during the period of rise are erratic, suggestions of recovery and decline appearing in the data. These vagaries may be due to the dis- tribution of local storms, which contributed largely to this somewhat slow rise in river levels. The general effect of the flood seems to be to depress the production and thus to deflect the apex or node and the mean of the curve of production to the left, that is. to an earlier date. The flushing effect of the floods of May, 1896, is apparently greater than that in 1S98. as shown by the plankton production. The flood of 1896 did not exceed bank height. Its diluent action is thus concentrated in channel waters. In 1898 the floods occur in overflow stages and are thus diffused over a large area. The chemical conditions show but little relation to plank- ton movement in this month. The maximum production fol- lows immediately upon a rise in nitrates, nitrites, and free am- monia, and coincides with a slight decline in the two first named. The decline in production during the rising flood takes place along with consideraljle increase in nitrates and nitrites. The average production in May, 1896 (1.30 cm.'), is less than that of the following years (.see table on p. 292), since it does not contain the vernal maximum, and also because it is reduced by flood action. The June pulse is not well differentiated in the volumetric data, and its delimitation here becomes largely a matter of conjecture though it stands out more clearly in the statistic- al results (PI. LI.). If we follow the latter the pulse termi- nates, at least so far as the chlorophyll-bearing organisms are 312 concerned, in the last week of June. If, on the other hand, we delimit the pulse here as heretofore by minimum volumes, we shall find its later limit to be July 6, giving it a total duration from June 1, of 35 days. Its greatest amplitude, 1.68 cm.^ oc- curs on the 11th, and its mean on the 14th—29 days after that of the preceding pulse. With the exception of the first three days this was a month of continuously falling river levels. The large proportion of silt in the catches and the fluctuations in the temperatures in the first ten days of the month suggest flood water of recent origin. Nevertheless, the maximum pro- duction of the pulse appears at the close of this disturbed peri- od, a slight decline with little subsequent fluctuation in pro- duction marking the remainder of the pulse. The average production in June, 1896 (.72 cm.^) is low in comparison with that of 1898 (3.96), the only other year in which the June production is sufliciently represented in our records. In both of these years there was rapid and prolonged decline from previous flood, but in 1896 the proportion of con- tributions from impounded backwaters was much less than in 1898. Greater time for breeding plankton is thus afforded as a whole in 1898, and greater production follows. The maxi- mum production coincides with the maximum of nitrates (PL XLITL), though it attends a depression in nitrites and free ammonia, The generallow production of this month occurs in the presence of an unusual quantity of nitrates, so that one at least of the important elements for production was not lacking. The Jiihi pulse has a duration of 19 days,—from the 6th to the 25th,—with a maximum amplitude of 2.24 cm", on the 20th. Its mean also falls on the 20th—36 days after that of the pre- ceding pulse. This is a month of considerable hydrographic disturbance, the total movement in levels being 7.7 ft.—a fall from 5.2 to 2.5 followed by an interrupted rise to 7.3. The pulse lies in the middle of this period and falls under the in- fluence of both fall and rise. During the period of decline .the recovery of the plankton from its minimum of .26 on July 6 313 progresses irregularly to a slight niaxinmm of 2.24 ou the 20th during a very rapid rise caused by a Spoon River flood (PI. 11.) which, while not invading the stream above its mouth to any great extent, held back the water from the upper river. The greater part of the rise in the latter part of the mouth was due to the access of water below the plankton station or in re- mote headwaters, and is thus a reflection of the rise in the lower river or distant tributaries. The freedom from silt apparent in the catches bears testimony to this fact. The pulse reaches its culmination and declines in this rising flood. The average production for July in 1896 (1.44 cm.") is less than that in any other year save 1898. and that, too, in what seem to be favorable hydrographic conditions. The sharp de- cline in nitrates (PI. XLIII.) from 2.8 to .4 parts per million may be a factor in the small production. The August puhc has a duration of 2? days,—from July 25 to August 21,— with a maximum amplitude of 3.90 cm.' ou July 30. Its mean falls on the 5th, 16 days after that of the preceding pulse. This was predominantly a month of falling levels. The culmination of the rise at 8.6 in the tirst .six days is followed by a steady decline reaching 5.8 on the 31st, broken only by the slight interruption in the middle of the month. The total movement is 3.7 ft., and the total at Copperas Creek (4.60) is somewhat above the average (4.06). The di.sa,strous effect of the local floods at the culmination of the rise which flushed out the rising plankton pulse is apparent in the de- cline from 3.90 cm.'' on July 3 ft., and a depth below low-water mark in the Illinois of 10 14 ft. It runs between almost vertical banks of alluvium (PI. XIV.), and has a hard gravelly bottom full of sunken logs which form treacherous snags at low water. The current at the point of collection at low water may be scarcely perceptible, while at times of sudden flood, due to local .storms in its water-shed, it is so strong that a boat enters it with difficulty. At such times its load of silt and drift is very great. During the heated term, and especially when the heat pulses occur and there is little wind to ruffle the surface, the green water-bloom on this stream is remarkable, exceeding—possibly because of protection from wind—that of the main stream in lividness and density. The turbidity of this stream (see Table IV. and p. 179) is gi-eater than in any other locality and serves as a general index of its silt content. It is. for example, in 1S9S, 31 cm. (_ average of disc readings), while in the Illinois River in that year the average is 40 cm. COLLECTIONS. All collections were made with the plankton pump. Ex- amination of the plankton of this stream was begun in August, 1896. and continued at a fortnightly interval until December of that year, and thereafter until the close of operations in March, 1899, at approximately a monthly interval. From the charac- ter of the curves of plankton production in the Illinois River we may infer that collections at this long interval in Spoon River will fail to give us any adequate or accurate delin- eation of the movement in production in this stream. Further- more, in the summer .season at least, the plankton of Spoon River is composed largely of those small planktonts—such, for exam- ple, as EiKjIena and TrarJielomotias—which almost wholh' escape through the meshes of the silk net. A comparison of the plank- ton of the two streams on volumetric data derived from the 342 catches of the silk net is to some extent misleading, owing to the relatively greater proportion which the escaping planktonts form of the production in the tributary stream. Another factor which prevents an equally accurate volumetric determination of the plankton of the two streams is the presence in Spoon River of a much greater proportion of silt. For example, in 1896 and 1898 the estimated ratios of silt and plankton in the average of the catches (Table IV.) is .007 to .349 and .029 to .796. In 1897, when low water and slight current and some probable invasion of channel waters increased the plankton production, the ratios are 1.257 to 1.173. The ratios of the first-named years are more nearly normal for this tributary, and in such ratios it is quite probable that the error in silt estimation to some undetermined degree tends to prevent any precise determination of the actual plankton production. Nevertheless, after a very wide margin is allowed for probable error in the data, the comparison of production in the two streams is instructive and significant, for it is the direction of change or contrast in production which is of greatest consequence, and this may be found even in the presence of a large but distributed error. It should be noted that the plankton ordinants in the Spoon River plates (XXII.—XXIV.) are plotted on a scale ten times that of all other stations in order to give an appreciable height to the plankton portion of the entry. PLANKTON PRODUCTION. 1896. (Table IV.; PI. XXII., XLVI.) For purposes of comparison I introduce at this point a table which gives in terms of monthly averages of plankton in cm.-' per m.' the relative production in the seven locali- ties examined by us. The number of collections entering into each average is stated, and the grand average of all collec- tions and of the monthly averages are given for each station. In 1896, nine collections were made in Spoon River in August-December, the average being only .007 cm.'' per m." COMPARATIVE EXHIBIT OF PLANKTON PRODUCTION IN ILLINOIS RIVKR ITS TRIBUTARV WATERS AND BACKWATERS, l894-l85<), Mo. 343 (Table IV.), with a maximum of .032 ou August 25. The amounts reported are all very small, aud the significance of their differences is questionable. The following correlations with environmental conditions may be noted. There is little plankton (.004) in the turbid (3 cm.) flood water of August 18 ; there is more (.032) in the clearer water (30 cm.) of August 25. The production following the rapid decline of temperature in September falls to a minimum (.002) as it does in the channel waters (.53 and .23), and like the latter rises again (.008) late in October, after a month of somewhat stalile temperatures (PL XXII.). The December production (.002 and .001), how- ever, shows no rise corresponding to that in the main stream. The average production in Spoon Eiver for the five months in 1896 in which collections were made there is only .007 cm. per m.', while that in the main stream is 97 times as great, the production there exceeding that in the tributary from 4-fold to 380-fold in each month (see table between pp. 342 and 343). Spoon River water is thus throughout this season a diluent of the channel plankton. The chemical conditions during this period reveal unutil- ized nitrates averaging 1.2 parts per million in Spoon River to 1.15 in the Illinois. Other forms of nitrogen are somewhat more abundant in the main stream. There is, however, plenty of food for the plankton in the tributary, and other causes than poverty of nutrition must be cited to explain its paucity of plankton. 1897. (Tables IV., XI. ; PI. XXVIII., XLVI.) There are 13 collections in this year, at intervals of two to six weeks. They average 1.257 cm." per m.\ and have a maximum of 7.296 on Septeml)er 11. The conditions attending the unusual plankton production in Illinois River channel waters in this year affect Spoon River also in much the same manner. The vernal overflow mingled impounded backwaters 344 to some extent with the channel water of Spoon River, and the prolonged drouth of the autumn cut down the run-off and re- duced the stream to a series of slack-water pools, in which, owing to the reduction in current, there was time enough for an abundant plankton to develop. The explanation of the contrast between the plankton con- tent of this stream on February 3 (.002) and 26 (.092) is to be found in the hydrographic conditions. The tributary shares the rise in production seen in channel waters (.03 to .05). The rising flood of the 26th forces the impounded backwaters away from the channel, and in their downward movement some of them get into Spoon River channel in the overflowed bottom- lands above the point at which our collection was made. Thompson's Lake waters contained considerable plankton (.39) at this season, and it seems probable that some of its richer waters may have entered and (PL II.) enriched Spoon River channel plankton at this stage of river. Under such circum- stances we find the tributary with a richer plankton (.092) than the storm-filled channel (.05)—an exceptional occurrence in the history of the two streams. The very slight production (.007) on March 22 is due to the fact that Spoon River itself at this time was rising rapidly, turbid (2 cm.) with silt, and invading rather than receiving contributions from the impounded back- waters through which it rushed to the channel. The collection of April 27 was also in flood waters (turbidity 5 cm., silt 4.75 cm.''), which are in part responsible for the check in the flood decline at that time (PL XXIII. ). This held back contribu- tions from connecting and impounded backwaters, and the .plankton content is low (.048), while that in the main stream (5.11) shows no such flood reduction. In the collection of May 25 (.44) we find the tributary waters as well as the main channel exhibiting a vernal rise in production, though its amplitude is 13-fold greater in the latter. The lower river level (8 ft.) then prevalent precludes the possibility of any considerable contri- butions from impounding areas, though accessions in small vol- ume are not improbable. On June 28 the silt-laden storm-water 345 ill the Illinois Eiver channel ( tiirbidit j'. 2 cm., silt, 26.33 cm.') contains liut .27 cm.' per iii.' of plankton, but the vernal pro- duction in the tributary does not suffer so marked a decline, remaining at .25 cm.'', so that its diluent action on this occasion was slight. As in the main stream, so also in the tributary there is a drop in production in August (.056) to about a fourth that in June-July. From this time throughout the remainder of the year the production in Spoon River is considerable, exceeding that in the main stream, however, only in the last two months of the year. The production rises on Aug. 26 to 1.248 and on Sept. 11 to the unprecedented record of 7.296 cm.^ while that in the main stream is only 2.77 on Aug. 24, and on Sept. 7 and 14 is 8.47 and 19.80 respectively. On the 21st and 29th of September production in the Illinois falls again to 8.00 and 4.(14 cm." re- spectively, while in Spoon River on the 30th it is only 2.96, production in the trilmtary thus remaining below that in the channel throughout this period. The low chlorine in Spoon River at this time (3 to 4 parts per million. See Table XL) as compared with the main stream l_21-50 parts) indicates that the Spoon River water is not contaminated by channel water, and that we are dealing with an indigenous plankton. None of our collections falls in the period in October (PI. XLVI.) in which chlorine in Spoon River rises temporarily. The three collections of November-December. 1.351, 1.99. and .599 cm.', are respectively 22.5, 1.6, and 20 times as great as the production in the main stream at the same time. The month- ly averages for this period are 1. and .56 for the Illinois, and 1.671 and .599 cm.' for Spoon River, so that the excess in the latter is apparently not more than 50 per cent. lender these conditions the tributary stream enriches the plankton of the channel instead of diluting it, but its discharge is slight. Hydrographic, thermal, and ice conditions are similar in the two streams, and we find the main difference in the chem- ical conditions. Aside from evidences of sewage contamina- tion in the last weeks of October, the Spoon River records 346 (PI. XLVL, Table XL) show somewhat uniform conditions in strong contrast with the instability in these particulars in the Illinois. With greater stability Spoon River apparently pro- duces a more abundant plankton in this low-water autumn. The exceptional production in this autumn stands in strong contrast with the poverty of this tributary in the same months in 1896 and 1898. As seen in the table between pp. 342 and 343, the maximum monthly production in the low water of 1897, as recorded in monthly averages, is from 285 to to 5,130 times greater than that in any month in this season in the other two years. A comparison of the data in Table XI. for the three years in question and of their plottings on Plates XLVL and XLVII. will show the great similarity in the chem- ical conditions which accompanies this remarkable inequality in production. The accompanying table summarizes the data concerning nitrogenous contents of the water and the plank- ton. NITROGENOUS SUBSTANCES AND PLANKTON, 1896-1898. MONTHLY AVERAGES— PARTS PER MILLION. Year 1896 , 1897, 347 relation to environmental factors that was found in the case of the Illinois. This is seen in the increased winter produc- tion, in the vernal rise, in the decline after the vernal pulse, and in the unusual autumnal development. The tributarj' stream, with but four exceptions, was acting as a diluent of channel plankton at each examination of its plankton content. These four exceptions—on Feb. 26, Nov. 2 and 30, and Dec. 28 — are due in the first instance to channel flood, and in the last three cases to exceptionallj' low water in the tributarj- and less stable chemical conditions in the channel. In the four years in which Spoon River was examined they are the only exceptions to the general rule that these tributary waters are dilu- ents of the channel plankton. The average production for 1897 (1.257 cm.') is 180 times that recorded in the last half of 1896, and 43 times that for 1898—as a result of the low-water con- ditions discussed above. 1898-1899. (Tables IV., XI. ; PI. XXIV., XL\-II.) There are 14 collections at intervals of four or five weeks in the 15 months included in this period, and they fairly rep- re.sent the contributions of this tributary in a year of consider- able flood and repeated access of storm water. In 1898 there is but a trace of plankton in the January ( .017 ) and February (.016) collections, whilethat in the March collection (.124)isthe maximum for the year. At this time the spring flood is nearly at its height (16.5 ft.), and the waters of Spoon Kiver are in quite free connection with the general overflow that spreads over the surrounding bottom-lands. On the day of the Spoon River collection there was .43 cm.'' of plankton in the Illinois and .79 the week prior in Thompson's Lake, three miles above Spoon River (PI. II.). There is thus three and a half times as much plankton in the main stream and six times as much in Thomp- son's Lake. With its maximum burden of plankton, the trib- utary is still a diluent, and its plankton content at this time is probably in large part derived from the run-off of the contigu- 348 ous impounding backwaters. No Spoon River collection falls in the week of the vernal maximum (see PI. XII. ), but the col- lections of May (.023), June (.096), and July (.036), all exhibit a con.siderable rise above the usual level of production, and all, moreover, were made during the run-off of the spring flood and receive slight contributions from impounded waters. It is in the dircriion of the movement in production that the tributary and main stream are alike at this season. In the nwpJitiide of the curve of production the diiference is very great, production be- ing respectively 491-, 41-, and 16-fold greater in the latter in the three months named. Throughout the remainder of the year 1898 plankton pro- duction in Spoon River is at a minimum, there being but the merest trace of living organisms in the catch. None of these catches was taken in rising flood water (PI. XXIV.), though they all show the results of the flushing action of the frequent floods which wash out with rapid current whatever plankton may have developed in the tributary, and at the same time af- ford little opportunity for its replenishment. The relative ab- sence of backwater feeders in the tributary stream at this stage of river levels serves also to emphasize the poverty in production of the tributary. The average for 1898 (.029 cm.' per m.') is exceeded over 73-fold by the yearly average of the Illinois (2.13). The tribu- tary waters are at all times—at least in so far as the data go — diluents of the channel plankton, reaching their lowest ratio, .124 to .38, in March, when they share most in impounded back- waters of the main stream, and at the same time are at the max- imum of their own reservoir action. This meager production occurs in waters almost as rich in nitrates (av., .67 parts per million) as the main stream (.809), and, save on rising floods (Pl.XLVIL), in normal chemical con- ditions. The potent environmental factor is ratherto be found in the recent origin of the tributary waters than in any availa- ble chemical data. In the three months' collections of 1899 we find that the low 349 level of production coutiiiues under the ice, which remains on the stream for about three months (PL XXIV. ). There is con- siderable fluctuation (PI. XLVII. ) in the organic nitrogen, free ammonia, and oxygen consumed, most if not all of which are traceable to the access of storm waters rather than to any con- siderable degree of stagnation. The catches are all full of silt, though the turliidity of the stream is not great under the ice in January and Feliruary. The silt at such times is mainly com- minuted vegetation brought in by the storm waters. There is a slight rise in the plankton production in March (^.026), when the river stands at 12.9 ft. and the plankton-rich waters of Thompson's Lake (see Table V.) are brought into connection slightly with Spoon River by overflow. During these three months the production in the tributary is but a small fraction—never more than a tenth—of that in the main stream. It continues to be a diluent of the channel plankton. SUMMARY. The average plankton in all of the Spoon River collections is .465 cm.^ per m.' of water. In the Illinois it is 2.19. or over 4.7 times as much. If we omit the low-water period, Aug. 26 to the end of 1S97. and compare only the remaining collections between Aug. IS, 1S96, and the close of operations, the ratio of production in the two streams becomes .044 to 2.19, or I to 50. As has been i-epeatedly pointed out in the preceding discussion, this contrast in production is not explainable on any difference in available chemical data. The tributary waters are fertile enough to yield a large production. The explanation is rather to be sought in the hydrographic conditions, in the recent ori- gin, from rains or springs, of the trilnitary water, and in the more rapid current, and consequently the less time for breeding a plankton in the tributary environment. That this is the proba- ble explanation is borne out by the large production in the only period of prolonged low water in the tributary in the fall of 1897, when time for the growth of the plankton was afforded in the slack waters of the tributarv. 350 The immediate effect of the access of the tributary waters of the stream is as a whole diluent upon the plankton content of the Illinois. A mixture of equal volumes of each would re- sult in a reduction in the Illinois to 1.33 cm.' per m.' from 2.19 —a falling off of 39 per cent., or even of 49 per cent, if we omit the low-water period of 1897. If we consider the areas of the drainage basins as an index of the relative volumes of water carried by the two streams, and determine the effect of Spoon River contributions, we find the net result, based on the aver- ages of our collections, to be a decline in plankton content in the Illinois from 2.19 cm.' per m.' to 2.00—a decline of 9 per cent. If the low-water period of 1897 is omitted, the decline is even greater, namely, from 2.19 to 1.96—a fall of 11 per cent, in the plankton content as an immediate result of the contri- butions of this tributai'y. Spoon River thus exerts in the econ- omy of the Illinois an immediate diluent function upon its pro- duction, which, qualitatively, is approximately 10 \)er cent. STATION C, QUIVER LAKE. (Tables v., XI 1. ; PI. II., IV., XV.-XVII., XXV.-XXIX., XLVIII., XLIX.) ENVIRONMENTAL CONDITIONS. This lake lies on the right bank of the Illinois (PI. II.), ex- tending parallel to the river for a distance of three miles, in- cluding Quiver Chute. This is the lower end of the lake, which is separated from the Illinois River only by a low mud bank submerged at levels of 4 ft. and crossed just below our plankton station in the main stream by two "cut-offs" which bring a va- rying volume of river water into the chute, the amount de- pending upon the relative levels on the two sides of the spit. The close connection of the lake and river makes the former responsive to all changes in level in the latter at all stages of water. This lower spit extends northward as a low bank 5 to 40 rods in width and generally less than 6 to 8 feet above low- water mark, lying between the lake and river. This is covered with low willows and, to the northward, with heavy forest (PI. 351 IV., XV., XVr.). The northern end of the lake is Y-shaped, and the western arm is known locally as Doghsh Lake. The lake from Quiver Creek to its mouth near Spoon River is about 3 miles long, and does not exceed a quarter of a mile in width at any point. The lake proper, that is, above the chutes and excluding the western arm, contains at low water about 230 acres (93 hectares), but approaches 500 acres if these contiguous bodies of water be included. As levels rise, its area increases rapidly, and at 6 to S ft. the demarcation between river and lake is obliterated, and extensive areas to the north- ward (PI. II.) come into connection with it, while at higher levels it quite loses its identity as a sepai'ate area (PI. III.) ex- cept as the tree tops and its cleai'er waters serve to differentiate it from contiguous channel waters. Its depth at low water (river levels about 2 ft. above low water of 1873) is throughout most of its area less than 2 ft., and in the deepest parts, at the narrows above the chutes, it does not exceed 4 ft. The bottom is of hard sand and bluish clay covered gener- ally by a soft alluvial ooze of 1 to 2 or more feet in depth. Its eastern bank is a sloping sandy bluff {V\. XV.), which abounds in clear springs of cold water, occurring the whole length of the lake and contributing not a little to its water supply. The western bank is of black alluvium, and the ooze along its mar- gin of considerable depth. The eastern arm of the lake receives Quiver Creek, a tributary draining 220 sq. miles of sandy upland and "second liottom." The vegetation, described on page 244. in low-water condi- tions frequently chokes the channel, which extends from the mouth of Quiver Creek in a tortuous course through the vege- tation along the western shore of the eastern arm towards the point between this and Doghsh Lake, and thence in an equally crooked and shifting course towards the mouth. It was in this channel in low water, and in its neighbor- hood at times of high water, that our plankton station was lo- cated (PI. II). It is simply a shifting path through the vegeta- tion, and is not generally marked by deeper waters than adja- 352 cent regions. It was only a few meters in width, and in 1894 and 1895 it was frequently necessary to clear it of encroaching vegetation in order to make feasible the 30-meter oblique haul. In high water and generally in years subsequent to 1895 it was only necessary to avoid with drawn net or pump the clumps of Ceratopht/Iluin which still dotted the bottom in this neighbor- hood. The point of collection thus lay at all seasons towards the narrowing end of the lake and in the path of the current maintained by Quiver Creek and the marginal springs. At times of high water it was in the direct path of the downward current of impounded backwaters thrust towards the channel by the encroaching eastern bluff (PI. II.). At times of flood the invading river waters extend for some distance, even to the middle of the lake, crowding the clearer lake waters to the eastern side. When the western bank was not submerged the backwaters entering by way of the chutes sometimes reached the plankton station. Our collections were always made in evident lake water unless otherwise stated (Table v.). The access of creek and spring water, the extensive areas of dense vegetation, and the shoal waters, which at all levels form the greater part of the area of this lake, all combine to make the temperature conditions subject to great local varia- tion, and to diversify the fauna and flora indigenous or adven- titious in the plankton of this body of water. COLLECTIONS. Our chronological series in this lake includes 115 collections, extending from June 6, 1894, to March 28, 1899. The collec- tions in the several years number 14, 13, 31, 24, 26, and 7 re- spectively. Their distribution by months is shown in the table between pages 342 and 343. In the earlier years the interval of collection was somewhat irregular, though with 6 exceptions every month is represented. From July, 1897, collections were made at least every fortnight and on the same day as at the other stations. The oblique-haul method was used—with a few 353 exceptions of repeated vertical hauls (TalileV. ) in the winter tiood ot lsy5-96—from the beginning till May •22, 1896. After this date all collections were made with the pump. This lake is a t5'pe of some other bottom-laud waters, spring- fed and rich in vegetation, and our collections suffice to show the relation which these bear to production in the adjacent channel waters. They also serve for compai'ison of production with that in other liackwaters less rich in vegetation, and since the quantity of vegetation in Quiver Lake varied from year to year they also throw some light on the effect of vegetation up- on plankton production in a single area. PLANKTON PRODUCTION. 1894. (Table V., V\. XXV.) The 14 collections in this year average 1.08 cm.'' per m.' to 2.49 cm.' in the Illinois. The maximum (3.50) falls on Septem- ber 6. There is a striking resemblance in the planktograph of Quiver Lake for this year [F\. XXV.) and that of the adjacent channel (PI. VIII). The amplitude is generally less in the form- er, but the direction of movement is the same in both. The June production (monthly average) is low in both Quiver Lake (.23) and the Illinois (.74): it rises in July (2.20 and 5.12); and it declines in September (2.12 and 1.36) toaminimum of .80 and .34, from which it recovers slightly in October (.95 and .61) to fall again in November (.02 and .10) and December (.03 and .10). The only exception to this parallelism in the movement in production is seen in August, when in Quiver Lake produc- tion drops to .74 but attains a seasonal maximum of 9.67 in the river. Aside from the fact that this is the season of greatest predominance of vegetation in the lake, owing both to growth and to low river levels, there seems to lie nothing in the en- vironmental conditions to be correlated with this contrast. While as a whole for this year the contributions of Quiver Lake (1.08 cm.^ per m.''), as shown by our data, only result in 354 an immediate dilution of the channel plankton of the Illinois, there is a season when its plankton content exceeds that of the river. The average of the collections in September-Novem- ber in Quiver Lake exceed by 70 per cent, the average of those in the Illinois. This is the season when some autumnal decay of vegetation takes place, and this vegetation-rich lake has a larger plankton prodtiction than the river waters which it thus enriches. 1895. (Table V. ; PI. XXVI., XLVIII.) There are 13 collections in this year, with an average of .78 cm.' per m.' as compared with 3.22 in the Illinois. The maxi- mum of 4.57 occurs on April 29, being but 1.26, or 22 per cent, less than the corresponding vernal maximum in the adjacent, but—owing to river levels in this season—non-contiguous, river. The similarity in the movement of production between this lake and the river noted in the previous year can be traced in 1895 in but two instances,—in the rise to the vernal maximum and in the increased production in December (cf. PI. XXVI. and IX.). Outside of these periods there is no resemblance between the planktographs of the two waters. From July to November inclusive the low level of production is broken only by two pulses, both of which attend a rise in river levels with increase in the impounding function of the lake. These changes in level shift the loosely attached vegetation, and are often followed by death and decay of masses of aquatic growths. The slight rise in the last week in August (PI. XXVI.) caused an invasion of muddy river water into the lower end of the lake. Decay of the vegetation and death of many fish, clams, and other ani- mals ensued in the invaded area. The flood early in Septem- ber (PI. XXVI.) came largely from up-river rainfall, and the lake waters, enriched by invasion, were impounded with result- ing increase in the plankton. It was not apparent that either of the large collections were made in invading waters, and I infer that the plankton was indigenous and not adventitious. 355 though the invasion resulted in the enrichment of the lake by the decay of vegetation and dead animals. It may also have "seeded" the lake with organisms whose subsequent multipli- cation caused these temporary increases in production. These same floods are attended by depi-essions in production in the main stream, so that these two pulses in Quiver Lake lie in these depressions, intercalated between summits of the curve of production in channel water (cf. PI. IX. and XXVI.). The inference is suggested that the run-off of this plankton-breeding impounded water of Quiver Lake and similar reservoirs else- where may have contributed to the increased production in channel waters following the flood. The plankton content of Quiver Lake water on July 26 (.71) and Sept. 6 (1.57)thus exceeds that in theriver on July23 and Sept, 6 (.68 and .99), and its contributions to the stream, if any were made, serve to enrich the channel plankton. In three other cases the lake production exceeds that of the river; on Feb. 23 (lake, .03, river, .01), April 9 (1.42 and .52), and Dec. 28 (.29 and .01). In the first instance there was stagnation under the long continued ice-sheet in both river and lake, as was .shown by the great mortality of fish in the latter. The plankton, how- ever, did not reach the degree of extermination in lake water that it did in the channel, since there was less sewage, more veg- etation, and access of spring water. In the April instance the silt burden of the channel waters (4.67 cm.-', PI. IX.) is much greater than that in the lake (1.43), and suggests the intercalation of storm water in the former, resulting in the slight rise in levels (PI. IX.) and the lessened plankton content of the channel waters as compared with those of the less dis- turbed lake. The great contrast on December 28 is also due to the flushing action of the great winter flood which depleted the channel plankton but increased the impounding function, and therefore breeding capacity and productivity, of the lake. Each of the three instances of greater production in lake than in river waters occurs with rising river levels, when the rising river checks the relative outflow from the lake or otherwise 356 increases its impounding function. Whatever run-off from lake to channel occurs under such conditions will result in a slight enrichment of the plankton content of the channel waters with which the tributary mingles. At all other seasons of this year our collections indicate that the immediate result of the access of Quiver Lake waters to the river is a reduction in plankton content of the main stream, on an average for the year for equal volumes of tributary and channel waters, of 38 per cent. ; or if the relative volumes of each based on areas of drainage basins are considered, the plankton content of the channel is reduced to 3.19 cm.^ per m.'*—a decline of about 1 per cent. This was a year of maximum development of vegetation in Quiver Lake. The low water of this and the preceding year and the absence of floods adequate to flush the lake of its loosely attached vegetation permitted an unusual and enor- mous growth of Ceratojjhi/llion and other aquatic plants, which choked the lake from shore to shore and from Quiver Creek far down the chute towards its mouth (PL XV.). The very slight plankton production in its waters during the summer is due, I believe, to this predominance of vegetation. The rise in pro- duction when river levels rose in July and September (PI. XXVI.) attends, among other factors favorable to production already discussed, a reduction in the relative abundance of vegetation. 1896. (Table v., PI. XXVII.) There are 31 collections in this year, with an average of 2.59 cm.' per m.' as compared with 1.16 in the Illinois. The maximum of 16.76 cm.'' occurs on April 24, exceeding by 7.37 cm.', or 78 per cent., the production in the adjacent river on that day. The similarity in the movement of plankton production in Quiver Lake and the Illinois noted as generally present in 1894 and but slightly so in 1895 is quite apparent throughout this 357 year, as will be seen ou a comparison of Plates X. and XXVII. With a few exceptions which will be noted in the subsequent discussion, the trend of the production is similar in the two bodies of water to a most striking degree month bj' month throughout the year. With rising temperatures in Feln-uary-^Iarch, production in the lake attains the unusual level of 1.75-1.85 cm.^ per m.\ an amount not equaled at this season in this lake in any other year, and exceeding by 88- and 26-fold the production in the adjacent flood-swept stream (see table l)etween pp. 342 and 34:3). This greater production in the lake is due, it seems, to the fact that Quiver Lake collections in these months represent the im- pounded Ijackwaters of the eastern bottom-lands forced through the lake by the configuration of the eastern l^luff (PL II.). Slight curi-ent and time for breeding permit in them a production not possible in the silt-laden rapidly flowing channel waters with which at our plankton station (PI. II.) they are contiguous during prevalent levels. The larger production in this year may be attributed to the enrichment of the water by the great mass of organic debris accumulated on the now submerged bot- tom-lands during the two preceding years of low water. The verual pulses in the two waters coincide in the posi- tion of their limits and maxima though not in amplitude at any time, as will be seen on comparing Tallies III. and V. and Plates X. and XXVII. The vernal development in the lake pro- ceeds more rapidly, appears earlier, and attains a greater am- plitude than it does in the river. Thus, on April 10 and 17 there is present in Quiver Lake 3.29 and 16.32 cm.^ of plankton per m.^ to 1.68 and 4.45 in the river. The rate of increase is 4.7 times as rapid in the lake, and attains on the 17th a volume 3.7 times as great as that in the river. The maximum (16.76) is almost twice that in the river (9.39). The large development (16.32) on the 17th indicates that the true maximum probably occurred a few days earlier in the lake than in the river. A partial ex- planation of this phenomenon, and also of the earlier and more rapid rise in production, may be found in the somewhat higher 358 temperatures in the shoaler and clearer impounded waters which are drained off through Quiver Lake. The temperatures of surface waters in the lake from February up to the time of the maximum are from 1° to 15° higher than in the river, as will be seen on comparison of the thermographs on Plates X. and XXVII. The May pulse in Quiver Lake attains 8.14 cm.^—more than twice the amplitude of that in the river, 3.56,—while the aver- age production for the month in the lake (2.99) exhibits a sim- ilar ratio to that of the river (1.30). The very sudden decline from 8.14 on the 8th to .51 on the 16th attends a decline of about 2 ft. in river levels at a stage which cuts off the lake from large impounding areas to the north, and also, at this season of the year, brings the submerged flora to the surface. These two factors combine in effecting this sudden drop in production in the lake before it appears in the stream (cf. PI. X. and XXVII.). The flood which wipes out the rising June pulse in the river (PI. X.) increases the impounding area and relative occupation of the lake water by vegetation and permits a pulse of some amplitude (2.60) to develop in the lake, while only a belated and slight development appears in the contiguous river. As levels fall in July and impounding areas are again cut off and vegetation anew occupies a relatively larger proportion of the lake, production declines to so slight an amplitude that a July pulse can hardly be traced (PI. XXVII.), and the average monthly production in the lake falls to a fifth of that in the stream, whose plankton content it had in previous months of the year exceeded. With the rise of the August flood, production again assumes a pulse-like character, lagging throughout its development a few days behind that in the adjacent stream (cf. PL X. and XXVII.), and lacking in the lake the cleft in the apex of the curve caused in the river production by the flushing action of local floods. The seven collections during the remainder of the year ex- hibit a similar direction of movement in production in every 359 instance but two, Oct. 14 and 29. In the first of these, silt-laden flood waters in the river, but not in the lake, interrupt the par- allelism. In the second instance the production of the lake de- clines and that of the river rises—again as a result of the prior flood conditions, as will appear on a comparison of the sequelse of the June and August floods in the two bodies of water. In these, as also in the October flood, there are indications that the rising plankton pulse common to both is temporarilj' suppressed in the river and continues undisturbed and reaches an earlier culmination in the lake, but only a delayed one of slight ampli- tude in the stream. The average production in the lake in the last four months of the year exceeds that in the river by 52 per cent, and in five of the seven collections. The comparison of production iu Quiver Lake and the Illi- nois River in 189G is very instructive in several important par- ticulars. In the first place, both the relative and absolute pro- ductivity of the lake has increased, rising from 1.08 and .78 cm.' in 1894 and 1895 to 2.59, an increase of two- to three-fold. The ratio of productivity in the lake to that in the river in 1894 was 1 to 2.3; in 1895, 1 to 4.1 ; while in 1896 it falls to 1 to .45. The low average in the river is, as has been shown, the result of the repeated flushing l>y storm waters. The increase in the lake is due to the higher levels and increased impounding function, and to the actual and relative decrease in its vegetation. The combined result of the operation of these factors is that in this year the lake waters cease to be diluents of the channel plank- ton and become sources of enrichment. Considering the areas of their respective drainage basins, and basing calculations on the yearly averages, the net result of the contributions of Quiv- er Lake is a rise in the plankton content of channel waters from 1.16 cm.' per m.^ to 1.18—an increase of a little less than 2 per cent. Not only was the average production in the lake (2.59) greater than that in the stream (1.16), but individual collec- tions upon coincident or approximate dates exhibit the same 360 relation in 22 out of 31 instances, and 4 of the 9 exceptions fall in the period of low water in July, during predominance of veg- etation in the lake. The monthly averages in the lake also exceed those in the stream in all months but July and Sep- tember. Higher levels, increased impounding function, and decrease in vegetation thus favor plankton production in Quiv- er Lake, and tend to raise it from a diluent to a source of imme- diate enrichment. In this connection it should be noted that the increased production of this year (2.59) still falls below that of the river in 1894 and 1897, and, as seen in the table on p. 292, below the general average of the river production (2.71) ; and also that the higher river levels of this year tend to lower the proportion which the tributary spring and creek waters form of the total volume of Quiver Lake. A second significant fact brought out by the comparison is rendered patent by the frequency in this year of the collections in Quiver Lake. The weekly interval from April to Septem- ber (Table V.) makes it possible to trace somewhat fully the movement of production, and demoustrates in Quiver Lake a pulse-like movement in production similar to that previouslij de- scribed in the Illinois River, and one, moreover, which exhibits a very striking coincidence of developmental succession. A superpo- sition of Plate XXVII. upon Plate X. will make this demon- stration apparent. There are exceptions, but these, as shown in the preceding discussion, are in most, if not all, instances to be correlated with local environmental factors confined to one or the other body of water. The return to parallelism with the cessation of the peculiar factor incident to the interruption serves still further to emphasize the significance of this simi- larity. The key to the parallelism must lie in fundamental factors common to the plankton of both areas or to their envi- ronment. 361 1897. (Tables v., Xlll. ; PI. XXVIII., XLVIII.) There are 24 collections in this year, with an average of 0.88 cm.' per ni.' as compared with 3.69 in the river, and a max- imum of 13.38 on April 27—more than twofold the production in the river (5.11) on that day. The collections of the tirst six months of this year are so infrequent that the course of production is but .slightly indicat- ed. In February the production in the impounded waters of the winter flood in Quiver Lake ( .19) is nearly fivefold that in the current-swept channel ( .04), while in March there is little differ- ence (.84 and .38) in their plankton content. The collection of April 27 probably falls near the presumably common vernal max- imum and in the midst of the decline of the spring flood. Since Quiver Lake at the stage of river (11.6 ft.) then prevalent contains the run-off of the impounded backwaters to the north, it is not surprising that its plankton content (13.38 1 is more than dou])le that of the river (5.11). The similarity in the movement of production thus far seen in this year is interrupted on May 25 by the decline in the lake to 1.29, while the river rises to 5.62. The decline in the lake may be attributed to the great reduction in impounding area due to the decline in levels to 8 ft., and to the gain in proportion in the lake of the contri- butions of creek and spring water and of the area occupied by the now rapidly appearing vegetation. The silt-bearing flood of June in the river yields le.ss (,.27) than the lake waters (1.26) impounded by the rise of the river. In the last six months of the year the collections are of sufficient frequency to enable us to trace somewhat the move- ment in production. This period is marked by a great depres- sion in plankton content as compared with that of the same season in the previous year, the average for each being 1.06 and .23. The parallelism in the movement of production can still be traced in the .slight tendency in Quiver Lake to increased production in July, September, and November at times of pulses 362 in the channel plankton. The amplitude attained in the lake is, howevei", but slight. The explanation of this marked decrease in production in this year as compared with that of 1896 lies, I believe, in the hydrographic conditions of the two years. In 1896 (PI. XXVII.) the average height of the river for the period August-Decem- ber is 5.89 ft., while in 1897 it is only 2.47 ft. The impounding action of the lake was at its minimum, and there was present in it in these months of 1897 only about a third the quantity of water that it contained in the corresponding season of 1896, and this consequently gave to vegetation in 1897 a relatively greater predominance in the lake, and also made possible a more frequent renewal of lake water by the contributions from the creek and tributary springs, thus cutting down the time for breeding. Both of these factors tend to limit plankton produc- tion. We find, accordingly, that the lake produces on an av- erage from August to December but .1 cm.' per m.'' to 4.0 in the adjacent river, that is, only a fortieth of the plankton content of the stream. The contrast between the lake and.the river in this year is heightened by the fact that owing to low water and increased fertilization by sewage the production in the river is much greater than usual. A comparison of Plates XI. and XXVIII. will reveal the fact that in only 5 instances out of 24 in 1897 does Quiver Lake contain more plankton than the river. These instances in Feb- ruary and April attend impounding action of the lake when it is not differentiated from overflowed bottom-lands as a separ- ate unit of environment, while those of June 28 and July 21 are caused by the flushing of channel waters by floods from which the lake is exempt. As a whole for this year, the net result of the contributions from Quiver Lake is a dilution of the channel plankton wdth which it mingles. Basing calculations upon the yearly aver- ages and areas of the drainage basins, the quantitative effect would be a decline in the plankton content of channel waters from 3.69 to 3.65, or a loss of 1.1 per cent. 363 The lake waters in October-December contain (Table XIII., PI. XLVIII.) but a fraction—from a third to a tenth—of the nitrogenous matter that is found in the channel. This is an index of the relative poverty of Quiver Lake waters when isolated as a separate unit of environment and dependent upon creek and spring waters, mainly of seepage origin, for its supply. This relative poverty, combined with the factors be- fore discussed, lies at the basis of the relatively small plankton production in this Ijody of water in this year. 1898. (Tables V., XIII.; PI. XXIX., XLIX.i « There are 26 collections in this year at fortnightly inter- vals, with an average of 2.44 cm.' as compared with 2.13 in the river, and a maximum of 42.14 on May 3 coincidently with the vernal maximum in the channel (Fl. XII.), which, however, falls 6.46 cm.'', or 15 per cent., short of that in the lake. The parallelism in the movement of production noted to a varying extent in prior years may be traced also in 189S. The most striking coincidence is the agreement in the location and relative development of the vernal pulse, and further resem- blance may be seen in the June pulse and the December rise in production. The small quantities of plankton in the lake at other seasons and the fortnightly interval of collection render the correspondences less obvious though perhaps not less sig- nificant. During the low water of January and in the subsequent Hood (PL XXIX.) there is little plankton in the lake (.02) as compared with the river (.45—see table between pp. 342 and 343), though an increase with a rise in levels and development of the reservoir function of the lake might have been expected. There is, therefore, no January-February rise in the lake cor- responding to that in the river unless the increase from .003 Jan. 11 to .04 Jan. 25 be held to be significant. The February flood, which depletes the plankton of the channel, is accom- panied by a rise to .58 on the 22d in the lake coincidently with 364 a slight but not equal rise (.10) in channel production. The March pulse in the river, with a maximum amplitude of .77 on the 22d, is attended by almost equal production in the lake (.67). The vernal pulse rises with like abruptness at both stations, increasing from April 1 to May 3 from 1.03 to 42.14 in the lake, and from 1.12 to 35.68 in the river. The decline of this pulse is much more abrupt in the lake, falling from 42.14 on the 3d to 4.7 on the Uth—a decrease of 89 per cent, in 8 days, while the decline in the stream is from 35.68 to 10.31, or 71 per cent., in 7 days. The more abrupt change in the lake plankton is due to the fact that the decline in levels of 1. ft. in the interim be- tween the two collections compared, is at the critical point ap- proaching bank height, when the bottom-lands to the north of Quiver Lake are beginning to emerge and cut off and divert some of the run-off of the impounded backwaters which at higher levels make their way to the channel through Quiver Lake (PI. II.). There is a very slight July pulse in the lake on the 19th coincident with the July maximum in the river. In the early part of August there is another maximum in the river, but no parallel developement in the lake, owing possibly to the low water then attained and the resulting dominance of vegetation and tributary waters—conditions not incident to these levels in like degree in channel waters. The rise at the close of August and again in September, and the low level and slight change in production in October found in channel waters are all to be traced coincidently, or approximately so, in the less complete records of the lake production. The silt-laden flood waters which cause rising levels in No- vember deplete the channel plankton (.25), increase the im- pounding function of the lake, and lead to greater production (.73) in the latter. There are coincident culminations in river and lake on Dec. 6, but the interval of collection in the lake does not permit comparison in case of the river maximum of Dec. 20. The large December production (1.74), six to eleven times that of July (.16), August (.22), September (.33), or Oc- tober (.23) is noteworthy. 365 There is thus a striking similarity in production in the river and lake in 1898, not only in the larger movements, such as the vernal pulse, the low level of midsummer, and the De- cember rise, but also in the minor details which differentiate movements at shorter intervals, suggesting in some cases, and demonstrating in others, the presence of coincident recurrent pulses of production with approximately similar locations but, it maj' often he, with more widely differing amplitudes. A part of this similarity is doubtless due to the fact that in 1898 for fully live months of the year, when the river was at 8 ft. or above, the lake was not, superficially at least, diffei'en- tiated from the general bottom-land environment, and there- fore shares more extensively the course of production elsewhere than it does when its emerging l)Oundaries delimit it as a sep- arate unit of environment. The similarity is not, however, con- fined to this period of aquatic continuity, but appears also in the season of delimitation, when local factors are relatively more potent. It is also true that even in the period of conti- nuity the environmental factors peculiar to the lake continue, though submerged or invaded,—as, for example, the chemical conditions, which even in flood periods exhibit a certain auton- omy in the lake, as will be seen on comparison of Plates XLV. and XLIX.,—to exercise some differentiating influence, which, in the presence of the apparent tendency towards similarity of movement in production, still produces modiflcations sufficient to stamp the seasonal planktograph with a characteristic facies, thus differentiating it from other localities. The average production for the year is 2.44 cra.^ per m.' as compared with 2.13 in the river, so that as a whole in this year the outflow from this lake enriches the channel plankton. On the basis of yearly averages and drainage areas the net result is an increase from 2.13 to 2.14, arise of less than .5 per cent. A more detailed analysis of the data reveals the fact that in 7 of the 12 months, in January, April, and June-October, the river ex- ceeds the lake in production. As will be seen on PI. XXIX., the remaining months are those of high river levels, when the im- 366 pounding action of the lake is most operative and its localiza- tion least pronounced. The largest production, in May and June, occurs when on declining flood the flow of impounded bottom-land waters from the north is greatest through the lake. If we omit from both records the months of May and June, we find that the averages of the remaining monthly averages (see table between pp. 342 and 343) are .91 and .50 respec- tively for the river and the lake. Thus for ten months of the year the plankton content of the latter is but five ninths of that of channel waters, and during this period the immediate result of the access of the run-off from Quiver Lake will be a dilution and diminution of the plankton content of channel waters, due, it seems, to the relatively more recent origin, from storm and seepage waters, of these tributary contributions, and to the greater prevalence of vegetation in the lake. Another factor operative in the diminished production of the lake is relative poverty in nitrogenous substances. For example, the average nitrates (cf. Tables X. and XIII.) for the year in river and lake are respectively .809 and .68 ; the nitrites, .121 and .029 ; the or- ganic nitrogen, .92 and .569 ; the albuminoid ammonia, .431 and .275; and the free ammonia, .95 and .138. The unutilized ni- trogenous substances in the lake are, however, of sufficient pro- portions to indicate the possibility of the support of a larger volume of plankton if greater time for breeding were allowed. 1899. (Tables v., XII.; PI. XXIX., XLIX.) The 7 collections in January-March average .67 cm.^ per m.' as compared with .41 in the river. As in the previous year, the direction of movement in pi'oduction is similar in the two regions. For example, the January pulse in both culminates on the 17th and that of February on the 14th and the 21st, while the March production is at low levels in both, and the apex of the pulse is not apparent in the lake records. The invasion of some channel flood water with the March rise and its speedy elimination may be traced in the chemical records (PI. XLIX.) 367 The production in the lake during this period is greater than that in -the river at all times of coincident examination excepting March 14 (.14 and .35). The average production in the lake (.67 cm.'') is 63 per cent, greater than that in the river (.41 cm.^). This percentage of increased production is a meas- ure, or an index, of the impounding or reservoir action of the lake under the hydrographic conditions of these months. The immediate result of the access of Quiver Lake waters to the channel will be a rise in its plankton content from .41 cm.^ per m.'' to .414—an increase of 1 per cent. The summary of the interrelations of production in this lake and the river will be made in conjunction with that of Dog- fish Lake, which is only an arm of Quiver Lake. DOGFISH LAKE. (Table VI.; PI. XVIll., XXX., XXXII.) ENVIROMMENTAL CONUITIONS. This so-called lake is only the westerm arm of Quiver Lake (PI. II.), separated from the eastern by Quiver Point, a low marshy spit covered with rushes and willows and lying but a few feet above low-w^ater mark. It is of elliptical form, about three quarters of a mile long and one third of a mile wide, contains about 150 acres at low water, and as levels rise it ex- tends northward and eastward over the low bottoms towards Mud Lake and Cartwright Slough, but it is only at highest levels that very much of a current makes its way down through this lake. As levels rise above 8 ft. the intervening ridge sep- arating this lake from the river is gradually submerged, and channel waters invade more or le.ss of the lake. It affords the natural channel for the run-off of the backwaters impounded in several square miles of bottom-land marsh and forest through the swale (PI. II.) which extends towards Mud Lake. Its shores are everywhere low and marshy, of black alluvi- um, and a soft black ooze of similar origin covers the bottom of the entire lake. In only a limited area towards the east- 368 ern side can a substratum of harder sandy clay be reached be- neath two or more feet of this deposit. With the exception of a narrow fringe along the eastern side, the vegetation lacks the lilies, rushes, sedges, and other emergent plants which charac- terize the eastern shore and northern end of Quiver Lake proper (PI. XVII.). It consists (PI. XVIII.) almost exclusively of Cer- atoplnjlhini, EJodca, and Potaiiiogefons, which, in the low water of 1895, represented in the plate, filled the lake from center to periphery. Irregular openings in this dense growth appear oc- casionally in the area, and are modified by the shifting of the lightly attached vegetation, by wind, and by flood water. Except at high water and during the rapid run-off of im- pounded backwaters no appreciable current traverses this area. It receives no immediate contributions of spring or creek water along its margins, but depends entirely upon backwater from Quiver Lake or flood invasions for its supply. The examination of the plankton content of its waters ac- cordingly affords an opportunity to test the effect of this im- pounding factor, and also serves to throw some further light on the effect of vegetation on plankton development in impounded waters. COLLECTIONS. The collections in this lake cover a period of two years — from April 29, 1895, to June 28, 1897 (Table VI.). They num- ber 48, and are distributed in much the same manner as those in Quiver Lake in the same period. The collections of 1895 and those of 1896 through May 8 were all made by the oblique-haul method with the single exception of that in the flood of Feb. 27, which was one of repeated vertical hauls. The collection of May 19, 1896, was made in the midst of rapidly growing veg- etation by dipping from surface waters, which then afforded no area suitable for an oblique haul. The oblique hauls were made for the most part near the center of the lake in a channel freed from vegetation a day prior to the collection. From May 21, 1897, all collections were made by the plankton pump in open stretches of water amid the vegetation. 369 In 1895 the lake was choked with vegetation which the winter flood largely removed and the recurrent floods of the following year reduced somewhat in extent, while higher levels lowered its relative occupation of lake waters. PLANKTON PRODUCTION. 1895. (Table VI., PI. .\.\.\.| There were 12 collections in this j^ear, from April to De- cember, averaging 3.25 cm.' per m.'' The average of the monthly averages (see table between pp. 342 and 343) is 3.3 cm.^ to .74 and 6.65—similar averages for the same period in Quiver Lake and the Illinois River. The maximum collection was made Dec. 19—a very unusual date for such production. A superposition of the planktographs of the river and Quiver and Dogfish lakes for this year brings out some in- structive similarities and differences in the movement of pro- duction. The vernal pulse of April 29. in so far as the data reveal it, is quite similar in all three localities, reaching its greatest development in Dogfish Lake (8.20), where im- pounding action is greatest, and l^eing greater in the river (5.S3) than in Quiver Lake (4.57), where, owing to low levels, the proportion of water of recent creek or spring origin is greater than in the channel of the adjacent river. The June-July pulse may be found in all three localities, but it is belated and much smaller in the lake waters. This pulse in Dogfish Lake (4.59 cm.'' per m.'') is less than a sixth of that in the river (30.42), where, in the semi-stagnant sewage- polluted channel waters of unusually low levels, Moina and other ('/(idocrrd caused the unusual production. Between the April and June-July pulses the river levels fell 2 ft., to mini- mum stages (PL XXX.), so that the proportion of creek and spring water in Quiver Lake is probably more than doubled at the later date. This may account in large part for the very low production in Quiver Lake (.02) on July 8, while on July 5 370 the contiguous but impounded and current-free waters of Dog- fish Lake contain 229 times as much plankton. The increased production following the September flood is apparent in all three localities, but reaches its highest level (4.65) in Dogfish Lake, the region where the impounding factor is greatest, while the least increase and quickest decline is in Quiver Lake, where tributary waters of recent origin are in greatest proportion. The low production in October is com- mon to the three localities, reaching a slightly lower level in the lakes (.13 and .52) than in the river (.57). The consider- able increase in production in November-December attains the highest level in Dogfish Lake (5.01 and 10.57), exceeding by 100- and 17-fold that in Quiver Lake (.05 and .68), and that in the river (4.37 and 2.60) by 1.2- and 4-fold. Here also the effect of the quieter impounding waters of Dogfish Lake is apparent in this relatively greater development. The large plankton content on Dec. 19 (10.57) seems to be due to a combination of several favoring environmental factors. This collection was made after a steady but slight rise lasting for over five weeks, fol- lowed by ten days of gentle decline in levels and contracting margins of the lake. The steady rise to levels which intro- duced no run-off currents through the lake established the im- pouiading function to its fullest, and invaded a considerable stretch of margins rich in dead and decaying vegetation. There is also at this season of the year less growth and more decay of the Elodea and Cerafophi/Uinii which abound in the lake. The collection was taken when the December flood had just be- gun to rise (about 2.6 feet) and with the combined action of wind and waves which attended the storm then raging had torn loose the vegetation and dislodged many of the smaller Crustacea and insect larvae which find shelter in it. There were at the time 572 A/lorchesfes per m.' adventitious in the plankton. A part of this large production is thus adventitious owing to disturbed hydrographic conditions. Nevertheless, there still remains after such contributions are deducted a con- siderable plankton of normal constitution (mainly Cladocera), 371 in amount certainly mucli in excess of the production at that time in Quiver Lake (.63) or the river (1.74). This large pro- duction in this locality is then, it seems, to be attributed to im- pounding and decaying vegetation combined with accession of adventitious planktonts. The average monthly production (3.3) in this lake is on the whole less than half that in the river (().65) for the same period, and it exceeds by over four-fold that in Quiver Lake (.74:), through which all its run-off passes to reach the river. Since the com- parison of the two lakes is based on coincident collections, these amounts may serve as a quantitative statement of the effect of the environmental differences. As vegetation is much the same in both lakes the difference in production must be attributed to some other factor presenting a difference which may be corre- lated with that in production. Such a factor is found in the impounding action, which is at a maximum in Dogfish Lake and is I'elatively much less in Quiver Lake at the point of our collections, where creek and spring water of recent origin cause a more rapid displacement of the contents of the lake and car- ry away the products of decay of vegetation before the plank- ton can reach the degree of development that it does in the more stable waters of Dogfish Lake. The run-off from this lake in this year would thus tend to eni'ich Quiver Lake, though not on an average of sufficient pro- duction to enrich the river even if it could reach it without mingling with that of Quiver Lake. However, owing to the fact that this lake receives no tributary creek or spring water, and except at high levels has no bottom-land current through it, we must infer that its run-off is confined—excepting only at stages of general overflow—almost wholly to stages of falling water. During rising levels and in fairly stable conditions its contribu- tions to Quiver Lake, and thus to the river, are practically nil. 1896. (Table VI., PI. X.XXI.) There are 30 collections in this year, with a distribution 372 similar to tliat of the collections in Quiver Lake for this year, the only exception to the coincidence of collections, actual or approximate, being on December 29, when on account of rotten- ness of the ice it was not possible to get the collecting outfit to the station. The maximum production occurs in the vernal pulse in the last fortnight of April, culminating at 20.35 cm.^ per m.^ on the 17th, though production is also large on the 24th (19.5). In Quiv- er Lake this maximum is on the 24th (16.76), though production is also large on the 17th (16.32). The maximum in channel waters (9.39) is also on the 24th. These differences in the time of the maxima may, I believe, be correlated directly with the thermal factor. For example, in both Quiver and Dogfish lakes the production is large and almost equal on the 17th and 24th, but is greater in Dogfish Lake on the 17th and in Quiver Lake on the 24th. This lag in the maximum is correlated with the fact that surface temperatures in Quiver Lake on the 1 7th and 24th are respectively 3° and .8° lower than they are in Dogfish Lake. On the 17th the latter is 8° warmer than the river. After all allowances are made for the time of day at which tempera- ture records are taken, it is still evident that the shallower waters of Dogfish Lake would warm up more quickly than the spring- fed waters of Quiver Lake or the deeper channel waters, and we have found that the thermal increase favors the earlier rise in plankton production. The coincidence of the dates of collection makes possible a precise comparison of the production in the two lakes, and facilitates the comparison with that of the river. A superposi- tion of the planktographs of Dogfish and Quiver lakes and the river (PI. XXXL, XXVIL, and X.) for this year emphasizes far better than any description the most striking similarity at the three stations of the movement of plankton production as shown by the direction of the differences in plankton content in suc- cessive collections. The correlation between production in Quiver Lake and the river in this year—discussed in detail on pages 357-360—is paralleled in every important detail by the 373 sequence of the changes in Dogfish Lake. Indeed, tlie corre- lation is, if anything, even closer, since the amplitude of the plankton pulses is greater in Dogfish Lake than in Quiver Lake, and the changes are here—as, for example, in August—the more readily followed and compared with those of the river. Since I have already compared in detail the production in Quiver Lake and the river I shall not repeat the comparison of these similar data from Dogfish, for the correlations are essen- tially the same in both cases, and it will suffice simply to empha- size the similarity of the course of production in the three local- ities. The similarity between the production in the two lakes is, however, even greater than that between either of them and the river. This results from the greater similarity of the envi- ronmental factors in the two lakes, with which the river con- trasts in matters of sewage and current. The similarity of en- vironmental factors lies in the amount and kind of vegetation, the depth, the character of bottom and shores—excepting the eastern margin of Quiver, the impounding function (modified, however, in the case of Quiver Lake by the access of creek and spring water), and the freedom from sewage. Under these cir- cumstances it is not surprising that the details of the course of production as well as its ciiscnih/r are so strikingly alike in the two lakes. To be specific, the similarity in details of the course of pro- duction in Quiver and Dogfish lakes lies in the fact that in the 31 coincident collections in these waters the plankton content rises or falls in both at the same time in 23 out of the total num- ber. The ampUtiiilc of the change is rarely equivalent, but its direction is identical—referring, of course, to the fact of its being an increase or decrease, and not to the particular angle which the lines forming a planktograph might take. The 8 exceptions to this similarity in the direction of movement in production are shown in the following table, and may without exception be correlated with differences in the environment. In this table the plankton contents of the two adjacent col- 374 lections deteriniuing the direction of the change in production are given under the date of the later collection, and the posi- tion of the entries also indicates the direction of the change. VARIATIONS IN THE MOVEMENT OF PRODnCTION IN DOGFISH AND QUIVER 375 on the 2d (4.24) and the larger one in Dogfish (13.39). On the 8th conditions are changed ; the decline in levels [.Q ft. ) has in- creased the run-off. and the recent contributions of tributary wa- ter have brought down into Quiver Lake an increased proportion of plankton-rich impounded backwaters which increase the con- tent at that point to 8.14 cm.', this being, how^ever, still below that of contributing and declining Dogfish Lake ( 13.39 to 13.()(i i. On Maj' lfi-19 the collections were not quite coincident, but, such as they are, they form another exception to the simi- larity of movement in the two lakes. In Quiver on the 16th the decline of levels brings the proportion of tributary waters at that point into greater prominence, while on the 19th in Dog- fish the impounding function is gi'eatly increased by the inter vening rise in levels (1 ft.). Examination of the collections also shows that the maximum of 18.4 cm.' in Dogfish Lake on the 19th is caused primarily by an extraordinary pulse, or possi- bly a local "swarm", of MrJasira with some Cladocera, whose apex and location the date of collection approximates. On the 21st the plankton content in the same locality fell to .36 cm.'—a decline of 98 per cent, in 2 days. On the 21st and 22d a quantitative survey of the local distribution of the plank- ton in the whole area of Quiver and Dogfish lakes was made, with the result that no development commensurate with that on the 19th at this point was anywhere detected. Since a sim- ilar sudden decline is to be seen in Flag Lake in this same week,! am inclined to the view that we are dealing here with a complex biological phenonenon in which the reproductive cycles of the organisms as well as external factors—such as possible tempo- rary decline of food supply, or encroachment of emerging veg- etation—are involved. This sudden decline is earlier and less marked in Quiver Lake than in Dogfish, possibly because of in- creasing differential environment, and thus occasions this tem- porary dislocation of the similarity in the movement of produc- tion on May 16-19, and again on May 21-22. On July 3 we again find lower levels reached and accom- panying decline in production in Quiver Lake when tributary 376 waters rise in proportion. The readjustment which this ne- cessitates so modifies production that its movement differs, though the difference is slight and consequently the less sig- nificant. On Aug. 15 and 29 we deal again with a phenomenon simi- lar to that of May 21, namely, a pulse of large production in Dogfish Lake with an accompanying one of lesser amplitude in Quiver. The pulse in Dogfish was again later, apparently, in reaching its culmination (on the 15th) than that in Quiver (on the 7th), and its decline (93 per cent.) on the 22d more complete than that which Quiver attains (63 percent.) on that date from its maximum (4.36) on the 7th. The conditions in Quiver Lake are further complicated by the fact that on the 15th river water was just beginning a temporary invasion at the point of collection (see hydrograph, PL XXXL). Save for this invasion the similarity of movement might perhaps have been preserved on the 15th. The abrupt and extreme decline in Dogfish Lake on the 22d, however, with the resulting inter- ruption of the similarity of movement on the 29th, is in some way related to the excessive development of the 15th, which may bring into operation again the factors above suggested in connection with the like phenomenon in May. It seems not improbable that the sharper localization in Dogfish Lake due to absence of current and tributary waters and the presence of these factors in Quiver tends to intensify environmental in- fluences or inherent tendencies of the plankton in the one area, and to minimize some if not all of them in the other, and that this differentiating influence of these purely local factors is fundamentally the cause of the dislocations and disturbances of the otherwise similar movement in production in the two lakes. These exceptions seem, however, to emphasize the es- sential similarity in the production in the two areas, a similar- ity founded on the common factors of the environment shared equally by both, and on the identity, in the main, of the con- stituent organisms of the plankton. The general hydrographic conditions of this year affect 377 profoiindh- the plankton production in this lake and in Quiver also. Although the average height for this year (6.975 ft.) is almost the same as in 1S97 (^6.903 ft. K the distribution of high water is such that the impounding function is exercised not only during the winter months, when production is low, but, owing to the recurrent floods, it is in operation to an unusual extent during the period fi'om June to October, when produc- tion is wont, as a rule, to run low in these waters. Thus levels (PI. XXXI.) are above 6 ft. fully half of this time and are at all times above 4 ft. with the exception of 10 days in July. Not only does this increase theimpounding function of these waters, but it decreases the relative occupation by vegetation in addition to reducing its actual extent by uprooting and removal. It also decreases the proportion which creek and spring waters form of the total content of the area, or impounds them long enough for the plankton to breed therein. The distribution of hi;/ li tenter is such in this year that it affords an opportunity for increased production in the lake. In comparison with 1895, when production averaged 8.25 cm.^ per m.' from April to the end of the year, we have 5.01 in 1896 for the year as a whole. The avei'age height of the river in 1895 was 3.61 ft. (p. 163 1. in 1896. 6.98—an increase of 3.37ft., or the equivalent of almost doubling the volume of water in the lake. So not only is the amount per cubic meter greatly increased, but the total run-off of plankton into the channel is multiplied by some undetermined factor. The net result of the hydrographic conditions of 1896 in Dogfish Lake is therefore an increase in its impounding func- tion at a time of large production (5.01). and its discharge tends to raise the plankton content alike of Quiver Lake (^2.59) and the river (1.16). but data are lacking which might enable us to compute its quantitative effect upon the plankton con- tent of either. Not only is the arerage production of Dogfish Lake greater than that of Quiver, but individual collections here exceed co/«- ci'Iriif ours there with the exceptions only of those on May 21 378 and 22 and Aug. 15, when phenomenal declines appeared in Dogfish lake. In a similar way its production exceeds that in the river in every case but one, that of May 21. Thus produc- tion is prevalently higher here than in Quiver Lake and the river, to which it contributes its run-oiJ, as a result of the im- pounding factor and, in this year, of the relative absence of veg- etation also. The impounding permits the growth of the plank- ton to utilize the nutriment derived from decay of vegetation and other sources before it is carried out of the lake. 1897. (Table VI., PI. XXXII.) There are but 6 collections here in 1897, in the first six months of the year, at approximately a monthly interval. The average production for this period is 2.23 cm.^ per m.', with a maximum of 8.18 on Apr. 27. Since the collections are coin- cident in the river, Quiver Lake, and this lake, a comparison of production is facilitated. The similarity noted in the previ- ous year may be traced here also, and the relationship of the three areas remains in the main unchanged during this half of 1897. Briefly, there is low production in all three under the ice in midwinter, with a slight increase in all in March, a ver- nal pulse in April followed by a decline in production in May in the lakes but not in the river, while in June the flood reduces the plankton content in the river but changes that in the lakes but little. The collections throughout the period show greater production in the lakes (Dogfish, 2.23, Quiver, 2.77) than in the river (average, 1.91) with the exception of the May collection (Dogfish, 1.94, Quiver, 1.29, river, 5.62). This drop in plankton content in the lakes below that of the stream occurs at the time of greatest increase in vegetation and rapid drop in levels, which increases the i-elative occupation by vegetation—a factor from which the river is relatively free. The flood of June, flushing the stream, obscures the relationships of production at that season. In all collections but those of May and Feb. 26 Dogfish Lake contains a more abundant plankton that Quiver 379 by from 25 to 100 per cent. This is apparently due to the pre- ponderance of the impounding factor in the former. Thus in this season also the similarity in the movement of production noted in the previous year can be traced, and the excess of production in Dogfish over Quiver continues in the main. Its run-off therefore serves in this period to enrich the plankton alike of Quiver Lake and the river. This is correlated with the high levels and consequent increase in the impounding factor and the relative diminution of vegetation in this area. GENERAL SUMMARY. RELATIONS OF PRODUCTION IN DOGFISH AND QUIVEK LAKES TO THAT IN THE ILLINOIS RIVER. The analysis of the data of production in these two lakes leads to the following conclusions. Plankton production is in a large degree a function of the time allowed for the breeding of the plankton. Thus at times of high water, when both lakes are filled principally with the impounded backwaters of overflow, production is greater, other things being equal, than at low water, when a greater proportion of water of the lake (Quiver) is of recent origin from tributary creeks and springs. So, also, areas such as Dog- fish Lake, in which by reason of absence of tributaries and springy shores the impounding function is greater, show a great- er plankton content than similar areas (Quiver) where by reason of access of tributary water the impounding function is de- creased. Vegetation of the character of that found in these lakes seems to exercise an inimical effect upon plankton production. Thus the season of dominant vegetation is generally one of low production in these lakes. Also during this period, as levels fall and occupation of the lake by vegetation becomes relatively greater, production generally declines, and, conversely, produc- tion rises when levels rise. Years of greater dominance of veg- etation, other things being equal, are wont to exhibit a decline in production, and, conversely, with lessened vegetation pro- duction rises. 380 These factors, combined with changes in production in the river, vary the relation which these lakes bear to production in channel waters. In general, in times of low water and domi- nance of vegetation the outflow from these lakes is a diluent of channel plankton, but during the run-off of impounded back- waters or in years of higher levels and less vegetation it serves to enrich channel waters. The river and the two lakes exhibit in common a very marked similarity in the seasonal movement in production. The recurrent pulses, which may be traced whenever collections are of sufficient frequency, coincide closely in their location but exhibit considerable local differences in their amplitude. This similarity is greatest when local environmental factors, such as vegetation and tributary waters in the lakes and sewage contamination and recent flood water in the river, are least op- ex'ative, and is diminished or obscured as these factors come more into action. The diversity, as shown in the differing am- plitudes of the pulses of production and in the divergences and interruptions in their rise and decline, can generally be traced to the preponderance of some local factor or factors above named. The similarity in the seasonal movement of production is all the more marked when the striking differences of the three localities in question are considered and the general instability of the whole environment is borne in mind. The changes in the plankton content of the river,—turbid and fouled by sewage, traversed continuously by a considerable current, and scoured repeatedly by flood,—of Quiver Lake,—with gentle current,clear spring-fed waters, and greater or less, but always considerable, vegetation,—and of Dogfish Lake,—with tranquil, almost cur- rentless waters, without access of tributary contributions, and also with considerable vegetation,—all exhibit a harmony that compels us to admit the potency of those general factors of the environment common to all—their climatic and geographical surroundings, which determine the amount and distribution of the light and heat, and the chemical constituents of the medi- 381 um in which the plauktou grows. The similarity iu the move- ment of production must also be correlated with the fact that these common environmental factors are responded to in the three localities by a plankton composed of identical or closely re- lated species in varying proportions. It is iu the main the re- sult of the response of similar organisms to the common fac- tors of an environment, factors, moreover, of fundamental sig- niHcance. FLAG LAKE. (Table VII. ; PI. XI.X., X.X.XIII., XXXIV.) ENVIRONMENTAL CONDITIONS. This is the local name for a marsh in the western bottom- lands opposite the location of our plankton station in the river (PI. II.). Together with its outlet, Flag Lake Slough, it ex- tends parallel to the river from north to south a distance of about -lA miles, and is generally less than | of a mile in width. It has no precise boundaries, since the fringe of willows which borders it, save for 2 miles along its northwestern margin where it joins Thompson's Lake, merges gradually with the marsh on the one hand and the bottom-laud forest on the other. It con- tains about 2^ square miles of permanent marsh, of which but a small area towax'd the lower end was free from vegetation. The depth depends upon the stage of the river or the extent of the run-off of the impounded water. Its bottom, if we may dig- nify the treacherous ooze from which the vegetation springs by this name, is generally, if not entirely, several feet above low- water mark in the river. In the autumn of 1897, during the prolonged low water of that season, the lake dried up and a road was opened across it to Thompson's Lake. Uenerally. how- ever, it retains sutticient water to tide over ordinary periods of low levels. The hydrographic conditions are such as to make this marsh exempt from all current save at times of most general overflow. Owing to the somewhat elevated banks along Flag 382 Lake Slough and along the west bank of the river, no access of river water is possible from the north or east until bank height is exceeded by the flood. At all levels below this, water enters the lake by the slough, which forms its outlet, or backs in from Seeb's Lake (PI. IT.). Another line of access is the low margin to the northwest between it and Thompson's Lake. The rank growth of living or dead vegetation which at all times fills this region, eifectually blocks any localized current here, and no channel has opened in this region. Probably much of the water, as indicated by the distribution of drift, enters the lake from its southern end. The same reasons which prevent access of water from the north also tend to restrict the flow through this area at times of general overflow, and the fact that Thompson's Lake (PI. II. ) affords for backwaters impounded to the north a channel where resistance is much less than in the shoal, forest- begirt, and rush-filled Flag Lake, tends also to divert the mov- ing backwaters to that region. Consequently, Flag Lake is in the main an impounding area whence the impounded water is drawn off as levels decline, but which is not generally trav- ersed by the waters of general overflow as are Quiver and Thompson's lakes. It is thus one of the most strongly localized of all plankton stations, and the unity of its environment is more continuously maintained than that of any of the localities thus far examined. Its vegetation has been described on pages 249-250, and it will suffice in this connection to call attention to the predom- inance of the emergent and succulent types in its waters, and to the fact that little, if any, of it is ever carried away by flood or currents as it is from Quiver and Dogfish Lakes. This is a large factor in maintaining the local fertility of this area. This is a favorite haunt of migrating water-fowl in fall and spring, and contains breeding grounds of the few summer residents. It is also much resorted to by the German carp, now one of the most abundant fish in the Illinois. Fish enter the lake in numbers when levels rise, but leave again before low water in the slough (PI. II,) prevents their departure. Evi- 383 deuces of the destructive work of the carp are seen iu their ac- tivity in uprootiug great patches of Sni/iffurid. Such condi- tions were prevalent shortly after the August flood of 1896 (PI. XXXIII. I. when the combined action of the change in levels and the invasion of Hsh destroyed not a little of the vegetation. COLLECTIONS. Systematic examination of the plankton in this area was begun Oct. 17, 1895, and continued until Aug. 16. 1897. Six summer collections were made in 1898 for the purpose of de- tecting Tnic/inspliirni. Owing to the surroundings of this region access to it ex- cept during high water was a matter of much time and con- siderable difficulty. Absence of roads and bridges made ap- proach by conveyance impracticable, and save at maximum overflow the elevation of the surrounding bottoms or the aluui- dant vegetation prevented the entrance of the steam launch- The drift in the slough i PI. II. ) and the matted, and in many places impenetrable, growth of Srirpus fimiatUis rendered ac- cess to the small areas of open watei- an arduous task. At low stages the only means of obtaining a collection was to wade out through the mora.ss to a suitable place. The difficulties of approach in winter were even greater, when ice and the emer- gent vegetation combined to interfere with rapid transit of any considerable load. For these reasons this station, though one of much local biological interest, was early dropped from our list. These same difficulties have enforced some variation in the methods of collection (^ Table VII. i and in the locality at which collections were made. In the autumn of 1895 they were made by dipping water in scattered areas in the vege- tation and in advance of the roiling of the water caused by wading. On Feb. 28, 1896. a measured quantity of water from successive levels was taken amid the standing but submerged vegetation with a pump. Other collections prior to May 23, 1895, were made by the oblique-haul method and thereafter by the plankton pump. 384 In the autumn of 1895 collections were made in the north- ern arm of the area marked as open water on Plate TL, and occupied at that time by a considerable amount of submerged and floating vegetation, largely of Nymplum. During high water, when landmarks were submerged, this location was ap- proximated as nearly as possible, but in the following spring the location of the plankton station was shifted to the lower arm of this open area (PI. II.), and as the vegetation emerged and blocked access the station was moved to the head of the slough in effluent waters. These changes in method and loca- tior^ impair somewhat the value of this series of collections for comparisons Inier se, but they still serve to throw important and significant light upon the relationship of such marshes and of their vegetation to plankton production in their own con- fines and in the channel waters to which they may contribute. There are 38 collections, extending continuously over 22 months in 1895-1897, with an interval of collection in the greater part of 1896 sufficiently short to enable us to follow the course of production. The scattered collections of the remain- der of the period and the six additional collections in 1898 throw but little light upon the movement in production, though they are of interest for comparison with other locali- ties. PLANKTON PRODUCTION. 1895. (Table VII., PI. XXXIII.) There are but four collections in this year, in October-De- cember, averaging 20.45 cm.' per m.' and having a maximum of 57.76 on Oct. 17, and declining to 6.38 on Dec. 19. This period was one of no marked changes in the hydrographic conditions. The lowered temperature and autumn rains had checked evap- oration and brought about a slight increase in the volume of water, as shown by the increase in depth from .25 to .45 m. The lake was choked with decaying vegetation, the product of two 385 season's growth without a flood exceeding 7 ft. Even the rise to 5.2 ft. Dec. 19 was only beginning to affect the conditions within the lake. In October the succulent vegetation, such as Niiiiiphd'd, Nehnii/xi, and Sfu/iffdria. was undergoing rapid decay, which was checked by falling temperatures, and we find plank- ton production declining (from 57.76 to 6.38). and the decline accelerated on Dec. 28 (3.26) with the invasion of flood waters. This large production, unsurpassed at any other station (cf. PI. XXXIII. with PI. IX., XXVI., XXX., and XXXVI.), is to be cor- related with the excess of decaying vegetation in this locality resulting both from the abundance and character of the vegeta- tion and its freedom from flushing by current due to access of tril)utary waters. The maximum in October is due almost wholly to SijHtini nhclhi, which declines in the later collections in which the ( 'Iddarcni and later the Citiippodii appear in in- creasing nunil)ers. Throughout this period there was no run- oii" until flood levels were reached late in December, and even then, owing to reasons above cited, the run-oft" from the area is relatively slight. There was consequently no direct enrichment of the channel waters from this area. Unfortunately, no chem- ical analyses of water from this area are available, and the chemical basis for an estimate of the relative fertility of this marsh is lacking. The data of production illustrate the great fertility of waters impounded where decaying organic matter abounds. Both the impounding factor and the local enrich- ment factor are apparently at a maximum potency here at this season, and production is correspondingly great. 189(3. (Table \'II., PI. X.X.XIII.) There are 27 collections in this year, with an average of 13.83 cm.'^ per m.', and a maximum of 203.52 on May 2. The weekly interval of collection in April-June eua))lesus to follow the course of production with some detail, but the fortnightly, or longer, interval prevalent during the most of the remainder 386 of the year reduces greatly the value of the data for such pur- poses or for comparison with other localities. The hydrographic conditions are such in 1896 that this lake maintains, throughout, a connection with the river. This is owing to the relative ahsence and brief duration of low levels, the run-off not being completed before a new invasion occurs as a result of a recurrent flood. Since falling levels prevail during more than two thirds of the year, a run-off from the lake continues during this portion of the time at least. The lake is therefore in this year a factor in the determination of production in channel waters, whose continuity is broken only when levels are such that no waters are draining off from the lake or passing through it during general overflow—which is the case in less than one fourth of the time. The average produc- tion in the lake for 1896 (13.83) is almost twelvefold greater than that in the stream (1.16), and the monthly averages also (see table between pp. 342 and 343) are from 2ito 218 times greater, while individual collections in the lake in all but three in- stances exceed coincident or approximate ones in the river. The exception on July 30 occurred, when the invasion of flood water was followed, as is usually the case in midsummer in vegetation- rich backwaters, by a semi-stagnation with great development of Oscill'in'a, and the formation of considerable gas with a strong odor of H-iS beneath the felt of Oscillaria which covers the bot- tom. Under these presumably abnormal conditions the plank- ton content reached a lower level in the lake (1.62) than in the river (3.9U), and this was at a time of influx rather than outflow of water. With the above exceptions the lake at all ob- served seasons contains a richer plankton than the channel, which its run-off directly enters, and under similar hydrographic conditions we are justihed in predicting at other times a similar relationship, though the exact ratio of production would proba- bly vary according as the vegetation by its growth or decay affected the fertility of the water. In the absence of any satisfactory basis for determining the amount of the run-off from this lake, a quantitative expres- 387 .siou of its effect in increasing the plankton of the channel can- not be given. Similar marshy regions are found, along the course of the river elsewhere, especially aliove higher bottoms which have been built up across the flood-plain l\v tributaries such as Spoou River, and such areas presumably share with Flag Lake this contributory function in the maintenance of channel plankton. In the discussion of production in Quiver and Dogfish lakes 1 have called attention to the similarity in the movement in production, these two lakes and in the river. In Flag Lake, we are dealing with a very different environment : bottom, shores, vegetation, hydrographic relations, especially in the matter of tributary waters, are all diverse. Indeed, the lake it- self includes several distinct types of environment. It is inter- esting to note that in .so distinct a unit of environment as this marsh we find so large a degree of similarity in the movement of production as can be traced between its seasonal plankto- graph and that of the river and of the lakes thus far examined. It should, however, be stated that the similarity is less precise here and is more evident in iSlXi than in other years, though this is probably in part due to the absence of sufficiently fre- quent collections. The degree of similarity may be seen in the following com- parisons. In 17 of the 27 possible comparisons between pro- duction in Quiver and Flag lakes (PI. XXVII. and XXXIII.) the direction of the change in production coincides. Most of the It) exceptions are due to slight differences in the location of apices of pulses, or occur at times of lowest water, that is. of most pronounced local differentiation- as, for example, at the drop in levels m ^lay and again in July. The same number of excep- tions similarly located occurs when production in Dogfish Lake (PL XXXI.) is compared with that in Flag Lake (PI. XXXIIL), and there are 11 exceptions in the possible 21 in the case of the river (cf. PI. X. and XXXIIL). In general terms, the similarity consists in the rise in pro- duction, probably obscured in Flag Lake by an overestimatiou 388 of silt on March 30—with inci'ease in temperatui-es in January- April, culminating in a vernal pulse in April-May, which in Flag Lake reaches a much higher level (203.52) than elsewhere, culminates later by 7 to 14 days and is not divided into two apices as in the other three localities, but in duration covers the period of two pulses elsewhere. It is further seen in the May-June and August pulses and in the fairly well sustained correspondence in direction of the changes in the September- December period. The most marked disagreement appears with the declines in stage of the river in May and July, when local environmental factors are most potent, and when, also, vegetation is at the height of its relative occupancy of the lakes in question. One of the most striking features in the production of this lake, and one not without parallels elsewhere in our records (PI. XXIX., XXXI.), is the very sudden decline in plankton content after the vernal pulse, namely, from 203.52 cm.' per m. on May 2 to 47.7 on the 9th—a decline of 77 per cent, in 7 days. On the 15th it reached the low level of .72, a decline of 98 per cent, in 6 days or of 99.6 in 13 days. The attendant hydro- graphic conditions are not without significance. This pulse (PI. XXXIII.) attains its growth between March 30 (1.02) and May 2 (203.52), in which period the net drop in levels in channel waters is only from 8.1 to 6.9 ft. and the total move- ment only 1.7 ft., while in this protected backwater the fluctu- ations are probably somewhat lessened, as will be seen in the fact that the depth in the lake changes only .5 ft. to 1.2 ft. in the channel. The pulse thus rises in stable conditions. The decline of the pulse takes place between May 2 and 23 from 203.52 cm.^ per m.'' to .12. In this time levels fall from 7.1 to 4.9 (see p. 159) on the 17th and rise again to 7.2 on the 23d. The decline in production from the 15th (.72) to the 23d (.12) is so small a part of the total that its significance in the present connection is slight, and the rise in levels has probably not had time to materially affect the lake. The hydrographic influences potent in the decline in production have been operative prior 389 to this rise, and consist in a fall of 2.2 ft. in channel waters though the depth at the station of collection changes only 1.6 ft.—equivalent to a reduction in volume of 25 percent, at the point of collection and 30-40 per cent, in the lake as a whole. It thus involves a considerable and rapid run-off of the rich plankton developed in these impounded waters. This factor alone is, however, quite insufficient to account for the total loss in plankton content in this period. Another factor which is correlated with this reduction in the plankton content is the increasing occupancy of the lake by vegetation. The decline in levels hastens the emergence of the emergent forms and in- creases the relative occupancy by submerged and floating spe- cies, while the vernal growth in all during these three w eeks in May, more than any other factor, transforms the broad expanse of open water into a vegetation-clogged marsh in which but few stretches of open water are visible. This phase of the growth of the grosser forms of the aquatic flora robs the water of some of its store of nutriment and cuts off the free access of light^ — both of which might interfere with the growth of the competing phytoplankton. Limnetic diatoms such as Asterionlla and Melo- sira are the principal synthetic organisms building up this re- markable pulse, and the ('/(idorcra, principally Bosiiinut &iid('ln/- rfor».s,appear in numbers with itsculmination. The composition of the plankton favors the inference that a temporary exhaus- tion of the food of the phytoplankton and zooplankton alike con- tributes to the sudden reduction in plankton content, while the additional and perhaps related factor of reproductive cycles may also have a large causal relation to the phenomenon. 1897. (Table Vll., PI. XXXIV.) There are but 7 collections in this year, at approximately monthly intervals in January-July. Collections were suspend- ed on July 16, when decline in levels made access even to the foot of the lake by boat impossible. With the further decline (PI. XI.) in river levels the run-off from the lake soon ceased, 390 and by the middle of September the water had practically dis- appeared within its boundaries. The 7 collections average 4.59 cm.' per m.''—about double the average production in the adjacent river, and in Quiver, and Dogfish lakes on coincident dates. Individual collections also exhibit in every case a greater plankton content in the lake than in the river. This area in this season thus contrib- utes to the enrichment of the channel waters, which its run-off enters, and its contributions exceed those of the lakes on the eastern side of the river. This higher production in this local- ity is, I believe, a corollary of the greater impounding function of Flag Lake, resulting from its freedom from tributary waters of recent origin, from its somewhat sheltered location—which checks the downward movement through its area of the gener- al currents of overflow, and from the enrichment of its im- pounded waters during this period by the decay of the abun- dant vegetation of the previous season, which, for the reasons just mentioned, is not extensively carried away by flood wa- ters! The fact that production appears to be so much less in 1897 (4.59 cm.'' per m.'') than in the corresponding months of 1896 (11.21) may be due to several factors ; to the greater dilution in the greater volume of overflow (cf. PI. XXXIII. and XXXIV.) in the winter and spring floods of the latter year, to the greater abundance in 1896 of decaying organic matter accumulated by the vegetation of two preceding low-water seasons, and, possi- bly, in a measure, to the infrequency of collections in 1897 and the probable omission of the maxima of pulses of production which would tend to raise the average. The similarity in the movement of production in this and other localities will appear at once on comparison of PL XXXIV. with PL XL, XXVIIL, and XXXII. The coincidence in the direction of the changes is precise in all of the 7 instances in the case of the river, in all but one for Quiver Lake, and in all but two in the case of Dogfish Lake. This is a period of max- imum overflow, when th» individuality of these several locali- 391 ties is submerged l\v the flood. It will be noted that the ex- ceptioi]s lie at the close of this period, when low water l)rings local factors into prominence. It is at this time also that the differences in the amplitude of production are most in evidence. 1898. (Table VII., PI. XXXIV.) Six collections made at the outlet of Flag Lake in this year in July-September for the purpose of detecting Troclio- spha-rn are introduced int:) J ::ble V^II., since they throw some additional light on productio.i ; e .. The four collections in July exhibit a very low level of plankton content, the highest being .62 cm.' per m.'. and the level is not raised in the single August collection. However, with the run-off of the slight rises of August and September we find a rise to 15.5-1:. At this time water of overflow was making its way from across the bottoms at the southern end of Thompson's Lake through the marshy swale into the foot of Flag, and thence out to the river. The similarity in the movement of production here and in other stations is seen in the general low level of production in July and the slight rise towards the end of the month. A de- cline early in August can also be traced, followed by a rise in the next month (,cf. PI. XXXIV., XII. and XXIX. ). " With the exception of the collection on Septemlier (i. the collections of this year indicate that the effluent of Flag Lake is a diluent of channel plankton. This may result from the low levels and consequent dominance of the vegetation in the lake at this time of low production there. SU.M.MARY. The data discussed in the preceding pages lead to the fol- lowing conclusions concerning Flag Lake. The average production of plankton in this lake, or, more properly speaking, mar-sh (11.46 cm.^ per m.'. or 9.23 on the basis of monthly averages), exceeds that in the river (2.19 or 2.71). This greater fertility appears not only in the averages 392 but in general throughout most of the seasonal changes. Its run-off therefore serves generally to enrich the channel waters. The greater production is due to the decay of the abun- dant vegetation which the lake contains, to the absence of trib- utary water of recent origin, to the relative freedom from the general current of overflow which largely takes the line of less resistance through Thompson's Lake (PI. II.), and, consequent- ly, to the greater time afforded for breeding an abundant plank- ton in this impounding area. The dominance of the abundant vegetation is inimical to large plankton production. Other things being equal, plankton production is greater when the relative occupancy of the water by vegetation is decreased. The movement in plankton production in this area is in the main similar to that in the river and in Quiver and Dog- fish lakes. Pulses of production tend to coincide, though their amplitude may differ widely in the several localities. This sim- ilarity is least when local environmental factors such as vege- tation, stagnation, or local exhaustion of the food supply are most potent. It is greatest when these are least, that is, during high water. STATION G, Thompson's lake. (Table VIII., PI. II., XX., XXXV.—XXXIX., L.) ENVIRONMENTAL CONDITIONS. This body of water lies in the bottom-lands on the right bank of the Illinois, above Spoon River, midway between the bluff and the main stream. It trends in a northerly and then a northeasterly direction, following somewhat the curve of the Illinois. It is about live miles in length at low water, with a width in three fourths of the distance of about two thirds of a mile, while the northern end is less than one third of a mile in width. At this stage it contains about 1,400 acres. As levels rise, its margins spread rapidly—owing to the slight gradient of the shores—northward to Grass and Slim Lakes, westward, through Mud Lake, towards the bluff, to the south, towards 393 Spoon River, while its connection with Flag Lake is eavly es- tablished over the low sandy bank which lies between them. At stages above six feet the "cut road" (PI. II.) and the marshy sw^ale above it fill, and connection with Flag Lake at its lower end and with the river is established. Its area is about doubled by the time the river reaches bank height and general overfiow ensues. The lake is of somewhat uniform depth in the middle half, but shoals toward either end. In prolonged low water, when the slough at the northern end is practically cut off from the lake, extensive mud-flats are exposed in the northern area, and a portion of the southern end for about half a mile is also left bare except when prolonged and heavy winds drive the water towards one or the other end of the lake. The depth at lowest river levels in the central region, which includes about two thirds of the total area, is 3 ft. Laterally the water does not shoal until within 10 rods of the shore. There is thus a large area (about two square miles) of water with uniform condi- tions in this particular. With slight exception the bottom is of the softest alluvial mud, several feet in depth, overlying a sandy blue clay. The shores along the southern, western, northern, and northeastern margins are also of soft alluvium and of a marshy character. The eastern shore, for most of its extent, and limited stretches along the western one, together with the spit which makes out into the lake on that side, are of sand and of a firmer consist- ency. All of the shores are bordered by a belt of vegetation, which has been described on page lJ46. This lake is the largest of the reservoir backwaters exam- ined by us, and is one of the permanent type, resembling in all important particulars except that of submergence in times of general overflow and its reservoir relation to the river a typ- ical lake of an alluvial prairie country. Its position in the bottom-lands brings it into intimate connection with the river, the source of most of its water supply, while at times of flood 394 its position is such that the backwaters from the bottom-lands up-stream sweep through it and out to the river through the "cut road," being deflected by the alluvial deposits of Spoon River (PI. II.)- Its relation to the river is a peculiar one in that its outlet, or "slough, lies at its up-stream end. At stages above six feet the current enters through this slough, and the run-off takes place at the lower end through the cut road. Below this level all the run-off must take place through the^ slough. The direction of the movement in the run-off of the lake is thus reversed as river levels pass this stage. There are no tributary waters of consequence which enter the lake, though a small rill and a few springs enter along the western margin. The main supply is drawn directly from channel waters through the slough, and when levels are stationary there is no in- terchange in either direction. The current sets in or out, at stages below 6 ft., according as the river rises or falls. The re- sult of this condition is that during the higher levels back- waters of overflow and the river water entering by the slough are impounded and drawn off slowly at the lower end of the lake. At stages below 6 ft. a run-off occurs only in the falling stages and in relatively smaller volume through the narrow and toi'tuous slough. The impounding function is accordingly more highly developed at lower levels, while at lowest levels all interchange ceases. This close and intimate dependence of this lake upon the river for its water supply in so far destroys the unity and inde- pendence of the lake as a separate unit of envii'onment, and tends to eliminate the dift'erences in plankton production be- tween it and channel waters. This tendency is counterbal- anced to a considerable degree by the large size of the lake and consequent increase in the time occupied in transit during over- flow, and by the impounding, at lower levels, of entering river water at the upper end of the lake, where it deposits its silt and soon permits the development of the lake plankton in its area. Here, as elsewhere, local factors are most potent at lower levels. The effect of the greater size of this lake is thus to equal- 395 ize envii'oumental fluctuations and to obviate their catastrophic results, which may be seen in their maximum violence in channel waters, and in a lesser degree in the lakes thus far ex- amined. COLLECTIONS. This station was opened June 7, 1 S94, and collections were continued until the close of operations on March 2S, 1899. In all, 99 collections were taken, distributed in the several years as follows: 5, 14, 27, 18, 25, and 7, with but few exceptions at approximately a monthly or fortnightly interval. It was only in the spring and summer of 189(5, when an interval of 7-10 days was adopted, that the interval is brief enough to enable us to trace the movement in production with any degree of fullness. At other seasons the data are suggestive, Imt not conclusive, of its course. The relatively smaller number of collections made at this important station is due to its distance from our center of operations, the round trip from Havana to the lower station in low-water conditions exceeding 25 miles. The difficulties of access were greatly increased when at low water it was necessary to make the trip from the outlet of the slough by rowboat, and to drag or push this over the soft mud and through the dense vegetation at the upper end of the lake, and when, in winter, at low water, the boat and outfit had to be dragged across the frozen bottom-lands. The locations at which collections have been made are principally the two marked on the map (PI. ID. The lower one was used exclusively in 1894 and 1895, and thereafter when access to the lake was had through the cut road. The location off Sand Point, at the upper end, was used when the lake was entered by way of the slough. Both were in the open central region, well out in the vegetation-free area, though in 1895 and 1896 the lower station was encroached upon some- what by shifting masses of Cmifoplii/llinn. In a few instances, owing to high southwest winds and the dragging of the waves in the shallow lake, it was not possible to maintain an anchoi-- 396 age in the unstable bottom, and refuge was bad under the lee shore, but still in usual depths and open water. In several in- stances in overflow stages, when the ice was too heavy to break and too light to carry our load, it was necessary to make the collection near the margin of the lake in effluent waters. These variations in the location of the point of collection introduce no error of consequence into the series, judging by the results of an examination of the local distribution of the plankton in this lake, the details of which cannot be given in the present paper. With the exception of the single pump collection on Feb- ruary 28, 1896, all collections prior to May 20 of that year were made by the oblique-haul method, and thereafter by the plank- ton pump. This lake is a type of the larger reservoir backwaters, such as Meredosia Lake, Clear Lake, and others found in the bottom- lands of the Illinois and maintaining a constant connection with that stream. An examination of its plankton content will therefore serve to throw light on the relation which lakes of this type bear to plankton production in channel waters. PLANKTON PRODUCTION. 1894. (Table VIII., PI. XXXV.) There are but 5 collections in this year, from June to De- cember, at an interval of a month or more, with an average production of 8.89 cni.^ per m." and a maximum of 24.92 on June 7. An inspection of the hydrograph (PI. XXXV.) of this year reveals the fact that only the first two collections were taken under conditions which permitted any run-off from the lake to the river, and both of them at times—that is, in falling levels below 6 ft.—when the run-off was largely, if not wholly, through the tortuous slough at the up-stream end of the lake. The pro- duction in the lake (24.92 and 10.74) at these times was 33- to 397 4-fold that in the river ( .74 and 2.39 1, so that the run-off at this season enriches channel plankton. With the exceptioQ of the August collection the plankton content in the other three col- lections in the lake exceeds that in the stream. The low con- tent in August (LOS) occurs at a time of lowest water, when vegetation bj- reason both of river stage and the season is at its maximum occupancy of the lake. At other times the effect of the reservoir function of the lake is seen in the relatively great- er production in its waters. The scattered data of this year are insufficient as a basis for any conclusions as to the correspondence in the movement of production in this and other waters. 1895. (Table VIII., PI. XXXVI.i There are 14 collections in this year, between April 1(1 and the end of the year, averaging 9.67 cm.' per m.\ and with a maximum of 61.44 on May 1—an amplitude nearly 11 -fold that of this pulse in the channel. The average production (9.67) is 3-fold greater than that of the river in this year (3.22), and the monthly averages (see table between pp. 342 and 343) are in 5 of the 9 months from 1.6- to 12-fold greater in the lake than in the river. In the remain- ing four months, June, July, August, and December, the ratios are respectively 30.42, 9.33, 4.03, and 1.14 (river), to 9.42, 4.83, 3.09, and 1.00. The lower production in June-August occurs at a time when, with the exception of three weeks, levels were low and vegetation at its maximum occupancy of the lake, and when, moreover, the current was greatly slackened in the river, and channel plankton in the richly fertilized waters had more than the usual time to breed, while the less production in the lake in December is, owing to the distribution of collections, more apparent than real. In the matter of individual collections on coincident or approximate dates the lake shows a greater plankton content in 9 out of the 14 instances, and of the 9 there are 5 in which 398 hydrographic conditions favor a run-off of this richer plankton of the lake into channel waters. There are two instances in which run-oft" occurs when lake waters are poorer than the channel, but they are both at low levels and during slow de- cline, so that the discharge and resulting diluent effect is but slight. Considering the average production, the times when run-off occurred, and the hydrographic conditions when the lake waters contained less than the channel, it is probable that even in this year Thompson's Lake, owing to its reservoir function' served predominantly to enrich the channel plankton. Though this relation predominated, the total contribution of the lake to the stream in this year was but slight owing to the hydro- graphic conditions. In the April-December period covered by our collections, the stage of river never exceeded 6 ft. until the December flood. There was, therefore, never any general cur- rent of overflow passing through the lake and carrying the im- pounded waters out from the southern end (PI. II.) into the riverandthusdischargingaconsiderable volume of plankton-rich water into the channel—a condition possible in both rising and falling levels above 6 ft. At the levels below this point which prevailed throughout this period, influx and efflux both can take place only through the slough at the northern end, so that con- tributions to the stream from the lake occur only during falling levels, and, moreover, owing to the tortuous course and clogged condition of the outlet, the volume discharged at these lower levels is very much less than at higher ones, across the broad outlet at the other end of the lake. Falling levels occurred in less than one half of the time in April-Decembei", so that the contributions of the lake to the river were not only slight in volume but limited in dm-ation and discontinuous. Collections were too infrequent to trace the movement in production with fullness or certainty. There are, however, a few suggestions of a similarity in the course of production here and elsewhere. The direction of the changes in the course of production in this lake and in the river in coincident or ap- proximate collections is the same in 9 out of the possible 13 in- 399 stances (cf. PI. IX. and XXXVI. ) ; in the case of Quiver Lake the agreements number 7 out of a possible 12 (ef. PI. XXVI. and XXXVI. ); in the production in Dogfish Lake the correspondence is found in 9 out of a possible 12 (cf. PI. XXX. and XXXVI.) ; while in Flag Lake there are 2 out of 3 (cf. PI. XXXIII. and XXXVI.). The agreement is lessened in this year, it seems, by the hy- drographic conditions. The low water affords less opportunity for a mingling of the waters of the stream and its backwaters, and also serves to bring out the local environments at each of the stations. Thus Thompson's Lake has but little connection with other backwaters at any time during the year, and ingress or egress of channel waters was but very slight during six months of the twelve in this year. Vegetation also gained more ex- tended possession of this lake in this year than in other seasons of our operations. Low v\'ater also tends to make the channel plankton more directly affected by its peculiar factors, such as sewage. It is noticeable that the agreement in production is most marked between Thompson's and Dogfish lakes, both back- waters of somewhat similar character in respect to tributary waters, relation to the channel, and vegetation. The most marked differences between production in this lake and the channel appear in the respective amplitudes of the pulses of production in April-May and June-July. In the lake the -rising vernal pulse attains the exceptional volume of 28.2 on April 10 to .52 in the river on the 9th, a difference which may in part be due to the earlier wanning up of the shoal- ef lake waters. The maximum (61.44) in the lake is 12- fold that observed in the stream. The June-July production in the river, on the other hand, is 3- to 5- fold that in the lake, the contrast being due on the one hand, it seems, to the temporary exhaustion either of the chemical substances utilized by the plankton or of the reproductive capacities of the planktonts of the lake waters, and, on the other, to the increased sewage con- tamination in the stream as a result of low levels. The direr- tion (if till' i-Ihiii(/cs in jiroihicfloii, however, remains the same in 400 both localities (cf. PI. IX. and XXXVI.) in the face of these contrasts in amplitude. 1896. (Table VIII., PI. XXXVII.) There were 27 collections in this year, at monthly inter- vals until April, and then every 5-11 days until the end of Au- gust, and thereafter every fortnight. The avei'age production in this year is 9 cui.^ per m.^ with a maximum at the vernal pulse on May 2 of 48.99 cm.' The hydrographic conditions of this year are such as to bring Thompson's Lake into intimate connection with chan- nel waters. The average height of the river for the year, 6.98 ft., is sufficient to maintain a run off from the southern end of the lake to the river, submerging the bottom-lands between to the depth of a foot. Indeed a run-off of varying depths was maintained for 241 days, in which stages exceeded 6 ft. This was due to the recurrence of 6 floods, so distributed as to keep the lake discharging through the southern outlet for 241 days with only 5 interruptions between May and December. Of the 125 days in which water did not flow through the lake from the northern to the southern end, there were 29 of rising water in which no discharge to the river occurred, 28 of stationary levels in which the movement of the water, if any, was declin- ing, and 68 of falling water, in which the lake discharged through the slough at the northern end. Thus, during 309 days of the year this lake was discharging to the channel, waters which had been impounded for a varying length of time within its boundaries. The importance of this impounding area is best shown by rough calculations which show that the run-off of a single foot from the lake proper, not including the expand- ing areas which join it with every rise in levels, will fill the channel of the river at Havana to a depth of 8 ft. (low-water stage) for about tJn-ci' miles. In 1896 the total depth of the run-off for the year computed on a single discharge after each 401 rise is 26.9 feet—sufficient to till the channel for 81 miles. When we add to this the consideration that at levels above 6 ft. water is continualh' pas.sing throiij^h the lake with brief im- pounding, the length of channel filled bj' the run-off of this area must l)e considerably extended. The relationship of plankton production in this lake to the plankton content in channel waters in this year may be in- ferred from the yearly averages. Thompson's Lake contained 9 cm." per m." to 1.16 cm. "* in the river. The net result would therefore l)e an enrichment of the channel plankton in a ratio dependent upon the relative volumes of the mingling waters. Xo quantitative statement of this ratio is possilde in the absence of data as to the run-off of Thompson's Lake. Not only is the net result an increase in the channel plankton, but the monthly averages (see table l^etween pp. 342 and 343) and the coincident or approximate individual collections (Tables III. and Vlll. ) in ercri/ iiistdiirt' exhibit a higher plankton con- tent in this lake than in channel waters. The monthly aver- ages range from 2 to 251 times greater in Thompson's Lake than in the river—ratios within which most, if not all, of those of individual collections fall. The data all indicate that this impounded water of the lake breeds a plankton whose run-off, without exception throughout this year, enriched channel waters. The effect of invading and plankton-poor river waters upon the plankton content of the lake is not conclusively apparent in the data, since we have also to deal with the phenomenon of pulse-like changes in plankton content which are combined with other factors in affecting the movements in production. It may be significant of the diluent action of invading river water that plankton content falls in the lake with the first en- trance of the May-June, the July-August, the October, and the November floods (PI. XXXVII.). The recovery in production follows promptly in each case with the impounding of the en- tering waters. Since, however, declines in content, as in June, July, and Augu.st, occur also when fiood waters are not enter- ing, we cannot conclude that the decline upon this entrance is 402 due solely and unequivocally to the diluent action of the in- vading waters, though their share in the phenomenon seems probable. I have previously called attention to the similarity in the movement in production in the several localities wherever col- lections were of frequency sufficient to permit the tracing of the fluctuations in production. The course of pi'oduction in Thompson's Lake in 1896 forms no exception to this similarity, though the parallelism is less precise than it is in some other in- stances. Thus the plankton content rises or falls together in Thompson's Lake and the Illinois River in 18 out of 26 instances of coincident or approximate collections ; in Thompson's and Dogfish lakes in 18 out of 26 instances ; in Thompson's and Flag lakes in 16 out of 25 cases ; and in Thompson's and Quiver lakes in 12 out of 25. The direction of the cliaiige thus agrees in a total of 64 out of 104 possible instances in the data. This is a some- what greater proportion of instances in agreement than chance would demand, and its significance is enhanced by the fact that the agreement with Thompson's Lake is greatest (64 and 69 per cent.) in the case of Flag and Dogfish lakes—impounding bodies similar to Thompson's Lake—and of the river (also 64 per cent.), which is in a measure and especially in this year a sum- mation of impounded backwaters. Quiver Lake, on the other hand, where tributary waters increase the local differentiation, has an agreement in only 12 out of 25 instances. In like manner months of high water, such as August, when local differences are to some extent submerged, exhibit greater agreement than monthsof low water, when they are emphasized. Thus in August (average river gage, 7.42 ft.) 92 per cent, of the changes in produc- tion are in agreement, while in July (average river gage, 4.55 ft.) only 58 per cent, exhibit this relation. Again, since the above comparisons ai'e based on coiticidence of changes in production it results that slight chronological dislocations of otherwise similar movements in production indicate a greater disagreement than really exists. This is especially true of the vernal pulses of April- June, where as a whole only 58 per cent, of the coincident or 403 approximate collections show this agreement. A comparison of Plates X., XXVII.. and XXXI. with XXXVIl.. will .show that much of thi.s disagreement is due to .slight variations in the positions of the apices of the several pulses in the different localities. In each locality we can trace three dimini.'^hing pulses in this period, pulses, moreover, which have much in common, barring variations in amplitude and time of culmi- nation. Their similaritj- is greater than the 58 percent, of agree- ment would seem to indicate. The most marked difference between the production in the river and in Thompson's Lake, as has been shown, lies in the amplitude of the pulses, which in the river never attain the height that they do in the lake. A part of this contrast is due to the fact that pulses of production are sometimes flushed out by floods in the channel while they continue to a normal cul- mination in lake waters, as. for example, the vernal pulse which culminates in the lake May 2. Similarly, in the flood of the last of May and July the plankton content is suddenly depleted in the channel waters, while the rising pulse continues to a later and much higher culmination in the lake. 1897. (Tdble\Ill , XIII.; Tl. X.X.WIII., L.) There are 18 collections in this year, at monthly intervals till July, and thei'eafter approximately every fortnight . The average annual production this year, 10.43 cm." per m." is the largest recorded for this body of water, and is due to the exces- sive development in the low-water period. August-Xovember. which reached an amplitude (35.35) over threefold that de- tected in the vernal pulse (10.38). (PL XXXVlll. i The hydrographic conditions are very different from those of the previous year, and change profoundly the relationship of the lake and river. As will be seen on Plate XXXVIII., the river levels were above 6 ft. from the beginning of the year until June 6, and thereafter from the 25th until July 15. a total of 175 days in which the lake received water throutrh the 404 slough at the northern end, impounded it for some time, and maintained a run-off at the southern end (PI. II.) of its plank- ton-rich waters. There are in addition 35 days in June, July, and August in which falling levels below 6 ft. afforded an op- portunity for a run -off through the slough at the northern end. Of the remaining 155 days, 10 are of rising levels below 6 ft., when the lake receives water from the river but does not dis- charge any into it, and 145 belong to the low-water period of the last 5 months, in which there was little interchange be- tween lake and river though the run-off continued in diminish- ing volume for a few days after stable levels were reached, early in August. About August 16 the channel discharge was so slight as not to float a rowboat in the narrow channel at the northern end of the lake, and connection with the river was not reestablished as the river rose in October-November until the level of 2.8 ft. was reached, Nov. 10. The slight fluctuations during the remainder of the year practically amount only to the reception of .4 ft. of water by the lake. For the last five months of the year—months of heavy plankton production in lake waters—there was no run-off to the river. On the average the lake produced this year 10.43 cm.^ per m.', about 3-fold that in channel waters (3.69 cin.^) and the net result of the run-off would be, it seems, an enrichment of chan- nel waters. The actual enrichment is, however, much less than these averages indicate. An examination of the monthly averages (see table between pp. 342 and 343) reveals the fact that the excessive production in the lake, when the plankton content rises to 5- to 16-fold that in the channel, appears in the low-water period when no run-off occurs. During the first 7 months, in which there is an almost continual run-oft', the production in lake waters is but 1.5- to 2-fold that in the chan- nel except in February and June, when flood waters in the latter increase the ratio to 1 to 7 and 13 respectively. The in- sufficiency of the collections in this period leaves in doubt the amplitude of the vernal pulse. The April and May collec- tions indicate only a low level of production as compared with 405 that in other j'ears, and this also tends to lower the relative productiveness in the lake. It is evident that the seasonal dis- triltntion of the period of Hooil waters and the resulting im- pounding function of the lake affect greatly its contributions to channel plankton. In this year Hood waters ai-e largely confined to the colder and less productive season, when the run-off contains little plankton and its contrilnitions are small, while in 1897 recurrent fioods throughout the year afforded a run-off in seasons of larger production, and this tended to greatly increase the enrichment of channel waters in that j'ear as compared with 1897. Plankton content in 1897 in Thompson's Lake was in ex- cess of that in the river in the case of coincident or approxi- mate collections in Ifi of the 18 instances, the two exceptions appearing in July and September, when pulses in channel plank- ton rise above the recorded production in the lake as a result of some undetermined factor. The similarity in the course of plankton production here and elsewhere is most marked in the first part of the year-, and decreases in the time of low water. Thus, on comparison of the planktographs of Thompson's Lake I PL XXXVIII. ) and the Illinois River (PL XI.) we find 14 out of 18 changes in the course of production coincident iu the two regions, the four ex- ceptions occurring in May ( U, July (2), and September (1). The environmental differences between Thompson's Lake and the river are much less than between this lake and Quiver, and we find a corresponding disagreement in their planktographs, only 10 out of 18 changes being in the same direction, and six of the ten are in the period of high water, when local differen- ces are submerged. In the cases of Flag and Dogfish lakes col- lections extend only to July, with agreement in 5 cases in each out of a possible 7 and respectively. In the year as a whole and including all the above localities we find 34 agreements to 15 exceptions, in January-June the ratio being 3 to 21 for 4 lo- calities, and in July-December, in low water conditions, 12 to 13 for from 2 to 3 localities. The effect of the common elements 406 of the environment which high water introduces, in unifying the course of plankton production in their several areas, and of low water in diversifying it, is well demonstrated by these compar- isons. In the planktographs of Thompson's Lake and the Illinois River there is a striking general agreement in the low vernal production and the increased and unusual autumnal production. There are also some indications of a pulse-like character of the planktograph in the lake, though the collections are too infre- quent to demonstrate it. 1898. (Table VIII., XIL; PI. XXXIX., L.) There were 25 collections in this year, at fortnightly inter- vals, with an average of 5.71 cm.' per m.^ to 2.13 cm.' in channel waters. The net result of the run-off from Thompson's lake in this year is thus an enrichment of the plankton of channel waters. This is true for all of the monthly averages (see table between pp. 342 and 343) with the exception of April, and this exception is due solely to the distribution of collections on the rising vernal pulse, and is more apparent than real (cf. PI. Xll. and XXXIX.). The relative plankton content in the two areas, as will be seen on a comparison of the planktographs, is not subject to great variations in this year aside from January, when the ratio of the lake to the river is 1 to 17, and, as above noted, in April, when the ratio apparently falls to 1 to .6. With these exceptions, it ranges in the first six months from 1 to 2-3 and in the last six from 1 to 3-5. These figures express quanti- tatively the striking similarity in the planktographs of the two areas, which may also be recognized at once in the plates (XII. and XXXIX.) in the low winter production, in the meteoric ver- nal pulse followed by a minor one in June, and in a low level of production during the remainder of the year with fluctuations within rather narrow limits. The cause of this close resemblance lies in the hydro- graphic conditions, which throughout this year favor constant 407 interchange between lake and river. The average height for this year is 8.02 ft., the highest during our years of record. From Jan. 22 to July 15 river levels were above 6 ft., and a constant inflow of impounded water from bottom-lands above the lake, or through the slough when overflow ceased, continued with impounding in the lake and subsequent discharge from its southern end to the channel. The same conditions again prevailed from Oct. 30 till the end of the year, with an inter- ruption of 6 days in December. During the remaining parts of the year there was a constant wavering in levels which fa- vored frequent—in fact, no less than 21—reversals in the direc- tion of flow in the slough connecting the lake with the river. During the 134 days of low water there were 56 of falling levels in which the lake was discharging its plankton-laden water through the slough to the river, making a total of 287 in which it contributes to channel plankton to 78 in which, owing to low levels, it merely receives an inflow from the river. Moreover, the periods of greatest plankton production in the lake, during the vernal pulse, occur at times when the run-off from the lake is at its height, so that in this year all the hydrographic factors combine with the distribution of the plankton production to render this reservoir lake a feeder of the channel plankton. Though the differences in the plankton content are such that the actual enrichment per cubic meter may be less than in other yeai's, the total run-off of plankton into the channel must com- pare favorably with that in any other year of our opei'ations. The comparison of coincident collections shows in all cases but three, a greyer plankton content in the lake than in river. The first of these is on April 5, at the height of the spring flood, when a considerable current sweeps through Thompson's Lake and shortens the period of impounding, and thus reduces the time for the development of the plankton. The second instance is on June 21, on the decline of the acces- sory vernal pulse, which reaches a lower level in the lake (2.47) than in the river (2.88). This is one phase of a not un- common phenomenon in the plankton pulses of the backwaters. 408 They have greater amplitudes, but are frequently followed by more sudden and complete declines. Thus, in this case the apex of the pulse is at 18.39 and 6.99 cm.'' respectively in the lake and river on June 7 and 14, v\rhile the decline has reached 2.47 and 2.88 on the 21st in the two localities—a fall of 86 and 59 per cent, respectively. The third instance occurs on Aug. 16 (lake, .45, river, .61), when a large silt content in Thompson's Lake, due to roiling of the water by heavy wind, obscures the actual quantity of the plankton. The similarity in the direction of the changes in plankton content in Thompson's Lake and the other localities continues in this year even to a greater degree than formerly, owing in part at least to the hydrographic conditions above noted and to the more complete and uniform records. In the case of Thompson's Lake and the river there ai-e 21 agreements in the direction of the changes to 4 exceptions, and in the records of Quiver Lake 22 to 3 in the possible 25. This is so far in excess of the degree of agreement demanded by chance that we may look with confidence for an efficient cause in the common fac- tors of the environment, in the similar reproductive cycles of the constituents of the plankton found in common in the sev- eral localities, and in the uniformity in the reactions of at least a predominant portion of the total plankton assemblage to the factors of the environment. The river levels average 8.01 ft. for the year and stood above 6 ft. for 8 months of the 12. The high water increases the area of the "open water," and causes a retreat of the shore-line and bottom, and a decrease in the relative ^occupancy of the bodies of water in question by the spheres of influence of the immediate environment. Thus the local differentiating char- acters of the several environments are in general progressively less potent as the open water increases in extent. The loca- tions of the 7 exceptions to the similarity in the direction of the movement in production are significant. All of them lie in the last five months, in the period of low water, and 2 in the lowest water in August, when local influences are more potent. 409 1899. (Table \1II., XII.; PI. XXXIX., L.) There are 7 collections at fortnightly intervals in the fii'st 3 months of the year, with an average of 1.21 cm.' per m.' to .41 in channel waters. With the exception of four days in Feliru- ary, river levels were above G ft. throughout the period, and con- sequently the lake was continually receiving water at the north- ern end and discharging at the southern, and contributing throughout the whole time, in this way or through the slough, to channel waters. The average result is an enrichment of the plankton of channel waters. The monthly averages (see table between pp. 342 and 343) in January and Fel)ruary in the lake exceed those in the channel by 9- and 2-fold respectively, while those of March, in highest flood waters, are respectively .28 and .21 cm.', owing, as will be seen on a comparison of Plates XIII. and XXXIX., to the distribution of the dates rather than to an actual smaller production. A comparison of all coinci- dent collections in lake and river exhibits likewise a larger plankton content in every instance in lake waters. The lake thus tends continually during this period to enrich by its run- off the plankton content of channel waters. The similarity in the movement in production noted in liS9S is interrupted in these winter months by dislocations of the apices of the slight pulses of production, due in part to the flushing action of sudden floods and its unequal distribution in channel and backwaters. Of 7 possible agreements in the di- rection of movement in production there are but 3 realized in the case of both the river and Quiver Lake. A comparison of the three plates (XIII., XXIX., and XXXIX. ) will, however, show that all, in common, exhibit evidences of a January and a February pulse and a common March decline. SUMMARY. The grand average of all the Thompson's Lake collections shows a plankton content of 7.94 cm.'' per m.-' in comparison with 410 2.19 cm.' for channel waters. The relative fertility of each is perhaps better expressed by the average of the monthly aver- ages, 8.26 and 2.71 respectively. The rnn-off of the impounded lake waters would thus tend to enrich the plankton content of the channel in some ratio dependent upon the relative vol- umes and plankton contents of the mingling waters. We have also seen that the enriching function of the contributions of this lake is continuous throughout a large part of the year, with a few interruptions dependent upon cessation of run-off in rising levels in low-water periods, and, rarely, to a lower plankton content in lake waters, due generally to increase of plankton in channel waters as the current slackens in low river stages. The following comparison of the averages of the monthly averages for the years of our operations, taken from the table between pages 342 and 343, is instructive in indicating the vary- ing relation of production in lake and channel waters. COMPARISON OF MONTHLY PRODUCTION IN THOMPSON'S LAKE AND ILLINOIS RIVER. 411 This is the time of greatest contrast, and also the time of highest levels (averaging 9.44 ft.) and therefore of continuous and largest run-off. It is also the season of largest plankton production, averaging 10.42 cm." per m.',—8.4 times that in channel waters,—and is accordingly the period of greatest en- richment of the channel plankton by the run-off from the lalve. The factors operative in producing this result are the high levels, with resulting increase in the impounded waters of the lake at a season of ri.sing temperatures favoral)le to plankton production and to the enrichment of the waters l)y decay of the vegetation of the previous year's growth. In June-September we have a period of falling levels, maximum temperatures, lowest water, and growth predominat- ing over decay in the aijuatic vegetation, which is relatively more abundant in the lake than in the river. It is therefore the season of greatest predominance of local environmental factors, and of run-oft' reduced to minimum volume and fre- quently interrupted. It is also the .season of least plankton production, averaging 6.74 cm.' per m."—only 1.54-fold that in channel waters. The midsummer sea.son is therefore one of least enrichment of channel plankton, as a result of both the decreased and interrupted run-off and the decrease in the relative production in the lake. This latter feature results both fi'oni the decline in production in the lake and the low-water condi- tions in the river, where increased fertilization by sewage and slackened current tend to raise its level of production at this .sea.son. Other factors tending to bring about the conditions of production prevalent in this season are possibly the greater relative exhaustionof the fertility in lake waters during the mid- summer and low water, as indicated iu our chemical analyses by the generally lower level of the various forms of nitrogen in the lake than in the river. (Cf. on this point Plates XLV. and L.) This greater relative exhaustion maybe attributed in part only to reduced interchange of river and lake waters at low levels and consequent reduction in intlux of sewage from the channel, and to the utilization of some of the constituents which support 41 '2 the phytoplankton by the rapidly growing aquatic vegetation. These factors are not, however, potent enough to overcome the effect of impounding and consequent time for breeding which prevail in the lake more than in the river, and thus to lower the plankton production in the lake below that in the channel. In October-December we find another season marked by rising water but not high levels, in fact, averaging only 4.22 ft.—a level insufficient to provide for any current through the lake or any considerable discharge in periods of decline. It is thus a season of slight and interrupted run-off. It is, however, a period of increased production, reaching 10.64 in October, declining to 3.08 in December, with an average of 6.70—a trifle below that of the midsummer period. Its relation to channel production changes decidedly, rising from a ratio of 1 to 1.54 in midsummer to 1 to 5. This five-fold greater plankton con- tent in Thompson's Lake makes whatever run-off occurs of con- siderable enriching effect upon channel plankton, though pre- vailing low levels and large proportion of rising levels tend to reduce the actual volume contributed in this season. The fac- tors operative in increasing the rehdive production in lake waters in this season are the influx of sewage-laden river water, and the decay of some of the succulent vegetation of the lake and its re-submerged margins at a season of plank- ton pulses of an amplitude increasing by virtue of other fac- tors, internal or external. Rising levels also bring about an increase in current in the channel, while marked changes in the bacteriological and chemical condition of channel, waters attend this and the fall in temperature. The com- bined effect of these factors, as shown by a comparison of the records of 1897 (PI. XI.)—when low levels continued and the autumnal decline in temperature was late—with those of other years, is to depress pi'oduction in channel waters more than it falls in the lake. This fact, together with the increase in the impounding function of the latter as levels rise, suffices to bring about the increased relatioe production in lake waters in the closing months of the year. 413 Thompson's Lake and pi-esumablj^ other bottDin-latid wa- ters of similar character, by virtue of their impounding func- tion, are reservoirs in which flood waters are stored for a great- er or less time, permitting the development in general at all seasons of the year of a plankton exceeiling in volume from l+-fold to 18- fold that coincideutally developed in channel wat&rs of the adjacent river. The run-off froin this and like areas elsewhere thus serves to enrich and maintain the river plankton proper. The slightly developed flood-plain of the Illi- nois and the consequent considerable area of such bottom-land waters—which equalize the floods, prolong the run-off. and favor the production of an abundant plankton in the impounded areas —become, accordiagly, factors of great importance in causing the richness, abundance, variety, and long continuance of the unusual plankton production of the Illinois River. The similarity in the course of plankton production in Thomp.son's Lake and elsewhere in our held of operations is shown in the following tabular summary, which gives the num- ber of instances of agreement and disagreement in the direc- tion of the changes in production in the four localities. SIMILARITY IN DIRECTION OF CHANGE IN PROUUCTlllN IN THOM I'.sO.N S LAKE AND AT OTHER STATIONS. 414 also years of higher water, of greater uniformity of environment —because of greater extent of open water, of greater interchange of water in overflow stages, and therefore of greater agi-eement in the course of plankton production. The similarity in the course of plankton production in different bodies of water is in a large measure a function of the similarity of their environ- ment and the resemblance of their planktons in the matter of constituent organisms. STATION F, PHELPS LAKE. (Table IX.; PI. XXL, XL.,-XLI1.) ENVIRON.VIE.STAL CONDITIONS. This body of water lies on the western side of the river about a mile below the city of Havana, in the elevated bottom-lands below the month of Spoon River. It trends northeast and southwest for a distance of seven eighths of a mile, has a width of 400-690 feet, and a total area of 50-60 acres. Its bottom lies about 6.5 feet above low-water mark, and the greatest depth re- corded in it at high water at the point of collection was only 10 ft. It is but slightly deeper toward the lower end. Its out- let is by a tortuous slough choked with driftwood, which runs for two fifths of a mile in a southerly direction to the river. The elevation of the bottom of this slough at its entrance to the river is 8-9 ft. above low-water mark, so that all run-off from the river drained by this slough ceases when it drains to this level, and it is not reinvaded by floods below this elevation. When the river falls below the level of the outlet and the lake drains as fully a.s the outlet permits, there still remain about 1.5 ft. of water from which no fui'ther run-off occurs. The vol- ume is then slowly reduced by evaporation or increased by sum- mer rains. The lake is not fed by springs or tributaries of any sort be- yond seepage from the level alluvial bottoms in which it lies, and which nowhere in the vicinity rise more than 10 feet above its bottom and generally very much less than this distance. At river stages of 11 ft. and above, backwater from Spoon River 415 makes its way through a now abandoned channel to the lake and thence out to the river through the slough. Below this level, the current of the gentle run-off of the great tract of adja- cent impounded backwaters with which this lake has then but a slight connection is the only movement in the area. The surrounding bottoms are heavily wooded for a narrow margin along the lake, though the forest giv^es way to cultivated fields on both sides within a short distance. Its bottoms and shores are of a rich black alluvium, which in low-water seasons such as 1895 beconies the soil of a cultivated field. The vegetation of this area is unique among our plankton stations in its character and relation to the plankton. In 1894 there was little vegetation present, and wdiatever aquatic growth had gained a foothold was eradicated by the dry au- tumn and by the cultivation of the soil in 1895. In 1896-1899 the occupation of the lake by water was more continuous, and Pofiiiii()(/('tt)iis, Naias, and even Ni'htmb:), gained a slight foothold along the margins. The principal vegetation was a dense mat of filamentous green alga?, such as Spiroiji/ra and Zi/ijiifiiia, which covered the margins for a considerable distance into the lake. During the heated term of midsummer a dense felt of Oaci/ldria covered the bottom of the lake everywhere at times. These algie were present during most of the summer, though most abundant in spring, and by their continuous and prompt decay they release into the lake waters a volume of nitrogenous and other substances which are utilized by the phytoplanktou. The cumulative action of the longer-lived aquatic phanei'Ogams in withdrawing from the lake large stores of food which are again released in the ensuing autumn or spring by the decay of the season's growth, is thus quite al)sent from this body of wa- ter. The rapidly growing and rapidly decaying algae permit a repeated Hux of nitrogenous and other substances utilized by the plankton as food in the course of a single season. This fac- tor, combined with the complete impounding function of this lake below river levels of 8 to 9 feet and the absence of tributary and spring water, is, I believe, the secret of the unusual plank- ton production in this area. 416 The absence of coarse vegetation, the sheltered situation in a rift in a dense forest, and the shallowness of the whole lake during much of the summer, permit an unusual range of diurnal temperature-changes falling but a few degrees short of the diurnal range in the air. The records (Table IX.) fre- quently contain readings of 90° to 95° in the summer season. This lake swarmed with the fry of various native fishes and the ' introduced German carp, all of which make great inroads upon the vernal plankton. It was also the favorite haunt of many fish-loving water-fowl. This abundant animal life served in turn to enrich the lake waters with its nitrogenous wastes, at once available for utilization by the phytoplankton. There are thus many chains of food relations in this lake, in most of which, if not, indeed, in all, the plankton forms many links. COLLECTIONS. There is a total of 67 collections from this lake ; 1 only in 1894, 29 in 1896, 9 in 1897, 22 in 1898, and (> in 1899. The single collection in 1894 was made by the oblique-haul method. The absence of collections in 1895 is accounted for by the fact that there was no water in the lake in that year. In 1896-1899 collections were made in various ways according to the conditions of access to the lake and the depth of the water. Owing to rafts of driftwood, access at any season through the slough is prevented. At high flood-levels, when ice did not prevent, it was possible to enter the lake by boat with our plankton pump and usual collecting apparatus. At all other seasons access by boat was impossible, 'and apparatus had to be carried across fields and through dense under- brush to the lake, and collections made by wading out into the lake or from a staging carried out from the shore for some dis- tance over the water. There are accordingly but 9 pump col- lections. The remaining 57 were all made by dipping water from the surface and pouring it through the plankton net. Most of the collections represent, therefore, surface waters, but owing to the exceedingly shoal water they are, nevertheless, in 417 tlie main representative of the plankton of the lake. Of the 67 collectious, 32 were taken from water which in the deepest part of the neighborhood of collections was less than 14 inches. Owing to the roiling of the water caused by our movements, it was necessary to dip from considerable areas in order to secure the desired volume for straining. Since a separate report on these collections is being pre- pared I shall only deal in this connection with those aspects of the data most intimately connected with the phenomena of the channel plankton. PLANKTON PRODUCTION. 1894. (Table l.\.) A single collection on June S yields a volume of 24.17 cm.' per m.^ an amount 7-fold that of the same season in 1890, al- mo.st equaled in 1897, and more than doubled in 1898. 1896. (Table IX., PI. .\L.l There are 29 collections in this year, extending from Jan. 8 to Nov. 17—when only a few scattered pools remained. This is the mo.st fully represented year of our series in this lake. The yearly average is 18.17 cm.' per m.'', with a vernal maxi- mum of 54.80 on Apr. 16, and an autumnal one of 51.60 on Oct. 15. This is the earliest vernal pulse recorded in our work, and should be correlated with the early rise in temperature in these shoal and protected waters. Thus, in Phelps Lake on Apr. 16 the surface temperature was 77° and the average for April 68.4° to 71" (Apr. 17) and 62.6° (average') in Quiver Lake, and to 66.3° (Apr. 17) and 68.2" in the Illinois Kiver. This lake was thus apparently 5° to 6° warmer on the average than these other localities, and the vernal pulse is accordingly accelerated. Owing to the elevations of the lake and its outlet, run-off from this area into channel waters practically ceases when river levels fall below 8 ft. In 1896 there were but 114 days of 418 stages above 8 ft., 90 from Jan. 1 to Mar. 30, and 14 and 10 re- spectively in the floods of May-June, and August. Of the 114 days there were 28 of levels above 10 ft., when, owing to run-oif from Spoon River, a current passes through the lake to the river regardless, as a rule, of rising or falling water. In addi- tion there were 43 days of falling water when a run-off might be expected, making a total of only 71 days in this year in which there was any run-off to channel waters from this lake. The remaining 43 days of levels above 8 ft. were times of sta- tionary (20) or rising (23) water, when discharge from Phelps Lake was diminished or cut off. Of the 71 days of discharge, 61 fall in the winter, in Janua- ry-March, and 5 each in June and August, at times of depres- sion in production (PI. XL.). Nevertheless, the plankton con- tent in Phelps Lake at all of these times greatly exceeds that in channel waters. The ratio of Illinois River and Phelps Lake plankton in January is 1 to 189, in February, 1 to 607, in March, 1 to 274, in the June flood, 1 to 4, and in the August rise, 1 to 7. These latter ratios are somewhat exceeded by those of the av- erage production for the year, 1.16 to 13.17 cm.^ or 1 to 11. During the months of little or no discharge, April-December, production in the lake as shown in monthly averages is 4- to 1600-fold greater in Phelps than in the Illinois, the latter figure being reached in November and the other months averaging only 11-fold. Thus, this lake contributed to the enrichment of channel plankton for a relatively brief part of the year, and at all times produced a plankton greatly in excess of that in channel wa- ters. The sharp contrast between the poverty of channel wa- ters and the wealth of this lake is due to the impounding func- tion in the latter, and to the repeated flushings by storm waters of recent origin in the former. Full time for the normal utili- zation of the resources for growth of the plankton is permitted in the lake but not realized in the constantly replaced river water. This is the only year in which collections were made in 419 Phelps Lake at weekly intervals for any length of time. Such intervals extend from the end of March to the last of August, and since similar series were made elsewhere we have an excep- tional basis for comparison of the course of production in the several localities. A comparison of the planktographs of the river and Phelps Lake (PI. X. and XL.) reveals certain general similarities. These are expressed in the three major fluctuations of the year, the vernal, midsummer, and autumnal rises in production, which, in the main, have coincident limits, but very divergent ampli- tudes, in the two bodies of water. The absence of the sudden diluent action of flood waters is noticeable in Phelps Lake rec- ords, though declines in plankton content coincide with the flood invasions of both June and August. The absence of col- lections in September in Phelps Lake at the time of the decline in plankton content between the midsummer and autumnal rises is due to the very low stage of water in the lake, per- mitting no collection. Replenishment by autumn rains is followed by the large development in October (51.6 cm.'). Not only is this general similarity between the movement in plankton production in Phelps Lake and the Illinois River traceable in 1896, but there is a more detailed agreement in the changes in the direction of movement in production in coinci- dent or approximate collections. This is most clearly seen in the months of April to August, when collections are of sufficient frequency to trace with some certainty the course of produc- tion. During these five months there are 16 agreements in the direction of the change out of a possible 20—a total of 80 per cent.—between Phelps Lake and the Illinois River. In the year as a whole the agreements number 19, or 68 per cent., out of a possible 28. Two of the exceptions in April-May are due to the dislocation of the vernal pulses in consequence of the higher temperatures in the lake above noted. When we take into con- sideration the marked differences in the local environment of the plankton in these two localities and the considerable inde- pendence of this lake as contrasted with other reservoir lakes, I 420 such as Thompson's and Quiver, this marked degree of resem- blance is the more striking. A comparison of the course of production in Phelps Lake (PI. XL.) and Thompson's Lake (PI. XXXVII.) in 1896 reveals 14 agreements out of a possible 25, or 56 per cent. Of the 11 disagreements 3 fall in the period of few^ collections in the au- tumn months, when Phelps Lake was reduced to shallow pools, and 4 occur during the vernal pulse of April-May. A compar- ison of the planktographs and thermographs of the two lakes shows that in Phelps Lake the temperature is from 2° to 8° higher than in Thompson's Lake for a period of six weeks dur- ing the rise of the vernal pulse. Hence this culminates earlier by a fortnight in the former, and in consequence a dislocation of the course of production in the two lakes occurs in this peri- od. The two planktographs are, however, strikingly alike in the fact that in the interval between March 31 and June 1 there are three pulses of regularly decreasing amplitude in both lakes. The similarity is thus greater than the percentage indicates. A comparison of the course of production in Phelps and Quiver lakes (PL XL. and XXVII.) reveals 15 agreements out of a possible 28, or 54 per cent.—but little more than chance de- mands. In this case the environmental diiSerences are greater, the effect of spring water, vegetation, and Hood invasion inter- fering in Quiver Lake with the course of production. A comparison with Dogfish Lake, where the disturbing fac- tors of spring water and flood invasion are less immediate in their action, reveals a slightly greater degree of similarity—16 out of 28, or 57 per cent. In the case of Flag Lake the agreement is still less, being only 11 out of 24, or 46 per cent. It is a noticeable fact that the disagreements are most numerous in Quiver, Dogfish, and Flag lakes, all of which are rich in vegetation, and these disa- greements occur in greatest proportion during the months of May-August, when with changing river stages the proportional occupation of these lakes by vegetation fluctuates greatly — 421 a variable factor from which Phelps Lake is to a large extent exempt. The degree of agreements, as a whole, in production between Phelps Lake and other localities is seen in a total of 84 instances out of a possible 141. or 60 per cent. ls<»7. (Table IX., PI. XLI.) There are only 9 collections in this j-ear. at approximately monthly intervals with the exception of the last collections, when the interval was somewhat reduced. Collections cease in August, when the water entirely disappeared from the lake. The average plankton content for the year is 10 cm.'' per m.\ the lowest annual average in which summer collections are in- cluded in this lake. The vernal pulse was not detected, if pres- ent, and the maximum record, 29.94 cm.'', was on Aug. 26, the date of the last collection. The hydrographic conditions were such (PI. XLI.) that a current from the tlooded l)ottoni-lands about Spoon Riv^er passed through the lake uninterruptedly from Jan. 7 to May 13, and the run-otf of impounded waters continued until the 25th, a total of 139 days. Throughout the period covered by our few collections they indicate that the plankton content of this area exceeded that in channel waters by from 1.6- to 11-fold with the single exception of Apr. 27, when the lake had 4.26 cm.'' to 5.11 in the channel. During the 139 days of run-off the production in the lake scarcely exceeds 5- fold that in the channel, but when discharge ceases the content rises to 10- to 11-fold that in the river—a phenomenon which illustrates the equalizing effect of general overflow on the one ha nd, and the effect of impounding in increasing production on the other. The run-off' from this lake in 1897 thus predominantly served to enrich channel plankton. The fact that produc- tion in Phelps Lake falls l)elow that of the channel on Apr. 27, when a vernal pulse might be expected of an amplitude greater in backwaters than in channel,—as indeed it is in Thompson's (PL XXXVIIL), Quiver (PI. XXVIIL), Dogfish (XXXIL), and 422 Flag (XXXIV.) lakes by 50 to 150 per cent.,—is to be attributed to the diluent effect of invading flood-waters from Spoon River, whose plankton content on Apr. 27 was only .05 cm." per m.' The entrance of these flood waters, indicated by the check in the decline of the hydrograph (PI. XLI.), was noticeable at the station on the day of collection, and is the cause of the increased turbidity ifTPhelps Lake on that day (Table IX.). The similarity in the movement of production in Phelps Lake in 1897 to that in the other bodies of water examined by us, is very close. In the case of the Illinois River, 8 out 9 pos- sible instances, or 89 per cent., are in agreement ; in Thomp- son's Lake 7 out of 9, or 78 per cent.; in Flag Lake all in- stances are in agreement; in Quiver Lake 6 out of 9, or 67 per cent. ; and in Dogfish Lake 4 out of 6, or 67 per cent. As a whole, 32 out of 40, or 80 per cent., of the changes in the direction of production in Phelps Lake accompany similar changes in di- rection in these other localities. All of the 8 exceptions to this agreement occur at levels below 8 ft., when local environments are more potent, and 5 of the 8 are found in Quiver and Dogfish lakes, where vegetation and access of tributary waters become proportionately more or less potent as levels fall or rise in May to August, when the 5 exceptions occur. This unusual degree of agreement in 1897 must be attrib- uted in large part to the hydrograph ic conditions in the period of comparison. For almost 5 months of the year levels were above 8 ft., when fluctuations have relatively but a slight effect on the various environments. Above this level the several lo- calities are more or less submerged in the general overflow, and all share alike in the wide stretches of open water in which some current exists, and the commingling to an increasing extent, as levels rise, equalizes and oblitei'ates local differences in pro- duction. The first 4 collections of this year were made under such conditions, and agree without exception in the course of production. The remaining 5 were taken at stages below 8 ft., in the very midst of the season of local diversification, and the proportion of agreements falls from 100 to 60 per cent., and is 428 greatest in the most differentiated localities, Quiver and Dog- fish lakes. It might also have been expected in Flag Lake (PI. XXXIV.) if examination had continued there beyond the mid- dle of July. It would seem, accordingly, that similarity in the course of plankton production in different localities is to a large extent a function of the community of environmental factors, and possibly also of the similarity of the constituent organisms. 1898. (Table IX., PI. XLII.) There are 22 collections in this year, at fortnightly inter- vals, in March-December. The average production for the year is the unsurpassed amount of 44.08 cm.'' per m.\ with a maxi- mum, also unsurpassed in our records, of 224.48 on Aug. 28. An unusually high level of production is also maintained from Aug. 9 to Dec. 13, averaging 63.54 cm.', and falling below 30 in but two instances. Water re-entered Phelps Lake with the flood of February, river stages passing 8 ft. on the 12th and 10 ft. on the 20th. Fi-om this latter date until June 28, with the exception of 4 days in May, levels continued al)ove 10 ft., so that a continued current of overflow from the bottom-lands to the north passed through the lake to the river. Declining river stages continued from June 28, pa.ssing 8 ft. July 9, thus permitting a run-off for a total of 138 days from the lake to the river in the period of spring and summer floods. To this must be added 7 days of de- clining levels above 8 ft. in the November-December rise, mak- ing a total of 145 days of contributions to channel waters from this lake. This is the most extended period of contribution in the years of our operations, and is a result of the unusually high and prolonged floods which l)rought the average height of the river up to 8.02 ft., almost bank height, for the year. The plankton content of Phelps Lake waters in the 10 col- lections made during the period of discharge above noted, is in excess of that in the channel in 7 instances by from 1.4- to 15- fold 424 and averages 7.3. In three instances, March Sand 1, (.01 and .02), March 29 (.20 and .43), and April 26 (10.72 and 15.81) the lake contains less than the river. All of these instances fall at times of high levels, exceeding 11 ft., when Spoon River floods invade this territory, and this deficiency in Phelps Lake is doubt- less due to their diluent effect. Since our station for collections was located in the upper end of the lake (PL II.), the full effect of the flood would be detected at this point, but would be di- minished by mingling with the lake waters and the adjacent impounded backwaters before it joined the channel. The first of these exceptions, on March 3, is not accompanied by increased turbidity (.45) in the lake (Table IX.), but the other two, March 29 and April 26, are attended by a marked rise in turbidity (.05 and .16). During this period of maxiuium spring flood in March and April, owing doubtless to this diluent action of Spoon River, the run-off from this area, as indicated by plankton content at the upper end of the lake, dilutes, or but slightly enriches, the channel plankton. This appears in the mouthy averages (table following p. 342), which for March are .33 cm.^ for the river and but .25 for the lake. In April they are 4.4 and 5.6 respectively. In later months, during the declines of the spring flood, and owing to absence of the flushing action of Spoon River floods and to the rise in impounding function with decline in levels and delimitation of the lake, we find a rapid rise in the relative plankton content in lake waters. The production in coincident collections is greater in the lake than in the river by 3- to 15- fold, and the monthly averages for lake and river respectively rise to 40.44 cm.' per m.^ and 11.30 in May; to 27.67 and 3.96 in June; and to 6.97 and .58 in July; that is, the production is from 3+ - to 12- fold greater at this season in the lake than in the river. During the run-off in these months this lake and its contributing adjacent bottom-lands serve to increase, in some unknown ratio dependent on their relative volumes, the plankton content of the channel waters with which they mingle. Although the frequency of the plankton collections is in- 425 sufficient to trace with accuracy the course of production in Phelps Lake in this year, they yield many suggestions of recur- rent pulses of production similar in duration, though of greater amplitude, to those more clearly dehned in channel waters. A comparison of Plates XII. and XLII. will indicate the presence of pulses of production in hotli localities, culminating in the ma- jority of instances at approximately monthly intervals. There are eight such culminations in Plielps Lake visible in the rec- ords of March-December, culminating in March, May. June, July, August, September, November, and December. In spite of the disparity in the records in this lake and the river, the similarity in the location of the pulses in the two localities is apparent in all of the above months but August and November —both of which are months of unusual hydrographic disturb- ances in channel waters. A detailed comparison in the movement in production in this lake and the adjacent river shows agreement in the direc- tion of movement in 14 out of 21 possible instances,or67 percent., 5 of the 7 exceptions falling in the hydrographic disturbances in August-September and November. In the case of Quiver and Thompson's lakesthe problem of comparison is madedifficult be- cause the fortnightly collections in Phelps Lake and these local- ities are not upon coincident, but alternate, weeks, and makes the the similarity or difference probable rather than precise. A comi)arison shows 16 agreements out of a possible 20, or SO per cent., in the case of Thompson's Lake, and 17 out of 20, or 85 per cent., in the case of Quiver Lake. Both of these lakes are af- fected by hj-drographic changes at lower levels which do not disturl) Phelps Lake, and we hnd that (5 of the 7 exceptions oc- cur in the period of floods at low levels. As a whole the move- ments in production in the lake in 1898 agree with those else- where in 47 out of 61 possible instances, or in 77 per cent. In view of the fact that the records cover also the low-water period this is a notable degree of agreement, and is to be attributed to the unusually high average level for the year and to the equalizing effect of high water. This factor is not, however, in 426 immediate operation during the last lialf of the year in so far as Phelps Lake is concerned, and other factors common to the whole environment or inherent in the common plankton must be responsible for the similarity in this period. 1899. (Table IX., PI. XLII.) There arebutG collections in this year,—in January-March, at fortnightly intervals. The hydrographic conditions are such that the lake is cut off from the river for 34 days during the 3 months, and of the remaining time there were only 32 days of stages above 10 ft. in which currents passed through the lake to the river, and 7 of falling stages at levels below 10 and above 8 ft., when the run-off continued, making a total of 39 days of contribution to channel waters. These times of contribution in January and March (PL XLII.) are also times of high plank- ton production for that season of the year. Thus the plank- ton content in Phelps Lake on Jan. 24 is 8.47 cm.'' per m.'' to .03 in channel waters. The run-oft" from the lake at that time is thus 286-fold richer in plankton than the water it joins. Again, in March, it is 3- to 9-fold greater. The monthly aver- ages of production are (see table following p. 342) from 6- to 26-fold greater in the lake than in the river. This lake thus serves, even in winter conditions and under a thick and long- persisting coat of ice, as a rich breeding ground for plankton whose run-off enriches the channel plankton. This is due to its impounding function, which results in high production, as, for example, during the decline of the January flood (PL XLII.). Proof of this is seen in the sudden decline in production (from 9.3 cm.' on Feb. 7 to .1 on the 21st) when flood waters from Spoon lliver were scouring out the lake beneath the ice. The movement in production in these months in Phelps Lake bears little resemblance to that elsewhere, agreeing with changes in channel production (PL XIII.) in only 1 out of 6 possible instances, and in 3 and 2 respectively oub of 6 in the case of Quiver and Thompson's lakes. This exceptional disa- 427 greement may perhaps lie due to the changes elsewhere, inci- dent to rising winter tioods. SUMMARY. Phelps Lake is the richest in plankton of all the localities examined by us, averaging 19.65 cm.' per m.'. the mean of all collections, or 22.35 cm.', the mean of the monthly averages. This is 8- fold the production in the river and more than twice that in any other impounding area examined. We hud. how- ever, that the lake does not contribute to the river at levels be- low 8 ft., and is therefore cut off for a consideral)le part of the time. In the years 1894 1899 inclusive, the days of run-off were 14, 0, 71, 139, 145, and 80 days respectively, or an average of 76 days. As a rule the plankton content of the lake waters during periods of run-off exceeds that in the channel in varying degrees, and the lake by virtue of the impounding function serves to increase the plankton content of channel waters. The exceptions fall mainly at levels above 10 ft., when the diluent action of Spoon River floods affects the production in the lake. The high records of production which indicate the great rela- tive and absolute fertility of this body of water are in the main found during summer and autumn months, when there is no discharge and the impounding function is at its maximum. This is conhrmatory evidence of the effect of impounding when the disturbing factor of tributary water is absent and coarse vegetation is of little extent. Owing to its small area, its early separation from the channel, and its relation to Spoon River at high levels, the total contributions from this area are relatively small as compared with those from Thompson's and Flag lakes, and at times from Quiver Lake, and its relative fertility diiriiN/ iiKiiiflis of ridi-ojf', as compared with these localities, is wont to rise above their level of production, especially at stages be- tween 8 and 10 ft., when run-off is slight and impounding function dominant. lUusti'ations of this will be found in the monthly averages of 1897 and 1898 in April-June, the sea- son of greatest run-off, when 4 of the 6 monthly averages 428 are considerably larger in Phelps Lake than in the other lo- calities. The course of production in this lake, as has been shown, is predominantly like that in the other localities. It frequent- ly ha'; similarly located pulses, though their amplitude, es- pecially in late summer and autumn, is often much great- er than elsewhere. Moreover, in the majority of instances the direction of the changes in production in coincident or approx- imate collections is also similar to that elsewhere. In a total of 260 possible instances there is agreement in 169, or 65 per cent. This excess of agreement over the demands of chance, combined with its recurrence in successive years and its occur- rence in the case of different localities, is confirmatory of the view that it is the result of the operation of common factors of the environment. The predominance of the disagreements at times of greatest local differentiation or disturbance, as in low water in summer or in rising floods, lends further support to the suggestion. GENERAL COMPARISONS OF YEARS AND STATIONS. It is my purpose to summarize in the following pages the results set forth in detail with respect to the individual locali- ties in the several years, and to make the comparisons and draw the conclusions which follow from such a summary regarding the relative production in these different years and localities and the factors operative in modifying production. The following table gives for the various localities the yearly averages of plankton, silt, and total catch, and the number of collections in each year. Station Illinois River. Grand av. Total Spoon River. Grand av. Total Quiver Lake , . Dogfish Lake. Flag Lake. 1894 1895 i8q6 1897 1898 1899 1806 1897 1898 1899 1894 1895 1896 1897 1898 1899 Grand av. Total.... 1895 1896 1897 Grand a V. Total 1895 1896 1897 Grand av. Total .... No. of collections lo 76 34 52 13 235 9 13 II 3 36 14 13 31 24 26 7 "5 30 6 48 4 27 7 6 Av. of monilily averages 2-53 5-91 1.05 3.28 2.03 .42 .71* .007 983 .029 .011 :^ .90 •65 2.19 1.62 1.96 .66 ••75 3-3° 3-99 2.65 3.16 66 31 34 31 Av. ol all collections Plankton Silt Total 2.49 3.22 1. 16 369 2.13 • 41 2.19t .007 1.257 .029 .oil 465 1.08 .78 59 88 44 67 70 325 5.01 2.23 4.22 20 13 4 45 S3 59 83 44 46 .28 72 2-55 1. 91 2. II 94 1-79 349 1.173 .796 2.2.6 939 1.17 .70 .20 .62 .40 43 52 2.15 •5> 77 94 71 60 24 35 98 .36 2-43 83 2.23 1.41 2.25 1.48 2.79 1.50 2.85 I. II 22 .88 3-45 45 69 30 23.90 15.28 5.28 7-13 90 5.40 5 5" 2 45 13 36 Thompson's Lake 1894 1895 1896 1897 1898 1899 5 14 27 18 25 7 Grand av. Total .... 96 Phelps Lake . 1894 i8g6 1897 1898 1899 I 29 9 2'* 6 8.89 13.3' 6.67 10.41 5.06 1.15 8.26 24.17 14.74 9.>5 37-34 374 Grand av. Total.... 67 22.35 89 67 00 43 71 21 94 II. 12 II. II 9.22 II. 71 6.50 1-53 86 8.79 24.17 >3^i7 10.00 36.31 374 trace 77 re 29 19.65 95 22 17 94 53 08 03 60 Grand average of all monthly averages, not of annual averages. tGrand average of all collections, not of annual averages. 430 1894. (PI. VIII., XXV., XXXV. ) Only three stations were established in this year : the Illi- nois River Station, with 10 collections ; Quiver Lake, with 14 ; and Thompson's Lake, with 5. The appended table gives the production in monthly averages of plankton per m.' for the seasons. covered by the collections. 431 hydi-ographic conditions. It is noticeable that production in the river is on the average 7- fold greater in Julj-August (PI. VIII.), during stable low water, when current is slackened and impounding most prolonged, than in June and September, when high levels and flushing by floods occur. The low level of production in the river in October-Decem- ber is e.xceptional. Similar hydrographic conditioas in 1S97 PI. XI.) yield a 5- to 10- fold greater production. It is not im- probable that the monthly collections of 1894 may be interca- lated in the depressions between plankton pulses of greater vol- ume, and thus inadequately represent the real production. Pro- duction in Quiver Lake in the months represented in our rec- ords in 1S91 e.xceeded that in channel waters only in September- October, when a slight run-off occurred, and that in Thomp- son's Lake in June, July, September, and December, but a run- off of any consequence occurred only in June and for a week each in July and September. Apparently the channel plank- ton in this year was largely independent and indigenous in origin, deriving l)ut little enrichment from impounding back- waters, and not infrequently diluted by their contributions. 1895. ( PI. IX., .XX\"I., xxx., XXXllI., XXXVl. ) To the river, represented by 50 collections. Quiver Lake, by 13, and Thompson's Lake, by 14, there are added this year. Dogfish Lake,repre.sented liy 12, and Flag Lake, by 4 collections in late autumn—a total of 93. The table on page 432 gives the monthly distribution of production in the several localities, and indicates their relation to the general averages. This is a year of lowest levels, averaging only 3.61 ft. above low water, and also one of high plankton content. On the basis of means of monthly averages, the average content in the river, 5.91 cm.^ per m.^ is the largest recorded, exceeding the average of all monthly averages, 2.71, by US per cent., and being sec- ond in the list if we base comparisons on the average of all col- lections. This high content in river watei-s must be attributed 432 to stable hydrographic conditions, abrupt changes being limited to less than 8 weeks in the year (PI. IX.), and to low water and consequent slackening in the channel with increase of time PLANKTON PRODUCTION IN 1895.* Station Illinois River Quiver Lake Dogfish Lake Flag Lake Thompson's Lake . Feb. 0.0 0.03 Apr. :i+ 3> 3.00 8.20 28.20 May + 61.44 June + 30.42 0.12 9.42 July 9-33 0-37 2.99 4.83 Aug. + 4.03 o.2r I . II 3 09 Station Sept. Illinois River — Quiver Lake + Dogfish Lake + Flag Lake ' . . Thompson's Lake — 1.52 0.94 3-iS 3.58 Oct. 0.57 0.13 0.52 57-76 3-'5 Nov. 3.02 0.05 5.01 14.40 5.07 Dec. 1. 14 0.46 5-32 4.82 1 .00 Mean of monthly Averages + 5-91 0.65 3-3° 25.66 '3-31 *The minus sign signifies below average and the plus sign above. for the breeding of the plankton. The larger amounts of plank- ton were found only during stable conditions, and floods inva- riably depleted the volume of the plankton. These periods of stable conditions occur in summer and late autumn, and we find the plankton content at such times 3- to 30- fold that in contiguous flood conditions. In the river the monthly produc- tion exceeds the monthly average for our records in 5 out of the 9 months represented, the exceptions being February, when stagnation under the ice prevailed, April, a vernal period of low water without overflow, September, a month of repeated floods, and October, when an unusually early decline in temper- ature occurs. The stable conditions which attend low water thus favor the increase in the plankton content per m.'', though by reason of the lower levels and slackened current the total volume pro- duced in the stream as a whole must be greatly diminished by such hydrographic conditions. The results of this low-water year upon production in Quiv- er Lake (PI. XXX.) are as a whole diametrically opposite to those in the river. Here in the lake, production falls below 483 the average, the mean of the monthly averages ( .65 cm."*) being 63 per cent, below that of all mouthh' averages of Quiver Lake, and but a ninth of the production in adjacent channel waters. The cause of this very marked contrast is to be found in the relative dominance of trilmtary waters of recent origin and of coarse aquatic vegetation in the lake,—a dominance increas- ing as levels fall,—and accordingly we find in this year of low- est levels the lea.st annual production (see table on p. 429 1. Production is not only low on the average but also lower than the average in every month of record save September, when it rises 22 percent, above the mean content for that month. This is a month of higher river levels, and a similar tendency to iu- crea.sed production is to be found in the August flood (^Pl.XXVI. i. In Dogfish Lake, production in this year averages for the 8 months represented 3.3 cm.^ per m.''—44 percent, less than that in the adjacent channel waters and 40S per cent, more than that in Quiver Lake. This is 4 per cent, more than the average monthly content, and 18 per cent, less than production in the same months of the following year. The deficiency below channel production may be attributed to the effect of vegeta- tion, and the excess over that in the contiguous waters of Quiver Lake to the absence of access of tributary waters of re- cent origin. Production is above the average for 4 of the 8 months, the exceptions being April. June. Augu.st. and October. The absence of overflow is apparently the cause of the suppres- sion of the vernal pulses in April and June, and the dominance of vegetation may be responsible for the low production in Augu.st and October, both low-water months. The months of plankton content exceeding the monthly average are 4; July and September—months of flood, and consequently of impound- ing and greater extent of vegetation-free water—and Xovem- ber and December—times of lessened growth on the part of the aquatic vegetation, of rising levels, and of some decay of organic matters from the summer's growth. Causes of like nature are the basis for the large production in Flag Lake in the late autumn months, when in October-No 434 vember the production is 10- to 3- fold the average, and in De- cember 9 per cent, above the mean for that month. In Thompson's Lake the mean of the monthly averages, 13.31 cm.' per m,' is 61 per cent, in excess of the mean of all the months of our records, though but 3 of the 9 months repre- sented in the records of 1895 are above the average. The first two of these, April and May, owe their predominance to a ver- nal pulse of unusual volume, andthe sniallnumberof collections gives these months abnormally high averages, while the large production in July may be attributed to the enrichment of the lake in this year by an invasion of plankton-rich river water from the channel (PI. XXXVI.). The deficiency shown in the records of the remaining 6 months, falling from 10 to 70 per cent, below the averages of our records for these months, finds its possible explanation in the relatively greater dominance of vegetation in the lake in this season, due to two successive years of low water and the prevailing low levels. Collections in this year were, moreover, taken near the margin of the veg- etation belt of the lake. Production in this, the second, year of low water, and the lowest in our term of operations in all the backwater plankton stations but Quiver Lake, is above the average in the year as a whole, though falling below it in 57 per cent, of the time rep- resented, The appai-ent suppression of the vernal pulse in the river and in Quiver and Dogfish lakes may be attributed to the ab- sence of spring overflow and the consequent elimination of vast impounding and breeding areas normally present at this season, and also to the direct delivery of tributary water to the chan- nel and increased relative diluent action of the slight April rise (PI. IX.) in both the river and Quiver Lake. The low levels also serve to bring the vegetation of the two lakes named into early dominance, and the relative occupancy is also increased by the second year of low water and no removal of the accumu- lated growth by flood action. Thompson's Lake, on the other hand, owing to its great extent of open water, is less affected 435 by the low levels, the cutting off of breeding backwaters, and the relative occupancy by vegetation, aud consequently a ver- nal pulse of unusual dimensions reaches a cuhiiination iu its area. The June-July pulse of the river, abnormal iu its location and relative size and apparently without equivalent elsewhere, is due to the unusual development of a stagnation plankton in the sewage-laden river in a period when rising temperatui-es hasten the decay of its unusual load of organic matter. The causes above enumerated render this the year of great- est fertility, iu so far as our records reveal pi'oduction. iu the river aud in Thompson's and Flag lakes, the next to the greate.st in Dogfish Lake, and the least iu Quiver Lake. The low levels preclude any extensive impounding of flood waters and,, moreover, the period of run-off is of slight extent. The rise in plankton content (PI. IX. I in the river following the April, July, and September floods is suggestive of the effect of impounding, the plankton content ( cf. PI. XXVL, XXX., and XXXVI. I being generally greater in the di.scharging backwaters examined than in the recipient channel during these run-offs. At all other seasons in our records for this year there is scant opportunity for enrichment iiy tributary backwaters, and but little suggestion of it. 1896. (PI. .\., XXII, .X.Wll., XXXI., XXXIII., .\XXVII., .XL.) This is the most fully represented year in our series iu the number of stations examined. There are 76 collections in the Illinois, 9 in Spoon River, and 31, 30. 27. 27. and 29 respectively iu Quiver, Doghsh. Flag. Thompson's, and Phelps lakes—a total of 229. It was a year of higher levels, averaging 6.98 ft. above low water,—almost twice the record of the preceding year.—and witnessed a series of recurrent floods approaching or surpassing bank height of the stream. The accompanying table gives a summary of the data of production. 436 PLANKTON PRODUCTIO.N IN i8q6.* Station Illinois River Spoon River Quiver Lake Dogfish L;ike Flag Lake Thompson's Lake. Phelps Lake 437 ol.9 ft. It is not so much the I'xfcnt of movement in levels as it is (lisfribiifioH which produces this depression in production. Eepetition of floods at relatively brief intervals is the cause of low production in r/ifi/nicl \\Riers in 1S9(>. Not only is the mean production in the river below normal, but all of the monthly averages ai*e likewise fi'oni 97 to 35 per cent, below their averages except those of April and December, which are 23 and 7 per cent, aliove. Hydrographic conditions in these two months of higher production are such as to favor in- crease in plankton, since in both cases there is a period of 6-8 weeks of slowly declining levels with little or no interruption in which a more abundant plankton becomes established. The relation of production in the channel and the back- waters in 1896 is also very different from that in 1895. While in 1895, owing to low levels in general and to the prolongation of rising levels, the backwaters were contributing Init a slight and interrupted run-off to the channel, and production, as shown in monthly averages, was in the case of backwaters examined predominantly lower than in the channel, we find in 1896. owing to higher levels, that there is more impounding, and, owing to the slow declines, a larger continuance of it and more run-off to the channel. There are. for example. 157 days of falling levels above 6 ft. distributed through Id months, while impound- ing and run-off continues for 90 days more at lower levels and in decreased volume. Not only are backwaters thus contribut- ing to the channel for a much longer period in 1896. but their plankton content is predominantly higher than that in the channel. An examination of the relative production (see table on p. 436) reveals but 3 out of the 58 monthly averages of pro- duction in backwaters, excluding Spoon Kiver, which are less than coincident production in the channel. These are for Dogfish and Quiver lakes in July, and for the latter in September—l)oth months of lowest water, and consequent predominance of creek and spring water in Quiver and of vegetation in l)oth lakes. This relatively greater pi'oduction in the backwaters is not due to increased absolute production as compared with 1895 except 438 in the case of Quiver and Dogfish lakes (2.59 and 5.01 cni.^) when we find for these lakes the highest mean annual production in our years of records. In the case of Flag, Thompson's, and Phelps lakes the annual mean falls below the average on the basis of monthly averages, though the average of all collections in Flag and Thompson's lakes is above the general average in these lakes (see table on p. 429). The former basis of compari- son is the better one, since it equalizes to some extent the ine- quality in the distribution of collections. The plankton content in backwaters in this year of recurrent floods is thus increased in some instances and but slightly reduced in others, with the net result of predominantly higher plankton content than in the current-swept channel, and a much greater total production. The effect of the run-off of the impounded backwaters upon channel plankton may be seen in the river planktograph for 1896 (P-1. X.), where the March, June, August, October, and November floods in each case reduce the plankton as they rise, and are attended by a noticeable increase as levels fall again. That other factors are involved may be seen in the rise in plankton attending the rising flood of August, and in the de- clines following the increases in plankton content in the midst of rapid run-off in apparently favorable hydrographic condi- tions in nearly every instance above cited. The effect of the midwinter flood following two years of low water, which permitted the accumulation in bottom-land forests, marshes, and backwaters of a great amount of vegetation, may be traced in the larye production in winter and early spring. In the table on page 436 it will be noted that in the months of February-April production is above the average in every back- water in each month except in Dogfish Lake in April. More- over, the largest records for this season of the year were ob- tained in this year in Dogfish, Flag, and Phelps lakes in all three months and in Quiver and Thompson's lakes in the first two, that is, in a total of 13 out of the 15 monthly averages- During the remainder of the year, from May to December, production in the backwaters generally falls below the average 439 production for those mouths iu the several localities. Out of the 51 monthly averages iu this period (see table ou p. 436) 37 are below the mean, and of the 14 above, 9 occur in Dogfish and Quiver lakes, where reduction iu vegetation increases the pro- duction. The cause of this sharp contrast in the relative pro- duction in the two parts of the year, is to be found in the hy- drographic conditions which affect the nutrition of the plank- ton. The rank vegetation which filled the forests, marshes, and mai'gins of the lakes during the two years of low water was submerged by December flood, and by this early submer- gence and subsequent decay increased the production in winter months. This early consumption of the products of decay and the relatively early run-off of the spring flood combined to make vernal production relatively low in 1896. A comparison of the planktographs of the 6 localities (see plates named at the head of this section) will indicate the suppression of the April-May, or vernal, pulse in every locality but Phelps Lake. No plausible explanation for its occurrence here when it is not found elsewhere is apparent. Subsequent floods by their bi'ief duration and frequent repetition tend to impoverish the back- waters by the removal of vegetation and organic debris, and by the run-off of nutrition in solution or suspension and of the de- veloping plankton. In more stable conditions or floods of longer duration, when the backwaters are impounded for longer times, —largely by the restraining action of high water in the Missis- sippi,—decay is longer continued, and there is more opportunity for the utilization of its products by a plankton not removed quickly by the rapid run-off of the flood. A comparison of the different regions even in this one year bears out this inference. Spoon River, scoured by repeated floods and swept by constant and relatively rapid current, con- tains only an insignificant amount (.007 cm.' per m.-') of plank- ton. Quiver and Dogfish lakes, rid to some extent of the accu- mulated growing vegetation and enriched by dead vegetation in their submerged borders, yield in this year the largest an- nual mean of monthly averages (2.19 and 3.99) in our records. 440 In these lakes at levels prevalent in this year the impounding or reservoir function is at its height. The higher levels reduce the relative proportion of creek and spring water, and the ab- sence of extreme high water cuts off to large extent the cur- rent of general overflow through the lakes. Accordingly, pro- duction here in the year as a whole exceeds the mean of all monthly averages by 25 (Quiver) and 26 (Dogfish) per cent., and exceeds the mean of the respective monthly averages in 8 and 7 of the^ 12 months. In the case of Thompson's Lake, on the other hand, the hydrographic conditions are such that its pre- dominantly reservoir function is interfered with as each recur- rent flood passes the level of 6 ft. and starts the current of channel water in at the northern, and out at the southern, end of the lake (PI. II.). Accordingly production in this lake falls below the average of all monthly averages by 24 per cent., and the individual monthly means are likewise deficient in 9 of the 12 cases. In Flag Lake the production (8.31 or 13.83 cm.' per m.^) is 10 per cent, below the mean of all monthly averages, or 21 per cent, above that of all collections, and exceeds the means of the monthly averages in 6 of the 12 months. The accumulated vegetation in and about this lake, and the moderate levels which develop the reservoir function of the area without per- mitting any current of overflow through it tend to keep up the level of production. It is in considerable excess of that in sub- sequent years, but falls below that of the exceptional condi- tions of the preceding autumn, discussed on page 385. The moderate levels free Phelps Lake from currents of overflow and increase its impounding function, and we find here, accordingly, the largest production recorded in any of the backwaters in this year. The fact that the production (14.74 cm.' per m.') falls 34 per cent, below the mean of all monthly averages for the lake and is deficientin 6 of the 12 months, may perhaps be due in part to the fact that in the previous year the lake had been a cultivated corn field and had not therefore been seeded by the spores and winter eggs of planktonts left on the 441 drying up of the lake. In September the lake did temporarily dry up only to be re-entered l)y the October flood, in which an unusual plankton (51.6 cm.') at once developed. 1897. ( PI. XL, XXIII., XXVIII., XXXIl., XXXIV., XXXVIII., XLI. ) This year is represented by 34 collections in the Illinois, and 13 in Spoon River, and by 24, 6. 7, 18, and 9. in Quiver, Dogfish, Flag, Thomp.son's. and Phelps lakes. It was a year of protracted winter and spring flood, a late June rise, and pro- longed low water in summer and autumn. The data of compar- ative production are given in the accompanying table. PLANKTON PRODUCTION IN l8g7.* Station Feb. March April May June July Illinois River Spoon River Quiver Lake Dogfish Lake Flag Lake Thompson's Lake . Phelps Lake +1 .04 .047 19 •15 .07 27 •19 + •38 .007 •34 .48 .83 .65 1.44 5. II .048 13.38 I 8.18 I 8.55 10.38 4.26 + 5.62 .440 1.29 1.94 10.61 7.88 22.58 27 1 + .250 1.26 2.48 4.87 3-59 .42 4.69 7.13 3-3' 9-49 Station 442 months represented also exhibit a plankton content above the average. The causes of this large production are to be found mainly in the prolonged lov\r water, slackened current, and sew- age contamination of the last half of the year. No large ver- nal pulse appears in the records of the river or its backwaters. It was either intercalated between collections, and thus escaped detection, or the early winter flood, as in the previous year, by its washing away sources of nutrition prior to the season and temperature of greatest plankton development, tended to depress production below normal at this season. It was to be expected, however, that a large plankton development took place in the stable conditions attending the three months of declining levels which followed (PI. XL) the crest of the spring flood. If such a development took place it would tend to raise still higher the level of production established by our records for this year. The relation which the backwaters bear to channel produc- tion in 1897 is correlated with the hydrographic conditions. In January-June, a period of continued high water, the plankton content in the backwaters exceeds that in the channel in 22 of the 30 monthly averages (see table on p. 441 ) , or, omitting Spoon River, which does not properly belong in the category of back- waters, in 21 out of 25. This was a period of extensive and long-continued impound- ing and of high levels, and, in the last three months (PI. XI.), of rapid decline and therefore of speedy run-off and rapid cur- rent in channel waters, factors which favor the breeding of the plankton in the reservoir regions and cut down the time for its development in the channel, in which barren tributary waters of recent origin and plankton-rich backwaters impounded in the more or less current-free areas for a greater or less length of time, depending upon the direction and rate of change in river levels, are mingled in varying proportions. In the low-water period, July-December, stability of hydro- graphic conditions continues throughout, while the extreme low levels maintained for so long a time make the channel wa- 44^ ters very largely independent of backwaters and tributaries by reason of cutting off of communication in some cases and ces- sation of run-off in others. Along with this independence goes increased fertility t)y virtue of greater relative contamination by sewage and longer time for breeding by reason of the slack- ened current. As a result, production in channel waters reach- es in July-November, 1897, a level unsurpassed in our records, rising above the monthly naeans of all years (see table fol- lowing page 342) 11 per cent, in July, 245 per cent, in Sep- tember, 250 per cent, in October, and 14 per cent, in Novem- ber, but falling behind by 6 per cent, in August. As a result of this increased development of plankton in the channel the pro- duction in backwaters becomes re/afireli/ less with respect to the channel production than in times of high water. Thus in July-December, backwater plankton exceeds that in the chan- nel in only 10 out of 19 monthly averages, or, omitting Spoon Eiver, in 8 out of 15—a marked change from the excess in the preceding six months, 21 out of 25. This is the only season in all our records in which the plank- ton content of Spoon River rises above the barren level of .1 cm.^ per m.^ or less. The production now rises to a level ap- proximating, and in November-December exceeding, that in the channel as a result of the practical absence of current and consequent increase in the reservoir function of the stream. The course of production this year in the various backwa- ters is in most instances strikingly similar to that in the chan- nel in its major outlines. Thus in all of them (PI. XXVIII., XXXII., XXXIV., XXXVIIL, and XLI. ) production rises grad- ually from the midwinter minimum to an unusually low ver- nal pulse in April-May and declines again in June-July. At this point collections were suspended in Dogfish and Flag lakes. Production in channel waters rises again in August and con- tinues at high levels till November, and in like manner and with even greater amplitude in Thompson's Lake (PI. XXX- VIIL), while a similar movement is initiated in Phelps Lake (PI. XLI.), only to be stopped by the drying up of the lake in 444 September. Quiver Lake, however (PL XXVIII. ), pursues a different coarse, production there dropping to a level rarely exceeding .5 cm.^ for the remainder of the year. This results from the greater relatice volume of spring and creek vpater in this lake. The discharge from Quiver Creek and the marginal springs continued with relatively much less diminution through the autumnal drouth than that from Spoon River, with the result of making this lake far less productive of plankton at this season than Spoon River (cf. PI. XXIII. and XXVIII. ). 1898. (PI. XII., XXIV., XXIX., XXXIV., XXXIX., XLK. ) This year is represented by 52 collections in the Illinois and 11 in Spoon River, and by 26, 6, 25, and 22 respectively in Quiver, Flag, Thompson's, and Phelps lakes. It was a year of normally located and fully developed spring floods, followed by low water in summer much disturbed by minor floods, with a subsequent autumnal rise of unusual proportions. This re- sulted in the highest average river levels in our years of opera- tion, 8.02 ft.—a level almost equaling bank height. The accompanying table gives the data of comparative production of the different localities in this year. 445 The average production in the Illinois for the year is 2.03 cm.' per m.^ or 2.1 3 cm.' if the average of all collections is taken instead of the mean of monthly averages. This is 25 per cent- belovv^ the mean of monthly averages, or 3 per cent, below that of all collections. This depression in production is due to the disturbed and irregular hydrographic conditions which throughout most of the year left insufficient time for the plank- ton to breed. As shown by the + and — signs in the table, production in channel waters is below the average in 7 of the 12 months, and 4 of the 7 deficiencies fall continuously in the disturbed period of August-November. The decline below the average produc- tion in this disturbed period ranges from 72 to 86 per cent. The other 3 months of deficient production are April, June, and Ju- ly—4, 45, and 86 per cent, lielow their averages. The April defi- cit is due to the delay in the vernal pulse, while those of June and July are due, possibly, to the after effects of the high ver- nal pulse of May. Production in excess of the average is found in Januarj'-March, during the unusually slow rise of the spring flood, in which the catastrophic effect alike of the sudden and higher floods and of stagnation under ice in low water is eliminated. Production is also high, by 86 per cent, in May, as a result of the delayed culmination of the vernal pulse —the largest one, moreover, found in our records in channel waters. It is also high by 40 per cent, in December, when declining levels (PI. XII.) afford the stability necessary for the breeding of the plankton. Production in Spoon River, as might be expected in a year of much flood water, falls to a barren level exceeding .1 cm.' in but a single instance, in March (.124), when waters of general overflow mingle with those of the tributary to a considerable extent. Production in the backwaters in 1898 again bears a striking resemblance, throughout, to the course it presents in the chan- nel. The higher levels conduce to greater unity in the envi- ronment, and to greater interchange between many localities, 446 and the plankton accordingly follows similar lines of develop- ment. This is noticeably prominent in the planktographs of the river and Quiver and Thompson's lakes, as will be seen by a comparison of Plates XII., XXIX., and XXXIX. The princi- pal features of the common course of production are the coinci- dence of the May and June pulses, the subsequent low level of development throughout the summer and early autumn, and the December rise. These three bodies of water were submerged in the common flood of overflow in February-June, and the succeeding minor flushes of summer and autumn caused re- current ingress and egress of water from and to the channel. The similarity in the course of production in these three local- ities and the lessened differences in the amplitude of produc- tion in this year are in no small measure the consequence of this equalizing action of this interchange due to floods. Phelps Lake (PI. XLII. ) is the only one of our backwaters which diverges from this marked agreement, and its divergen- ces are increased by its intimate connection with Spoon River during high levels and its isolation during the remainder of the year. The vernal pulse of this year is noticeable for its amplitude, its meteroric appearance and disappearance, and its coinci- dence in different localities. It follows a prolonged period of extreme overflow, and a very gradual and somewhat tardy rise in vernal temperatures. It appears, moreover, at levels of 10- 11 ft., just when great stretches of bottom-lands are contribut- ing their last run-off to the channel. The submergence of the bottom-lands did not occur until late in February in this year, so that the period of vernal increase in the plankton was not preceded by a long interval of flood, as in 1896 and 1897, which might carry away in suspension or solution those organic sub- stances in the vast amount of vegetable detritus which covered the bottom-land as a result of the low water of the preceding autumn, and which may have been utilized by the plankton in this extraordinary vernal development as a result of the juxta- position of flood and vernal growing season. U1 The comparison of backwater and channel production in 1898 is in some contrast with that in 1897. In 1897, omitting Spoon River, production in backwaters exceeded that in the channel in 29 of 40 monthly averages, or in 73 per cent., while in 1898 the excess occurs only in 26 of 37, or in 67 per cent. The excess, moreover, is frequently of less amplitude in the lat- ter year, as is seen in the relation of the means of the monthly averages of backwaters and channel in the two years. Thus in 1897 production in Thompson's Lake (10.41) was 217 per cent, in excess of that in the channel (,3.28), while in 1898 (5.06 and 2.03 ) the excess was only 149 per cent. In Phelps Lake, on the other hand, production rose to the unparalleled height of 37.34, the mean of the monthly averages, 67 per cent, above the mean of all monthly averages, and fourfold that in the previous year, when the last 4 months were cut off l)y the drying up of the lake- Production in Quiver Lake in this year is 1.96 cni.^ mean of monthly averages, or 2.44, average of all collections—12 per cent, and 44 per cent, above the mean i-espectively of all years. This larger production is due to the excessive production in the vernal pulses in May and June and to the high levels of pro- duction in November and December, rising 138, 78,217, and 176 per cent, above the average respectively for these months. The hydrographic conditions in these months in Quiver Lake are favorable to increased production. The May and June pulses are at levels (11 ft.) when impounded run-off from slightly submerged l)ottom-lands to the north was rapidly draining to the channel through the lake. In November and December there was at least douhle the usual volume of water in the lake, due to a 34 per cent, increase in river levels, with a considera- ble reduction in the jiroporiion of tributary water and increase in the impounding function. In the remaining 8 months of the year, then, average production is 79 per cent, below the general average for those months in Quiver Lake. This very considerable depression in production falls in the main in the period of greatest hydrographic disturbance. This body of water, owing to its frequent invasion by channel waters and to 448 its own influx of tributary water from Quiver Creels, is the most liable of all the backwaters examined by us to hydrographic disturbance. It is therefore not surprising that in this year of extreme disturbance we should find marked depression for a long period in this lake. The total movement in levels in 1898 is 67.2 ft. (see table p. 163), 44 per cent, above the average. Of this, 50.8 ft. fall in the 8 months of depressed production, that is, 76 per cent, of the movement occurs in 67 per cent, of the time. To this relative excess of fluctuation in levels, and proba- bly to large access of local flood and spring water, we must at- tribute the low production in Quiver Lake in these months. Thompson's Lake has an average production of 5.06 cm.', or, if all collections are averaged, 5.71 cm.', 39 and 28 per cent, below the respective averages for all years (see table, p. 429). Not only is the general average below normal, but all of the monthly averages, save only those of January (7.22) and De- cember (3.58), are likewise deficient by from 2 to 88 per cent. The large January production is the largest plankton content in this month in any year or locality, and accompanies an invasion and impounding of sewage-laden river waters in the lake (cf. PI. XLV. and L.). The cause of the low production throughout the remainder of the year is again to be found in the hydrographic conditions. During 8 months of the year (at levels above 6 ft.,seePl.XXXlX.) the lake is swept by a gentle current entering at the northern end and discharging to the channel at the lower. There is, thus, in this year more than the usual run-off, not only of or- ganic matters in solution and suspension, received with the waters of ingress, but also those developed in its impounded waters or about its shores. This tends to impoverish the waters, and interferes with the accumulation and flux of or- ganic matter in the plankton which manifested itself in such amplitude in the low water of the preceding year (PI. XXX- VllL). To a much less extent than in Quiver Lake is the de- pressing effect of flood waters seen in the broader expanse of this body of water. While in the former the production in 449 months of flood disturbance falls 79 per cent, below the mean for that season, we find in the 10 months of depression in Thompson's Lake a falling off of only 46 per cent. The differ- ence is due to the greater proportion of creek and spring water of I'ecent origin in the former, and to the greater reservoir ca- pacity and consequent longer impounding, as a rule, of the sewage-laden channel waters which predominate in the latter backwater. Production in Phelps Lake is 87.34 cm.^ per m.\ or, if all collections are averaged, 36.31—67 or 84 per cent, above the means for all years. In keeping with these facts we find that the monthly averages equal or exceed the means for their months in 7 of the 10 months of record, the greatest excess oc- curring during the period of complete i-solation of the lake. 1899. (PI .\I1I., .\X1V., X.\IX., XXXIX., XLII.) This year is represented by 13 collections in the Illinois and 3 in Spoon River, and by 7 each in Quiver and Thompson's lakes and 6 in Phelps Lake, all in the first three months of the year. This was a period of a slow rise of the river in January to bank height, with an equally slow decline in the next month followed by an alirupt and well-sustained March Hood. The data of comparative production are brought together in the following table. PLANKTON PRODUCTION IN 1899. Station Illinois River I — Spoon River ! — Quiver Lake ] -|- Thompson's Lake — Phelps Lake -|- 450 This is the largest production for this period of the year in our records, exceeding the average, .24 cm.\ by 75 per cent, and ap- proximating or exceeding the monthly average in each instance. The cause is to be found in the relatively stable conditions at a level sufficient to prevent sewage stagnation beneath the ice- sheet vi'hich covered the stream prior to the March flood. Spoon River continued to discharge barren waters (av. .011 cm.''), while the backwaters, with the exception of Quiver and Thompson's lakes in March, produced a more abundant plank- ton than the channel. Quiver Lake produces .66cra.^—an ex- cess of 20 percent, above the usual production for this season, — and in the first two months has 2- to 3- fold the usual plankton content as a result of the moderate levels which make the lake a reservoir without greatly increasing its current. When, how- ever, the general current of overflow passes through it with the March flood, production drops to one fifth of the mean for that month. In Thompson's Lake the mean production is 1.15 cm.^ 56 per cent, below the mean production for these months. It also falls below in January and March, when hydrographic con- ditions are such (PI. XXXIX.) that channel water is diverted through it, and rises above the mean by 25 per cent, in Febru- ary, when the run-off is diminished by falling levels. In Phelps Lake the mean production, 3.74 cm.'', is 23 per cent, below the mean for these months. This lowered production, which also falls below the mean in the last two months, is due, in part at least, to the invasion of Spoon River water with the higher levels. The various years of our operations may be briefly charac- terized as follows. 1894. A year of low water and fairly stable hydrographic conditions, with nearly average production in channel and open backwaters and deficiency in the vegetation-rich Quiver Lake in the months of our records. 1895. A year of continued and but slightly interrupted low water, with stagnation destroying the winter plankton and tending to abnormal production in early summer in channel 451 waters. Production is deficient in vegetation-rich backwaters (Quiver Lake ) and in excess in open reservoir lakes (Thomp- son's Lake). 1896. A year of recurrent floods of modei'ate height and gi'eatly diminished production in channel waters. The repeated summer invasions of the vegetation-rich backwaters by flood destroyed and removed much of the season's gi'owth, and we find this the year of greatest plankton production in these areas (Quiver, Dogfish, and Flag lakes). The more open waters, poor in vegetation (_ Thompson's and Phelps lakes), have vernal pulses of considerable magnitude, but as summer and autumn production is not up to the average there is as a whole a defi- ciency in these areas. 1897. A year of prolonged winter-spring flood, followed by continued and abnormally low water in the last 5 months of the year. The early flood apparently reduces production everywhere, while the prolonged low levels lead to abnormally high production in the channel and in vegetation-poor back- waters (Phelps and Thompson's lakes i. In Quiver Lake the relative dominance of vegetation and tributary waters of recent origin is so increased that production falls to a low level. Even tributary streams (Spoon Eiver) so decline in run-off in the low-water period that an abundant plankton develops in their waters. 1898. A year of normally located spring floods, followed by repeated minor flushes in summer and autumn. Thei'e is a meteoric but normally located vernal pulse, with a low level of production throughout the remainder of the year in the channel and in liackwaters intimately connected with it (Quiver and Thompson's lakes i. and an abnormally high and well-sustained level of production in backwaters isolated from flood contact and free from vegetation (Phelps Lake). The different localities may be briefly characterized as fol- lows. Illinois Birer. Channel waters contain a plankton which in constituent organisms, character of the course of production, 452 and relation to environmental factors shows marked similari- ties to that found in the backwaters. The river is, however, more immediately and directly subject to ihe effect of floods, sewage, and stagnation, and exhibits less uniformity and regu- larity in its planktographs. High water and repeated floods are wont to depress its production, while stable conditions such as prevail more fully in falling levels and low waters often lead to increase in production. The run-off of impounded backwa- ters where plankton breeds, which attends falling levels, and the concentration of sewage in low water likewise conduce to increased production. Spoon River. This is always plankton-poor save at lowest levels, when the slackened current renders the stream an im- pounding area. At high levels in spring impounded backwaters of the bottom-lands join it and tend slightly to increase its plankton content. It is immediately and predominantly a dil- uent of channel plankton. Quiver Lake. This is an area subject to great vicissitudes of production by virtue of the variety of environmental factors operative and their changing efficiency with fluctuations in river levels. Influx of creek and spring water of recent origin and relative dominance of vegetation are increased at low lev- els, and production falls. Backwaters from the channel in floods below 6-8 ft. increase its impounding function, and de- crease current, proportion of tributary water, and relative oc- cupancy by vegetation, and production rises. In floods of higher levels the general current of overflow from submerged lands up-stream courses through the lake, and local factors are great- ly reduced in their effect. As a whole it is the least productive of all the backwaters examined, but like the others has a rich plankton when it is filled with impounded backwaters, espe- cially at the season of the vernal pulse. Dogfish Lake. This resembles Quiver Lake in many fea- tures, but is freer from immediate access of tributary water and invasion of channel waters at lower levels. It has accordingly a higher and better sustained level of production. 453 Fhif/ Lakr. This, like Quiver Lake, is an ai-ea of great con- trasts in production, resulting from its varying occupancy by vegetation as levels rise and fall and from the changes in nutri- tion available for the plankton attendant upon its growi:h and decay. In so far as the data go they indicate a production sev- eral fold greater than that attained in Quiver and Doghsh lakes. Thompson's Lake. In this area, by virtue of its considera- ble size and freedom from immediate access of tributary water of much volume, the fluctuations of the hydrographic, and to some extent of the other, factors of the environment are in a measure equalized. Production is therefore less disturbed than in the backwaters previously named, pursuing what may be called a more normal course, and is accordingly greater than theirs,—excepting perhaps that of Flag Lake in seasons of high water.—being three-fold that in the channel, from which its main water supply is directly drawn. Owing to this intimate connection with the river and freedom from dominating local influences the planktograph of this lake is more like that of the river than are those of other localities. Comparative freedom from vegetation, highly developed impounding function, free- dom from access of tributary water of recent origin, and close connection with channel waters rich in organic matter, all com- bine to cause the high production found in this area. Phelps Lake. In this area production reaches the highest level found in any of our localities. It exceeds that in the chan- nel eightfold, and that in the most productive backwaters else- where examined by us by two and one-half- fold to threefold. Ingress of flood water fi-om Spoon River depresses production at high levels. At lower levels the freedom from vegetation of the coarser, more cumulative, and permanent sort, the abun- dance of alga? whose decay releases nutrition for the summer plankton, and the absence of tributary creek or spring wa- ters, all favor the high production found in this area. Per- haps most potent of all factors is the isolation of the lake at high levels and consequent retention within its boundaries of 454 the organic substances in solution and suspension in its waters, substances which in the other localities run off with the de- clining flood. During last stages of low water in Phelps Lake there is usually some dying off of the fish and other aquatic animals, possibly as a result of extreme temperatures, and the lake becomes a favorite resort for fish-feeding water-fowl. The organic substances thus released for immediate solution in the water and utilization by the phytoplankton may be of sufiicient quantity to materially increase production at these low levels. It may be that the more complete access of light in these very shoal waters and the condensation caused by evaporation and seepage as the lake dries up are contributory to the increased plankton content, but none of these factors seems adequate to explain the excessively large production found in the summer and autumn in the shallow pools which form the remnant of the lake. RELATION OF ENVIRONMENTAL FACTORS TO PLANKTON PRODUCTION. The detailed discussion of the course of plankton produc- tion as defined by volumetric data found in the preceding sec- tion of this paper, has afforded many specific instances of the relationship existing between the movement in production and a number of factors in the environment. Prominent among these are hydrographic conditions, temperature, light, chemic- al conditions, vegetation, and the reproductive cycles of the constituent organisms of the plankton. It is my purpose in the following pages to summarize, with a few references to specific illustrations, the conclusions as to general tendencies and the effect of these various factors upon the course of production in the river and its backwaters. HYDROGRAPHIC CONDITIONS AND PLANKTON PRODUCTION. This is a comprehensive designation for a great variety of major and minor influences which continually impinge upon the plankton as a result of its environment in water. In the case of the plankton of the Illinois River and its backwaters it 455 is the unique factor which more than any other differentiates it from lake plankton. To a less degree it seems to differentiate this aquatic environment fi-om other streams, as a result largeh' of the imperfectly developed flood-plain, and conse- quent unusual proportion of reservoir backwaters. It is, moreover, an exceedingly varialjle factor, operating with almost constant change in each locality, and from season to season in the same locality. It is this element of fluctuation and the resulting chaos in the movements of production which particularly characterize the river as a unit of environment, and in a large measure differentiate it from the more stable lake. The changes depend primarily upon the unequal distri- bution of the rainfall and its run-off and the consequent fluctu- ations in levels with attendant changes in area, depth, condi- tions of ingress and egress of water in any given area, cun-eut, and age of the water. The effect of the area of the body of water upon its plank- ton production in our situation is so masked by combination with other factoi's that the available data are inconclusive. Our largest vegetation-poor backwater. Thompson's Lake, pro- duces less (8.26 cm.'') than the smaller one. Phelps Lake ( 22.35 cm.'t. On the other hand, the largest vegetation-rich area, Flag Lake, produces 9.23 cm.'—considerably more than the smaller Quiver and Dogfish lakes, 1.75 and 3.16 cm.'—and these differ- ences are. moreover, in all probability to be attributed to other factors than mere area. The relative development of the shore-line is a corollary of the form of the body of the water, and is thus related to its area. In the case of the river, the development of the shore- line, as shown on page 2S4, is 78.3—a disproportionately large element in the environment of the plankton. This factor of area has been introduced here in order to emphasize the fact that mere size in itself has apparently little to do with plankton production. Plankton is present in small as well as large bodies of water, with, of course, an increasing proportion of littoral influences as areas contract. Within 456 our euvironment the submergence of all tbe localities, large and small alike, in the major floods, tends to obliterate areal differences, and to unify their plankton and to give them all much the same initial start in the season's course of production. This is noticeable in the marked similarity in the vernal plank- tographs of the several localities in years of high water. The areal differences in later months and lower river stages are, however, very considerable (cf. PI. XX. and XXI.). Thus in Thompson's Lake there is a broad expanse of open water sev- eral square miles in extent, while in the last stages of Phelps Lake there are only a few pools, a few acres in extent (PI. XXL). Under similar climatic conditions production in both runs high, higher in fact than in other localities, and the same plank- tonts are dominant, though in varying proportions in the two. In general the plankton content per cubic meten-uns very much higher in Phelps Lake, the smaller body of water. There is no evidence that the smaller size has anything at all to do with this larger production, but this instance suffices to show that a typical plankton with large production may be found in small areas. On the other hand, Quiver and Dogfish lakes, next in size to Phelps Lake, are the least productive of the backwaters. Thus, area in itself tends apparently neither to deter nor to pro- mote production. The effect of depth, in all of its relations, upon plankton production is manifestly not demonstrated in our data, since all of our collections have been made in shoal water of less than 10 meters in depth. Our deepest waters are Spoon and Illinois rivers, where spring and flood water of recent origin tend to depress production, and the shoaler waters are found in the backwater reservoirs where impounding favors larger plankton content. The significance attaching to the fact that our lar- gest plankton production (224 cm.' per m.^ on August 24, 1898) is found in Phelps Lake, the shoalest of all our localities, and in depths no greater than 20 centimeters, is a matter of conjec- ture, since many other factors are also involved. The highest production in this lake was, as a rule, found during the periods 457 of shoaler waters, from 50 to 20 cm. Flag Lake also, in the shoal waters of the autumn of 1.S95, yielded a large iJroduction ; and Thompson's Lake, in the low water of the autumn of 1897, gave the largest production on record for that season of the j'ear in that body of water. The vernal pulses, noted for their large production, usually fall at the time of the run-off of large areas of sl/(//it/i/ submerged bottom-lands. Shoalness, then, does not prevent large production. Indeed, there are some important reasons why shoal waters should, other things being equal, produce more plankton. Light pervades the water more com- pletely, aeration by wind and waves preserves the gaseous equi- librium more perfectly than in deeper waters, and upon sedi- mentation the suspended organic matters in the water are not removed from the immediate proximinity of the growing phy- toplankton. Their decay and solution renders them imme- diately available, while in deep waters only the slow process of diffusion, in the absence of vertical currents dependent upon temperature or hydrographic changes, brings them within the field of surface-dwelling plankton. It is well established that the plankton is relatively more abundant in the surface than in the deeper waters of the ocean. The researches of Reighard ('94), Birge ('95), and Ward ('95) show conclusively that the surface waters of our larger lakes contain the greater part of the plankton. Thus, Reighard ('94) finds in Lake St. Clair from 1.2 to 37.2 times as much plankton in the surface stratum of 1.5 meters in depth as in the remain- ing bottom layer in depths of 2.2 to 8.4 meters. Birge ('95) finds 50 per cent., or more, of the Crustacea of the plankton in the upper 3 to 4 meters, and over 90 per cent, in the upper 9 meters in depths of 18 meters, and Ward ('95) reports 64 per cent, of the plankton in the upper 2 meters. Of our total 640 collections, 389 were made in water over 2 meters in depth. In the Illinois there were 208 such out of 235; in Spoon River, 34 out of 35; in Quiver Lake, 60 out of 115; in Dog- fish Lake, 27 out of 48; in Flag Lake, 8 out of 44; in Thompson's Lake, 48 out of 96; and in Phelps Lake, only 4 out of 67. Thus 458 about 40 per cent, of our collections were made in water less than 2 meters in depth and at all river stages, and at all sea- sons of the year a considerable proportion of the Illinois and its backwaters has a depth not exceeding this limit. Moreover, less than 12 per cent, of our collections are in 5 or more meters of water, and these are confined almost entirely to the two streams. Our collections as a whole are therefore within the limit of depth of the surface stratum, which in deeper wa- ters contains by far the greater proportion of the plankton. This feature of our environment has a tendency to increase the plankton content per cubic meter and to lower it when production is stated in volumes per square meter of surface. Thus, if we compare production in the Illinois River in August with that of Lake Michigan, as given by Ward ('95), we find that in quantity per m." of water the river (3.88) exceeds that of Lake Michigan. The average of Ward's August collections is 3.69 cm.' per m." by the gravity method, or 1.23 cm.^ if we reduce it to centrifuge basis (see Ward '00). Production in our environment in this month (see table following p. 342) in all localities but Spoon River and Quiver Lake is from 1.5 to 47 times greater per cubic meter than in Lake Michigan ; but if the average per sq. meter of surface be taken as a basis, Lake Michigan exceeds the most fertile of our localities at this season. The amount per sq. m. in Lake Michigan I find on computation to be 37.98 cm." estimated centrifuge measure- ment, while in the Illinois at average August river-levels (4 ft.) it is about 15 cm.^ in Thompson's Lake only 9.29, and in Phelps Lake approximately 14 cm." The age of the water is, other things being equal, one of the most vital of all the hydrographic factors environing the plank- ton. Rain and spring water of recent origin entering the river or its backwaters in tributary floods or by seepage from under- ground storage beds, is practically barren of plankton. This same water impounded in the reservoir backwaters, seeded by the spores and resting eggs of the planktonts deposited on the submerged territory on the recession of antecedent floods, and 459 mingled more or less with the residual and plankton-rich wa- ters of reservoir lakes and sloughs, soon develops an abundant plankton ; or, retained in the channel and mingled with plank- ton-bearing waters from a thousand sources, it develops a phyto- and then zoo-plankton as soon as the requisite fiincfor brecd'uKj has elapsed. Illustrations which demonstrate the operation of this hydrographic factor permeate all of the data of the variable environment with which we are dealing. Attention has been called repeatedly in the detailed discussion of the movement in production to specific instances of the barrenness of flood waters of recent origin which invade plankton-rich areas. By dilution and replacement they lower plankton content in channel waters pre-eminently and to a less extent in backwaters, where their diluent action is less pervading and replacement less com- plete. The most perfect illustration of this will be found in the August flood of 1S9G (PI. X.). Tributary streams such as Spoon River, whose waters ai-e of recent origin, contain but little plankton. Quiver Lake falls in production as spring and creek waters preponderate in its area at low-water stages, while the phenomenally rich waters of Phelps Lake are poverty-stricken only so long as Spoon Riv- er Hoods maintain a current through the lake. The time requisite for the development of the plankton in such waters of recent origin must necessarily vary with the at- tendant circumstances, such as climatic conditions, previous state of the plankton, and the proportions of the mingling wa- ters. The roliiuu'trir pulses of plankton are frequently so located as to suggest the agency of floods in determining, to some extent at least, their location and amplitude. The apices of these pulses are found at varying intervals after the pre- ceding depression occasioned by invasion of water of recent origin. Thus the August pulse in 189(5 in channel waters makes full recovery in a week from the scouring effect of the flood [F]. X.). In the backwaters this same Hood (Pl.XXVIL, XXXI., and XXXVII.) was attended by a rising pulse which culmi- nated after an interval of 2 to 3 weeks. In a general way, 10-20 460 days suffice in most cases for production to recover from the dis- aster of, or respond to the stimulus of, the flood. It is evident, how- ever, that this matter is greatly complicated with the rythmic, pulse-like character of the movement in plankton production. In passing, attention should be called to the fact that re- covery from the flood is not merely replacement by up-stream waters rich in plankton, as might be concluded regarding chan- nel waters. It takes place also in the impounded backwaters under conditions of absolute independence of channel, as in Phelps, Dogfish, and Flag lakes. In these cases it is an indig- enous development in impounded flood waters, and, by infer- ence, thei"e must be a process of like import in channel waters and areas more or less intimately connected therewith. This necessity of some lapse of time before accessions of tributary flood and spring waters can produce a plankton of any considerable volume makes all important the impounding function of the backwaters of the Illinois River. They act not only as storage reservoirs whence the floods are drawn ofiE in reduced and equalized volume, but they serve also as nurseries, where under favorable conditions the flood waters are seeded with planktonts whose progeny utilize the organic materials in suspension and solution in the invading flood or derived from the invaded territory. The plankton thus developed, or the product of its decay, is carried away by the run-off of the flood into channel waters unless utilized by the larger and more per- manent residents of the backwaters as food, or sequestrated in some land-locked pool or lake as levels fall. Intimately connected with the age of the water are the as- sociated factors of current and rate of renewal. In and of itself alone, current has little demonstrable influence upon plankton production, but conjoined with other factors it becomes potent for ill. Thus, in narrow confines, as within the banks of a stream, by its mechanical action it fills the water with silt, which diminishes access of light and thus tends to decrease photosynthesis by the phytoplankton. The silt likewise, as has been pointed out on page 185, seems by adherence to im- 461 pede the movements of the Enfomostraca, and perhaps also of the Rotifem, and thus to exercise a deleterious eifect upon them. The excess of moribund and dead individuals of these groups in flood waters is evidence of the destructive character of the silt-laden current. When the existence of a current in a body of water involves the run-off of its contents and their replacement by water of more recent origin, the result tends to lower production, and the proportions of the depressing influence will rise with the rate of the current and the contraction of the volume, or with rapidity of replacement, in the main in proportion as they shorten the time for breeding of the plankton. Illustrations of this tendency appear in contrasting our various localities. In Spoon River, where current is rapid and renewal with re- cent water complete and frequent, production (.256) continues in minimum quantity and rises only when current slackens. In the Illinois River the current is perhaps as a rule less rapid than in Spoon River, replacement because of connecting back- waters less complete and frequent, and the replacing water has a greater proportion of older water by virtue of the longer wa- ter-course and the greater development of contributory im- pounding backwaters. Production is accordingly greater (2.71 ) in the larger stream. In Quiver Lake conditions in these par- ticulars are extremely varied. As the lake emerges from the general overflow the extent and rate of the current declines, but the proportion of creek and spring water of recent origin is increased with resulting depression in production. When by reason of backwater from the channel at low levels the current is reversed, the run-off checked, and the flood and tributary waters are impounded,—as in the summer floods of 1896,—pro- duction rises. Current with attendant run-off and renewal of water in this lake constitutes one of the important factors in loweringitsproduction(1.75)belowthat of the other backwaters. In Dogfish Lake the factor of current and renewal by recent water is largely eliminated, and production (3.16) exceeds that in the contiguous waters of Quiver Lake. 462 In Flag Lake a current is only slightly present in high water and during the run-off of impounded floods. It is per- haps less noticeable here than at any other station examined, owing to the sheltered location and freedom from tributary re- lations. Renewal is therefore least rapid and production ac- cordingly high (9.23). In Thompson's Lake, owing to freedom from tributaries, to the large volume of the lake, and to the relatively small and frequently interrupted run-off, renewal is relatively infrequent and incomplete. Moreover, the water which replenishes the lake is drawn from the channel, and has a considerable propor- tion of old and plankton-rich contributions from impounding areas up-stream. Production is accordingly high in this area (8.26). In Phelps Lake, current, other than that of run-off of con- ta,ined waters, is absent at levels below 10 ft., above which Spoon River flood waters of recent origin find their way through the lake and depress production. At other seasons current is in- considerable and soon ceases, and replacement no longer occurs. Impounding and production (22.35) are at their maximum here. The net effect of the continued run-off and renewal of the water in a stream or lake upon the fertility of the locality will depend largely upon the inflow and discharge of nitrogenous and other food materials or living organisms developed at their expense. It is evident from the chemical data contained in this paper that a vast amount of organic matter is continually carried away from the drainage basin of the Illinois, and that the amount in the streams is usually greater than that in the backwaters. In general, production is less where this impov- erishing process is facilitated by continuous and relatively rapid replacement (Spoon River, Quivei' Lake, and the Illinois), is greater where the rate of renewal and total run-off is less (Dog- fish, Flag, and Thompson's lakes), and is greatest when and where impounding is most complete (Phelps Lake at low wa- ter) and the organic matter in suspension and solution is re- 46B tained within the lake and augmented l)y the synthetic activi- ties of the phytoplankton and other vegetation. The Jfiicfiiafioihs in In/drof/raphic conditions consequent upon changes to river levels constitute the one pre-eminent factor peculiar in the fluviatile environment in the extent to which it is developed and to which it influences and controls the course of plankton production. These fluctuations opex'ate by bring- ing about changes in area, depth, and volume, in current, in age of the water, in rate of renewal and period of impounding, in relative proportions of tributary and impounded water, in chemical contents and sewage contamination, in I'elative dom- inance of vegetation, and in the interrelations of channel with backwaters and of the backwaters with each other. Illustra- tions of each and all of these results have been cited in the de- tailed discussion of the course of plankton production in the channel and the various backwaters. It will therefore suffice in the present connection to cite briefly certain prominent fea- tures, and to deal particularly only with some of the general phases of the problem not readily followed through the maze of details of the previous discussion. The rise in levels results usually from access of tributa- ry flood waters in local or up-stream territory, and rarely from backwater due to entrance of storm water in lower reaches of the river only. These flood rises occur in both high and low river stages, though they are more frequent and of shorter du- ration at lower levels, since in the much contracted volume of the stream at such stages slight increases in the run-off which would scarcely cause a ripple in the hydrograph in overflow stages, now cause considerable change. The river channel itself is most immediately affected by this access of flood waters of recent origin, since tributaries, with few exceptions, discharge directly into the channel, and even when these courses across the bottom-lands ai-e submerged in general overflow the tributary current is always maintained, in part at least, along its old path. The result of this invasion is always a dilution of the plank- 464 ton-rich channel water, or its partial replacement by the plank- ton-poor and silt-laden flood. Typical illustrations of its action appear in the planktograph of 1896. The rhiiig stages of the flood are the most disastrous. They carry the heaviest burden of silt, are formed of the most recent water, and usually, because of their sudden inroads, most com- pletely replace the previous channel contents. The December flood of 1895 is a typical illustration. Rising suddenly from low levels (3 ft.) to overflow stage (12.6 ft.) in 12 days, it de- pletes the channel plankton from 2.6 cm' in the initial stages of the flood on the 20th (PI. IX.) to .08 on the 25th, if not, in- deed, earlier. Not only does it thus depress plankton content in channel waters but, with a less catastrophic completeness, that also in the backwaters. Thus in Thompson's Lake the plankton falls from 1.87 cm." on the 19th to .13 on the 28th, in Quiver and Dogfish lakes, from .63 and 10.57 to .29 and .06, and in Flag Lake, from 6.38 to 8.26. The effectiveness of the depletion is greatest where overflow currents are best es- tablished, as in Thompson's and Dogfish lakes, and least where the currents are slight and impounding greatest, as in Flag Lake. With the culmination of the flood the proportions of in- coming storm water of recent origin decline rapidly, the effects of the run-off of impounded backwaters begin to appear, and recovery in production, other things being equal, marks this stage of the flood. Typical illustrations will be found in the June and August floods of 1896 (PI. X.) and in the June flood of 1897 (PI. XT.) and 1898 (PI. XIL). The results of flood in the backwaters are similar to those in the channel wherever currents of overflow are established and replacement of the plankton-rich contents of the lake by flood waters ensues, as has just been shown. Usually there are accessory bottom-lands, not swept by current, where a part of the original lake waters are retained, and where plankton breeds abundantly. As a result of this, the recovery from the depletion by flood—and this is usually not so complete as in 465 the channel^s more prompt and attains a greater amplitude. A comparison of after effects of the June flood in 1896 in the river (PL X.^ and in Quiver (PL XXVII.). Dogfish (PL XXXL). Flag (PL XXXIIL). and Thompson's (PL XXXVIl.) lakes will conclusively demonstrate these conditions. While the recov- ery from flood effects in the river is delayed till about June 10, in the vai-ious backwaters it occurs from a week to 10 days earlier. It is evident from the pulse-like character of the movement in plankton production, to which attention has been called re- peatedly in the detailed discussion, that the recovery from flood conditions is complicated with this phenomenon. Flood con- ditions in backwatei's, by influx of waters laden with organic detritus, and by destruction, submergence, and decay of vegeta- tion, tend to accelerate the appearance and, it may be, increase the amplitude, of the pulses. When, however, the invading waters have little plankton and largely replace the previous contents, they tend to delay and depress the pulses, and, as in the case of the May-June and August pulses, the flood may even obliterate the apex of a normally developing plankton pulse. Flood conditions thus affect these pulses profoundly, but they do not seem to be fundamentally their cause. Periods of declining river stages stand in sharp contrast with rising levels. It is for the plankton a period in which re- construction and growth are possilile. The decreased propor- tion of recent water and the prolongation of impounding in backwaters, characterize this as a time of relatively stable con- ditions as contrasted with the chaos caused by the rising flood. In a general way there is some correlation between the to- tal annual movement in river levels and the average plankton content per cubic meter of water. In the table on page 466 the available data are tabulated for the channel plankton. The arbitrary division by years results in one misleading presentation of the data. The December flood of 1895 (12.6 ft.) occurred in the last days of the year, while its effects continue beyond that limit. If we subtract the 9.8 ft. of this rise from 466 the total movement (51.9) of 1895 and add it to that of 1896 (45.7), the resulting figures (42.1 and 55.5) will more nearly characterize the hydrographic conditions in which our plank- ton collections were made. After making allowances for the AVERAGE CHANNEL PLANKTON AND ACCOMPANYING MOVEMENT IN RIVER LEVELS. Year 1894. 1895- 1896. 1897. Total movement in river levels—in feet 32-9 5l-9(42.l)t 45-7(55-5)t 44.8 67.2 Plankton per m.' —in cm." * 2-53 5-91 1.05 3.28 2.03 *Mean of monthly averages. tAfter transferring December rise to 1896. fact that our plankton averages imperfectly repi'esent the actual production, it is still apparent that plankton production (per m.'') is inversely proportional, in a varying ratio, to the extent of the total movement in river levels. In not a few of the years (PI. VII.) the hydrograph falls into two periods dominated respectively by high- and by low- water conditions. This is seen typically in 1897, and with less contrast in 1894, 1895, and 1898, while in 1896 the recurrent floods obliterate almost all traces of such a division. In their totality these two hydrographic extremes pi'esent strong con- trasts which must bear important relations to the plankton production. They are, moreover, so involved with other fac- tors that no simple analysis of their results seems possible. For example, the high-water period is predominantly, as a rule, of low temperature, considerable ice, dilution of sewage, decreased light, increased access of recent water, increased current, de- creased occupancy by, and lessened growth of, aquatic vegeta- tion, and of greatly increased area and intimate and free con- nection with reservoir backwaters. The low-water period, on the other hand, is one of higher temperatures, no ice, concentration of sewage, increased light, reduced volume of tributary water, — which itself is of less recent origin than at high water,—slack- ened current, greater domination by, and growth of, aquatic vegetation, and of greatly contracted area and restricted con- 467 nectioiLS with and coutributions from the remaiuiiig backwaters. The complications arising from the combination of these vari- ous factors in varying degrees and the seasonal shifting of the two periods in the several years, render any sweeping general- izations impossible. The following table gives in parallel columns the monthly means of plankton per m.'' and of river stages in feet above low- water mark. PLANKTON PRODUCTION AND RIVER LEVELS. 468 recent origin of the flood waters of these months, from the di- lution of sewage and increased i-apidity of run-off, and from the reduction in time of impounding under these hydrogra|Dhic con- ditions. March-May is a period of rising temperatures, maximum flood levels, and increasing plankton production. The data, in so far as they go, indicate that high levels tend to increase pro- duction and low levels to decrease it, in some instances at least. These months witness the maximum and the initial de- cline of the spring flood, as a rule, and the greatest volume and principal run-oil of flood waters whose impounding has been more or less prolonged. • The proportion of impounded water is greater in higher than in lower levels, and we find, accordingly, production increased in the former and decreased in the latter, as a general rule. It is noticeable that levels in May slightly exceeding bank height, as in 1897 and 1898, yield much greater production (5.62 and 11.30) than levels not attended by overflow, as in 1896, when production falls to 1.80 at a level of 6.58 ft. The data of the June period are somewhat aberrant, in part as a result of insufficient data in some years, as 1895, and in part because of the relatively great irregularity of hydrographic conditions in this month in different years. The high produc- tion (30.42 cm.^) in low levels (1.88 ft.) in 1895 attends sewage concentration. The data for the remaining years conform in the main to the conclusions concerning production in the three prior months, namely, that high levels favor and low levels depress production, and for the same reasons above cited. A comparison of production in 1898 and 1898 in June yields con- firmatory evidence on this point. With July begins the low-water period proper, which con- tinues during the remainder of the year. Levels do not rise, in the means of the monthly averages, above 5 ft. in this pei'iod. The relation vfhich existed between production and high water in March-June is reversed in the period of July-November. An inspection of the table shows that in 20 of the 25 months in- cluded in the table in this period this reversed relation obtains; 469 that is, levels aborr the average are accompanied by a pill in production to an amount below the average, and those below, by a rise in production above the average. Thus, in this period higher levels depress production and lower levels tend to in- crease it. Two of the 5 apparent exceptions are in October and November, 1894, when insufficient data are available, and one is in July 1897, when the customary vei'ual conditions (PI. XI.) encroach upon the low-water period. The cause of this changed relationship of levels and produc- tion in these months of predominantly low water is to be found in the relation which summer rises in levels bear to the im- pounding function of the backwaters. These summer rises are rarely above bank height. They flush the channel, are not ex- tensively impounded in the backwaters, run off quickly, and accoi'dingly depre.ss production. The months of lower average levels are more stable, and, owing to slackened current, a more abundant plankton breeds, other things being equal, than in the more rapid current in the summer months of higher levels- The relations in December between production and levels are again reversed. Indeed, suggestions of this reversal appear in November. In this month minimum temperatures are again reached and higher levels prevail, and production now is higher in the years of high levels and falls below the monthly mean in every year of low levels. The cause of this relation does not seem to lie in hydrographic conditions. It may possiblj- be in- volved in the changed sewage and bacterial content of the channel that accompany the increased current and the decline in temperatui'es. The submergence of the summer's growth of vegetation in the margin of the river and its connecting back- waters in years of higher levels may also be a contributory factor. We thus find that in channel Avaters higher levels favor production only when they increase the impounding function and by long duration afford time for production and run-off of the plankton, and when they make available additional sources of nutrition. They depress production when they first appear, 470 and when they are of short duration and merely flush the chan- nel, as is predominantly the case during the low-water period. Lower levels depress production when they introduce stagna- tion conditions—as in the winter under the ice, and when they cut off contributory backwaters or otherwise reduce the run- off of impounded waters of long standing. They increase pro- duction when they lend stability to hydrographic conditions, increase the relative fertilization (sewage) of the stream, and by slackened current afford time for breeding. In general terms, production in the backwaters exhibits relations to levels similar to those we have described for chan- nel waters so long as the backwaters retain an intimate con- nection with the channel, that is, generally during high water, and for longest periods when, as in Thompson's Lake, the con- nection with the channel is most intimate. The diversification, as levels fall, of the several regions examined by us, renders generalizations impossible with respect to all of the backwaters, since one or another local factor sooner or later comes in to modify conditions. Moreover, some of the backwaters, as Phelps Lake, are cut off from the channel early and are not affected by changes in river levels, and, in general, the effect of the changes in channel levels, especially the minor ones, is reduced, equalized, or even obliterated, before it reaches the backwaters. The season at which the initial stages of the major flood of the year occurs, affects the subsequent production. Thus in 1896 and 1897 floods begin early in the year. The result is the carrying away in the run-off of great quantities of organic matter in suspension (Table X.) before they have had time to decay and yield up in solution their nitrogenous and other con- stituents for the support of the plankton. Temperatures are low in these months, decay is not rapid, and the plankton is not produced in large quantities. The result is that the stream is locally impoverished by this early run-off of matters in sus- pension and to some extent in solution. In 1898, on the other hand, the flood does not reach overflow stages till late in Feb- 471 ruary and is continued well into the early summer. Thus, while in 1896 and 1897 the vernal pulses of plankton production (9.89 and 5.62) are not large, in 1898 the production in this season rises to 35.68. Late high water, with decay and solution of organic substances increased by higher temperatures, occurs at the season of rapid plankton increase, and food matters which run off in the winter floods are here utilized and increase the amplitude of the plankton pulse. Winter floods thus tend to locally impoverish the plankton, and spring floods to in- crease it. Enough has been said to indicate the supreme importance of hydrographic conditions in the fluviatile environment in de- termining the amplitude of plankton production and in differ- entiating local areas in our environment. It is the prime fac- tor which distinguishes the fluviatile from the lacustrine envi- ronment, stamping the former with an instability as character- istic of the river as stability is of the lake. TEMPERATURE AND PLANKTON PRODUCTION. On pages 168-177 will be found a discussion of the temper- ature conditions at the various plankton stations and their gen- eral relations to the larger phases of plankton production. In the present connection the more detailed comparison will be made. To facilitate this comparison of production and tempera- ture conditions I have prepared the accompanying table (see p. 472), which gives the monthly means of production and of sur. face temperatures recorded at the times of collection in the river. This table in conjunction with the one following page 342 suggests the cooperation of temperature in controlling in a large way the seasonal fluctuations in production. In general, in the colder months less plankton is produced than in warmer months. Thus in the river the mean production in the 5 months below 45° is but little more than 9 per cent, of that in the 7 months above this temperature, in Phelps Lake, only 40 per cent., and 472 PLANKTON PRODUCTION AND TEMPERATURE. u 473 of rising and falling temperatures at the beginning and close of the season, there comes, as a rule, a decline in production from that of the vernal season. In channel waters this amounts to 16 per cent, of vernal production, or, omitting the single aberrant datum of June 1895, to 44 per cent. In the backwa- ters, owing to the comliination with various local factors, such as tributarj' waters and vegetation, the change from vernal pro- duction in midsummer vai'ies greatly in different localities. Thus, in Quiver Lake, where vegetation and the proportion of tributary waters is increased in summer, the decline in that season amounts to 87 per cent., while in Dogfish Lake, where veg- etation alone is the main disturbing factor, the decline is 74 per cent, of the vernal production, as seen in the April-May aver- ages. In Flag Lake, where also vegetation enters as a disturb- ing factor, the decline is 80 per cent. In Thompson's Lake, where disturbing local factors are less in evidence, it is but 69 per cent. In Phelps Lake, in contrast with all the other localities, pro- duction during the period of maximum heat exceeds that in the vernal season by 68 per ceiat. Thus the period of maximum heat in most localities attends a depression in production, but the exception in Phelps Lake is so striking as to preclude any conclusion that summer heat is necessarily inimical to large production, or that it is of necessity the most potent of the co- operating factors. The omission of the averages for Augu.st and September in 1898 from the Phelps Lake data would make the average production in the period of maximum heat 33 per cent, below that of the vernal months, and bring this locality into agreement with the other stations as to the depressing effect of summer heat in plankton production. It should be emphasized in this connection that these conclusions apply to catches of the silk net only, that the summer temperatures of our waters approximate 80° on the average and frequently rise above it,andthat temperature is only one of the factors involved. Following the period of maximum summer heat is that of decline in October-November—including also a part of Sep- tember, or even December in some seasons—to the winter min- 474 imum. In general, this is a period of declining plankton in channel waters, where production in these two months falls 71 per cent, below that in the preceding four months of maximum heat, and in Quiver Lake, where it falls 48 percent, below. On the other hand, in the rest of the backwaters there is a slight in- crease in these two months as compared with the production in the period of maximum heat. In Dogfish, Flag, Thompson's and Phelps lakes the October-November increase in percent- ages over the average summer production in each of these sev- eral localities is 33, 362, 26, and 11 per cent, respectively. In view of these divergent tendencies in production under similar temperature conditions it is evident that other factors are operative, or at least more potent, in controlling autumnal pi'o- duction. The October production is as a rule higher than that of November, and suggests a tendency towards an autumnal pulse comparable with the vernal pulse but of lesser amplitude. The vernal pulse occurs in rising temperatures of 60°- 70°, and this autumnal one in falling temperatures of 60°~50°. The month of December does not on the average quite at- tain the minimum winter temperature, though in some years, as in 1897 and 1898, it approached closely to it. Neither does the plankton production drop to so low a level on the average or in individual years in channel waters as during the two colder months which follow. In general the same relation ex- ists in the backwaters, though exceptions occur—principally in Thompson's Lake. Thus, in a large way, temperature plays an important part in controlling plankton production. Additional proof of its potency is to be found in the correlations between production and exceptional divergences from the normal course of tem- perature changes, such, for example, as early or late vernal rise or autumnal decline. The accompanying table (p. 475), kindly furnished by Mr. W. G. Burns, Section Director for Illinois of the U. S. Weather Bureau, gives the vernal air temperatures for 1896-1898, and permits a comparison with the course of plankton production. 475 MONTHLY MEANS OF VERNAL TEMPERATURES FOR ILLINOIS AND OF PLANKTON PRODUCTION—CM.' PER M^ 476 So also in 1897, the means of our records of water tempera- tures for September, October, and November of that year are 2.9°, 7.6°, and 2.7° above the average of the monthly means for all years. This maintenance of high temperatures into the pe- riod of normal autumnal decline is apparently o)i(' of the fac- tors tending to make production in these months of this year greatly exceed that of the same season in other years. In chan- nel waters in these months of 1897 (see table following p. 342) production is from 13 to 250 per cent, above the mean of all years, and often 10- to 20-fold that in other years. In Thompson's Lake the excess in 1897 is even greater, ranging from 87 to 233 per cent, of the mean of all years, and from 1.6 to 28 times that in the same months in other years. The higher temperatures do not suffice, however, in the case of Quiver Lake, to overcome the other factors tending to depress production there in these months, and we must conclude that, although all-pervading and potent, temperature is nevertheless not always pre-emi- nent among the environing factors of the plankton. We thus find that in a general way, in conjunction -with other factors, rising temperatures tend to increase, and falling to decrease, plankton production, and that in the same locality the warmer months generally yield more plankton than the colder ones. On the other hand, minimum temperatures when once established are not of themselves inimical to a considera- ble plankton production. Evidence of this is to be found in the not infrequently increased production in December over that of several months preceding. This is perhaps most notice- able in the records of 1898. Thus in channel waters the am- plitude of the December pulse (PI. XII.) exceeds that of all other months since the last of June, and the December maxi- mum in Phelps Lake (43.14) exceeds in amplitude all ofIier pro- duction in our records for 1898 in all of lie)' localities ssive only the single apex of the vernal pulse (51.39) in Thompson's Lake. It is, however, only about one fifth of the August maximum (224.48, PI. XLII.) in Phelps Lake itself, so that the depressing effect of lower temperatures is still apparent if we limit com- parisons to a single locality. 477 The effect of the autuiiiual decline, and. iu general, of low- ered temperatures, in depressing production is apparent in not a few instances in our records. It can be seen in the October- November thermographs and curves of plankton production of channel waters iu 1894-1898. and in those of Quiver Lake for the same j-ears; is much less apparent in Thompson's Lake, especially in 1897. even »when temperatures have fallen : and is often but feebh* developed in Dogfish and Flag lakes in 1895-1897, while in Phelps Lake iu 1896 and 1898 there are pulses of considerable magnitude (^51.6, and 99.86) in this period of decline of temperature. The minima demarking these pulses are. however, of less than the usual amplitude. This depre.ssing effect is thus traceable in all localities, but is bet- ter developed in stream than in lake waters, appearing most clearly in the channel and Quiver Lake, where, at this season of the year, tributary waters are present in considerable propor- tion. Our water temperatures and the records of the United States Weather Bureau at Havana and elsewhere in our lo- cality reveal many instances of heat pulses at various seasons of the year. There is little regularity iu their duration or amplitude. When plotted from the means of the tri-daily readings of the air temperatures at Havana they do not ex- hibit delimitations as well defined as those, for example, of a fully observed plankton pulse. Their amplitude, except in winter months, rarely exceeds 20° between extremes, and their duration is usually less than a fortnight between minima. That these fluctuations affect the course of plankton production can- not l)e doubted. A detailed comparison of the course of produc- tion iu 1896 and the thermograph of that year will show that, predominaiithj, rises of temperature attend or precede rising pro- duction, while declines in heat are often correlated with de- creased production. This may be largely coincidence, or. iu some cases, the common effect of cooler, barren flood-waters, especially in the case of the recoixis of channel production. A close comparison, however, of the plauktograph in Phelps 478 hbSl 479 Lake—where flood factors are largely excluded—and the ther- mograph (air) for 1896 will serve to suggest the possibility of a causal nexus between the two phenomena of fluctuations in heat and some of the movements in plankton production. The many exceptions to any close correlation emphasize, however, the fact that Jtntt is only one of the many factors involved in the problem, and also indicate the necessity for much fuller plankton data, with closer interval and the proper quantitative representation of the minute forms now lost Ijy leakage through the silk, for any adequate discussion of the problem. The present data serve only to suggest the problem for investiga- tion. The effect of the ice-sheet upon the course of plankton pro- duction is apparent in a number of instances in our records. The most noticeable case was the extermination of the plank- ton in the channel in February, 1895, by the ice-sheet of two months' duration; but this catastrophe was not repeated else- where in our records in this or other years. Indeed, owing to the fact that the period of the ice l:)lockade is usually one of lower levels and more stable conditions, we find generally that production under the ice, even at minimum temperatures, rises above prior or sujjsecjuent levels. An inspection of the plates, especially those of 1898, will show repeated instances of this phenomenon in both channel and backwaters. One of the most striking phenomena in all our records is this winter pro- duction under the ice-sheet 'in 1898-1899, a production which in the river attained an amplitude in December (.99) not equaled since June, and in February (.81) one surpassed only by the August (.91) and December means. In Quiver Lake likewise, the December (1.74), January (.77), and February (1.05) means are all considerably in excess of the June-Novem- ber production, the average of the winter months (1.19) being over threefold greater than that of the warmer months (.33) named. In Thompson's Lake also the midwinter production in this season was large, reaching an average of 1.94 for the winter months above named, and only 1.96 for the five preceding 480 months, while the amplitude of the December and January pulses was surpassed but once from June to December. Con- ditions under the ice at minimum temperatures were thus in these years and localities quite as favorable to the quantita- tive development of the plankton as were the conditions prev- alent in summer and autumn. As a whole, then, temperature changes bear an important relation to the course of plankton production, but at times they are not more potent than other factors. An abundant plankton may develop at any temperature within the normal seasonal range provided other factors favor it, but generally the ampli- tude is less in lower or in falling temperatures, and greater in higher or in rising ones. The relations here discussed between the volume of plank- ton and temperature depend primarily upon adaptations of particular species to temperature—a subject which will be dis- cussed in another connection. LIGHT AND PLANKTON PRODUCTION. There are at hand no adequate data on this subject, and it is, moreover, complicated with the thermal and other forms of solar energy and with the problem of turbidity in the water itself. No detailed comparison is afforded by the data, espe- cially since the more minute forms are not adequately repre- sented by the catches of the silk net, and it is largely these synthetic organisms, chlorophyll-bearing algae and flagellates, which are most dependent upon light for their growth and re- production. Our data alike of light and plankton are thus deficient. Nevertheless, in the chain of relations, the catch of the silk net—largely of animal plankton—is, at most, but a few links removed from these synthetic organisms, and it must therefore in some measure reflect their quantitative fluctua- tions. Our data suggest a few inferences concerning the rela- tion of light and plankton production. The period of greatest illumination lies between March 20 and September 22, and owing to the proximity «f these dates 481 to the ends of the mouths it will be possible, for the purpose of utiliziug our data in monthly totals and means, to divide the year into two periods, April-September and October-March, of greater and less illnmiuation respectively. The contrast in illumination is further heightened by the fact—derived from the following table of cloudy days—that the number (at Havana, 159) of cloudy days between the vernal and autumnal equi- noxes is only about one half that (311) between the autumnal and vernal. On the avei-age, the season of greatest light is also the season of greatest production. Thus, in the channel waters average monthly production in April-September (4.76) is seven- fold that in October-March (.67 ), and in the backwaters, such as Quiver, Thompson's, and Phelps lakes, it is respectively 5-, 2.2-, and 1.6-fold greater. The records of individual years in all of the localities will Ije found to exhiljit a similar relation- ship. We may infer, accordingly, that the increased light be- tween the vernal and autumnal equinoxes tends to increase production, and that the decreased amount in the remainder of the year tends to lower it. It operates, of course, in conjunc- tion with other factors, and our records contain not a few in- stances where production in the period of less illumination ex- ceeds that in the period of greater light. For example, on December 20, 1898, in the ni'niiniiiiii illumination of the year and under an ice-sheet 21 cm. thick, which still further reduced the light, the plankton production in Thompson's Lake reached an amplitude (2.58) exceeding that on June 21 (2.4:7) in the same lake in iiKi.rliiunii illumination, an amplitude, moreover,sur- passed but once from June to October. Other factors are thus, at times, at least, more potent than light in controlling production. A phenomenon of like import exists in the conditions of illumination and production in Quiver and Thompson's lakes. Both lakes are of approximately the same depth, but the former, especially in low-water conditions, has renuirkably clear water, the bottom being generally visible, while the latter is always more turbid, and light penetrates the water far less completely. Nevertheless, the lake, with most illumination, 482 yields least plankton. The factors of vegetation of the coarser sort and of tributary waters serve here to modify and overbal- ance light as a controlling factor in production. From the data of the U. S. Weather Bureau at Springfield, Mr. W. G. Burns, Section Director, has kindly furnished me the records of the number of cloudy days per month in 1894-1899 observed at Peoria, Havana, and Springfield. These are given in the accompanying table. The records for Havana have also been plotted on Plates VIII.-XIII. in the uppermost row of squares. 483 impossible. That the reduction in light due to clouds does in a measure affect production might be inferred from the August -October records in 1896 and 1898. In the two years named, cloudy days and production in August are 2 and 8, and 1.12 and .91 cm. ^ per m.^ respectively ; in September they are 11 and 2, and .38 and .69; and in October 3 and 16, and 1.11 and .24. Hy- drographic conditions are not remarkably different in the two years, and while their differences in this respect are doubtless potent, ORusing differences in production, it still seems prob' able that the fluctuations in light are also operative. In any event in these three months the mean production runs higher in the year of fewer cloudy days and lower in the year of less sunshine. Similar relations will be found to exist generally in the production of the backwaters for these months (see table following p. 342). The statistical data of the synthetic organisms to be discussed in Part II. of this paper still further serve to demonstrate the correlation of light and plankton pro- duction. The necessity of light for the process of photosynthe- sis on the part of the phytoplankton places this factor at the very beginning of the chain of relations whose later links are the larger animals of the zooplankton which constitute the greater proportion of the volume of the catch of the silk net — the basis of the present discussion. VEGETATION AND PLANKTON PRODUCTION. It is evident that our investigations afford a unique oppor- tunity of determining the effect of vegetation (the word being here used to refer to the coarser aquatic growth as distin- guished from the microscopic phytoplankton) upon the course of plankton production with reference to both its volume and constitution. The conclusions to be drawn from our observations with reference to volumetric production, already suggested in the detailed discussion of production, will be summarized and dis- cussed here, though some of the data upon which they rest lie outside the scope of the present paper. 484 1. Other things being equal, bodies of fresh water free from vegetation (sul)merged macro-flora) produce more plank- ton than those rich in such vegetation.* Thus, the amount of plankton produced (as indicated by the averages of all of our collections in the several localities examined) in our open vpaters is from tvs^o to eleven times as great as it is in our lakes closed by vegetation. As shown in the table on page 429, the average planktons in Thompson's and Phelps lakes are 7.94 and and 19.65 cm.^ perm.^ respectively, while in Quiver and Dogfish lakes the quantity is only 1.70 and 4.22. Flag Lake, Avith an average of 11.46 cm.^ is an interesting exception to this con- trast which will be discussed in another connection. The con- trast is even more striking if the averages of the monthly averages for all the years are made the basis of comparison, as in the following table and diagram. COMPARISON OF PLANKTON PRODUCTION IN VEGETATION-POOR AND VEGETATION-RICH WATERS. Month January February.. March April May June July August . . .. September . October. ... November. December. Gr'd av. of monthly av. Vegetation-poor Quiver Lake ^."f^f • 27 .67 •77 7.26 6.85 1.25 .78 •77 •77 .69 •23 •63 175 •53 1 . 10 i.g6 10.50 5^79 ••75' 1.95 2.51 2^39 3^05 2.64 3-76 3^i6 Vegetation-rich Tho/^Pf"'s Phelps Lak 37Q 1.27 2.96 14^49 29.59 10.66 4^74 6. 19 5^37 10.64 6^39 3.08 8.26 3^29 5.68 5.68 11.77 25-33 11.40 8.50 58.12 47-25 27.68 41-57 21 .96 22.35 Ratio :9 4 3 '•5 4 7 5 20 17 10 17 :6 1:6 On this basis, the waters full of aquatic vegetation pro- duce throughout the whole year less plankton than waters free from such gi-owths. Relatively few exceptions to this rela- * This relation of vegetation to the plankton may be formulated as follows: The amount of plankton produced by bodies of fresh water is, other things being equal, in some inverse ratio proportional to the amount of its gross aquatic vegeta- tion of the submerged sort. 485 OulverLahe, Dogfish ijke. ThompacnLakc Phelps Lahe,- Fig. D.—Seasonal distribution of plankton production in veKetation-poor and vegetation-rich waters, based on the averages of the monthly averages for all years of collection. tion will be t'oimd in the individual collections recorded in Tables V. and IX. and VI. and VII., or in the monthly averages of the table following page 342. This striking contrast is still more enhanced by the statement of the monthly ratios of pro- ductivity in waters rich and poor in vegetation. The.se range from 1 to 1.5 in April to 1 to 20 in August. The fluctuations in the ratio are of themselves very significant. During the period from February to July inclusive the ratio is at its lowest, ranging from 1 to 1.5 to 1 to 7. Excepting only the month of July, this is the period of high water, in which the vegetation, if present, occupies a much smaller propoi-tion of the volume of the lake, and is therefore to a proportionate degree restricted in its effect upon the plankton. Under such flood conditions these several localities are more or less merged in the general 486 overflowed district, and are to a varying degree traversed by waters from the bottom-lands above and adjacent to them, and the purely local factors of their environment, such as vegeta- tion, thus become less potent. Again, it is not until the latter part of this period that the vegetation attains the development which continues throughout the remainder of the summer. The relative barrenness (in plankton) of the vegetation-rich waters is thus least striking when the vegetation is least in evidence. During the period from August to November inclusive the ratios are very much higher, rising to 1 to 16 or 20. This is the low-water period, when the vegetation in the vegetation-rich lakes is at its maximum development both in quantity and in the relative volume of the lake occupied by it. It is also at such times that these several bodies of water are more distinct units of environment, with their local factors no longer merged by flood conditions. The relative barrenness of the vegetation- rich waters is thus greatest when the vegetation is at its maxi- mum development and is most emphasized as a factor in the environment. The conclusion from this comparison of the mean production of plankton in vegetation-rich and in vegetation-poor waters in our locality is thus inevitable that vegetation (in the usual sense of the word) is inimical to the development of an abundant plankton. It may also be said that the contrast would be considerably heightened if it were possible to elimi- nate from all the collections on which this comparison is based the adventitious organisms—such as small insect larvag, mol- lusks, oligochaetes, Hijdm, etc., which form a considerable vol- ume of many of our catches in the vegetation-rich waters. On the other hand, it must be maintained that the vegeta- tion is only am of the factors concerned in the phenomenon pre- sented by this contrast. It is quite probable that other fac- tors, especially the current, tributary waters, and the chemic- al constituents of the water, affect the problem in hand. Dur- ing high water both Thompson's Lake and the Dogfish-Quiver region are traversed by a considerable current from the bottom- 487 lands above. The elevated deposits of Spoon River and the consequent crowding of the channel of the river to the east bluff at Havana force all of the water of overflow (at stages below about 16 feet) to seek the main channel. The configura- tion of the low-lying bottoms above is such (see PI. II. I that the lakes in question form natural channels for the movement of a large body of impounded water. This movement is well marked at stages above eight feet. So far as I am alile to judge from field observations, the current conditions in Thompson's Lake and the Doghsh-Quiver area ai'e not greatly different. The current continues in both lakes as levels fall to six feet, at which level Thompson's Lake loses its connection with the river through the "cut road" (PI. II.), and movements in it at lower levels are confined to those due to ingress and egress of water through the slough, and are consequently inconsider- able. On the other hand. Quiver Lake continues to be traversed by the discharge from Quiver Creek, and onr collections were usually made in the channel in the vegetation. In Dogfish Lake at low stages there is no current traversing the lake. Phelps Lake lies at so high a level that only the Hoods exceed- ing eleven feet bring it into connection with the general cur- rent of overflow, which in this case generally comes from Spoon River. Below this level the only movement in its water is the gentle one due to the receding flood. So far, then, as the current is concerned, it is a common though not equally distributed factor at high-water stages in all areas compared, while at low water it is an important feature in the environ- ment in Quiver Lake but is practically absent in the other three localities. This fact undoubtedly accounts in part for the barrenness of the waters of Quiver Lake (1.53 cm.'' per m.'. or only .55 for the average of low-water periods—i. e. below 5 ft. ) as compared with those of Dogfish (4.22), Thompson's (^8.13) and Phelps (19.44) lakes. This current does not. however, traverse or appreciably affect the waters of Dogfish Lake, and their bar- renness still remains for contrast with the productiveness of the vegetation-poor waters of Thompson's and Phelps lakes. 488 Data are not available for a full comparison of the chemic- al constituents of all the waters here under consideration. No data whatever are available for Dogfish and Phelps lakes, and only sanitary analyses for Quiver and Thompson's lakes. These shed no light on the relative amounts of phosphates and carbon dioxid in the water, both important elements in the growth of plants. On the other hand, data for a comparison of the ammonia and the nitrates are found in Tables XII. and XIII. and Plates XLVIII. and XLIX. The nitrates, in so far as they are concerned,—as shown in the accompanying table, CHEMICAL ANALYSES, SEPTEMBER, 1897, TO MARCH, l8gg. AVERAGES OF ALL ANALYSES—PARTS PER MILLION. 489 It is different, however, iu the matter of nitrites, which are about twice as alnindant in Thompson's Lake as in Quiver, and the same ratio also holds approximately for the free and al- buminoid ammonia. These are all substances indicating or- ganic matters iu the process of decay, or available for decay and thus for plant nutrition. These data show clearly that iu these particulai's the waters of Thompson's Lake are much richer than those of Quiver, and this difference is undoubtedly one of the factors on which the contrast in plankton produc- tion depends. The more al)undant plankton of the former lake may itself be one of the sources contributing to the or- ganic decay here indicated. This contrast in the chemical constituents in part at least follows from the sources from which the waters in the two lakes are derived. River water impounded from receding floods and more or less charged with sewage and industrial wastes constitutes the principal source of the water in Thomp- son's Lake. Spring and creek waters replace this very slowly, and every rise in the river introduces a new supply of richly fertilized water which at levels alcove six feet trav- erses the whole lake. Quiver Lake is subject to like invasions, but its more abundant supply of creek and spring water coun- teracts their influence to some extent and soon replaces their contributions. A consideration of these other factors, cun'ent and chem- ical constituents, makes it probable that they also are efticieut in causing the contrast in plankton productivity in the two lakes. How much of this contrast is due to vegetation and how much to their agency is a matter upon which conclusive evidence is needed. Experiment in the field may yield con- clusions that will be final. Some evidence corroborative of my contention that the vegetation of Quiver Lake is inimical to the development of its plankton is afforded from two soui'ces; [D the examination of Matauzas Lake, and {'2) the comparison of the plankton pro- duction of Quiver Lake in years of abundant and scant vegetation. 490 Matanzas Lake (PI. II.) more than any other body of water in our field of operations resembles Quiver Lake in the various factors of its environment, only upon a somewhat smaller scale. Like Quiver Lake it has free communication with the river at all levels, is subject to the same conditions of invasion and submergence, has an eastern sandy and springy shore with lit- toral vegetation, a western one of alluvium, and, between, a bottom changing from sand to mud. The depth and bottom configuration are very similar, and there is a supply of creek and spring water roughly proportionate to the size of the lake. The two lakes are thus strikingly alike save only in the matter of vegetation. The vegetation in Matanzas Lake is confined to a narrow belt of the littoral zone along the greater part of the eastern margin and to a little Ceyatoplujlhuii adjacent to it and fringing the western shore in places. Less than 5 per cent, of its area is thus occupied. Quiver Lake, on the other hand, has at all times a more abundant flora, which even in the years of its least development holds possession of not less than 30 per cent, of its area. Under these circumstances a comparison of the production of the two lakes should throw some light upon the effect of vegetation upon the development of the plankton. No chronological series of collections has been made by us in Matanzas Lake. A few isolated collections have indicated that it is rich in plankton, and two thorough tests of the local distribution of the plankton, made in 1896, afford a basis for comparison with Quiver Lake at that time. Fifteen collections with the plankton pump were made in various parts of Matan- zas Lake on July 9, and twenty-five, similarly distributed, on August 14. The averages of the plankton per m.'' of water in these collections and the amounts found in Dogfish and Quiver lakes on the days following (July 10 and August 15) are given in the accompanying table (p. 491). Averages for the months of July and August in the several years are also given for Quiv- er and Dogfish lakes. The production of plankton in Matanzas Lake on the dates 491 PLANKTON PRODUCTION IN MATANZAS LAKE COMPARED WITH THAT IX QUIVER AND DOGFISH LAKES. Date 5 \ July Q, 10 ' ] August 14, 15 1894, \ July-... / August. i8q5.5l"'y--;- -" I August. i896,.iJ,"'y--;-^ ' I August. 1897, Uulv.... I August, 1898, -^J"'y--;- ^ I August . ^-"s-.)ffiusi: Matanzas Lake 2.40 6.20 Quiver Lake •30 3.42 2.20 0.74 0-37 0.21 0.30 2.46 89 21 .16 ,22 .78 77 Dogfish Lake 1. 18 2.90 2.99 I . II 0.91 3-91 1.95 2.51 of collection above indicated is approximately twice that of Quiver and Dogfish lakes, where vegetation was at that time somewhat more abundant. In 1S9C Quiver Lake was freer from vegetation than at any other time in the period of our opera- tions, and the contrast between the production of the two lakes appears greater if we consider other years or the average for all collections in the months named. On the latter basis the ratio rises to 3 to 1 for July and S to 1 for August in the comparison of Matanzas and Quiver lakes. In the case of Dog- fish Lake the contrast is less striking, but still evident. Ma- tanzas Lake, similar in its environment to Quiver Lake save in the matter of vegetation, thus produces a more abundant plank- ton, and we may infer that the vegetation of the latter is in- imical to the development of plankton in its waters. A second line of evidence bearing upon the question under discussion is to be found in the production in Quiver Lake itself under different conditions of vegetation. In LSy4. and still more in 1895, owing to low water in early summer, vegetation was very abundant in Quiver Lake. The growth of Cerafo- pln/IIiiiii and Elodci choked its waters from shore to shore and from bottom to surface except in a narrow poorly dehued chan- 492 nel found in the lower end of the lake. This part of the lake is shown in Plates XV. and XVI., which portray the conditions as they appeared in 1894 and 1896 respectively. The upper end of the lake and its western arm, Dogfish Lake, are shown in Plates XVTI. and XVIII., the latter having been photographed in 1896, when the center of the lake was not so full of "moss" as during the preceding year. The repeated floods of 1896 swept the lake of much of its vegetation, and during the three following summers it never recovered the abundant flora which it presented in 1895. In 1897 and 1898 there was also much less vegetation than in 1895, though somewhat more than in 1896. The plankton production, as shown in Table V. and graphically presented in Plates XXV.-XXIX., does not uni- formly rise and fall as the vegetation decreases or increases. The phenomenon of its fluctuations involves many other fac- tors, among which the effect of vegetation may perhaps be de- tected. The average production for the years of vegetation, 1.08 and .78 cm.-' per m.'' of water, is surpassed in 1896 (2.59) and 1898 (2.44) but notin 1897 (.88). The marked increasein 1896 over the production of 1895 parallels the great change in vege- tation, and is also accompanied by higher water, the average for the year being over three feet above that of 1895. This differ- ence in levels also tended to decrease the relative extent of the vegetation in 1896. In Dogfish Lake also the contrast in vege- tation in the two years, 1895 and 1896, is well marked, and the average plankton production rises from 3.25 to 5,01 cm.'* per m.' The omission of winter collections in 1895 makes the contrast less striking. Allowing for this, it is probable that the plank- ton production is practically doubled in the year of decreased vegetation. This is approximately the ratio of increase in Quiver Lake in 1896 and 1898. Other causes, such as current and chemical conditions, doubtless share in producing this change in the plankton, but it seems highly probable that the reduction in vegetation caused a considerable part of this doubling in the plankton production. A, comparison of the plankton production of the same body of water (Quiver and 493 Dogfish lakes) in different years thus shows that more plank- ton is produced in j'ears of little, than in years of much, vege- tation, and tends to confirm the view that abundant submerged vegetation is inimical to the production of plankton. An insjiection of the ])lanktographs in Plates VIII.-XTII. and XXV.-XLII. shows the frequent occurrence of an autumn maximum, often well defined. In the planktographs of Quiver and Dogfish lakes, this autumn maximum is usually depressed or mi.ssing. The spring maximum occurs, as a rule, while the lake is full of water from the general overflow, and it is therefore not purely a local phenomenon. The midsummer and autumn plankton, on the other hand, is entirely a local product, and the depression of the autumn maximum must be due to local influence.s. In 1896, in both Quiver and Dogfi.sh lakes the autumn maximum occurs in two or three sharply marked prominences, that of Octolier 14 (3.52 and 6.(50) being • a Mclosini-Sijiu-lni'fa assemblage, typical for the autumn season. This was a year in which there was little vegetation and high (for autumn) water, the vegetation 1)eing, consequently, at a minimum as a factor environing the plankton. In other years this autumn maximum (.see Tables VI. and VII.) is less evident. In 1894 the apparent maximum on September 5-6 is almost wholly due to the development of Oscil/firia at a time of local stagnation consequent upon l)ack\vater. In 1895 there was in Quiver Lake a maximum onSeptember6 (1.57), due in part to au- tumnal plankton and in part to adventitious organisms. Dogfish Lake exhibits a somewhat larger maximum (4.65) on Septem- ber 17, which is mainly normal in its components. The No- vember-December maximum of 1895 in this lake is wholly due to adventitious organisms, and may he disregarded in this con- nection. At their best, these maxima in vegetation-rich years are but one half to one third the magnitude of those of 1896, a vegetation-poor year. In Quiver Lake in 1897 and 1898 the autumn maximum is again depre.ssed almost beyond discern- ing. Although vegetation was not abundant in the lake in these two years, the period of the autumn maximum was one 494 of prolonged low water in both years, so that whatever vegeta- tion was present occupied relatively a large proportion of the area and volume of the lake, especially as contrasted with the conditions in 1H96. The available data thus indicate that veg- etation is inimical to the production of plankton, as shown not only in the general averages but also in these maxima, which may be regarded as the expression jjar excellence of the produc- tive capacity of the lake. There still remains for consideration, with reference to the effect of vegetation upon plankton production, the result of our examination of Flag Lake. As before stated, this is a marsh choked with a rank semiaquatic growth whose extent, abundance, and relative occupation of the area of the lake equals or exceeds that in any other body of water examined by us. If our thesis that vegetation is inimical to the production of plankton be true, we might expect to find here, of all places, barren waters. This is not, however, the case ; for, as shown in the table of comparison of plankton production on page 429, Flag Lake is very productive (1L46 cm.^ perm.^), being ex- ceeded only by Phelps Lake (19.65). The only indication that vegetation is in the least inimical to the plankton in the lake is suggested in Plate XXXIIL The amount of plankton present from May 15 to October 1, the growing period of vegetation, is only 2.87 cm.^ per m.-', while in spring and late autumn (April 1 to May 15 and October 1 to December 30) it is 32.89. In Phelps Lake, which, save for vege- tation, is much like Flag Lake, the plankton during the period of dominance of vegetation in 1896 averages 7.64, and in 1898 52.43 cm.'', 3 to 18 times as much as in the vegetation-rich waters of Flag Lake. In the character of the vegetation in Flag Lake lies, I be- lieve, the explanation of its fertility in plankton. Two kinds are predominant, neither of which is present in like abundance in Quiver Lake. These are (1) succulent vegetation, such as Sagiftaria, Potitederia, Nijinphxea, and Nelumho, which die down and undergo considerable decay in the early fall, and (2) the 495 emergent vegetation, principally Scirj^is, which, on account of its growth and structure, does not reach an advanced stage of decay until ice and winter floods have broken it down. With rising spring temperature it yields to decay and releases a great store of nitrogen which the phytoplankton can utilize. Both of these types of vegetation are rooted in the humus and allu- vial deposits of the lake, and both are to some degree emergent. They thus draw their supply of food I dissolved salts and gases) largely from soil waters and the air. and less from the supply in solution in the water of the lake. The submerged and non- rooting vegetation {CeratophijUnm and Elodea) is not abundant in Flag Lake, so that the food supply in the lake waters is not drawn upon to any great extent by the aquatic vegetation, and it thus becomes available for the phytoplankton, which, in turn, supports the zooplankton. The products of decay of the succulent and emergent vegetation, on the other hand, are in large part released directly into the lake waters, and at times (^fall and spring) when the plankton i-eaches its greatest devel- opment in this region. Owing to its character and to the pro- tected situation of the lake the vegetation is never swept away by floods, nor is the lake traversed by any marked current as are both Thompson's and Quiver lakes. The fertilizing effect of the decaying vegetation is thus more localized in this region than in the otiier bodies of water examined by us. The data from Flag Lake thus throw light upon the effect of emergent and rooted vegetation—which is typically of the littoral type—upon the plankton. They indicate that this kind of vegetation favors the development of the plankton by add- ing to the food materials in the water, while at the same time it does not to a large degree compete with the phytoplankton in the consumption of the food thus released by its decay. In 1896 a series of examinations of the local distribution of the plankton in Quiver. Matanzas, and Thompson's lakes was made by the pumping method, and since the collections were made in the areas of vegetation as well as in the open water they might also be examined to determine, if possible, the effect 496 of vegetation on the distribution of the plankton. Only the quan- titative data are at present available, and the results are con- flicting. In some cases the plankton is greater in the vegeta- tion than in the adjacent open vv^ater ; in others the reverse is true. These examinations w^ere made at times of unstable river levels, and the movements of water consequent thereupon make any satisfactory analysis difticult. The general conclu- sion that lakes full of vegetation (Quiver) are everyvphere poor in plankton, while those relatively free from it (Thomp- son's and Matanzas) support generally a more abundant plank- ton is in all cases upheld by these examinations. This poverty of the plankton in vegetation-rich lakes vras one of the surprises of our investigations, and, so far as I have been able to ascertain, it contradicts the general expectation among observers of aquatic life. It has its parallel in the pau- city of life in tropical forests and among the pines and red- woods of the Sierras. It is fundamentally a problem of nutri- tion, and inheres in the utilization of the available food supply by a single type, or a few types, of plants which do not them- selves in turn afford support for an abundant or varied animal life. Wherever the depth of the water, the currents, the winds, or other factors, prevent the development of a submerged aquatic flora, the nutrient materials for plant growth^the oxy- gen, the carbon dioxid, the nitrates, phosphates, sulphates, and carbonates dissolved in the water—are utilized by the phyto- plankton, which, in turn, supports the zooplankton. The en- tire production of such a lake takes the form of plankton and, in turn, of those larger species, insect larvae, mollusks, and flsh, which are directly or indirectly supported by it. When, on the other hand, the conditions are such that a submerged non- rooted aquatic flora obtains possession of a lake,—as, for exam- ple, CerafopJii/lluiH and Elodea in Quiver Lake,—these nutrient materials are appropriated by it to the great reduction, even practical exclusion, of the phytoplankton. In the struggle which must ensue between the phytoplankton and the sub- 497 merged aquatic flora for the possession of a body of water capable of supporting either, the greater duration and perma- nence of the larger plants which constitute the submerged flora must in the long run inure to the advantage of the latter, hence they predouiinate over the phytoplankton wherever other conditions favor their appearance. This coarse sub- merged vegetation cannot in its living condition be utilized by the minute organisais of the zooplankton, and only such as feed upon it in decay cau tiud sustenance in the vegetation-rich lake. The absence of an abundant phytoplankton and of the greater part of the zooplankton may thus be accounted for in waters rich in sul)merged and non-rooted vegetation. The total production of such a body of water consists mainly of a large amount of coarse aquatic vegetation, which but few ani- mals can utilize in its living condition as food, and a much re- duced plankton, largely of animal constituents, together with such larger and often attached species as And food in these elements. Some light on the relation of vegetation and plankton to certain of the chemical constituents of the food of the aquatic flora can be gained from a compari.son of Plates XLV., XLIX., and L.. and Tables X., XII., and XIII., which show the results of analyses in 189S. The appended table also gives the average AVERAGE OF ALL ANALYSES—PARTS PER .MILLION. Station 498 A comparison of these two lakes indicates more nitrates in Quiver than in Thompson's (.68 to .53 parts per million)— a phenomenon which may be explained by the proximity of the former to the river and thegreaterinvasion by its richer (.81) wa- ters. In the matter of free ammonia Thompson's Lake is much the richer (.352 to .138 parts per million), though it falls con- siderably below the river (.95) in this particular. The .striking feature of the diagrams and tables is the marked reduction in nitrates and free ammonia during the period of growth, from June 1 to October 1, in both lakes as contrasted with that of quiescence, from October 1 to June 1. The former period is one of higher temperature and less flood water, thus favoring the process of decay and the concentration of its products. The marked decrease in both the free ammonia and nitrates during this period may be explained by the utilization of these prod- ucts of decay by the chlorophyll-bearing organisms, which presumably are much in excess of those of the colder period. In Thompson's Lake the phytoplankton would be the principal consumer, while in Quiver Lake submerged vegetation assumes this role. The uniformity in the nitrates throughout this period, and the reduction to a similar amount (about .2 parts per mil- lion) in both lakes are significant of some sort of an equi- librium between the supply furnished by decay and its utiliza- tion in the growth of plants. This phenomenon of reduction of nitrates to a summer equilibrium is to some extent manifest in the analysis of soil waters (see Palmer, '97), and may in like manner be attributed to utilization of the nitrates by vegetation. At first thought the volume of submerged vegetation seems large in comparison with that of the phytoplankton, which it replaces ; but when the permanence and persistence of the con- stituent cells of CeratophijlhiiH are contrasted with the many generations of the algaj and diatoms of the plankton which arise during a season's growth, the difference is less evident. Furthermore, a much greater proportion of the cells of the phytoplankton contribute directly to the growth of the animal life of the lake. 499 The submerged vegetation—.such as that found in Quiver Lake—aiJects the conditions of nutrition in other ways than those above indicated. The absence of roots and the slight hold which its lowermost stems can obtain upon the soft bot- tom facilitate its removal by floods and seines, and the nutri- ment stored in its tissues is thus taken from the lake, and its waters are impoverished to that extent. Again, both Ccrafo- plijllhiDi and E/odca are perennial, continuing beneath the ice from year to year and never wholly yielding to decay. The lit- toral vegetation of Flag Lake, with its large annual growth and well-marked periods of decay in autumn and spring, contrib- utes more generously to the enrichment of the water. Thus, while robbing the water of its food material, the submerged vegetation often fails to make equivalent returns. The submerged vegetation also interferes with the free operation of certain other factors which affect the plankton of open water. It shuts out the sunlight, and effectually modi- fies the temperature thereby. Thus, on a midsummer day the water in Thompson's Lake rarely shows a difference of more than three degrees (Fahr. ) between surface and bottom in two meters of water. In the vegetation, on the other hand, the temperature contrast is much greater and within much narrower limits. On July 15, 1897, when surface waters were at 88.2°, the temperature was but 80° at 15 cm. below. The diurnal range of temperature is thus much less in vegetation than in open waters. The growing portion of the submerged vegetation is usually at or near the surface, while the deeper portions are older and often moribund. This vegetation thereby enjoys the full benefit of the sunlight, so essential to the growth of chlorophyll-bearing plants, while its occupation of the water —especially at the surface—shuts out the light to a consid- erable degree from the more open deeper waters, and in this way adds another effective barrier to the growth of the phyto- plankton in surrounding w^ater. The dense growths of the CerntophtjlJitm also interfere with the movements of the water, and thus tend to establish and 500 maintain local units of environment within a body of water. Lakes full of vegetation, like Quiver Lake, exhibit greater variations in the local distribution of the plankton than are found in open ones, such as Thompson's Lake. Greater differ- ences in the component organisms also appear. The vegeta- tion thus acts as a barrier, isolating differing assemblages of organisiiis. Thus, in Quiver Lake in one instance local aggre- gations or swarms of Volvox, of Copepoda, of Oscillaria, and of Melosira were detected in the examination of the local dis- tribution of its plankton. To this isolation by the vegetation may also be attributed the considerable irregularity in the sea- sonal fluctuations of the amount of plankton, which is some- what more evident in the planktographs of Flag Lake (PI. XXXIIL and XXXIV.) and Dogfish Lake (PI. XXX.-XXXIL) than in those of other stations. Such fluctuations, for ex- ample, as those in May, 1896, in Flag Lake, when the plankton fell from 203. 5i^ cm.' to 0.72 in 13 days, or the fluctuations in Dogfish Lake in 1895, which do not seem to be correlated with any fluctuating feature of the environment, may be referred in par-t to the isolation resulting from vegetation and the modifi- cations of food supply and reproduction consequent upon it. The maximum-minimum contrast in Flag Lake was due to an excessive local development of Bosmina followed by its sudden disappearance. The cycle of changes in the succession of life are thus accentuated, and run a more rapid course in the midst of vegetation than they do in the larger unit of environment, the open water, where minor differences are quickly merged by the turmoil of current and waves. The plankton catches made in vegetation-rich lakes usually contain a lai'ger proportion of littoral and bottom-loving spe- cies than those from open water. There are the RhizoporJa— often those with the heavier shells—the attached diatoms, cili- ates, and rotifers, together with many bdelloid and Ploiman rotifei's not found in open water, the aquatic insects, both adult and larval, the oligochsetes, the smaller mollusks, Hyalella, and Hydra. They materially increase the volume of the 501 catches recorded in the tables, and show in the plates of the plankton of Quiver, Dogfish, and Flag lakes. The sessile or- ganisms above named, with the liri/ozoa, which often occur on CefdtopJnjIJion, avail themselves of the plankton as food. Hi/dra, especially, increases when the plankton is more abundant. In Quiver Lake on May 8, 1896, Hydnt was taken in plankton at the rate of over five thousand per m.' of v^'ater. These organ- isms which find a substratum and shelter on the aquatic veg- etation must have some important effect on the plankton, and their presence is doubtless one of the minor factors in the suppression of the plankton in lakes rich in submerged vege- tation. The economic aspects of the question of vegetation in bodies of water arise from the relation which it bears to the production of marketable fish. Quiver and Thompson's lakes are both seined by local fishermen, and their relative produc- tivity as fishing grounds may be expressed in the market value of the leaseholds of the fishing privilege. Quiver Lake is so blocked with vegetation that clearing it for seining is at times an expensive operation, and this has a tendency to lower its market value. Thompson's Lake, on the other hand, is less accessible, and some clearing out of the littoral lielt of vege- tation is always necessary before seining, the operating ex- penses being thus somewhat increased. For years the lease- hold of Quiver Lake has been purchased for a merely nominal sum, not exceeding $100, and it has often lacked a purchaser. Thompson's Lake, on the other hand, has been, in recent years at least, an object of increasing value, and brings over ten times this amount for a portion of the lake only. Thompson's Lake has an area of about 1,200 acres, while Quiver has only 2i30. Their market values are thus out of proportion to their re- spective areas. Capt. J. A. Schulte, of Havana, whose knowl- edge of the fishing industry in the Illinois River is extensive and accurate, estimates that in the same area Thompson's Lake will produce five times as much fish as Quiver, and production of fish thus stands in somewhat the same ratio as the average 502 plankton production (1.75 and 8.26 cm.-' per m.'). The produc- tivity of the lake full of submerged vegetation, is, it seems, less than that of one free from it, whether measured in cubic centi- meters of plankton or returns for marketable fish. The data here presented concerning the inimical effect of submerged non-rooted vegetation upon the plankton suggest an interesting subject for field or laboratory experiment. In- deed, experimental proof is desirable for the generalization here advanced. Hovv^ far it will find support in the examina- tion of other localities remains to be seen, for no investigation bearing upon the question seems to have been made elsewhere. It should be noted that it is not maintained that all vegetation is inimical to the development of the plankton, but only such as successfully competes with the phytoplankton for the availa- ble plant food, and thus brings by its decay no additional sources for plant nutrition into the water. These conditions are approximately realized where the submerged non-rooted type of vegetation prevails. Where, however, by reason of the local conditions or the nature of the constituent plants, the aquatic vegetation adds by its decay to the fertility of the water owing to its utilization of sources of food in the soil and the air not available to the phytoplankton, we may expect to find the development of the plankton fostered by such vege- tation. These conditions are realized wherever rooted, and especially emergent, vegetation prevails and contributes by its decay to the enrichment of the water. A belt of littoral vege- tation of this sort may thus be of considerable effect in main- taining the plankton in a body of water. INTERNAL FACTORS AND PLANKTON PRODUCTION. Under this head attention will be called to certain phases of plankton production with which in the present state of our knowledge no environmental factors stand in apparent corre- lation. From this point of view, which lays emphasis upon the reacting organism rather than upon the stimulating environ- ment, most of the relations and adaptations of the plankton to 508 enviroumeutal factors might be treated under this head. But this has not been my method nor is it now my purpose to adopt it. The phenomena of growth and reproduction of the con- stituent organisms of the phmktou, on the other hand, owing to our ignorance of their controlling factors, can at present be treated only under this head. The volumetric data in them- selves contain little evidence bearing directly upon the prob- lem, but in the light of the statistical results the fluctuations in the plankton become dependent upon fluctuations in the rate of growth, and especially in that of the reproduction of its constituent organisms. These fluctuations are often concur- rent, or, at most, shoi'tly consequent, in many species at the same time and in several different localities, and give rise to the coincident volumetric pulses to which attention has so often been called in the preceding pages. Somewhat regular alternations of growth and rest, of fission and spore formation, or of parthenogenesis and sexual reproduction, are funda- mentally the basis of the cyclic movement in production. The amplitudes, and to some extent the location and duration of the pulses, are plainly affected by the various factors of the envi- ronment discussed in preceding pages—l)y light, temperature, vegetation, tril)utary water, various hydrographic factors, and by food supply, and, possibly, also, by chemical conditions not directly concerned in nutrition, l)utthe available data fail com- pletely to afford any satisfactory environmental factor or group of factors which stands in correlation, even remotely obvious, with this cyclic movement in production. I therefore class ih.\s periodic groirfh, these se.nia/ cijrh's which cause volumetric pulses, under the head of internal factors. The element of periodicitij in itself does not seem to be consequent upon any known external factor. NORMAL REGIMEN OF PLANKTON PRODUCTION. The records of plankton production in the Illinois River, its tributaries, and backwaters, contained in this paper raise 504 the question whether there is in this fluviatile environment a normal regimen of production. Is there in the course of pro- duction an orderly sequence, of any sort, of sufficient stability and of sufficient frequency in occurrence in successive years to justifiy its designation as a normal regimen? A cursory inspection of the planktographs in the plates, of the data in the plankton tables, and of the table of monthly means following page 342 reveals at once an apparent state of chaos that accords well with the instability of most of the en- vironmental factors of the plankton, notably the hydrographic. For example, the production in the same month^in different years or in the different localities examined by us is exceeding- ly variable. Taking at random the month of August, we find that the mean production for this month in the years of exam- ination ranges in the channel from .91 to 9.67 ; in Spoon River from .002 to .652 ; in Quiver Lake from .22 to 2.46 ; in Dogfish Lake from l.ll to 3.91; in Flag Lake from .03 to 3.74; in Thompson's Lake from 1.08 to 19.40 ; and in Phelps Lake from 8.80 to 139.85 cm.' per m.'; and, furthermore, that the extreme range in these means—.002 to 139.85— is found coincidentally in the same year, 1898 (see table following p. 342). This does not afford a very satisfactory basis for predicting the probable August production in cubic centimeters of plankton in any of these localities. It is evident that there is little regularity in the actual amplitude ofproduction in a given season and locality in successive years. If the problem be approached from the standpoint of rela- tive production in different localities at the same time, or in the same locality at different times, more semblance of order is traceable, though not equally so in all localities or in all months of the year. The relative rank of each locality in mean monthly production, as seen in the table following page 842, is tabulated below. For example, in the case of the Illinois River in the total of 51 monthly means there were 5, 6, 16, 12, 4, 10, and 1, instances when its production attained first to sev- enth rank respectively among the seven or less localities repre- 505 RANK IN PLANKTON PRODUCTION. Rank 506 in the riv^er than in Phelps Lalie. A further inspection of the table indicates that Phelps and Thompson's lakes and Spoon River exhibit the most stable relations in productive rank. It is in these bodies of vpater that we have found environment- al conditions most uniform. In Flag, Quiver, and Dogfish lakes the divergences in rank are much greater, and it is in these localities that conditions of high and lovp vpater and of vegetation afford sharpest contrasts. In the Illinois River itself, v^rhere hydrographic fluctuations are most immediately effective, we find apparently the widest divergences in produc- tive rank. Even in these instances of greatest divergence the tendency towards a certain rank in production in each locality is sufficiently evident to warrant the statement that in the main the relative rank in production in the several localities examined by us is well established and generally maintained. Still more pertinent to the question of the existence of a normal regimen in production is the question of sequence in the course of production in successive years and in different localities in the same year. Do the planktographs form curves which may be superposed in successive years in the same lo- cality or in different localities in the same year? A single glance at the plates which accompany this paper will suffice to reveal the chaotic complex of lines which such a superposi- tion would produce. There is no such unity or similarity if we base the comparisons on the actual volumeti'ic production. If, on the other hand, we disregard coincidence in amplitude of the curve and consider mainly the direction of the changes, sim- ilarity becomes increasingly apparent. It rarely approaches to the condition of complete parallelism, however, owing to the great variety in the amplitude of production in various years and localities. If the data in the table following page 342 be analyzed with reference simply to the upward or downward movement in mean production from month to month and year to year in the various localities, we find certain tendencies appearing which may afford a basis for predicting the probable course of pro- 507 duction to a slight degree in the localities examined and in the region as a whole. The following table exhibits the instances of upward and downward movement in the monthly means for each locality for each month of the year. DIRECTION IN MOVEMENT OF MEAN MONTHLY PRODUCTION. Station 508 Fig. E.-Planktographs o£ all stations^ means. , plotted from averages of monthly 509 tion again drops to a low level, and this movement appears in 60-70 per cent, of the instances in each month. In December the decline in production is checked, and in a few cases the upward teudeucj' reappears, only to give way again in most cases to the January minimum. There is thus a general seasonal regimen apparent in the totals of all localities, and an approximately parallel seasonal routine for each of the localities, the degree of approximation varying with the locality and the season. But this regimen can be outlined only in the most general terms, and is evei-ywhere subject to divergences that frequently reduce it well nigh to a semblance of chaos—a condition arising from the instability of this aquatic environment. The course of production above outlined considers only the movement of the average monthly production, and consequently deals only with the lai'ger and more general seasonal move- ments. It masks almost completely the minor fluctuations, and especially obliterates all consideration of the phenomenon of pulses discussed in the treatment of the course of produc- tion in the several localities. It is in this matter of pulses, or, in other words, in the cyclic movement in plankton production, that the nearest approach to a normal regimen appears in our volumetric data. I have shown that wherever our collections were made at intervals of a week, or less, this movement is generally distinctly traceable. In the backwaters, where collections were usually less frequent, the cyclic character of planktographs is still apparent, even in fortnightly collections. This phenomenon is therefore.it seems, a constant feature of the movement in plankton production in our waters as it appears in the purely volumetric data. Sug- gestions of the occurrence of a similar movement in the plank- ton of other waters may be found occasionally in the data of other investigators, but nowhere, to my knowledge, is there at present a chronological series of collections of sufficient dura- tion and brief enough interval of collection for comparison with the data presented in this paper. The main basis for the 510 emphasis laid upon the subject m connection with my discus- sion of the volumetric data will be found in the statistical data of Part 11. of this paper. In this connection it may be noted that the statistical data of other investigators lend some sup- port to the probability that this cyclic movement in the plank- ton will be found to occur in other localities as well as in those examined by us. Instances suggesting this occurrence may be found in the statistical data of Amberg ('00), Steuer ('01 ), and the more recent work of Cohn ('03). This cyclic movement consists in the repetition of rise and decline in production—a repetition broken in our records only by the imperfection of our data. Wherever collections are of sufficient frequency it is possible to trace it—sometimes, how- ever, only with the aid of statistical data—continuously through all the vicissitudes of changing seasons, of summer heat and winter's ice, of the vernal rise and autumnal decline in tem- perature, and of high and low water. Moreover, it appears in all of our localities whenever collections have been made with a weekly, or even fortnightly, interval. These pulses vary in duration from 2 to 7 weeks, though the majority occur in limits of 3 to 5. Their amplitudes vary greatly, and are plainly influenced by various environmental factors. Their limits are also frequently modified by these factors, though the evidence is not clear that any of the envi- ronmental factors we have discussed are correlated directly with the cyclic movement itself. In the discussion of production in the backwaters the- uni- versal approximation in time of these pulses in the various localities, or even their precise coincidence in many instances, was recorded. The cause of this tendency toward uniformity in the direction of movement in production in these various localities at the same time is not apparent from the data at hand. Contrilnitory to it are the facts of a plankton largely composed of the same species of planktonts, of a connection and commingling of waters in all of the localities in flood con- ditions, and of the operation of environmental factors com- mon to all of the localities. 511 The normal regimen is accordingly not delineated by a planktograph of marked detiniteness in its course, Ymt is one formed by a sequence of recurring pulses of approximately a month's duration and varying amplitudes, low in winter floods, rising with the temperature to a vernal maximum of consider- able magnitude, often, but not always, declining during the summer mouths, frequently rising again in late summer or au- tumn, and in some localities and years to an extent exceeding that of the vernal season, and falling with autumn tempera- tures but increasing in stable winter conditions after the min- imum winter temperature is reached,—such is the general reg- imea of plankton production in the Illinois River and its back- waters. If this cyclic movement-in production be characteristic of the plankton generally, fresh water and even marine, it must follow that scattered and irregular collections, or those at in- tervals exceeding a week or at most a fortnight, may fail en- tirely to give an adequate representation of the course of pro- duction or relative fertility of a body of water. Chronological series throughout the whole seasonal range of climatic condi- tions and at close intervals—of one week or less—are neces- sary for any accurate delineation of production and fertility of water by the plankton method. SOURCE AND MAINTENANCE OF THE POTAMOPLANKTON. The existence of a very charactei'istic and abundant plank- ton in the Illinois River at once raised the question as to its source and maintenance. We hud at Havana a stream which year in and year out carries by a burden of life, microscopic as to its individuals, to be sure, but in the aggregate a volume of surprising extent. This stream of life exhibits a routine of seasonal changes in constitution and quantity which neither flood, the ice of winter, nor the drouth of summer wholly inter- rupts. It recurs year after year with a regularity which stands in strong contrast with the fluctuations of the environment. Although the water in the stream is subject to continuous re- 512 newal,—the channel from Utica to the mouth discharging in from 5 to 25 days, according to the rate of the current,—the plankton is continuously maintained, and the seasonal routine is run in the face of this continuous renewal of the water. Furthermore, the plankton product of the stream is discharged at the mouth of the river practically in its entirety, for the or- ganisms of the plankton cannot maintain their place in the stream against the current. The only organisms of the pota- moplankton which remain, are those used as food by fishes and other animals which are not carried away by the current, and such as may be lodged—usually in encysted, and thus heavier, condition—along the bottom or banks of the stream. At times of flood the receding waters leave some of the plankton side- tracked in the reservoir backwaters—in the lakes, lagoons, bayous, and marshes of the bottom-lands. As river levels fall it may be slowly drawn oil into the channel of the stream, or cut off from connection with the river. This continual dis- charge of the plankton, never to return, makes the problem of the maintenance of the potamoplankton, quantitatively at least, very different from that of the maintenance of the plank- ton in a lake. Three suggestions arise in explanation of the perennial character of the plankton of the river: (1) The plankton en- ters with the tributary waters, in which case the problem is only removed a step; (2) it is autonomous, developing in the stream while the waters are in transit, in which case the solu- tion lies in the river audits environment; or (3) the two elements, contribution and autonomy, are combined, in which case the share of each will appear on a comparison of the plankton of the river and of its tributary streams and backwaters. The water in the channel of the river comes from three sources ; from springs and seepage along the banks, from the impounded backwaters of the bottom-lands, and from tributary streams. RELATION OF SEEPAGE WATERS. The contribution from springs and from seepage are incon- 513 sidej-able in comparison with that fi-om the other sources. Wherever the river encroaches upon the bluffs, as, for instance, below Havana, seepage areas of some extent and springs of some size are to be found. Such banks, however, in immediate contact with the main stream itself, are of very limited extent. Furthermore, their contributions to the plankton are relatively still smaller. These springy banks abound in life—planarians, amphipods, isopods, oligochagtes, and rhizopods, mainly limic- olous species, which rarely leave their habitat to enter the river with the spring water. Such springy banks, exposed to midsummer's heat during low water, do at times teem with species common in the plankton. For example, a bank of this sort on the levee at Havana was covered with a brownish scum composed principally of Si/iiedni aci/s, a diatom abundant in the plankton at cooler seasons of the year. The temperature of spring waters along the bluffs in midsummer is about 60°. Elsewhere in the warmer waters of this springy shore are to be found patches of green and i-ed scum, where Enyk'na ciridis and E. sangtiinea were abundant, both species being common in the plankton at that time. Tiny rills of cool water traverse the oozy bank and carry stray individuals of these various species into the river, but their clear waters are poor in com- parison with the brown water of the stream which they join, turbid with plankton. Their contributions are thus insignifi- cant in amount and, while adding a trifle to its diversity, their main action is that of diluents of the potamoplankton. RELATION OF TRIBUTARY STREAMS TO CHANNEL PLANKTON. The relation of tributary streams to the potamoplankton in the channel of the Illinois is a much simpler problem than that presented by tlie backwaters. Their contributions enter the river in well-dehned channels, and the areas of their respective basins are an index to the quantity of water they bring to the river. Their share in the formation of the potamoplankton can thus be more readily tested and estimated. Under conditions prior to the opening of the Chicago drainage canal the river 514 received at Utica, the upper end of the basin under considera- tion in this paper, the drainage of 10,365 square miles of catch- ment-basin, over one third of the total basin of the stream, to which was added at La Salle, a few miles below, the water from the Illinois and Michigan Canal containing the sewage from the Chicago River. A large volume of water, richly fertilized, is thus provided for the reception of tributary streams, no one of which, with the exception of the Sangamon, has more than one ninth of the drainage basin above the point of its union with the river. The Sangamon has at its mouth a basin one fourth as large as that of the river above. This presence of a considerable initial volume and the distribution of tribu- tary waters in relatively small streams at intervals along the course are conditions which favor the mingling of the constit- uent waters of the Stream and tend to maintain the uniformity of the plankton in the main channel. The tributary waters, with a few minor exceptions, such as Quiver Creek, enter the main channel directly. The elevated deposits built up across the bottoms by their agency confine their floods within their banks except during the maximum stages of overflow, when they contribute directly to the back- waters, but even under these conditions a strong current is still maintained along their customary channels directly to the main channel of the river. Their contributions are thus, as a rule, caiTied directly to the river and mingle with it without any period of impounding, and their effect upon the plankton is direct and immediate. Two tributaries. Spoon River and Quiver Creek, were availa- ble for examination at Havana, in both of which collections were made by us which throw consideral^le light upon the character and quantity of their plankton contributions. QUIVER CREEK. In Quiver Creek we have a small tributary with a basin of only 220 square miles, largely of alluvial second bottom with more sand and less heavy , loam and clay than the adjacent 515 prairies. Its waters are discharged into the upper end of the eastern arm of Quiver Lake, and are impounded for a varying length of time before reaching the river. No channel defined by the configuration of the bottom traverses the lake, and since its area is relatively large in comparison Avith the discharge of Quiver Creek, the tributary waters are subject to considerable impounding when the lake is free from vegetation. When, however, vegetation is abundant, a fairly well-defined channel, through which the discharged waters make their way with per- ceptible movement, is kept open through the matted growths. The impounding period is thus reduced for the channel water under such conditions. Collections were made in Quiver Creek, near Topeka. 111., above McHarry's mill-pond, from September 1. 1S96. to April 20, 1897, at intervals of ten days, by Mr. W. E. Deverman. who kindly volunteered this service for the Station. A tow-net of No. 20 silk was used, but no exact quantitative method was adopted, so that these collections are available only for quali- tative comparisons. The catch consisted largelj^ of silt in the form of quartz grains and coarsely comminuted vegetable debris, with rela- tively few plankton organisms. The several catches were uniformly diluted, and the plank- ton organisms counted in a uniform fraction of a cubic centi- meter of the dilution from each catch. The various species detected and their monthly averages in numbers are given in the appended table. The figures have but slight quantitative value, though they will serve to illu.strate in a general way the composition of the plankton and its seasonal changes, and will also afford asuflicient basis for a comparison of the constituent organisms of the plankton of Quiver Creek, Quiver Lake, and the I'iver, though not for a comparison of their relative num- bers per cubic meter in each of the three situations. The plankton of Quiver Creek, as shown in the table, may be characterized as largely tycholimnetic, that is, composed of littoral species, shore-loving and liottom forms. This is seen 516 PLANKTON OF QUIVER CREEK, SEPTEMBER, 517 in the predominance of the diatoms, the Bliizopoda, and Chiron- onius larvae, and in the character of the few rotifers present. It is also evidenced by the erratic distribution and often small numbers of manj' of the species occurring in the creek waters. There is little evidence in the data of marked seasonal changes. The diminished number, both of species and individ- uals, in the winter and spring months is due in part to the flood conditions prevailing at that time. The disappearance of the desmids in the colder months is apparent, as is also the decline in the diatoms, the effect of which is heightened liy the greater proportion of dead and moribund individuals in the winter months. There is also some slight evidence in the ta- ble of a spring increase in March and April and of a late au- tumn maximum in November. From conditions observed in Spoon River and Quiver Lake during the summer months not included in the period of our Quiver Creek collections it seems probable that the plankton of this stream at that season of the year is more diversified by the addition of flagellates and rotifers. The flow of the stream is, however, much reduced at this sea- son, and its contributions correspondingly small in quantity. The inter-relations of this creek plankton are very patent. The diatoms are the predominant members, having one third of the species and nine tenths of the individuals. The rhizo- pods include one fifth of all the species and almost two thirds of the animal individuals. We have here a rich diatom flora supporting a rhizopod fauna. The remainder of the species, about one half, belongs to diverse groups and is numerically, at least, an insignificant element in the plankton assemblage, constituting but 4 per cent, of the total population. A comparison of the plankton here delineated with that of the river at the same seasons of the year leads to the follow- ing conclusions : 1. The creek organisms all occur in the river plankton. 2. The facies of the river plankton is quite different from that of the creek in that it has a greater abundance and variety of organisms, a greater proportion of limnetic species, and a 518 much greater proportion of limnetic individuals, with a cor- responding decrease in the relative numbers of the littoral in- dividuals. This is very apparent throughout the whole season covered by the collections. Thus in autumn months, when Meloslra, Si/niira, the ciliates, Si/)ieli(i4a, and various Brarhi- onidce characterize the potamoplankton, we find them but sparingly represented, or, as in the case of SynvhMa and the ciliates, wholly absent from the creek plankton. So, also, in the winter months Brachionus dorca^ and C'l/clops hicuspkkdus, so characteristic of the river plankton, form no part of the popula- tion of the creek. The perennial and abundant Polyaiiha playfijpteya was at no time found in the creek. That limnetic species are not wholly absent from the creek is shown by the presence of Melosira, Synura, Notholca jugosa, Eudorina, and TrachelowoHus, but their occurrences are isolated and their numbers few. The creek is thus not a center of distribution for such planktonts as these, their presence and numbers in the river being practically independent of their appearance in the tributary. The abundance of diatoms in the creek waters suggests that these may find centers of distribution here. Most of the species, however, are quite as abundant in the river, with the exception of the SurireJlas. These are present in small num- bers in the potamoplankton, and are often moribund. 3. The creek waters act as diluents of the potamoplank- ton. The character of the Quiver Creek plankton and the quantitative studies on Quiver Lake support this view. The contribution of Quiver Creek to the potamoplankton is thus largely of littoral species and of small quantity, and its effect is that of a diluent of the potamoplankton. The following table, which gives the relative number of species in the plankton of Illinois and Spoon rivers and Quiver Creek, demonstrates the small number of species found in the creek, and the monotony of its composition as shown in the pre- dominance of the Rkizopoda and the diatoms. The relative pau- city of the Mastigophora in its fauna may be due in part to the 519 RELATIVE NUMBER OF SPECIES IN THE PLANKTON OF QUIVER CREEK AND SPOON AND ILLINOIS RIVERS. Group 520 . pear in large numbers in the creek waters after they mingle with those of the impounding bottom-lands or the slowly mov- ing current of the main stream. No chemical analyses have been made of the water in Quiver Creek, but those made of samples taken at our plankton station in Quiver Lake are in a measure applicable to the creek itself, especially at low-water stages. In Tal)le XIII. are data derived from weekly analyses from September 24 to December 3, 1895, and fortnightly analyses from October 19, 1897, to March 28, 1899, made by the Chemical Survey. A summary of the averages will be found in the table on page 521. These show a smaller amount of residue on evaporation in the lake (268.9) than in the Illinois River (367.5), which may be due in part to the deposition of suspended silt owing to the impound- ing action of the lake and to the sandy nature of the catch- ment-basin. The loss on ignition is somewhat less, and the oxygen consumed very much less (5.9 to 10.4), in the lake than in the river, indicating a smaller amount of organic matter in the former. The small amount of chlorine (4.8 to 21.6) ex- hibits the freedom of the lake from sewage contamination. The albuminoid ammonia, representing undecomposed organic matter, is also present in small quantity in the lake as com- pared with the river (.25 to .48), while the free ammonia, indic- ative of the first stages of decomposition, is still less (.165 to .860). The nitrates, the final products of decomposition, are not at all abundant in the lake (.66 to 1.58). The lake waters, and by inference the tributary creek waters also, are thus deficient, as compared with the river, in organic matters and the products of their decay. These prod- ucts are fundamental constituents in the nutrition of the phy- toplankton, which in turn supports that part of the zooplank- ton which does not depend upon the organic detritus in sus- pension for food. The chemical condition of the water of Quiver Creek is thus unfavorable to the development of a plankton quantitatively as great as that of the river. Further- more, its waters, poorer in the plankton itself, not only dilute 521 the plankton of the river, but even diminish the productivity of the river water by lowering the relative amount of its nutri- ent constituents. SPOON RIVER. In Spoon River we have a typical tributary of the larger type, from prairie country, with no unusual contamination by sewage or industrial wastes, draining 1,870 square miles, a little more than one tenth of the basin above its mouth, and dicharg- ing directly into the main channel. A detailed discussion of the environmental conditions in this stream and of its plankton production will be found on pages 340-350. It will suffice in this connection to recall the facts that the recent origin of the tributary waters, its greater turbidity, and bux-deu of silt, all militate against plankton pro- duction in this tributary. A consideration of the chemical conditions in Spoon River and in the Illinois throws much light on the nutrition availa- ble for the support of the plankton in tributary and channel waters, a very important factor in the matter of plankton pro- duction. In the following table the averages of all analyses in these two streams and in Quiver Lake are given. CHEMICAL EXAMINATION OF WATER FROM THE ILLINOIS RIVER AT HAVANA, FROM SPOON RIVER NEAR ITS MOUTH, AND FROM QUIVER LAKE—PARTS PER MILLION.* 522 It is noticeable that Spoon River carries the lai'gest amount of matter in suspension, both absolutely and relatively, as shown by the high total residue on evaporation (522.3 to 367.5 and 268.9) and by the smallest amount of the residue in solu- tion (167.1 to 304.1 and 248.2). This is further shown by the fact that the solids removed by the army filter in 8poon River average 1,755 cubic centimeters per cubic meter to only 592 in the same year in the Illinois. The large amount of organic matter undecayed and undissolved, and therefore not available for the support of the plankton, is partially indicated by the high oxygen consumed (14.1 to 10.4 and 5.9), the high albumi- noid ammonia (.60 to .48 and .25), and the high total organic nitrogen (1.29 to 1.03 and .61), when considered in conjunction with the small amount of residue in solution. On the other hand, the waters of Spoon River are quite deficient in forms of nitrogen more available for the phytoplankton, the free am- monia (.24), nitrites (.039), and nitrates (1.01) being in each case less than in the Illinois (.86, .147, and 1.58), while the chlorine (3.8), an index of sewage contamination, is less than a fifth of that in the channel (21.6). Spoon River has therefore great resources, in so far as or- ganic matters and the products of their decay are concerned, for the support of the plankton. Not all of the matter is in solution for immediate utilization, but there is still sufficient for a large plankton development, time for which has not been allowed in the tributary stream. The immediate effect of the access of the contributions of Spoon River to the channel is, in the average, a dilution of its inorganic nitrogen per m.' of water, which is in some unknown measure made good by the contributions of silt, in part of organic origin. The net result is, of course, a large addition to the iotal resources of the chan- nel waters available for the present and future development of the plankton. The amount of nitrogen in its several forms which Spoon River carries is not small as rivers go, for this stream drains a plain unsurpassed in fertility by any other part of the catch- 523 meut-basiu of the Illinois. It also receives a moderate amount of sewage from the cities of Lewistown and Canton, and the drainage from a consideral)le number of towns. Its diluent effect upon the plankton of the Illinois is thus not due to the poverty of its own waters but rather to the excessive fertility of the main stream, a fertility resulting from the sewage and industrial wastes received by that stream from the cities of Chicago and Peoria. The contrast in fertility as indicated liy tlie analyses tabu- lated above is not as great as the differences in the plankton production of the two streams. The ratio of nitrates which perhaps most fully expre.sses their relative fertility is 1.01 to 1.58, while the ratio of the plankton production as expressed in the average of the monthly means of the catches by the silk-net method is 0.256 to 2.71. The failure of Spoon River to develop a more abundant plankton is thus apparently due to some other cause than the lack of nutritive elements in the water for the support of the plankton. The development of a considerable volume of plankton in it at times of low water and slack current makes patent the probability that the lack of time for breeding is at least one of the important factors in the relative paucity of the plankton of this tributary stream. QUANTITATIVE COMPARISON. A comparison of the quantities of plankton taken by means of the silk net in the two streams affords a fair contrast of their relative productivity. Certain sources of error are. how- ever, present in the data of comparison, and as they are not equally distributed in the ca.se of both streams they invalidate to some undetermined extent precise comparisons. These sources are the leakage of the plankton through the silk and the presence of silt. The plankton escaping through the silk is largely made up of the M(it when penned up for the market, upon "moss" or CcratoplnjIJiim, though it may be that the insect larvie and mollusks found in the vegetation constitute the more important elements of the food. It is only when this growth of vegetation decays that it releases into the water the elements which conduce to its fertility. The phytoplanktou, on the other hand, multiplies very rap- idly and is immediately available for the support of the micro- scopic animals of the zooplankton, and this, and to some ex- tent also the phytoplanktou itself, is the immediate food of most young fish upon hatching and the customary food of some adult fishes,—such as Foli/odo/i (Forbes, '88, and Kofoid. '99 ) and many minnows,—of the bivalve mollusks, and of many other or- ganisms of sessile habit. The plankton is thus the prime source of food of fishes and of many other organisms utilized by fish as food. The chain of food relations, for example, between the food elements of the water and the black bass is in the main a short one, with the plankton as the principal link. Pro- fessor Forbes ('SO) has showu that 86 per cent, of the food of the game fish consists of other fishes, principally Dorosoina with 564 perch and minnows. In its youngest stages the black bass was found to be a plankton [Entomostraca) feeder, and later chang- ing to a fish diet. It is interesting to note that its principal fish food, young Dorosoma, is itself in its younger stages a plankton feeder, secondarily adopting a limophagous habit. The bottom slime thus eaten by the growing and adult Doro- soma is a food largely because it contains so many organisms normal or adventitious to the plankton. Thus at all seasons the plankton forms an important link in the chain of food relations leading to the black bass. The buffalo-fishes and the German carp are likewise to a large degree dependent for food upon the plankton in early stages of growth, and, like Dorosoma, subsequently adopt the limophagous feeding habit. The organisms of the plankton thus at all times enter largely into the sources of their food supply. The contents of the digestive tracts of these impor- tant food fishes examined by me at Meredosia during the spring months of 1899 were found uniformly to contain com- minuted vegetable debris, which constitutes the greater part of the unstable ooze or slime which abounds in the backwaters of the river, and, associated with this, many Entomostraca, rotifers, rhizopods, and unicellular or colonial algae, belonging to species common in the plankton at that season of the year. The mori- bund, the spore-forming, the egg-laden organisms of the plank- ton sink to the deeper strata, and together with the normal denizens of the bottom slime, which are everywhere adventi- tious in our plankton, form the food of these fishes of greatest commercial importance. A very striking instance of the adaptation of the breeding seasons of fishes to food conditions is found in their nice ad- justment, for the most part, to the seasonal course of plankton production in our waters. Most of our minnows, the dogfish and gar, the Catostomidcp, the carp, Dorosoma, and the Etheosto- miche spawn in central Illinois during April and the first of May, while the bass, the sunfishes, and many of the Siluridoi follow in May. This brings the maximum number of young fish, re- 565 cently hatched and generally plankton feeders, just at the sea- son of the vernal plankton pulse, which is often the maximum production of the year. It may well be that the abrupt dimi- nution of plankton following the May and June pulses is accel- erated to a large degree by the plankton-feeding habits of the fry of these fishes. The plankton thus enters directly into the food of most young fishes and of some important adult fishes, and indirectly it is the primal source of food of most fishes. A knowledge of its local and seasonal distribution and of the environmental conditions which favor or impede its development is fundamen- tal to any scientific utilization of the present resources of this stream or any future development of resources now unproductive. The data at hand afford an opportunity of comparing the plankton production and the annual output of marketed fish. The reports of the Illinois River Fisherman's Association for 1894-1898 give the following statistics based upon estimates and partial I'ecords of leading men engaged in the fisheries. They are not exact records, but the error involved is no greater than that in the plankton data. PLANKTON AND FISHERIES. Year 566 therefore the weight, of the marketed catch. Some correspond- ence in plankton and the products of the fisheries might there- fore be expected, though on account of the complexity of the problem and the limitations of our data it is difficult to demon- strate it in every case. In 1894, when our data indicate a plankton production be- low the average in channel, in backwaters, in their sum, and in the estimated discharge, we find the total production of mar- ketable fish also below the average (for 1894-1898). In 1895 our collections indicate an increase, approximating 50 per cent, in plankton production as exhibited by each of the methods tabulated, and there is also an increase in the product of the fisheries, though it amounts to only about 4 per cent. The direction of the change in production is the same in all cases. In 1896 plankton production falls in channel, in backwaters, in their sum, and in total discharge, the decline in all but the back- waters being greater than the increase from 1894 to 1895, and we find, accordingly, that the decline in the product of the fish- eries is also greater than its antecedent rise to the level attain- ed in 1895. In 1897 plankton production again rises in the channel, in the sum of channel and backwaters, and in total discharge, but not in the backwaters. This apparent decline in backwater production may be due to the elimination from our data, for a part of the year, of Phelps, Flag, and Dogfish lakes, thus giving undue weight to the depressing effect of the Quiver Lake data. The fact that in Thompson's Lake plankton pro- duction rises from 6.67 in 1896 to 10.41 in 1897 is an indication that plankton production in the open backwaters in 1897 rose above the level of that of 1896. In correspondence with the increased plankton production in channel, and total discharge, we find the product of the fisheries rising from 7,252,811 pounds in 1896 to 9,703,298 in 1897—a change not exceeded in any other year of the records. In 1898 the product of the fisheries continues to increase, reaching 10,647,466 pounds, but plankton production rises only in the backwaters and in the sum of chan- nel and backwaters, falling in channel and total discharge. 567 There is thus, in general, a correspondence between plank- ton production and the product of the fisheries in that the di- rection of moveinenf in both is usually the same. The_y rise or fall together. If we compare the changes of the product of the fisheries with those of the sum of plankton production in chan- nel and backwaters, as given in the table on page 565, we see that the direction of the rliaiKjc is the same in both from year to year in every instance in 1S94-1898. If similar comparisons are made of the product of the fisheries and plankton produc- tion in channel or backwaters alone, or in total discharge, we find that in three cases out of four the dii-ection of the change is the same in both. It is also generally, true that years in which plankton production is below the average are also ones in which the product of the fisheries falls below the mean. Plankton production at Havana, provided a similar plank- ton content is maintained until the run-off reaches the mouth of the river, would result in an average discharge of 67,750 cubic meters of plankton, equivalent in weight to somewhat more than 149,050,000 pounds, or 15 times the annual produc- tion of fish. To this wastage of organic matter, which in great part is permanently lost to the drainage ba.sin of the Illinois, should be added the unutilized nitrogen and other food elements in suspension and solution which e.scape with the run-off, es- pecially of flood waters (see Table X. ). How shall this waste be prevented and the plankton be turned into marketable fish ? The problem is a complex one, but the results of this investigation should contribute towards its solution. The first step will be to impound the richly fer- tilized flood waters and thus to afford time for the utilization of their food elements by the developing, plankton, which by various chains of food relations is joined to marketable fish. An illustration of the productive possibilities of impounded Spoon River floods is seen in Phelps Lake, our richest plankton station, and also the home of great numbers of young nsh. Thompson's Lake, another impounding backwater, not only breeds an abundant plankton but contributes no insignificant 568 portion of the 1,000,000 to 1,500,000 pounds of fish marketed annually at Havana. The development in recent years of extensive systems of levees in the bottoms of the Illinois River for the purpose of protecting farm lands from untimely floods increases the impor- tance of, and necessity for, the reservoir backwaters. In con- nection with these systems it might be feasible from an engi- neering point of view, and perhaps even profitable from the commercial standpoint, to convert some of the adjacent low- lying marshes, swamps, bayous, and lakes into reservoirs in which invading and richly fertilized storm waters might be impounded and retained as river levels fall. The increased vol- ume of water thus provided should— in the light of our results —yield an abundant plankton, and support a large fish popula- tion. Under present conditions of abundance of most of our valuable food fishes in the Illinois, stocking such reservoirs is relatively a simple matter. If propeVly protected from es- cape at high water, such an area once stocked with the now rapidly disappearing roli/odon, whose roe is much sought for the manufacture of caviar, might become a very profitable in- vestment. As a basis for further development of the fishing industry it seems desirable that public and private waters should be more accurately defined, a,nd that fishing privileges for market purposes in the former should be matters of license or franchise to responsible parties, so that legislation concerning methods and seasons of fishing could be more easily controlled. With the ever increasing industrial development in the drainage ba- sin of the Illinois River, especially in Chicago and the minor cities along its banks, there is great danger that industrial wastes will so accumulate in the river waters that not only the plankton but also the fish and other animal inhabitants will be driven out or exterminated. Legal supervision over the dis- charge of such industrial wastes may soon become imperative for the Illinois River as it has for some European streams. With the legal status thus clearly defined, and with wise legis- 569 lation which should eflSciently prevent pollution of the stream by deleterious industrial wastes, protect the most desirable food and game fishes fi'om depletion, and, at the same time, per- mit the full utilization of the annual crop of matured and mar- ketable hsh, there is no apparent reason why the Illinois River and its backwaters should not become an increasing source of wealth to the state, and the great waste which now occurs be utilized to a considerable extent in future development. Conclusions. The following are the more important conclusions arrived at from this examination of the plankton and its environment in the Illinois River and its backwaters, based upon the study of 645 collections made in 7 localities in 1894-1899. 1. There is little correlation between the seasonal flux in chemical conditions (as shown in data of sanitary analyses) and the seasonal course of plankton production (as shown in the catches Of the silk net). The nitrogenous matters are in- fluenced by the plankton pulses, especially when diatoms are multiplying rapidly, but the changes are not uniform or pro- portional. 2. The plankton in the Illinois River is distributed with a uniformity approximately equal to that found in German lakes and in Lake St. Clair. The average departure from the mean in short distances (3 miles) probably falls within ±: 10 percent. Chronological catches in periods of 2 to 15 days in 14 series yield an average departure of ± 14.1 per cent. In the river, in 205 miles of the course the average departure in flood conditions was ± 51 per cent., or ± 43 per cent, if the river is divided into four sections, or ± 29.7 per cent, if computations are based on total catch of the net. 3. The average departure from the mean plankton con- tent in two tests in a cross-section of the river is ± 27.2 or ± 23.2, or, omitting marginal collections, ± 21.9 or ± 12.1 per cent. 4. The plankton method can be applied to a stream as legitimately as to a lake. 570 5. The mean of the monthly averages of 235 collections in the Illinois River is 2.71 cm.'' of plankton per m.^ of water. 6. The plankton of the river channel is subject to great seasonal and annual variations. The monthly averages of all collections indicate a period of minimum production of plank- ton in January-February, of rising production in March, of maximum production for the year in April-June, usually cul- minating in a vernal maximum about the end of April and often declining rapidly to a low level in June. The average monthly production declines gradually during the remainder of the year to the winter minimum in December. 7. Individual years vary greatly from these averages as a result of hydrographic, climatic, and other environing con- ditions in varying combinations. 8. The waters of Spoon River contain but a very small amount of plankton (.4(55) except at very lowest stages, when the flow is at a minimum. Its production at other times (.044) is less than one fiftieth of that in channel waters which it joins. Chemical conditions in this tributary are apparently such as to support a large plankton. The recent origin of the water is the cause of the low production. Its diluent effect on the plank- ton content of the channel is about 10 per cent. 9. Quiver Lake produces less (1.75) than channel waters. At high stages its production is relatively larger, and most re- sembles that in channel waters, while at low levels, when sub- merged vegetation is dominant and access of tributary waters of recent origin relatively great, its production is both rela- tively and actually low. It is predominantly a diluent of channel plankton. 10. Dogfish Lake produces (3.16) more than channel waters, freedom from access of tributary waters permitting a higher production than in the contiguous waters of Quiver Lake. 11. Flag Lake also produces more (9.23) than channel waters, the freedom from access of tributary waters, the im- pounding function, and decaying vegetation favoring high pro- 571 duction. The plautographs of this lake are marked by extreme changes in bi'ief time, and by depression in production with the emergence of vegetation. 12. Thompson's Lake also produces more (8.26) than channel waters, and maintains its higher level of production more generally. Its impounding function and the freedom from access of tributary waters contribute to this result. 13. Phelps Lake produces the most abundant plankton (22.55) of all our localities, freedom from vegetation and from access of tributary waters, and the highly developed impound- ing function contributing to this result. The maximum pro- duction in our records, 224.5 cm." per m.', was found in this lake on Aug. 28, 1S9S. This lake is marked by relatively and abso- lutely high production in summer and autumn. 14. The course of plankton production in channel and backwaters throughout the year exhibits a series of recurrent pulses, culminating in maxima and separated by minima, which give the planktograph the appearance of a series of "frequency of error" curves of varying amplitudes. These pulses generally have a duration of 3 to 5 weeks, and tend to coincide in their location in all localities coincidently exam- ined by us. This similarity in the direction of movement ia production amounts quantitatively to 65 per cent, of the possi- ble comparisons in our records. This cyclic movement in pro- duction is plainly influenced, accelerated or retarded, or its amplitude extended or depressed, by environmental factors, but is not itself traceable to any one or any combination of them. A brief interval of examination—not more than one week—is essential to a demonstration of the existence of these pulses. 15. Area and depth, witlim limits of our environment, show little relation to plankton production. 16. Age of the water is an important factor in determining production in streams. Young waters from springs and creeks have but little plankton, and even such tributaries as Spoon Eiver (drainage basin 1.S70 square miles) contain but little plankton, principally of more rapidly developing organisms. 572 This barren water, impounded for 10-30 days in backwater res- ervoirs such as Phelps Lake, develops an abundant plankton. The rate of run-off and replacement of impounded waters de- termines to some extent the amplitude of production. This is greatest where run-off is least and rate of renewal slowest. 17. Fluctuations in hydrographic conditions constitute the most immediately effective factor in the environment of the potamoplankton. Rising levels usually witness a sharp decline in plankton content (per m.') as barren storm watei's mingle with or replace plankton-rich waters of channel and reservoir backwaters. Falling levels are periods of recovery and increase in plankton. Stability in hydrographic conditions conduces to rise in production at all seasons of the year, and instability is always destructive. Winter floods tend to lower plankton production; spring floods increase it. 18. Temperature affects production profoundly. Below 45° the plankton content in the river is only about 9 per cent, of that present above this temperature, and in backwaters but 29 to 40 per cent. Minimum production is at times of mini- mum temperature. The vernal pulse in production attends the vernal rise in temperature and culminates at about 60°-70°. With the establishment of the midsummer temperatures (about 80°) production falls from 44 to 87 per cent, in channel and backwaters. It rises, however, 68 per cent, in Phelps Lake, so that other causes than temperature may be operative in pro- ducing the midsummer decline. The autumnal decline in tem- peratures is accompanied by decline in production in the channel and in Quiver Lake, but by an increase in other backwaters, which exhibit a tendency toward an autumnal pulse. The de- cline to winter minimum occurs in December. An early spring accelerates, and a late spring retards, the vernal pulse, and a late autumn prolongs the autumnal pro- duction. Summer heat pulses often attend plankton increases. Minimum temperatures are not prohibitive of large plankton production. The December production in Phelps Lake in 1898 (43.14 cm.^) exceeds the vernal maximum elsewhere in all local- 573 ities but one, but falls much below the suniinei- production in Phelps Lake. The ice-sheet is not inimical to a considerable plankton production unless stagnation conditions occur. 19. Light affects plankton production. The half year with more illumination and fewer cloud}- days produces from 1.6 to 7 times as much plankton as that with less illumination and more cloudy days. Seasons of unusual cloudiness are accom- panied by depression in production. 20. Lakes rich in submerged vegetation produce less plank- ton than those relatively free from it, in an annual ratio of 1 to 6 and a monthly ratio varying from 1.5 to 20 to that of 1 to 20. The higher ratios generally prevail in periods of dominance of vegetation. Quiver Lake produces more plankton when free from vegetation than when it abounds in it. The emergent and rooted vegetation of Flag Lake conduces by its autumnal and vernal decay to large plankton production, but tends to de- press production in summer.. 21. The normal regimen of the course of plankton pro- duction in the Illinois River and its Itackwaters does not form a definite seasonal planktograph. but consists rather of a series of recurrent plankton pulses, whose varying amplitudes are largely determined by the fluctuating environmental factors of the unstable fluviatile environment. Hence planktographs of the same locality in different years and of the different sta- tions in the same year show resemblances only in such funda- mental features as the winter minimum and the vernal pulse. The relative productive rank of the several localities is gener- ally maintained, and more completely in the more stable en- vironments. 22. The plankton of the Illinois Kiver is largely autono- mous. Seepage and creek waters are diluents of its plankton and add little to its diversity. Even Spoon River is generally a diluent, reducing the plankton content 10 per cent, and add- ing but few diversifying species to its population. The reser- voir backwaters, on the other hand, generally contain a more abundant plankton than the channel, the amount, on the bases 574 of monthly averages, being from 1.3 to 17 times as great. At all levels, waters from impounding areas in the bottom-lands are drawn into the channel, mingled with the plankton-poor contri- butions of tributaries, and further enriched by the growth of in- digenous channel plankton. The reservoir backwaters are thus of great importance both as a source of the channel plankton, of the Illinois River and in its maintenance. 24. The total annual production of plankton in the Illi- nois River, on the basis of normal discharge and a plankton content at the mouth of the river equal to that of our average record at Havana, is 67,750 cubic meters. 25. Filter-paper catches indicate the presence, on an av- erage, of a plankton 3.3 times the volume of that taken by the silk net. Leakage thi-ough the silk is therefore a matter of some volumetric importance. 26. The annual production of plankton and of the fisher- ies of the Illinois River show some correlation in their changes from year to year. University of California, August 23, 1903. 575 a z 5 c < -j--2fe2t3g 1 5;i 9S3A101 3S3I(SlIj KSJS£.??,':?,t:,5 5>- =5=tS = 5SSt5- luaniOAOn :: C; y ~- » a H 5 5 c ^' 5! %i 55 aSaitjj ;5i^~.':'-':^^.'^=:^.*^"»'-5^S35? = ;.; ) S isaAvo'j r^ ~ 'i 5 '^^ ^ '-': "''^ -; '" '": ^ - * "5 - S § S s5 I S I jsaq. isaivoq 4saqiiifj i i^ X X X X 3t X -A X >: X x 3r X ^ x X X xX X X X X X 576 o z S < I W O « S 00 o 579 580 I I M £ 581 I 582 583 I S 584 585 c3 < 586 587 > 588 I > < 589 6 h3 590 m ED 591 592 is" = '"e iS-S S H 3 S ' 5 s ^3, . ", a bD M eO O •Oa-C S — 'S =" : SS QQ a- OSxC tie's *i - =1 0£i£ 2 li qo^'B?) iB^ox ocoxt^^^2:i--i-ii-iccw:»i f O C-l -^JH Cl "* tt< 5! M in -* * -* lO ©OOOi J9d ins'TOTOA Oi-iO-i-SCi-iinoi^io * 510 0t-00-3 lUD 111—JO i ao3>inT;[d JOA 00 00 00 _i, ^.-~.—,—,.-, ^.„,. ^ i^iHC. COCOOX>Cii->C\I'*COOtDCO-*:'3«!f cioooxi'Tiret-t-^ CO O O O 00 9gTixaaoj9tT '^ m ifj ifi t- ( - ooino E ;?> Emofiomr Ci O O r -iOO'-'-H.-.-t t^ 'lift inmm C O O O! OO C Csl ca (M (M w e oooo inininm C.I w ^i ot laaiogjaoo oiMj^cOtnroM t-H CO m H M «1 1-1 -^ Q o < Sag: a 8: —s.inoq f^. o" a^. (MC-[r-nHrtOO<7tC>!'MCtOT-«p--lO:Oi.-n-Hi-(OOOr-(r- {O 'jagj ui —.I3Ali oocao + ++ ooooooooooooooooooooooo + 1 + 1 I I+-I- I M++I I I I I ooo + 1 3000 I++T - -js lO I- 1- oci- 1—p -*- 1- X' t- '-D ^* 'X) ift -XI c; T*< ootno oi ci -^ <* - = OO-^ifSC C'tOIrH.H'Ht-li-H-H-l.-H-tC'I—n— OOt- - .-I OO 1-n-n-l T-I'X) CM C'l -^ lyi C ^ fl & c & ';.^ o k3h " o " 593 I > 594 595 596 597 OS 598 « bo .-5 c— rf ,r=* E^us S Ct •a a 2 s as-0-- 5- s.° •mo ut—aa^ -T?M fi'ia J9(l O Or-lCJOr jad jiis 'lOA -.ooooo^oo'jin^oi'-Hira^T-iooo t- oooo oo -VIA E'lti .laci OXOlO-Hr-l *ns JO aSB-juaoaaj t- ojt-os 00 c^ii-H er3i~.— 1 w CO oc -^ ^ O^MCOt« OOOOr^Cl O'-0T-lrHOt^QCin^C0'-C0ve0Olf5O>-. O- sja-jix UI —pauii^jis J91BM 'lOA p.- O' P-- o eiH o '493.+ Ul —sanoi( ej Ul aaiiBHf) oooo + 1 ooooooooooooooooooooc I ++ I I + I I I I + I + M +++ ooo + r+ 5, St) J. Ul —J3Ali t- lO -M '-( H ?! 'X>* 0» O .-O - -(M:^-*-*-^-*i u .J u -^ niog^oa QOT^jans MY l>>^ " fO in ci 01 "^ i- wrj< cc 7} -^ Gi Ti^ «D o -+ 00 T"! in oi cr: t^ O -t t- .H 1-- - jaqoiniij suoissaoov ?( o *r; oOt-OjO in lA in :o (M !MW (M ooiO'-i'M -Hot-oi r;:o(Mi— oio: -* o'-Oi-t'-octx- 'r^ t- 1- 00 O". o 1-1 r-t w 51 cc *J3 "S i-^ t- 'X. oi a; oo i-i -^ CD ^ X> :3 :o t— b- 1- t~ t- (- 1- I- 1- 1- 1- 1— l^ X CO 'x CO (M O! -M (?l O ( 1 1 Of O! (M 1! Qi (71 71 Ol (M Ol Ol CJ CI O! O^ (M in o I— -t o; m 1 .13 CO * ** in QOx CO 00 Zoo 01 IM O! (M Ot C-[ CM !M (M »!W (M 599 H H p^; o 2 !3 J < —J9l'BA\, (-"ni J9Cl -JtlS -[OA g'rao ai—J91 -BAV e'lii aad aoi>iuB[d '[O V sa^-BJiiN sa^iJlIIQ 'Biaoni -uioanqiv "Btuora painnsuoD U35jCxo aaijojqo rt K » rt ®-^ c - S»!Sffi2E=rr -H Oi * in C0 7CC1'^ = t- M CC t- •?( s> " -J •d tc *#* » ******** "* -^ X X re — i-1 in '.o re '^( 'J c<; ift t-^r-iooo — i?»-*-iTldM3: x^- # # * # * "*M ro 0-. -* 3: inX •*« 'Ji O! a; i^ -^ * * # „ * * C! i'-x^^s;t-o:? -J! X X --t O X «* ^ i^ x; C -i«^r-!«t-TOC'!C!"*"*»ft * * # * „ * 000000000000 cCintceiiftM-^-^C'tTjteoM 000 — »-'000<-iOOO 5J — X X ^ X ^( 'i>^ -^ »x*-So^— ot— t— -^ j:,^X"» r-.^-. ^WfTI*! in CO — 31 rct>.x 71 rt-* — oo-. M iri x" in in iri >n :o tri in t^ x — x"* 31 -j5 c'r- x'Jf'^to o^7I'ft•-^^>^^-3- 7l cn-T-»i©»tf5»inininift'- vn I- X :c ^XOi-it-^ -?! CO on .-^1 M Xm t- X 01 ffi Si tt tO ci t- tft t-^X X ao i-t o o I- in x" '^ t-^ X in oxsioc'i'Trocs'-^Cixw TjiOO'O O oox *o o' 1-J o t- ?t rM o o o' *( T-mi ^ i-i '?» —1 ?i i-> — M ^ - ^^0»I5> •*:£>» OOX'^QO fi-^-^x C! :D'~lOln;rl'^>«3. 01^0'yiSjfneor^mmccxp;^*?^ lO I— ?> , E fo iM CO :c I— a; c! r?t x o *^ Ji' -i 3. o co >-' in Ci t— c. ff(o i-- cr.xxxxxgooj 'Xf — x-fc^xinrroi- 600 -UAi i;-ia jad U055|UBldlOA. OS n o o Oh-' I > o T— < 601 'S a 5 o H C 1 P< ^ <5 S iii J 5 J O O 602 P3 P3 s o •S ^ 1 '.i ;^, - J f>03 < z s X ^ i: 604 K< I > Eh P3 rt s Q 0) d£ O E-( !zi nnild -lOA sa5K,ii!fij saii.niN p.puadsns psAtossid IT^ox paAIOSSIQ iniox BiuotnaiT! eaj^j t-i -^ X to CJ c» .ra a; tn as :o 00 w inootn .re i^ CO ire o CO o! -^ -^ -^ co ti koeO=Ot-^n^O.HOO^-HOIOO:iO?^0HOOOOi-lOOOOOOOOO— o _oo Ti M .n -3< t- -J -1 CO 7i -iJ ro -^ ^ r ooooocoosooo -o jatniod papuad paAjossip imox 9auo[qo p3AI0SST([ imoj, p.puadsng paAtossig l«»oi joioo ^uacatpag jfjipiqjnj, c3 ,_, ^ r-i i-i i-'i-iiH':!'—f'M'MTi'-" rcwci'—iiH oi X oo 00 ifiM -x) lo -Tp rj K S! r' ^ 3! S ^. 2^ ?: T ! c t ; I -* oc '^ '; - O -! iri '^ *M ^! X l-^ '-S '^ "-D O -^ -dH X 3 o CO '-0X O"' t— X c! -H >-H i;.t c; cc —> o" "~ ICOCITI ' ^^ . , . , ...,._- CJ --1 c^[ ro ro cciri Tt re ro era cc "M CO re :>';!0«'-'»it-^ctxc.x*-DooxTtt~.x H 'is o t- X --S --s '-C -* t- '-D o --o^ re ^ cr- £55m t- W£i » QD lA W M in * "* Ot OJ »ft --( CO ri ro m --1 71 'Ti cTD -^ oi ?i i?i N to CO 3C500000=iOO W -OCl^X fco7'it-;i"*if30Jaotso>ift"* 3 XCO 1 O iC CO ifS Ol CO O •£ fO "* 7 [ C-l CO T-4.-H xxxxi:oM i©-*M^x»ftT*,-(Oioocvirox CI c»i 7i -^ ro ei CO T-i CO rH C'i >H -^ 71 CO CO X ^i^ 5T CD X) X •*X C! <* CD X 7t X 7! --C 71 CD T*X-X OOX 1^ CDO ' 71^ CO I>^ '-' 7-! O '-CXX O (Mooieo — ac in en 'Si cf3 CI in ^' -" "* 7i 7 ! 7 ! "* CD r(^ 608 < o H P3 -H .— 610 ns y. '£^ pq 5 < d p Hi o :^ w o 611 a ŝ o 612 pq < cq O o O M Q o S Z P-l OS « >; P )£. o < Z; < < P ij SI W o e'oio ni 613 w ta p b O gl r> IS a pq X 3 O Q z; K z'mo ui —J91BM p-m 614 TABLE XIV. Relative Nujibbk of Planktonts in Illinois River and Spoon River. Date AhGM DIATOMS RHIZOPODA MASTIGOPHORA INFUSORIA 1896 18 VIII 25 VIII 16 IX IX X X XI XII 28 XII Average 30 14 29 17 3 1897 3 II 26 II 23 III 27 IV 25 V 28 VI 10 VIII 26 VIII U IX 2 XI 30 XI 28 XII Average 1898 I II III V VI VII 5 VIII 12 IX 4 X 2 XI 6 XII Averag-e 18U9 3 I 7 II 7 III 83. 200 4,840 5,446 10, 840 4,560 6,080 576 1,200 2,800 396 3,280 1,160 480 480 480 8, 890. 000 1,025,200 J50, 489 682, 200 816, 400 1, 064, 500 9U3, 916 165, 9110 ll,2s-0 2i), ;ir6 3, 3li(l 6, 240 10, 080 19, 200 480 7,200 14,420 40 1,495 2,700 1,600 68,400 49, 000 74, 680 166, 760 406, 400 28.980 940 1,200 7,500 2,500 3,800 24, 000 816, 000 145, 000 3,600 1,745,583 231,900 339, 480 1, 231, 900 328,414,320 79, 827, 200 1, 937, 600 2,318,160 847,320 282, 440 148,880 90,725 65, 400 9,688 240 240, 000 102,000 3, 048, 000 85, 200 54,0011 4, 800 662, 400 744, 000 24, 000 36, 000 66, 725 387 3,159 5,400 68, 800 17, 200 50, 040 308, 040 57,060 25, 000 18. 500 520 85,534 6,240 2,400 18, 000 16,800 3,700 1,440 5,500 34, 644, 610 174, 901 211,653 132, 140 314, 224, 400 30, 803, 600 3, 772, 000 360, 240 1,217,000 837, 200 981, 000 251,250 416, 920 16, 800 4, 800 27, 320 778, 400 31, 200 7,200 50, 400 8, 400 37, 4-10 15, 000 9,120 i,200 873 120 5,960 768 2,880 8,160 2,640 5,280 3, 360 500 480 1, 508, 000 110,000 75. 845 175, 000 1,111,200 3, 360 960 88,800 179, 606 64, 880 480 480 480 240 120 23, 800 11,200 40, 106 27,672 65, 200 44,000 10,083 1,440 3,540 192 240 2,400 1,200 1,920 7,200 480 240 16, 825 300 10,700 2,240 11, 520 25, 760 232, 600 8. 400 19, 20(1 118,000 3, 9U0 800 16,640 3,336 414, 166 720 37,440 105, 600 370, 200 40, 900 13, 200 1,200 :8, 800 2,400 42, 420 46, 600 172, 100 27, 987, 400 9, 029, 360 148, 006 439, 600 408, 160 1, 624, 800 1, 100 11,100 15, 400 160 3,840 9,600 43, 201. 3,700 18, 000 53,00(1 5, 030, 400 33, 600 116, 400 54, 000 27, 938 .3,400 18, 800 18, 240 515, U20 29, 600 :, 186, 000 27,600 ISO, 000 12,000 26, 500 244, 430 45, 600 1,935 2,880 19, 200 42, 000 98, 400 1,300 5,000 249, 600 55, 200 4,900 19, 200 50, 373 4,916 140 32, 087, 762 10, 300 66, 750 59, 120 180,553 3,840 800 37, 505 66. 338 141, 524 1. 400 49, 800 23, 600 19, 360 16,81:0 28. 001 12,580 32. 060 1,000 3, 577, 171 447, 158 19, 200 21,600 1,760 14, 880 21,600 10, 800 14, 400 22. 800 4,840 2, 500 4811 84, 22, 059 227,448 324, 800 967, 600 597,000 536, 800 252, 400 19,500 265, 600 5, 500 708, 000 55. 7. 51, 112. lii: 3. 4, 1, 109, 0O7 190, 017 69. 498 42. 020 129,600 ,516,000 20,440 129, 600 56,640 35, 900 69,060 7,620 41,473 39,000 14, 400 18, 880 11,520 63,600 108, 000 3,600 6,000 1,440 16, 800 3, 480 35, 678 220 1,125 3,200 12, 260 960 1,200 38,400 143, 020 117,000 899, 200 3,360 7,200 206, 036 108, 940 6,000 30, 820 26, 065 1,920 1,200 Average Grand At. Ratio - 53 43, 539 30,513 45, 390 23,031,820 78 1,547 293, 788 1,515 289, 637 13, 520 23, 281 .386, 407 3,921,328 3,520 161, 149 48, 586 468, 051 1,040 615 TABLE XIY —Coticluded. Relative Nuiibbr of Planktonts in Illinois River and Spoon Rivek. Date HOTATORIA EKTOMOSTRiCA ilSSECT LARYil MISCEIUSEOUS rOTlL SPECIES TOTll PUKSTOmS 189B 18 VHI 25 VIII 16 IX 30 11 29 17 i) IX X X XI XII 28 XII Averag'e 3 26 •» 27 25 28 1897 II II III IV V VI 10 VIII 26 VIII U IX 2 XI 30 XI 28 XII Average 1898 25 I 22 II 29 III 10 V 7 VI 5 VII 5 VIII 12 IX i X 2 XI 6 XII Average 1899 3 I 7 II 7 III 361, 185, 4)S. 131. 172, 86, 100, 2,7: 76^, 3,B0(i SWO S.Slili 480 2,m. 840 241^ 10,4.60 26,200 47. 180 1,276,000 2, 287, 160 351,900 658, 120 2, 059, 360 1,744,25(1 8. 911(1 109,84(1, 9. 04UI 1,5611 50« 2.880 20,000 34, 800 217,400 27,900i 15, 3001 1.3:W.2(I0 2. 3(i2. 4(K' 1,(172.8(10 1,965.600 1.135.200 68,160 31,480 K),070 16, 400 9,120 4,920 9,340 26, 140 15,701 682,083 126, 603 48,649 115,880 2,663,400 903,000 153, 000 1, 294, 240 197, 960 105, 020 156,300 64,280 529,848 41,300 112,310 108, 860 em mi. B2Q 480 900 400 20(> 400 88C( 600 ,280 6,720 '80C( 3, 520 2,100 21.320 67. 000 84. 720 22, 70U 118,100 86. 240 618.75010. 19.5001 7. 55. 340f 5.720!10, 60rt 660 3,360 560 480 2,560 401 130 160 80 4,120 1,520 2,607 2,400 4,640 5,360 1,663 520 680 216 24(f '-1 1,880 114 100 85 85 68 63 120 42 10,973,920 1,388,120 600,005 977,434 2,173,440 5, 408, 460 1,252,579 362,720 34,460 34,704 14,520 20,080 20,920 30,680 17,880 9,820 1,360 440 320 100 160 91,918 4, 3, 22 235^ 4a'i, 4, 22 24^ 33, 8, 3.98; 600; 7oa 3.16Ci 2.800 14. 700 920 1.2CK1 160 720 100 880 600 740 73,498 2,840 13, 976 18.500 80 400 400 7tai 80 'iool 120 400 2,10li 3. 700 2,00(1 1,400 160 :,o68 40 167 887 2,854 1,040 6,400 2, 080 4, 000 10, 240 4(1, 900 24, 400 15,760 26, 000 1,900 7,000 1,760 655 5,280 19,200 2S, 800 ' 500 200 !,400 11, 757 7,661 3,285 1,940 13,200 2:i,600 5,040 3,360 2,420 2,700 4,360 680 77 44 45 81 67 90 71 102 80 62 65 46 4,8 1,000 200 720 960 200 200 6,204 2,640 861 2,140 12, 000 1 74 35 100 78 67 96 81 89 79 75 40 2,892,085 29.3,080 450,2.80 1.498,:;00 35.><, 278. 080 91,362,440 14,968,406 3,6ti7,220 3, 782. 800 4.832.(>4(1' 239,6li(ij 520. IG5! 16(1, 7rtlj 20,492 1,720 52, 800 423, 200 591,600 3, 460, 600 136, 700 93. 400 1.418.(100 ll.-4i;7.6IX) J.O'U.OOO 2. li:i.9tK) 1.258.800 4(.), 004, 478 1,923,277 592, 831 708,501 645,840 402,352,600 .34,323,200 4,562,36(1 2.a86.920 1,603,300 1,287,720 1, 275, 380 2, 043. 090 41,071,06' 309.280 318,0; 1,121,980 85,600 41,800 70,400 1,8S6,360 165.900 209, 100 261,200 48,700 65,000 5.3,300 24.800 265. 560 16.840 10, ,"00 52.000 Averag-e Grand At, Uatiu - 87.490 465, 06' 2,50; 238, 828 11. 63,983 2,2551 104 200 616 6,805 2.8 2,430 41 69 2.9 3t.3,094 28, 283, 295 38 26,547 750, 429 616 TABLE XV. Solids in Suspension at Station E, Berkefeld Filter. Accession number 617 TABLE XY.—Contimml Solids in Suspension at Station M, Bekkefeld Filter. Accession number 618 TABLE XY.—Continued. Solids in Suspension at Station G, Bkrkefeld Filter. Accession number 22B32_.. 22636... 22640.... 22645... 22653... 22657... 22661... Z2667.. 22672.... 22678... 22683.... 22694... 22699_.- 22703... 22713.. . 22719.... 22724.... 22732..., 22750.... 22764... 22769. . 22776... 22781.... 22787.... 22792... 22798.... 22803... 22815.. 22820.. 22827. 22832. 22840., 22846.. 22852., 22857.. Pate XI, XI, XII, XII, I, II, II, III, III, IV, IV, V, V, V, VI, VI, VII. VII, VIII, via, VIII, IX, IX, X, X, XI, XI, XII, XII, I, I, I, I), II, III. Ill, 1897 1897 1897 1897 1898 1898 1898 1898 1898 1898 1898 1898 1898 1898 1898 1898 1898 1898 lh98 1898 1898 1898 1898 1898 1898 1898 1898 1898 1898 1899 1899 1899 1899 1899 1899 1899 Cu.cm. strained 5,000 5.000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,0!i0 5,000 5,000 5, 000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5,000 5, 000 5, 000 5,000 5,000 5,000 5,000 Gu. cm. silt 7.85 5.36 7.50 1.41 2.10 2.26 1.92 3.23 4.01 4.68 3.18 1.66 1.37 3.10 .96 1.25 1.48 2.42 3.80 4.14 3.80 4.23 2.06 6.86 4.12 2.85 1.45 1.08 1.55 2.66 3.88 3.52 3.18 8.91 13.86 11.65 Silt per CO. n 1,570 1,072 1,600 282 420 452 384 646 802 936 636 3:12 274 620 192 250 296 484 760 828 760 846 412 1,372 824 570 290 216 310 632 776 704 636 1,782 2,772 2,330 Eiver irage 3.2 3.4 3.2 6.8 7.1 10.7 11 14.1 17.6 13.1 11.1 10.1 13.6 12.5 10.8 8.7 4.7 2.6 3.7 3.9 4.2 4.9 3.9 4.3 6.7 8.5 7.2 6.9 6.8 8.2 8 6.6 10.2 13.1 13.5 Average for 1898.. 556.48 TABLE XV.—Concluded. Solids in Suspension at Station F, Berkefeld Filter. Accession number BIBLIOGRAPHY. Amberg-, O. '00. Beitriig-e zur Biologie des Katzensees. Inaug'. Diss. 78 pp., 5 Taf. Zurich. Also in Vierteljahrschr. d. naturf. Ges. Zurich, Jahrg-. 45, pp.59-136. Apstein, C. '96. Das Siisswasserplanktoa, Methode uad Resultate der quan- titativen Uatersuchung-. Mit 113 Abbildung-en. 200 pp., 5 Tab. Kiel uad Leipzig. Birge, E. A. '95. Plankton Studies on Lake Mendota. L The Vertical Dis- tribution of the Pelagic Crustacea during July, 1894. Trans. Wis. Acad. Sci., Arts, and Letters, Vol. X., pp. 421-484, PI. VII .-X. '97. Plankton Studies on Lake Mendota. II. The Crustacea of the Plankton, July, 1894—Dec, 1896. Trans Wis. Acad. Sci., Arts, and Letters, Vol. XE., pp. 274-448, PL XV.-XLII. Bokorny, Th. '97. Ueber die organische I5rn;ihrung griiner Piianzen und ihre Bedeutung in der Natur. Biol. Ceotralbl., Bd. XVII., pp. 1-20, 33-48. Brandt, Karl. '99. Ueber den Stoffwechsel im Meere. Rede beim Antritt des Rektoratesder KimiglichenChristian-Albrechts-Universitiit zu Kiel am (> Marz, 1899. 36 pp. Kiel. Burckhardt, G. '00. Quantitative Studien iiber das Zooplankton des Vierwald- stattersees. Separatabdruck, 309 pp.; Mitteil. d. naturf. Ges. Luzern, 1900, Heft 3, pp. 129-439. Cohn, Ludwig. '03. Untersuchungen iiber das Plankton des Liiwentin und ein- iger anderer Seen Masurens. Zeitschr. f. Fischeri, Bd. X., Heft 4, pp. 201-331, Taf. 1-12. Cooley, L. E. '89. The Illinois River Basin in its Relations to Sanitary Engi- 6ig 620 neering-. Prelim. Rep. 111. State Bd. Health, on Water Sup- plies of Illinois and the Pollution of its Streams, Appendix, II., pp. 49-81. ['91.] The Lakes and Gulf Waterway as related to the Chicago Sanitary Problem. xv+92pp. Appendices, 34 pp. With 3 maps and 1 profile. Chicago, (1891.) Forbes, S. A. '80. The Food of Fishes. Bull. 111. State Lab. Nat. Hist., Vol. L, No. 3, pp. 18-79. • '88. Studies of the Food of Fresh-Water Fishes. Bull. 111. State Lab. Nat. Hist., Vol. II. , pp. 433-474. '96. Biennial Report as Director of the Illinois State Laboratory of Natural History for 1895 and 1896. 31 pp., 20 pi. Fuhrmann, O. '00. Beitrag zur Biologie des Neuenburger Sees. Biol. Centralbl., Bd. XX., pp. 85-96, 120-128. 1 Fig. Garbini, A. '93. Primi materiali per una monografia limnologica del Lago di Garda. Mem. Accad. Agric. Art. Com. Verona, Ser. 3, Vol. LXIX. 73 pp. Greenleaf, C. E. '85. The Water-power of the Mississippi River and some of its Tributaries. Tenth Census of the U. S., Vol. 17, Pt. II., pp. 119-276. Hensen, V. '87. Ueber die Bestimmung des Planktons, oder des im Meere treibenden Materials an Pflanzen und Thieren. Fiinft. Ber. d. Komm. zu wiss. Untersuch. d. deutsch. Meere, pp. 1-107-f- I-XVIII, Taf. I. -VI. '95. Methodik der Qntersuchungen bei der Plankton-Expedition. 200 pp. Mit 114 Fig. im Text, 11 Taf. und 1 Karte. Kiel und Leipzig. (Ergebnisse der Plankton-Expedition der Hum- boldt-Stiftung, Bd. L, B.) Huitfeldt-Kaas, H. '98. Plankton in norwegischen Binneuseen. Biol. Centralbl., Bd. XVIII., pp. 625-636. 2 Fig. Jordan, E. O. '00. Some Observations upon the Bacterial Self-purification of Streams. Journ. Exper. Med., Vol. V., pp. 271-314, PI. XX. (521 Juday, Chancey. '97. The Plankton of Turkey Lake. Proc. Indiana Acad. Sci., 1896, pp. 287-296. 1 map. Kofoid, C. A. '97. Plankton Studies. I. Methods and Apparatus in Use in Plankton Investigations at the Biological Experiment Station of the University of Illinois. Bull. 111. State Lab. Nat. Hist., Vol. v., pp. 1 25, PI I.-VII. '97a. On some Important Sources of Error in the Plankton Method. Science, N. S., Vol. VI., pp. 829-832. '99. Notes on the Natural History of Polyodon Science, N. S., Vol. XL, p. 252. Kramer, Aug. '97. Ueber den Bau der Korallenriffe und die Planktonverthei- lung an den Samoanischer Kiisten. Mit Anhang: Ueber den Palolowurm, von Dr. A. Collin. 174 pp., 1 Karte. Kiel und Leipzig. Lemmermann, E. '98. Der grosse Waterneverstorfer Binnensee. Eine BiologiscRe Stiidie. Forschungsber. Biol. Station zu Plon, Theil VI., pp. 166-204, Taf. V. 1 Karte und 4 Fig. im Text. Leverett, Frank. '96. The Water Resources of Illinois. Seventeenth Annual Re- port of the U. S. Geological Survey, Part II., pp 645-849, PI. CVIII.-CXIII.; also as separate. '97. Water Resources of Indiana and Ohio. Eighteenth Ann. Rep. U. S. Geol. Surv., Pt. IV., pp. 419-559, PI. 35-37. Lockyer, Sir Norman, and Lockyer, W. J. S. '00. Sunspots and Rainfall. Science, N. S., Vol. XII., No. 311, pp. 915-918. '01. On Solar Changes of Temperature and Variations in Rain- fall in the Region surrounding the Indian Ocean. Proc. Roy. Soc, Vol. LXVIL, pp. 409-431. Loew, O. '96. Das Asparagin in pflanzenchemischer Beziehung. Chem. Zeit., Jahrg. XX., pp. 143-147. Lohmann, H. '01. Ueber das Fischen mit Netzen aus Mtillergaze Nr. 20 zu dem Zweck quantitativer Untersuchungen des Auftriebs. 622 Wiss. Meeresuntersuch., Abth. Kiel, N. F., Bd. V., pp. 45-66. 1 Taf. '03. Neue Uatersuchuag-ea iiber dea Reichthum des Meeres an Plankton und iiber die Brauchbarkeit der Verschiedenen Fang-methoden. Zugleich auch ein Beitragf zur Keuntniss des Mittelmeerauftriebs Wiss. Meeresuntersuchung-., Abth. Kiel, N. F., Bd. VII., pp. 1-88, Taf. 1-4. 14 Tabellen. Mag-nin, A. '93. Recherches sur la Veg-etation des Lacs du Jura. Rev. g'en. de Botan., T. 5, pp. 241-303. Marshall, W. L. '90. Annual Report upon the Improvement of the Harbor of Chicago and Calumet, and Illinois and Calumet Rivers. Lo- cation of Illinois and Mississippi Canal, and Operating and Care of LaGrange Lock on the Illinois Ris-er. Ann. Rep. Chief of Engineers, 1890. Appendix JJ, pp. i-viii-|- 2399-2605. Maxwell, W. '96. * » * * Report of the Hawaiian Experiment Station, 1896. Palmer, A. W. '97. Chemical Survey of Water Supplies of Illinois. Prelimina- ry Report. 98 pp., 3 pi., 1 map. Champaign, 111. Pieters, A. W. '94. The Plants of Lake St. Clair. Bull. Mich. Fish Comm., ' No. 2. 12 pp., 1 map. '01. Contributions to the Biolog-y of the Great Lakes. The Plants of Western Lake Erie, with Observations on their Dis- tribution. Bull. U. S. Fish Comm., Vol. XXL, pp. 57-79, PI. 11-20. Reighard, J. E. '94. A Biological Examination of Lake St. Clair. Bull. Mich. Fish Comm., No. 4. 60 pp., 2 pi., and 1 map. '98. Methods of Plankton Investigation in their Relation to Practical Problems. Bull. U. S. Fish Comm., Vol. XVII. , pp. 169-175. Rolfe, C. W. '94. List of Altitudes in the State of Illinois. Bull. 111. State Lab. Nat. Hist., Vol. IV., pp. 36-137. 623 Russell, I. C. '98. Rivers of North America. XV+ 327 pp., 23 fig-., 1 table. New York. Schorler, B. '00. Das Plankton iler Elbe bei Dresden. Zeitschr. f. Gewasser- kunde, Bd. III., pp. 1-27. Schroder, B. '97. Ueber das Plankton der Oder. Berichte das deutsch. botan. Ges., Bd. XV.. pp. 482-492, Taf. XXV. Seligo, A. '90. Hydrobiolog-isehe Untersuchung-en. I. Schriften d. naturf. Ges. Danzig, N. F., Bd. VII., pp. 43-89. '00. Untersuchungen in den Stuhmer Seen. Nebst einem An- hang-e : Das Pflanzenplankton preussischer Seen, von B. Schroder 88 pp., 9 Tabellen, und 10 Taf. Danzig. Steuer, A. '01. Die Entomostrakenfauna der " alten Donau" bei Wien. Eine ethologische Studie. Mit einem Anhang : Zur Frage iiber Ursprung und Verbreitung der Entomostrakenfauna des Siisswassers. Zool. Jahrb., Abth. f. Syst. Geog. u. Biol. d. Thiere, Bd. XV., pp. 1-156, Taf. 1-12. 20 Abb. im Text. Yolk, R. "01. Die bei der Hamburgischen Elbe-Untersuchung ang-e wand- ten Methoden zur quamitativen Ermittelung des Planktons. Mitth. a d. Naturhist. Mus. in Hamburg, Bd. XVIII., pp. 137-182, Taf. I.-III. '03 Hamburg-ische Elb Untersuchung. I. Allgemeines iiber die biologischen Verbiiltnisse der Elbe bei Hamburg- und iiber die Einwirkung- der Sielwasser auf die Organismen des Stromas. Mitt'n. a. d. Naturhist. Mus. in Hamburg-, Bd. XIX., pp. 65- 11.4, Taf. I. -VII. Ward, H. B. '96 A Biological Examination of Lake Michigan in the Trav- erse Bay Region. Bull. Mich. Fish Comm., No. 6. 100 pp., 5 pi. Ward, H.B., assisted by Graybill, H. W., and others. '00. A Comparative Study in Methods of Plankton Measurement. Trans. Am. Micr. Soc, Vol. XXL, pp. 227-247, PI. XV.- XV [I. 624 Whipple, G. C. '98. Classification of Lakes according- to Temperatures. Am. Nat., Vol. XXXII., pp 25-33. 3 fig-. '99. The Microscopy of Drinking-Water. XII+300 pp., 20 pi. New York. Whipple, G. C, and Jackson, D. D. '00. A Comparative Study of the Methods used for the Measure- ment of the Turbidity of Water. Tech. Quart., Vol. XIIL, pp. 274-294. Yung, E. '99. Des Variations Quantitatives du Plankton dans le Lac Leman. Arch. d. Sci. Phys. e. Nat, Ser. 4, T. VIII., pp. 344-364, PI. II. Zacharias, O. '95. Ueber die wechselnde Quantitat des Plankton im Grossen Ploner See. Forschungsber. a. d. Biol. Station zuPion, Th. IIL, pp. 97-117. '96. Quantitative Untersuchungen uber das Limnoplankton. Forschungsber. a. d. Biol Station zu Plon, Th. IV., pp. 1-64. Zimmer, C. , u. Schroder, B. '99. Das Plankton der Oderstromes. Forschungsber. a. d. Biol. Station zu Plon, Th. VII., pp. 1-24. Zumstein, BE. '99. Zur Morphologic und Physiologie der Euglena gracilis Klebs. Inaug. Diss. 50 pp., 1 Taf. Leipzig. Also, in 1900, in Jahrb. f. wiss. Botanik, Bd. XXIV., pp. 149-198, Taf. VI. EXPLANATION OF PLATES.* Plate I. Map of the Illinois River Basin, modified from a map in Cooley's " Lakes and Gulf Waterway," facing p. 58. Boundary of catchment-basin of whole system shaded, those of individual tributaries marked by dotted lines. Plate II. Map of field of operations of the Illinois Biological Station at Havana, III., i8g4-i8c)g. Locations of plankton stations in Illinois River (Ei, Spoon River (M), Quiver Lake (C), Dogfish Lake (L),Flag Lake (K-i, Thompson's Lake (G|, and Phelps Lake (F). Plate III. Illinois River bottoms at high water in spring flood in March, 1898, looking westward. River gage about fifteen feet above low-water mark. Taken from base of the eastern bluff, just below field headquarters (see Plate II.) on Quiver Chute. En- tire bottoms submerged, Quiver Chute and Illinois River united, Seeb's Lake show- ing dimly through the forest on west side, beyond this the broad expanse of Flag Lake, with low forest intervening between it and Thompsons Lake. Western bluff visible. Plate IV. Illinois River at low water during summer of 1894. Taken from same point as Plate III. River gage about 2 ft. above low water, showing minimum levels since erection of dam at LaGraiige. Mud spit between Quiver Chute and Illinois River exposed. Summer foliage and atmospheric conditions obscuring bottom-land waters to westward. Plate V. West bank of Illinois River a short distance below plankton station, looking northeastward. Taken during low water in midsummer of 1S94. Sloping shore of black alluvium covered by low vegetation. Narrow marginal belt of vegetation visi- ble. River about 400 feet in width. Plate VI. Sun-spots, rainfall, and river levels. Upper section of figure taken from Lockyer ('01). Middle section gives fluctuations in average rainfall in Illinois above and be- low the mean, as given in records of U. S. Weather Bureau. Lower section gives fluctuations in mean annual river levels, compiled from records at Copperas Creek dam, 1878-1899. Average of all annual means shown at the left. Plates IV., XV., XVIL, and XXI. are from the Biennial Report of the Director of this Laboratory for 1893 and 1814 : plates II., V., VIII., XVIII., XIX., and -SIX., from that for 1895 and 1896 ; and plates III. and XVI., from that for 1897 and 1898. 626 Plate VII. Hydrographs of Illinois River, 1879-1899, from records of State Canal Com- missioners published in reports U. S. Army Engineers, taken at lower gage on Cop- peras Creek dam, and from records of U. S. Army Engineers, taken at lower gage on LaGrange dam, 1883-1899. Mean hydrograph at the right based on means of monthly averages. Plate VIII. Seasonal distribution of plankton in Illinois River, Station E, in 1894. Volume of plankton in cm.'' per m." of water sliown by heavy black ordinants, the diagonal- lined apices of which indicate the estimated proportion which silt forms of the total catch. Thermograph in dotted lines, from records of surface temperatures made at the times of plankton collection. Hydrograph in continuous line, plotted from rec- ords at Copperas Creek. Heavy black areas at top of plate indicate the relative number of cloudy days per month at Havana, the vertical space equaling seven days. Plate IX. The same for 1895. Hydrogfaph from Jan. i to Aug. 8 is that at Copperas Creek, and thereafter in the main from Havana records. Relative thickness of ice- sheet indicated by black area at bottom of plate, i mm, equaling 6 cm. of ice. Plate X. The same for 1896. Hydrograph entirely from Havana records. Plate XI. The same for 18^. The same for 1898. The same for 1899. Plate XII. Plate XIII. Plate XIV. Spoon River near its mouth, looking toward southwest from first bend in the stream. Plankton station (M) located near trestle. Taken at moderately low water. Plate XV. Quiver Lake in midsummer, 1894, at low-water levels, looking northward from Station C (see PI. II.) toward the mouth of Dogfish Lake. Littoral vegetation in foreground. Driftwood indicating high-water margin. Lake rich in vegetation. Plankton station located in narrow strip of open water in middle of lake. Plate XVI. Quiver Lake, from same location, in low water of 1897. Only a small amount of marginal vegetation visible. Dogfish Lake also largely free from vegetation. Plate XVII. Western shore of upper end of Quiver Lake, looking northward, showing rich- 627 ness of vegetation. Emergent Nelunibo liitea Pers., with leaves, flowers, and seed pods. Submerged Cera/ophy/lum demersum L. Taken in low water of summer of 1894. Plate XVHI. Dogfish Lake, looking northeastward, in low-water summer conditions. Lake full of Ceratophylluvi, Elodea, and Potamogeton. Plankton station (L) near center of lake. Plate XIX. Flag Lake in autumn of 1895 •'t plankton station (K), looking north-northeast- ward. Scattered dwarfed clumps of 6'£-!;-/;/j- and an abundance oi NymphcFa consti- tute the principal vegetation in this open area. Plate XX. Thompson's Lake from shore station (G), looking southwestward, in low-water conditions of midsummer. Lotus bed in distance, and broad belt of submerged veg- etation, principally Ceratophyllinn, along shore. Plankton station (G) in open water to the right (northward). Plate XXL Phelps Lake, looking southwestward from plankton station (F), in midsummer in 1894, just as the lake was drying up. Plate XXI L Seasonal distribution of plankton in Spoon River (Station M) in 1896. Scale of plottings of plankton o. i cm.'' per vertical unit, instead of i cm.\ as in case of all other stations. Dotted portion of ordinant indicates estimated proportion of silt in total catch. Thermograph plotted from surface temperatures of water at times of collection of plankton, and hydrograph from gage-readings in the adjacent Illinois River at Havana. Ice indicated by black areas below diagram, I mm. equaling 6 cm. of ice. Plate XXIII. The same for 1897. The excess of plotted plankton-silt urdinants over limits of diagram is indicated by figures at top. Plate XXIV. The same for 1898-1899. Plate X.XV. Seasonal distribution of plankton in Quiver Lake (Station C) in 1894. Scale of plotting of plankton-silt is 0.4 cm.^ per vertical unit. Hydrograph is that of the Illi- nois River at Copperas Creek. Thermograph is that of surface temperatures at times of plankton collections. Plate X.WL The same for 1895. Hydrograph from Jan. i to Aug. 8 is that of the Illinois River at Copperas Creek, and thereafter, from river gage-readings at Havana. Plate XXVII. The same for l8g6. Hydrograph from gage-readings in the Illinois River at Ha\ana. 628 Plate XXVIII. The same for 1897. PtATE XXIX. The same for 1898-1899. Plate XXX Seasonal distribution of plankton in Dogfish Lake in 1895. Hydrograph from Jan. I to Aug. 8 is that of the Illinois River at Copperas Creek, and thereafter, at Havana. Plate XXXI. The same for 1896. Hydrograph is that of the Illinois River at Havana. Plate XXXII. The same for 1897. Plate XXXIII. The same for Flag Lake (Station K) for 1895-1896. Plate XXXIV. The same for 1897-1898. Plate XXXV. Seasonal distribution of plankton in Thompson's Lake (Station G) in 1894. Hydrograph is that of the Illinois River at Copperas Creek. Plate XXXVI. The same for 1895. Hydrograph from Jan. i to Aug. 8 is that of the Illinois River at Copperas Creek, and thereafter, at Havana. Plate XXXVII. The same for 1896. Hydrograph is that of the Illinois River at Havana. Plate XXXVIII. The same for 1897. Plate XXXIX. The same for 1898-1899. Plate XL. The same for Phelps Lake in i8g6. Plate XLI. The same for 1897. Plate XLII. The same for 1898-1899. 629 Plate XLIII. Seasonal distribution of chemical data and plankton in Illinois River in 1895- i8q6. Chlorine, oxygen consumed, free ammonia, albuminoid ammonia, total or- ganic nitrogen, nitrites, and nitrates, in parts per million, plotted according to scales specified at the left, and plankton in cm.^ per m.^, according to scale at the left, in the form of a continuous planktograph. The hydrograph, with scale at the right, is plotted in the usual form as a continuous curve. The planktograph, and the chlo- rine and nitrite plots are also in continuous lines, but, owing to distribution of data are more angular. Nitrite scale should read o.i to 0.3. Plate XLIV. The same for 1897. The same for 1808- Plate XLV. Plate XLVl. The same for Spoon River (Station M), for 1896-1897. Nitrite scale should read o.i instead of 10. Plate XLV II. The same for 1898-1899. Plankton scale at the left should read 0.1 to 0.4 in- stead of I to 4. Plate XLVIII. The same for Quiver Lake (Station C) for 1895 and 1897. Plate XLIX. The same for 1898-1899. Plate L. The same for Thompson's Lake for 1897, 1898, and 1899. Nitrite scale should read 0.1 instead of i. ERRATA AND ADDENDA Page 9Q, line 6 from bottom, for {'S/) read i'Sj); line 5 from bottom, alter J?o/fi, read, ('97). Page 100, line 3, page 132, line 10, page 264, line 11 from bottom, page 457, lines 7 and 15 from bottom, page 458, line 14, and page 541, line 8, for /F^r/v/l'pj) read ll'ani ('96). Page 159, Feb. 18, for 6.S read S.S; June 30, for j.j read j.j. Page 160, Oct. 9, for i.ig read /.p. Page 161, Nov. 24, for 6.6 read 8.6. Page 169, line 14, for Tal'le I. read Tables III. -IX, Page 170, line 15 from bottom, for X'lII, read /'//. Page 202, line 9, for ('g6) read (97). Page 253, line 15 from bottom, for ('97) read Cgya). Page 263. line 12, for iSgg read ]8g6. Page 282, line 11 from the bottom, for—43 per cent, read ± 43 per cent. Page 288, line 2 from bottom, transpose j/.i' and 28.8. Page 290, line 2, for 2 to j read / lo /j. Page 295, line 18, omit the first eight words; line 20, for 2 I2, 1.08, and -/.oi, read respectively, J, P2, o.y2, and j.p./. Page 311, line 14, and last line page 313, for node read iiiode. Page 319, line 14 from bottom, for HI. read XLIV.\ line 4 from bottom, tor the read an. Page 323, line 9 from bottoui, read in iSgy 'dher catc/ies. Page 332, lines 16 and 17, lur exceeded read p?'e<:eded. Page 343, line 14, for cm. read C}n:\ Page 350, line 17, lor qualitatively read quantitati7>ely. Page 357, line 6 from bottom, after 16.^6 read on the 24th. Page 358, line j8, after and read decreases the. Page 367, after heading Dogfish Lake read Station L. Page 371, line 4, before decaying read to. Page 372, line 16, after ma.xiinuni read in Quiver Lake. Page 381, after heading Flag Lake read Station K. Page 385, line 14, for uhella read uvella. Page 403, line 17, lor flood rea.A floods; line 18, before////)' read of. Page 405, line i, after relative read anttual; line 8, for j8gy read i8g6. Page 414, line 12 from bottom, for river read area, and for drains to read 7-eaches. Page 416, line 12 from bottom, for /.y read was. Page 422, linej from bottom, forj readj-j. Page 424, line 12, for (./j) read (/> c?n.}; line 13, for (.oj) read Ij ctn..)\ line 16 from bottom, lor flood read floods. Page 429, line 4 from bottom, for 22.JJ read 22.JJ. Page 439, line 2 from bottom, for the read their. Page 440, line 8, for mean read means and for the read their. Page 463, line 4, for to read in\ line 5, for peculiar in read peculiar to. Page 484, in table, transpose Vegetation-poor and Vegetation-rich. Page 501, line I. for show read shown. Page 505, last line of table, column 4, for .4 read 4. Page 510, line 9, add, and Vo/k ('oj). Page 546, line 15 from bottom, after is read in the main. Page 549, line I below heading, page 551, line i, and page 556, line 3, for Vgyb read {'gya). Page 556, line 7, before 'oj read '0/ and. Page 560, table, second column, line 4 from bottom, for ('00b) read ('00),; line 6 from bottom, for /z<«^ read Yung; line 16 from bottom, for ('pj) read Cpj),- line 17 from bottom, for ('gg) read (00); line 19 from bottom, for ('pj) read ('96). Page S84, at head of second column, for iSgy read i8gS ; under remarks, line 4, for above read along. Pages 597 and 598, below table, read *Bottom visible. Page 598, eighth column, line 7 from bottom, lor 0.16 read 0.28. Pages 599-603, 606-610, 612, and 613, below table, read *Plankton not collected on same date as sample for water analysis. Pages 599-613, columns 2-4, meaning of symbols, abbreviations, etc., as follows: -\- — rising river level. v. d. = very decided. — — falling river level. v. m. = very much. ± = stationary river level. Decimal in color column= volume c. = considerable. of standard ammonium chloride d. = distinct ; decided. solution required to develop the f. = filtered. same tint when diluted to fifty cu- 1. — little. bic centimeters with ammonia- m. = much. free water and treated with the n. f. — not filtered. usual amount of nessler reagent s. = slight. < a. o OPERA! Plate II. Q'TsLSq ^-O-O :«?„' Q \n nv LVDEA M HART ^= Wagon Road tSuhsrat.on.. ^caie of At'laa '^V'? ForesT z "•Tf ;-tv-» jiji»ii ^j _, .-n*!!- i 3 < »= a. + I LOWER GAOE AT COPPERAS GREEK DAM. General Hydbograph of Illinois River. 1879 to 1899 m CO u i J -a T >< X X "l F=i . > < < H M O > < 3 wsmmwjmjmm V j VHL [ ^ B > d ' ! , i > ^ ^ X |1 ' 1 '—> III 1 . X XXX mm .3 iJl a a D B HT XX mm o I J -8 .9 ^ ' E3 J .-> s ^ i -^T 1— 1 ^ RiterQaqeFl. =|a=l^'^«^>»«>|^^-^^-| > — K.erC»,en :^^=t5^.^«^,,^.^ Seasonal Distribution of nl Station C 1B9 5. ^ f ff.VerGrfffcn S^^lS^fo-l-^l^loi