Bulletin STATE OF ILLINOIS DEPARTMENT OF REGISTRATION AND EDUCATION NATURAL HISTORY SURVEY DIVISION STEPHEN A. FORBES. Chief Bulletin OF THE Illinois State Natural History Survey URBANA, ILLINOIS, U. S. A. VOLUME XVII 1927-1928 PRINTED BY THE AUTHORITY OF THE STATE OF ILLINOIS STATE OF ILLINOIS DEPARTMENT OF REGISTRATION AND EDUCATION A. M. Shelton, Director BOARD OF NATURAL RESOURCES AND CONSERVATION A. M. Shelton, Chairman William Trelease, Biology Henry C. Cowles, Forestry Edson S. Bastin, Geoloyy William A. Noyes, Chemistry John W. Alvord, Engineering Charles M. Thompson, Representing the President of the University oj Illinois THE NATURAL HISTORY SURVEY DIVISION Stephen A. Forbes, Chief H. C. Obstbeling. Editor ScHNEPP & Barnes, Printers Springpibld, III. 1929 13455—600 / 7 , 14-Aa^ >iA^ CONTENTS ARTICLE I. EPIDE.MIC DISEASES OP GRAIN CROPS IN ILLINOIS 1922-1926. By L. R. Tehon. (102 figures) October, 1927 1-96 Introduction 1 General methods used in the survey 2 Analysis of data 5 Data on the year-to-year prevalence and destructiveness of cereal dis- eases 15 Cereal rusts 18 Cereal smuts 43 Scab of cereals 56 Leaf spot and leaf stripe diseases 61 Summary of all data showing prevalence and destructiveness 71 Some relations of disease to weather and climate 79 Relation to mean annual temperature and total annual rainfall 80 Relation to seasonal temperature and rainfall 83 Geographical occurrence of dates of first spring infection, and the temperature relations governing the subsequent development of an epidemic 90 Summary and conclusion 95 ARTICLE II. A MANUAL OP WOODLOT JIANAGEMENT. By C. J. Telford. ( 11 figures) November, 1927 lui-194 Introduction and summary 101 Land classification 103 Total area of woodlot 105 The woodlot as a source of fuelwood. posts, and lumber for the farm 105 Woodlot protection 109 Grazing 109 Fire 109 Insects and fungous diseases 110 Adaptation of trees to soil Ill Even-aged system of silviculture 123 Development of an even-aged stand 125 Harvesting the crop 126 Reestablishment 126 Regeneration from sprouts 126 Regeneration from seed 128 Application of the even-aged system 129 Species suitable for production of piling 129 Intolerant species, exemplified by Cottonwood 129 V Illinois Natukal History Suuvky Bulletin Selection system of silviculture 131 General cultural measures 131 Fixing the diameter limit 135 Advantage of larger diameters 135 Planting to reentorce hardwood stands 138 Black walnut 138 Tulip poplar and basswood 139 Oaks 140 Methods of planting 140 Growing hardwood planting stock 141 Planting cleared areas 143 Commercial aspects of reforesting cleared land 143 Plantations for production of saw-logs 144 Growing coniferous planting stock 145 Transplanting 147 Conifer plantations on sands 148 Conifer plantations on loams 150 Christmas tree plantations 150 Cottonwood plantations on bottomlands 152 Period for most profitable harvest 156 Plantations for post production 156 Catalpa 157 Black locust 160 Osage orange 162 Summary of points on seed and tree planting 163 Measuring and marketing woodlot products 164 Choice of products 164 Sawlogs 168 Estimating standing timber 168 Scaling logs 171 Markets, specifications, and quotations 172 Lumber 173 Posts 175 Ties 175 Piling 177 Mine timbers 179 Cordwood 180 Appendices — A. Preservative treatment of fence posts 181 B. List of concerns dealing in tree seedlings and transplants.... 183 C. List of consumers and dealers (arranged by counties) 184 D. United States Department of Agriculture publications relating to woodlot management and tree culture 192 E. Illinois State Natural History Survey publications on forestry.. 193 P. Services offered to woodland owners by Illinois State Natural History Survey 194 Contents v ARTICLE III. AN EPIDEMIC OF LEECHES ON FISHES IN ROCK RIVER. By David H. Thojipson. November, 1927 195-201 Introduction 195 Extent of the epidemic 195 Seasonal habits of the leeches 196 Description 199 Color variant 199 Other species of Piscicola 200 Effect on fish yield 201 Acknowledgments 201 ARTICLE IV. THE PLANKTON OF LAKE MICHIGAN. By Samuel Ei)i>Y. November, 1927 203-232 Introduction 203 Methods and materials 203 Physical conditions 204 General character 205 Seasonal aspect 207 General distribution 207 Notes on constituent organisms 218 Cyanophyceae ( blue-green algae) 218 Bacillariaceae (diatoms) 219 Cblorophyceae ( green algae ) 221 Protozoa 223 Coelenterata 225 Nemathelminthes 226 Rotatoria 226 Cladocera 228 Copepoda 229 Summary 230 Acknowledgment 230 Bibliography 231 ARTICLE V. SOME PROPERTIES OP OIL EMULSIONS INFLUENC- ING INSECTICIDAL EFFICIENCY. By L. L. Enolisii. (8 fig- ures ) March, 1928 233-259 Foreword, by W. P. Fi.ixt 233 Introduction 235 Theory of wetting 236 Angle-of-contact measurements 238 Relation between wetting ability and toxicity to aphids 240 Relation of chemical property of oil and stability of emulsion to effec- tiveness against aphids 244 Relation between wetting ability and toxicity to San Jose scale and oystershell scale 246 Relation of volatility and viscosity of oil to effectiveness against scale insects 248 vi Illinois Natural History Survey Bulletin Relation of chemical property of oil and stability of emulsion to effec- tiveness against scale insects 249 Injury to plants 253 Conclusions 255 Acknowledgments 256 Bibliography 256 Appendix — A. Definitions of terms 258 B. Experimental methods 259 C. Analysis of tap water used in experiments 259 ARTICLE VI. SOME CAUSES OP CAT-FACING IN PEACHES. By B. A. Porter, S. C. Chandler, and R. F. Saza.ma. (10 figures) March, 1928 261-275 Introduction 261 Nature of the injury 261 Losses 262 Procedure 262 Some insects that cause cat-facing 264 The tarnished plant bug 264 Stink bugs 268 The plum curculio 269 Other possible causes 269 Relative importance of the different insects 270 Host plant relations 271 Hibernation of the tarnished plant bug 271 Control of the tarnished plant bug and the stink bugs 272 Insecticidal control 272 Cage tests 272 Control in the orchard 273 Repellents 273 Control of the curculio 274 Thinning 274 Summary 275 ARTICLE VII. THE BIOLOGICAL SURVEY OF A RIVER SYSTEM- ITS OBJECTS, METHODS, AND RESULTS. By Stephen A. Forbes. March, 1928 277-284 ARTICLE VIII. THE "KNOTHEAD" CARP OF THE ILLINOIS RIVER. By David H. Thompson. (14 figures) October, 1928 285-320 Foreword 285 Introduction 285 Description of knothead carp 288 Distribution of knothead carp 300 Relation to pollution 301 Changes in the food supply of the carp 304 Rate of growth of knotheads 306 Contents -vii Behavior of knothead carp 313 Etiology of knotliead disease 316 Conclusion 320 ARTICLE IX. METHODS AND PRINCIPLES FOR INTERPRETING THE PHENOLOGY OP CROP PESTS. By L. R. Tehox. (18 fig- ures ) July, 192S 321-346 Introduction 321 Graphic depiction of climate and weather 322 Weather in relation to the late blight of potatoes 323 Thermohyetics of the beet leafhopper and sugarbeet "curlytop" 327 Thermohyetics of the cucumber beetles and the bacterial wilt of melons 331 Significance of hythergraphs 339 Bibliography 346 ARTICLE X. AN ACCOUNT OP CHANGES IN THE EARTHWORM FAUNA OP ILLINOIS AND A DESCRIPTION OF ONE NEW SPECIES. By Fhaxk Smith. August, 192S 347-362 Introduction 347 Upland-soil species 348 Woodland species 349 Stream-bank species 349 Additions to the list of Illinois earthworms 353 Helodrilus heimburgerl n. sp 353 External characters 354 Internal characters 355 Circulatory system 355 Reproductive organs 356 Distinguishing characters of Illinois species 357 Generic names of Lumbricidae 361 Literature cited 361 ARTICLE XI. THE HESSIAN FLY AND THE ILLINOIS WHEAT CROP. By W. P. Flixt and W. H. L.MiUiMEH. (13 figures) Au- gust, 1928 363-385 Introduction 363 Life history 364 Control measures 368 Date-of-seeding experiments 368 Results 369 Winnebago and Boone counties 369 LaSalle and Bureau counties 370 Hancock county 373 Champaign county 374 Macoupin county 376 Marion county 378 Randolph county 379 Jackson county 381 Pulaski county 382 viii Illinois Natural History Survey Bulletin Summary 384 Acknowledgment 385 ARTICLE XII. THE BOTTOM FAUNA OF THE MIDDLE ILLINOIS RIVER, 1913-1925. By R. E. Richardson. (1 figure) December, 1928 387-475 Foreword, by Stephen A. Forues 387 Collections, dates, apparatus, river levels 391 River reaches covered, main hydrographical features, and principal collecting stations [with map] 393 Load of sewage and industrial waste and zones of pollution [with chart] 398 Classification of the species with reference to degree of tolerance 403 Some principal points regarding index value 410 Changes in the number of species taken and missing 414 Changes in the dissolved oxygen supply 420 Major changes in abundance of the more important groups 425 Changes in valuation as fish food and in per cent composition by weight 432 Temporal succession of the leading groups of poUutional or unusually tolerant forms in the polluted reaches below Chillicothe 437 Competitive relations 442 Accessibility, quality, and extent of use as fish food 444 Changes in abundance of the principal groups during and just after the continuous summer floods of 1924 448 Comparative abundance in the channel and extra-channel areas, 1920 to 1925 453 Changes in the mussel fauna of Peoria Lake, 1912 to 1925 454 Explanation of general tables 458 Summary 469 Appendix 473 INDEX 477 ERRATA Page 117. third line in table, for Sofe read Soft. Page 118, transpose Heavy soils and Light soils in columns 3 and 4 of Table VIII. Page 170. third line above figure, for S3 read Id!). Page 200. fourteenth line, tor sand read stand. Pages 20S, 211, 214. 219. in headings, for Bacrllariaceae read Bacillariaceae. Page 209. middle of table, for olirKutus read oligactis. Same correction on page 213, first line of table, and on page 225. second line from bottom. Page 211, fifth line in table, for Aiihanotheca read Aphanothece. Same correc- tion on page 218, sixth paragraph. Page 215, fourth line in table, lor acuminata Ehr. read acumi7tatiim (Kiltz.) CI. Same correction for acuminata on page 220, fifth paragraph. Page 224, fourth line from bottom, for Paiidoiiii read Pandorina. Omit last line and read Traverse Bay rcfjion. Page 228. fifth line from bottom, for Antario read Ontario. Page 267. for top row read bottom row. and vice versa. Page 327, second line, for Eutettic read Eutettix. Page 348, third line in second paragraph, for ro.^esus read roseus. STATE OF ILLINOIS DEPARTMENT OF REGISTRATION AND EDUCATION DIVISION OF THE NATURAL HISTORY SURVEY STEPHEN A. FORBES. Chief Vol. XVII. BULLETIN Article I. Epidemic Diseases of Grain Crops in Illinois, 1922-1926 The Measurement of Their Prevalence and Destructiveness and An Interpretation of Weather Relations Based on Wheat Leaf Rust Data BY L. R. TEHON PRINTED BY AUTHORITY OF THE STATE OF ILLINOIS URBANA, ILLINOIS October, 1927 NATURAI LiSiORi SURVEY SEP ' 1969 lIBnaqy STATE OF ILLINOIS DEPARTMENT OF REGISTRATION AND EDUCATION DIVISION OF THE NATURAL HISTORY SURVEY STEPHEN A. FORBES. Ch,er Vol. XVII. BULLETIN Article I. Epidemic Diseases of Grain Crops in Illinois, 1922-1926 The Measurement of Their Prevalence and Destructiveness and An Interpretation of Weather Relations Based on Wheat Leaf Rust Data BY L. R. TEHON PRINTED BY AUTHORITY OF THE STATE OF ILLINOIS URBANA. ILLINOIS October, 1927 STATE OF ILLINOIS DEPARTMENT OF REGISTRATION AND EDUCATION A. M. Sheltcix. Director BOARD OF NATURAL RESOURCES AND CONSERVATION A. M. Sheltox. Chairman William Trelease, Biolofty John W. Alvohd. Enyineering Henry C. Cowles, Forestry Charles M. Thompson, Reinesentinp Edson S. Bastin. Geology the President of the University of William A. Noyes. Chemistry Illinois THE NATURAL HISTORY SURVEY DIVISION Stephen A. Forbes. Chief SCHNEPP & Barnes. Printers Springfield, III. 1927 68727—1,200 CONTEXTS PAGE Introduction 1 General methods used in the survey 2 Analysis of data 5 Data on the year-to-year prevalence and destructiveness of cereal diseases.. 15 Cereal rusts 18 Cereal smuts 43 Scab of cereals 56 Leaf spot and leaf stripe diseases 61 Summary of all data showing prevalence and destructiveness 71 Some relations of disease to weather and climate 79 Relation to mean annual temperature and total annual rainfall SO Relation to seasonal temperature and rainfall 83 Geographical occurrence of dates of first spring infection, and the temperature relations governing the subsequent development of an epidemic 90 Summary and conclusion 95 VoLrUE XVII Article I EPIDEMIC DISEASES OF GRAIN CROPS IN ILLINOIS, 1922-1926 The Measurement of Their Prevalence and Destructiveness and An Interpretation of Weather Relations Based on Wheat Leaf Rust Data L. R. TEHON As the first of the necessary steps in a study of the epidemiology of crop diseases, the Natural History Survey ptiblished a report in 1924 on the occurrence and distribution of the common diseases of crop plants in Illinois*, which was based chiefly on specimens and notes secin-ed dur- ing the three years 11>21, ]!l"22, and 111-';!, but which also contained brief histories of each disease. The second step consists of the accumulation of data which will show the differences in abtmdance and severity of disease from year to year, from season to season, in diverse regions, and even in diverse localities. In order to ftilfill these requirements, the data must be both precise and as acctn^ate as circumstances governing the work will permit. Compara- tive words, such as "mild", "light", "moderate", "heavy", and "severe", which have been used commonly, are not sufficiently precise when applied to diseases, for their meanings are as varied as the experiences of the persons who hear or speak them. The most satis factor)- statement, and the one most likelv to be tinderstood by everyone, is therefore the one which presents the facts of disease attack in numerical terms. This im- plies nut only that it is possible to measure the attack but also that the measuring can be done bv means of quantitative units ciuite as definite in their particular application as are the feet, jiounds, and gallons of com- merce. Since the beginning of the survey in l!i"21, methods have been adapted and devised In- which disease abundance can be measured with a con- siderable degree of accuracy, and large quantities of data have been se- cured each year for many diseases. In both the accumulation and the analysis of data, problems have arisen which have made it necessary to adopt standard arithmetical meth- ods, the use of which is quite as important in plant disease surveying as *A pi-eliminary repcirt on tlie oocurrf in-c and distribution of the common bac- terial and fungous diseases of crop plants in Illinois. By L. R. Tehon. Bull. III. State Xat. Hist. Surv. Vol. XV. pp. 173-325. 1924. Illinois Natikal History Sirvky Biixktix are the calculations made by the civil engineer with the data he secures by transit and chain. The detailed account given in these pages of the means used to mea- sure diseases, of the methods by which the data are analysed, and of the results obtained, is confined to the diseases of cereals. It is presented, not merely because it follows so directly the task accomplished by the first report, nor yet because it represents the first intimate and extensive study of crop diseases growing naturally in Illinois fields, but especially because it will enable the grower of crops to understand, better than he ever has, the destructive etifect of diseases which he all too often ignores. GENERAL METHODS USED IN THE SURVEY The survey of crop diseases was begun in midsummer of 1!)21. The first task, that of determining the kinds and the distribution of diseases attacking crop plants in Illinois, was undertaken at once. At the same time, a start was made toward securing quantitative records of the amounts of each disease, for it was clearly seen that a means of arriving at definite comparisons could be found only in measurements of disease abundance. One of the first difficulties encountered was that of handling diseases difificult to identify in the field because of the lack of distinctive symptoms. This problem, after six years of work, has not been solved entirely, for the certain identification of some diseases requires equipment and a per- sonnel for routine work that is beyond our resources. However, for diseases more or less readily identifiable by the eye, the hand lens, or the compound microscope, a system was developed by the beginning of the growing season of 192;5, which, though requiring variation in its application to different types of disease, has the following general plan. Whenever the identification of a disease can not be made with abso- lute certainty in the field, a typical, representative, and copious sample of it is taken at the time the field is examined. The specimen is given a temporary "collector's" number, which is placed on a small paper label along with other identifying and suggestive information, such as the name SPECIMEN LABEL Collector's No. T IG Host MHicat—Kanred mixed. Disease Node cankers. Cause Examine for pycnidia of Spp/or/o nodortim. Town Fixher County Champaiyn Date Collected July 29. 1!I2S. Diseases of Grain Ckops, 1922-1926 3 of the crop, a brief note of the type of injury, and the place and date of collection. The specimen, with its label, is then packed in a small pressing and drying device carried by the observer. As often as convenient, the collected samples are sent to the laboratory where, when the process of drying is completed, they are arranged in sequence in the order of their inmibers and stored awa}' to await further examination. As a matter of convenience, and to assure the inclusion of sufficient information with each specimen, the labels have been printed and ar- ranged in pads. The label with ty]iical notes, shown at the bottom of the ])receding page, furnishes an example of its use. For recording the amount of disease in a field, standard blanks mea- suring 8 by 11 inches are used. They also are printed and arranged in pads convenient to carry. The reproduction on the following page shows the tvpe of blank that has been developed and the use that was made of it in recording the leaf rust infection in one wheat field. A complete statistical suminaiy of the information carried on these blanks would show the number of fields in which a disease was found, the number of acres comprised by them, the particular varieties of the crop on which a disease was mild or serious, whether or not preventive treat- ments were used consistently, the average percentage of diseased plants, and the average amount of disease per plant. There would also be data bearing on the abilitv of disease to reduce yield, on differences in the dates of first appearance from season to season and in various localities, on common sources of infection, local histories, and kinds and numbers of diseases encountered in a held, as well as special notes on local weather relations. Finally, the number given ])y the observer to the sample he collected is noted in the lower right hand corner of the blank. It connects the record blank to the disease sample preserved in the laboratory. The filled-in record blanks are sent, at suitable intervals, to the lab- oratory, where they are sorted according to crops and diseases. They are kept in a permanent file, the records for the various diseases being allotted individual folders which are arranged in the alphabetic order, first of the crop and then of the disease names. Separate folders are maintained, as a rule, for each year's records of a disease. In securing the information required for the field record, it has been found that the greatest care must be exercised, both in the identifying and estimating of amounts of disease. The observer must not be over-con- fident of his abilitv to distinguish diseases in the field but must be con- stantly watchful to ]irovide adequate evidence, by means of representative samples, of the accuracy of his judgments and the reliability of his esti- mates. In estimating disease abundance, difterent procedures are necessary for different types of disease. An estimate of the amount of loose smut in a wheat field may be secured by determining the number of smutty heads among a representative number in the field, but an estimation of rust 4 iLLiJfols Natural History Survey Bx-ixetin 1026 PLANT DISEASE RECORD BLANK Crop Wheat Disease Leaf rust (Puccinia triticinaj County Adams Locality Camp Point Variety of crop Kanred (somewhat Size of field 20 acres impure J State of development Heads just now emerging front the boot. Control measures {^^'t^apjijed x ^''"^ ^^^"^ When used How applied Infection: % of culms diseased 2o.S Amount per plant 20.S j;e?- cent Data: 362 in SOO Scale classes: > 10 25 },0 65 100 5S0 in 3000 F. 3 19 37 26 1 'i 7 1 348 in 1200 c. X F. 95 370 650 560 J,5o 100 1290 in 5000 S. C. F. 2230 = = 20.S per cent S. F. 107 Type of injury Estimated damage 20.S% X 2J.«% = 5.30% Date first observed Source of infection Past history Association with other diseases Stem rust, loose smut, bunt, scab, speckled leaf spot. Septoria nodorum on culm nodes. Additional notes (weather, phenology, etc.) Date of observation June 19 Specimen No. T Ji21 Observer L. R. Tehon. DisEjVsks of Grain Crops. 1922-1926 5 infeclii)n requires not only an examination of numerous jjlants to prove the presence of the disease l^ut also a close examination of the diseased parts of the plants to determine the average amount of rust on those parts. To obtain reliable data from any field, the number of plants examined must be large enough to give an average which will be representative of the entire field. There must be standards, also, for measm"ing the diseased area of leaf and stem, which can be carried and used with ease in the field. One of these standards is already widely used ; it. and the others which we have devised, will be described in connection with the diseases with which they are used. ANALYSIS OF DATA in order thai the oliservations made in fields by the methods outlined briefly above may be of use, they must be subjected to an analysis which will reduce them to brief and concise statements. The first ste]) is to make certain that the diagnosis made of the disease in the field is correct. In most instances, a microscopic examina- tion of the sample indicated by the collector's number is sufficient. The sam])le is labeled with its proper scientific name ; this name is noted at the top of the field- record ; and the serial "accession" number given to the specimen replaces the collector's number on the record blank. The sample is filed according to a system (the details of which are immaterial here) and serves as permanent and substantial evidence particularly of the accuracy of the diagnosis but also to some extent of the severity of the attack. Reference can be made between record sheet and specimen at any future time, and the two together constitute the entire record. In many cases, two items enter into a calculation of the amount nt disease present in a field. One is the prop(n-tion of diseased plants, and the other is the amount of disease jiresent on each diseased plant. The latter is expressed as the area of leaf or stem destroyed b}' disease, or as the proportion of the head injured or destroyed. With cereals, data are secured for the entire area oi a field by making counts of a representative number of stems. In taking data on the leaf rust of wheat, for examjile, a hundred stems may be counted in one place, two hundred in another, and other quantities in other places, until a large number—10,000 if necessary—have been examined. The same proportion of diseased stalks is not found, as a rule, in each group, and an average must be secured for all the counts. This is done in either of the following ways: 1. If 5000 stalks are examined, the counting may show that 362 stalks among SOO bore disease 580 " " 3000 " 3-18 " " 1200 " 1290 " " 5000 " 6 Illinois Natural History Survk.y Bulletin The totals show that 1,290 of the 5000 stalks were diseased. The pro- portion of diseased stalks, expressed finally as a percentage derived from the totals, is in this case 25.8. 2. If, in the same instance, the group counts had heen recorded directly as percentages, the following data would appear on the record blank. 45.3 per cent among Calculation Made with Disease Data The county acreages for a given crop, as estimated by the Agricultural Statistician, are combined with the county indexes of prevalence and destruc- tiveness computed in the first tabulation, and indexes obtained from them for the total acreage directly represented by the sample fields. These data apply to the leaf rust of ivheat. Diseases of Grain Crops, 1922-1926 13 Specifically, the sums of the acreage-percentage products for prevalence and destructiveness each are divided by the total acreage of the crop in the counties listed. The resulting quotients—.SG.G and 32.6 in the ex- ample—are weighted percentages indicative of the prevalence and de- structiveness, res])ectively, of the disease in question in the entire ter- ritory adjacent to the fields from which data were obtained by examina- tion. In order to compare prevalence and destructiveness of diseases year by year for the entire State or for smaller districts, it is necessary to go a step further and compute indexes which will apply to the State's total acreage of a given crop in each year, and to the acreages of the sta- tistical districts (see Jiage 11). Satisfactory indexes can be obtained for this purpose by extending the observed data, first to the county acreages, as was done in the foregoing table, and ihen to the total acreages of the agricultural districts and of the State. An example of the method is given in the following table. Taule III A.x Example of the Tabulation and Calculatiox Used to Obtaix Indexes OF Prevalence ANM) DicsTRroTiVEXEss for Each of the Nine Districts and for the Entire St.\te These data apply to the leaf rust of tcheat. For the sake of brevity, it is assumed in this example that the State is composed of the three districts listed. 14 Illinois Natihal History Sikvky Billetin With the exception of the district acreages shown in cokinin 7, which must be secured from the reports on agricultural statistics, this table is made up entirely from Table II. First, the counties appearing in the two previous tables are arranged in proper district groups in accordance with the districting in the statistical reports (see page 11). They then are listed in column 1. In columns 2, 3, and 4, the acreage of the crop per county and the acreage-per- centage products relating both to prevalence and destructiveness are en- tered, the figures being transferred directly from Table II. Totals are taken of the items in these columns for each district, and the weighted district percentages shown in columns 5 and 6 are obtained by dividing the two sums of the acreage-percentage ptoducts by the total county acreage. The resulting quotients represent prevalence and destructive- ness by districts. In coKimn 7 are the acreages reported by the Agricultural Statis- tician to be devoted to the croj) in question in each of the districts. The acreage-percentage products in columns 8 and 9 are based on these dis- trict acreages. They are obtained by multiplying the district acreages by the corresponding weighted percentages entered in columns 5 and 6. When all of these items are entered, totals are taken for columns 7, 8, and !). The totals of columns 8 and 9 are divided by the total of column 7, which is the State acreage of the diseased crop. The quotients thus obtained are placed, for convenience, at the foot of columns 5 and 6. These final weighted percentages—88.5 and 22.1 in the example—are concise expressions of the average prevalence and average destructiveness of a disease throughout the State. It probably has been noticed that in the sample tables no calcula- tions have been carried farther than one decimal place. Because mosi of the computations involve percentages, which are arithmetically really decimals extended to the second place, the usual mathematical procedure would be to stop the calculations with the first decimal and then either increase the unit place by one or discard the decimal, as it was either greater than 0.5 or less than 0.6. However, as round numbers occur very often in the acreage statistics, it is possible to retain the decimal with- out adding greatly to the labor. The result is that the final percentages, though not greatly influenced, are a trifle more accurate than they other- wise would be. In practice, computations similar in all respects to those just de- scribed are made whenever possible for every disease for which field data are secured, and each year's folder of field records for each disease contains these statistical analyses, comprised of the three complete tables, as a part of the permanent record. Diseases of Gkain Crops, 1922-1926 15 DATA ON THE YEAR-TO-YEAR PREVALENCE AND DESTRUCTIVENESS OF CEREAL DISEASES In order tu obtain useful data on disease prevalence, fields must be examined in as many parts of the State as possible. Traveling by auto- mobile, observers go from county to county, visiting fields along the way. This amounts in practice to a random sampling of disease conditions in crops throughout the State, the quantity of sainjjles taken of each disease on each crop depending on the intensity of cultivation accorded it, the number of observers at work, and the length of time elapsing between the ap])earance of a disease and the arrival of harvest. The procuring of the data presented in this paper has been influenced also by the fact that the observers at the same time were gathering similar information on diseases of the tree and bush fruits and forage and truck crops. Through the summer of 1922, four men acted as observers; in the summers of 1923, 192-1, and 192(;. two men; and for the greater part of the summer of 192."). only one man. This, to- gether wiih the large size of the State, the number of crops to be examined, and the number of diseases—more than 200 of which are serious—to be seen, contributed to what may appear an unnecessarily haphazard grouping of the grain fields subjected to examination. But the wide prevalence of diseases and the com])arative uniformity of their destructiveness in regions having the same latitude compensate in a large measure for these apparent but unavoidable faults. Cereal crops are grown throughout the State, some intensively in all parts, others only in restricted regions. The acreages devoted to par- ticular kinds vary considerably in difl:erent regions, as shown by Figure 2 which presents gra])hically the acreage statistics reported for 1925 by the Agricultural Statistician. The lines in this figure which divide the State into four roughly equal parts follow county boundaries, separating the counties into groups characterized in the main by the intensity with which certain cereals are cultivated. Corn predominates in each of the four sections, with a total of 9,240.000 acres in the State, or almost 5li per cent of the total acreage for all grain crops. Oats rank next to corn except in the south where the oat acreage is exceeded by the wheat acreage. Wheat has large acreages everywhere, although it is exceeded by barley in the north. Barley is im]X)rtant only in the north, and rve is unimportant when compared with the others. In the two central sections, which together have (i2 [jcr cent of the State's grain acreage, the order of importance of the various crops is: (1) corn, (2) oats, (3) wheat, (Ij rye, and (5) barley. In the northern section, the order of importance of crop acreages is : (1) corn, (2) oats, (3) barlev, (4) wheat, and ( •") ) rve. 16 Illinois Natukal Histouy Sukvky Bullktin Corn Z,^'i7,ooo W/iea^ 2/8,-500 Sa/^/lete than those given here are desired, they may lie found in the report referred to at the beginning of this paper. Cereals are grown in fields comprising, as a rule, rather large acreages. The plants are small and grow closely crowded together. These two facts have an important bearing on the obtaining of disease (lata, for even in a verv small field an examination of every plant is an impossibility and. furthermore, the crowding of plants makes it extremely difficult to distinguish one from another without digging out and care- fully separating them—a procedure which would not, and should not, 1)6 tolerated by the owner of a field. Hence, it is necessary, in securing data representing prevalence, to use the stem instead of the plant as a unit. The degree of prevalence is indicated by the {proportion of disease-bearing stems. Counts are made of a number of stems sufficiently large to give a dependable representa- tion of the infection in the field. Because of variations in soil and topography, disease prevalence is apt to vary noticeably from one part of the field to another. In order to equalize such variations, parts of the field count are apportioned to dift'erent places. The selection of the places for the counts and the decision as to the comparative number of stems to be counted in each place rests upon the judgment of the observer, but his aim must be to secure as true a representation as possible. Inasmuch as each disease destroys a particular pan of its host, the measure of destructiveness must be suited to that part. The device used is described in the discussion of each disease. It should be explained at this point, however, that although neither the devices for measuring destructiveness nor the methods of comput- ing values from the data obtained by using them were employed ex- tensively or consistentlv in the first years of the survey, the collection and preservation of adequate and representative samples enabled us to measure, at a later time, the destructiveness of diseases during those years. 18 Illinois Natitral History Survky Bio.ktin CEREAL RUSTS The most constantly prevalent type of disease in our grain fields is that known as rust. No cereal crop grown in Illinois escapes at- tack, and with the exception of corn, each is subject to two distinct rust diseases. The rusts are almost unique, in that they are not amenable, under present agricultural practices, to treatments designed either to prevent their attack or reduce their destructiveness. They present, there- fore, the outstanding opportunity for a study of the relation of disease to the na'.ural environment. Rusts do not kill their hosts. Individual infections are limited to very small portions of the ])lant, seldom extending over an area twice that of the rusfy pustule which is the outward sign of the disease; but a multiplicity of individual infections results in the conversion of a large part of the host tissue to the uses of the parasite, giving the appearance of a heavily rusted plant. Measuring the destructiveness of a rust attack thus becomes a matter of ascertaining the relative abundance per plant of the individual infections, or—for convenience—the relative amount of plant surface occujiied by pustules. An actual measure- ment of this area would be im])racticable, of course, under field condi- tions, and a substitute must be used. A standard was devised by N. A. Cobb in 1892, which he used extensively and successfully while studying rust diseases of grains in Australia. An adaptation of it which has been used extensively in ex- perimental work has been found equally useful for field work. As in- dicated in Figure 3, it consLsts of a number of diagrammatic pictures Diseases of Grain- Crops, 1922-1926 19 which illustrate a series of typical rust attacks ranging from slight to heavy by easily distinguishable intervals. The classification of the units in the series is based on the relative abundance of rust pustules. The heaviest attack illu.strated is the heaviest infection possible and is given an arbitrary value of 100 per cent, the lesser infections being graded as percentages in their proper relation to the heaviest. A copy of this scale is carried by the observer in a small notebook. As he makes his counts in a field to determine the number of stems carrying rust, he selects from time to time and quite at random a copious sample of the disease, which, either then or later, is compared with the scale bv fitting its individual parts, one after another, between the two diagrams which they most closely resemble. In order to elim- inate errors of judgment as completely as possible, each part of the sample is classed as the lesser of the two units of the standard between which it falls, rather than being assigned an estimated intermediate value. From one sample so measured (see the typical record sheet on page 4 ) , the following data were secured. Scale classes of Number of parts of the Class value rust abundance sample in each class times frequency per cent 3 5 " •• 19 95 10 " •• 37 370 25 " 26 650 40 '• " 14 560 65 " " 7 455 100 ' • 1 100 107 2,230 Destructiveness (mean for the sample) of rust on disease-bearing 2230 stems, • = 20.S per cent. 107 These items are recorded on the field record sheet, and serve as a complete and permanent record of the examination made in that field. Since the individual values given above have no adaptability as a usable figure, they must be reduced to a single value representative of the whole sam])le. 1 he simplest way to obtain it is to compute the arithmetical mean. This process will require some explanation. Since it is obvious that the larger numbers of moderately rusted sample parts will have a greater significance than the lesser number of heavily and lightlv rusted parts in determining the average of the sample, it is necessary to give each class of infection its proper weight with respect to the other classes, as determined by the number of parts of the sample falling in it, by multiplying the value of the rust-class by the frequency with which it occurred in the sample. The product in each instance, in- dicated in the above tabulation as "class value times frequency," shows approximately the total rust infection in that class, and the sum of these products is approximately the total rust infection in the sample. To find the average rustiness of the sample, it is now necessary only to divide the sum of the products of the class values and frequencies by the number of parts in the sample. The value resulting. 20.9 per 20 Illinois Natural Histoky Survky Buixktin cent in the above example, is considered to re])resent the destructive- ness of the rust on the disease-bearing stems. For the field as a whole, destructiveness is calculated by the methods outlined previously (p. S ) . In the main, the rusts of cereal crops are classified as "stem" rusts and "leaf" rusts, according to the plant part which they customarily inhabit. This characteristic demands that the destructiveness of stem rusts be measured as diseased stem surface and that of leaf rusts as diseased leaf surface. It is customary to speak of stem area and leaf area, rather than "surface." Wheat L,eaf Rust Caused by Piicciiiia triticina Erikss. The period during which wheat leaf rust has been under observa- tion e.xtends from the spring of 1922 through the harvest of 1926. In each of these five seasons, fields of many varieties of wheat were examined in many parts of the state, and careful records were kept and sunnuarized in accordance with the statistical methods previously explained. Each year's records represent as accurate a statement of the prevalence and severity of leaf rust as could be obtained by sampling, with the means at hand. During the early summer of 1922, before wheat harvest, leaf rust was observed in 139 fields, so distributed as to furnish data applying to the territory shown in black in Figure i. This being the first season of attempted quantitative recording, it was found later that only 43 of these fields had been sufficiently well examined to allow statistical consideration of the observations. These fields, which were distributed among 17 counties, contained 1S62 acres. In them, an average of 95.4 per cent of the wheat stems bore rust-infected leaves, and an average of 50.6 per cent of the entire leaf area was occupied by pustules. An exceedingly large number of fields were examined in 1923, rec- ords relating to leaf rust being secured from a total of 10,963 acres dis- tributed among the 52 counties shown in Figure 5. In this acreage, the average proportion of stalks bearing rust-infected leaves was 92.0 per cent ; in the entire acreage of the counties represented, 91.0 per cent; and for the state as a whole, S9.T per cent. It was de- termined that the rust pustules occupied an average of 46.6 per cent of the leaf area of every stem in the acreage examined, 33.8 per cent in the county acreage, and 31.3 per cent in the state's acreage. Records were taken on the prevalence of leaf rust in 1924 from fields containing 8,686 acres. They were distributed among the 68 counties shown in Figure 6. For the acreage involved, an average of 74.6 per cent of the wheat stems bore rust-infected leaves, and 49.5 per cent of the leaf area per stem was occupied by pustules. When the data were extended to the county acreages, the index of prevalence became 67.4 and that of de- structiveness 19.1 ; while for the entire state the index of prevalence was 68.3 and of destructiveness 19.1. DiSEASKS OF Git.\iN Ckops, 1922-1926 21 Figs, 4-8. Sources of wheat leaf iu-st data. 1922-1!1_;6 The black regions on these maps show the territory from which data were taken each year. Nearly 16,500 acres of wheat were examined in the five years. These maps do not represent the distribution of the disease, which is supposedly state-wide every year, but simply show the counties in which our observations were made. (See page 20.) 22 Illinois Natiral Histouy Sirvey Billetin In 1925 records of leaf rust were taken in fields aggregating 3,225 acres. The fields in which the observations were made were distributed among the 36 counties shown in Figure 7. When the data that were secured were subjected to analysis, the ratio of stalks bearing rusted leaves in the acreage examined was found to be 63.5 per cent and the average amount of rust-occupied leaf area per plant 6.5 per cent. The index of prevalence for the total coimty acreage represented by the sample fields was 46.3 and the index of destructiveness was 6.0, while for the entire state the indexes were 52. (S and 17.1, respectively. Fields aggregating 1 ,-145 acres and distributed among the 31 counties shown in Figure 8 furnished leaf rust data in 1!t26. Since agricultural statistics have not yet been provided for this year, it has been impossible to carry out all of the usual computations ; but the data secured show in their preliminary analysis, that in the acreage examined 57.3 per cent of the stems carried rusted leaves and 11.2 per cent of the leaf area was occupied by pustules. As a very close agreement has been evident between the indexes se- cured directly from field data and those secured for district and State acreages by indirect computation, the direct indexes for 1926 will be compared in subsequent pages with the computed indexes of previous years. The summarized data for the five seasons are brought into closer comparison in the following table : Year Acreage examined Prevalence: Calculated per- centages of wheat stems bear- ing: diseased leaves For the acreage examined For the county acreage repre- sented For the State's wheat acreage Destructiveness: Calculated percentage of leaf area occu- pied by rust For the For the i county acreage acreage examined ! repre- t sented For the State's wheat acreage 1922 192.3 1924 1925 1926 862 10,963 S,6S6 .i,225 1,445 95.4 92.0 74.6 63.5 57.2 94.9 91.0 67.4 46.3 94.1 89.7 68.3 52.6 50 46 49 6 11 53.9 33. S 19.1 B.O 50.3 31.3 19.1 17.1 There is a considerable difference in both the prevalence and the destructiveness of leaf rust between any two years, and there has been a marked and continued decrease in both through the five-year period. The first two of the five years were characterized by very heavy attacks, while in each of the last three years the attack was moderate or light. When considered from the point of view of destructiveness alone, these contrasts are esj)ecially evident, but they are borne out in very certain terms also by the corresponding decrease in prevalence. Diseases of Grain Crops, 1922-1926 23 Oats Crovvx Rust Caused bv Piicciiiia coronata Corda. The crown rust oi oats is essentially a leaf rust, at least under average Illinois conditions, for though it often attacks the stems, infec- tion as a rule does not occur there until a day or two before harvest and usually results in but little damage. Quite early in the season, however. its red pustules appear on the leaves, multiplying rapidly until harvest. .Measurements of the destructiveness of crown rust are obtained in e.xactlv the same way. and by employing the same standard, as are those of the leaf rust of wheat (see page lit and Figure 3). In ]!I22 crown rtist was seen in oat-fields distributed among the 44 counties shown in Figure !•. Data were secured from TOo acres, the re- maining fields not having been sufficiently well examined to give usable information. In this acreage. 7M.7 per cent of the stems bore rust infected leaves. oT.U per cent of the leaf area being occupied by pustule.-. The field data when apiilied to the total acreages of the counties which they represented, gave a prevalence index of G!l.2 and a destructiveness index of 28.0. The indexes for the State's acreage are respectively (>7.0 and 27.3. The e])idemic of 1922 was preceded by an abundant and heavy fall infection on volunteer oats throughout at least the southern half of the state. This infection is known to have extended on the eastern side of the state as far north as Champaign County where specimens of actively sporulating rust were collected as late as the latter ]iart of October. In 1!I2;5 crown ru.st infection was observed in fields located in the 28 counties shown in Figure 10. Fields totaling 1,;^06 acres were ex- amined, and the records secured in them indicate a prevalence of 93.5 per cent and a destructiveness of 41.2 per cent. For the total county acreage to which the data applied, the indexes were 83.3 and 28.8 for prevalence and. destructiveness. respectively, while for the State's acreage they were 83.1 and 30.(5. The epidemic of 1923 was early in starting. By June 4 an infection was observed in Jackson County involving an average of 25 per cent of the leaf area on 50 per cent of the stems, and two days later infection was seen in Saline County involving 1.5 per cent of the leaf area on about 10 per cent of the stems. The infection, thus started, spread and in- creased until, by harvest, the heavy epidemic indicated by the figures given above had developed. In 1924, crown rust was observed in fields located in the 33 counties indicated in Figure 11. It is remarkable that the epidemic of this year, as suggested bv the map. was confined largely to the northern section of the state, the crop in the south almost completely escaping attack. F'ield data were secured from 1.080 acres parcelled among 29 counties. The analysis of them indicates that in those fields 79.9 per cent of the oats stems bore infected leaves and that 14.8 per cent of the leaf area was occupied by pustules. The data, when applied to the total county acreage of which they are typical, show a prevalence index of (i5.7 and a destructiveness index of 10.7. For the State's acreage, these indexes are 64.3 and 11.9, respectively. 24 Illinois Natukal History Suhvky Eitlictin Figs. 9-13. Sources of data on oat.s crown rust. 1922-1926 The black regions on each of these maps show the territory from which data were talten in each year of the period. The fields examined contained more than 4,000 acres. (See page 23.) Diseases of Grain Ckohs. 1922-1926 25 The rust was late in getting started. All the observations were made during July and August, the earliest July fifth in Adams and Brown Counties, at which time the crop was already ripening. One isolated and very earlv report from the north (June 'i1 at Rockford) showed an exceedingly light infection apjjarently originating from a crown rust infected buckthorn hedge nearby, but elsewhere the usual seasonal de- velo])ment prevailed. Crown rust was not a very important disease in 1!I25. Records of its jiresence were secured only from the 13 counties shown in Figure 12. which lie chiefly in the ncirthern half of the state. The earliest infection was seen near Greenville, in Bond Comity, where, by June 26. an aver- age of .03 per cent of the stems bore rusted leaves, one leaf on each of these stems having one or two pustules. The fields from which data were secured contained only 2-K acres In them, an average of 14.8 per cent of the stems bore rust, only .17 per cent of the leaf area being occupied by jnistules. These data may seem to be insufficient for further use. but in view of the very mild attack it is worth while to evaluate them for the greater acreages. For the total county acreages of which they are typical they indicate a prevalence of 2.S.2 per cent and a destructiveness of .4 per cent, while for the State's acreage the corresponding figures are 27 per cent and .6 per cent, respec- tively. In 192(i. crown rust appeared early. It was found first June 2 near Marshall, Clark County. Throughout the southern half of the state this original infection was so slow in spreading that it never reached heavy proi)ortions ; but farther north, after becoming established, it spread rapidly and became exceedingly abundant. .\s a result, records were obtained chiefly in the north. The fields from which records were taken were distributed among the 20 counties indicated in Figure 13. aggregating (597 acres. In them, 70.4 per cent of the culms bore crown rust and U.li per cent of the leaf area was occupied by pustules. Owing to the lack of acreage statistics, the data can not be carried farther to show the prevalence and destructiveness of the disease throughout the state. For closer comparison of the difterences in crown rust infections in the years concerned, the data are brought together in the following table : Prevalence: Calculated per- centages of oat culms bearing I diseased leaves Acreage i examined For the acreaee examined For the county acreage repre- sented For the State's oats acreage Destructiveness: Calculated percentages of oat leaf area occupied by rust pustules For the For the county acreage acreage examined repre- sented For the State's oats acreage 1922 26 Illinois Naturm. Histokv Survey Bili.ktin No two years show exactly the same indexes, and even when the indexes of prevalence are similar, as in \'J22 and li)24, those of destruc- tiveness are quite different. With respect to prevalence, the five years listed fall into two groups, 1923, 11)23, 1924, and 1926 showing great prevalence and 1925 slight ; but with respect to destructiveness, there is an even more distinct grouping, 1932 and 1923 ranking high and the subsequent years very low. Barley Leaf Rust Caused by Puccinia simplex (Koern.) E. & H. The method of measuring the leaf rust of barley is the same as that used in the two leaf rusts just discussed, but since the host crop is confined in a very large measure to the northernmost third of the state (see page 16), the data must be held to apply particularly to that region. Owing to the limited distribution of the barley fields, which are concentrated in the north but occasional elsewhere, data for this disease have not been obtained in the same abundance as for the ones previously discussed. Although in some cases they are too few to be treated statistically, they are summarized here very briefly in order to afford such comparison as they will permit in the latter part of this paper. In 1923, in a single field examined in Boone County 25 per cent of the stems carried rusted leaves, about 5 per cent of the leaf area being rust-covered. This indicates an average destructiveness of about 1.2 per cent. In 1923, a single examination, similar to the one above, made in Ogle County showed a prevalence of 100 per cent and a de- structiveness of 5 per cent. Single fields were examined in 1924 in Kendall, McHenry, Lee, and Ogle Counties. They contained a total of 40 acres, in which 100 per cent of the stems were so lightly rusted that only 4.4 per cent of the leaf area was occupied. Only 1 field furnished data in 1935. In 15 acres near Rockford 5 per cent of the leaf area was occupied by pustules, but the disease was confined to 34.6 per cent of the stems. The data for 1926 were more complete. In the north, records taken from fields in Boone, DeKalb, DuPage, Kane, and McHenry Counties showed 100 per cent of the stems to be rust-infested and 10 per cent of the leaf area rust-covered. Other records were secured in Jackson and Mason counties. Only a trace of rust was found in the first county, but in the second 1 or 2 infected leaves, each bear- ing from one to five pustules, occurred on 13 per cent of the stems. Averages of these data indicate a ijrevalence of 76.1 per cent and a destructiveness of 7 per cent. The several years compare as follows : Diseases of Gr-\ix Crops, 1922-1926 27 vear to year in prevalence, but not equally in destructiveness, for the indexes of the latter were closely similar every year. Rye Leaf Rust Caused by Puccitiia dispcrsa Erikss. Leaf rust data can be secured more readily for rye than for barley, not only because the rye acreage is g^reater but also because the fields are more widely distributed. Nevertheless, in comparison with wheat and oats, the acreage is small, and its wide distribution results in fields of small size, except in the small regions especially adapted to its culture, which are characterized especially by sandy soils. In I'.t'i'i data were secured from fields located in the seven counties shown in Figure 14. .\lthough taken from a small number of fields, they have considerable weight, both because of their relative uniformity and because of the representative distribution of the fields. A consider- ably greater acreage of rye was examined for leaf rust in 1923, the fields being distributed among the 1:5 counties shown in Figure 15. The exanfinations yielded data of considerable significance, especially because a fairly large number of the fields were located in regions of intensive rye culture. The season of 1924 provided data much more complete than any other year included in this report. The 4iS counties shown in Figure 16 each furnished data from one or more fields, the fields examined reaching T:! in number and containing a total of 393 acres. Fields located in tlie H counties shown in Figure IT furnished data on rj^e leaf rust in lOSo. With the exception of Morgan County, where three fields were examined, observations were made in one field in each county. The wide distribution of the counties, however, gives weight to the data, particularly since important rye sections are included. While very little data could be secured in 1926. several small fields were examined in the 4 rather centrally located counties shown in Figure 18. The results of the field examinations, as shown by an analysis of the data, are briefly as follows: Year Acre.ige examined Prevalence: Calculated per- Destructiveness; Calculated eentages of rye stems bearing percentages of rye leaf area rust infected leaves I occupied by rust F\)r the acreage examined For the county acreage repre- sented For the State For the acreage examined For the county acreage repre- sented For the State 1922 1923 1924 1925 1926 97 217 393 230 23 7: 59, 39 75.0 19.1 97.7 54. S 45.0 30.9 99.1 50.3 ••22.2 44. S 46.4 12.0 6.3 1.9 1.7 34.0 9.2 S.2 2.0 0.9 Here, as in the cases of the rusts previouslv discussed, there is variation in both prevalence and destructiveness from year to vear. 28 Illinois NAxrHAi. Hismuv Strvky Bitlktim Fig. 17 192S Figs, 14-lS. Sources of rye leaf rust data, 1922-1926 The black regions on these maps show the territory from which data were taken in the years making up this period. There were, in all, 960 acres in the rye fields examined. (See page 27.) Diseases of Grain- Crops, 1922-1926 29 Fk;. 19 1922 \-H.. 20 /V2V Fig. 21 192S Vh,.22 1926 Fk;.s. 19-22. SorRcKS of corn risi ma ta The black regions on these maps show, respectively, the territory from which data were taken in 1922 and in the period 1924-1926. In 1923, data were obtained in Douglas, Effingham, and McLean counties only. (See page 30.) 30 Illinois Natiiual Hi.story Survey BriimN This rust has been, on the whole, much milder than the other leaf rusts, the season of 1923 being the only one in which a really destructive at- tack occurred. The yearly variation appears to have been concerned with prevalence more than with destiiictiveness. Corn Rust Caused by Piiccinia sorghi Schw. Corn, the most important both in yield and acreage of all the cereals grown in Illinois, is subject to the attack of but one rust. It is a leaf rust. In securing data on corn rust, the only phase of the attack that has been studied is prevalence. The intensity of the attack is exceed- ingly hard to estimate, first because of the large size of the corn leaf, and second because the pustules occur very sparsely on the leaf and only in limited spots. No means for estimating the severity of the attack has yet been devised. In 1922, corn rust was not very prevalent. .\s a result, data were secured only from the (i counties shown in Figure 19, the small acreage represented being located entirely in the northern half of the State. The data, though meagre in comparison with other years, were nevertheless representative. Even less abundant in 1923 than in the previous year, rust was recorded in only three fields, one in Douglas, one in Effingham, and one in McLean Counties. If the infection in these fields was typical of that throughout the State,' the prevalence would be expressed, as 3 per cent. An abundance of data was obtained in 1924. In the 60 counties shown in Figure 20, data were taken from 84 fields aggregating 3762 acres. For 1925 the data are less copious than for the preceding year, but the IT counties shown in Figure 21 contributed records from fields aggregating 450 acres. Exactly 70 fields, distributed among the 63 counties shown in Figure 22, furnished data in 1926. Rust prevalence was exceedingly variable, being very high in some districts and very low in others. The data for the 5 years are brought together in the following table for more intimate comparison. Diseases of Giiain Citors, 1922-1926 31 Although prevalence only has been recorded, a difference from year to year is evident. It is remarkable that the greatest prevalence occurred in 1924, succeeding the year of least prevalence. Comparison of all leaf rust data In the preceding pages the yearly indexes of prevalence and de- structiveness shown by data taken directly from cultivated fields have been given for the several leaf rusts. As they were considered, it be- came apparent that, though year-to-year variation was a common charac- teristic, the particular years of greatest and least intensity of attack were not the same for all. This fact is brought out more clearly in Figures 23 and 24. The indexes of prevalence for the entire state, as listed in the foregoing tables (see pp. 22-30), are shown graphically tor all the leaf rusts in Figure 23. The year 1922 appears to have been very favorable to the leaf rusts of wheat, oats, and rve, but unfavorable to those of Fig. 23. Pkev.\lence of cfrk.yl le.\f rists, 1922-1926 The prevalence indexes computed for the leaf rusts from field data and given in tables in the text are shown graphically here. The wide differences in the prevalence of these rusts in each year and in the trend of prevalence from year to year suggest that there is a particular combination of weather conditions most suitable for the maximum prevalence of each rust. (See page 32.) 32 Illinois Natural Histoky Survky Bulletin corn and barley. The following year, 1!I2;5. was somewhat less favor- able for wheat leaf rust and very much less favorable for the corn and rye rusts, while the oats rust found better conditions and the bar- ley rust attained its maximum prevalence. The general trend in 11I24 was downward, as compared with li)2;>, although the barley leaf rust maintained the high prevalence of the previous year, and corn rust found the best conditions of any of the five years. Rye leaf rust, alone of the five, was able in 1925 to increase its prevalence over that of the preceding year, the four others showing very marked decreases. Definite increases in prevalence were shown by the barley, oats, and wheat leaf rusts in 1926, while those of corn and rye decreased beyond their low points of the year before. Through the five years to which these records apply, the same trend of prevalence has been followed by the leaf rusts of barley and oats in one instance, and of wheat and rye in another ; but corn rust has had an individual trend, quite distinct from either of ihe two other groups. It is especially noteworthy that in no one year did the prevalence indexes of all move in the same direction. In contrast with the foregoing is the fact, illustrated in Figure 24, that through all of the four years ])revious to 1926 the general trend of destructiveness was downward, ranging from severe attacks in 1922 and 1923 to mild and insignificant attacks in 192J: and 1925. With the exception of wheat leaf rust, a general upward trend is apparent for 1926. Pig. 24. Destuuctivrness of the cereal leaf rusts, 1922-1926 The indexes of destructiveness computed from field data and given in tables in the text are shown graphically here. Since 1922, the general trend has been downward; but crown rust increased in 1923 and 1926, and barley leaf rust in- creased in 1926. DiSEASKS OF Grain Cnoi-s, 1922-1926 33 The Stem Rust of Cereals Caused by Piicciiiia (/raiiiiius Pers. Although each of the cereal crops grown in Illinois has an in- dividual leaf rust, the stem rust is the same for all, attacking all but corn. Special races of the stem rust exist, however, which are so lim- ited in their ability to attack that each occurs on one crop only. The varied requirements of each crop in length of season and planting and growing periods present diverse conditions for the development of stem rust epidemics on each crop. Hence, it is not likely that the same phe- nomena of prevalence and destructiveness will be exhibited by this rust upon any two crops in a given year. It is necessary, therefore, to treat the stem rust on each crop as though it were a separate rust. The method of measuring the jirevalence of stem rust is similar to that used for the leaf rusts. Instead of individual plants, individual stems are taken as units, and prevalence is computed according to the method described on pp. ti-S. Destructiveness is expressed in terms of stem area occupied by the rust pustules, and the standard shown in Figure '^ is used as described on pages 18 and 19 as a measure of stem area. The STE>r Rust ox Wheat At the time field work was begun in V.)22. stem rust infection was already widely ])revalent on wheat. Its occurrence was demonstrated in the 30 counties shown in Figure '.;">. Although no definite records of ])revalence or severity were secured from the southern territory marked on the map, data were taken from 1' fields in the northern part of the state, and the destructiveness indicated bv them agrees well with fnai shown by the specimens collected in the south. The earliest record of ir.fection in 192;? was made June 4 in lack- srn County. Iso'ated infections such as those subsequently found T^jj^g 7 in Bond County, June 12 in Franklin County, and June H in Hamliton and White Counties served as foci from which, about June 16, disease dissemination began ; and by June 22 the epidemic had become general. Data were secured from 72 fields distributed among the 39 counties shown in Figure 20. The extent of the territory from which the data were secured is iiarticularly fortunate, in view of the intensity of the rust attack. During 1924 stem rust was luore ])revalent in the north than in the south. Data were secured from (il fields parcelled among the 31 counties shown in Figure 27. The original infections took place at least S days later than was the case the year before, and the period of moderate temperatures was so curtailed in the south that only a mild epidemic resulted there. So rare was stem rust during 1925 that only six records of it were obtained. The fields from which these records were taken were located 34 Illinois Natural History Survey Bulletin Figs. 25-29. Sources of stem rust data fob wheat, 1922-1926 The black areas on these maps show the regions from which the data dis- cussed in the text were talien in each year. The fields from which records were taken during the five years contained 3,851 acres. (See page 33.) ' Disi-.ASKs OF Grain Crops. 1922-1926 35 in the counties shown in Figure 28. A recrudescence of stem rust oc- curred in 1920, which, by comparison with the Hght attack of the previous year, gave rise among wheat growers to grave fears of damage, fortunate- ly not reaHzed. During the season data were taken from 27 fields located in the li) counties shown in Figure 29. As in previous years, abundant infection was found less readily in the south than in the north. The variations in prevalence and destructiveness alluded to in the foregoing paragraphs are illustrated in the following table, in which the final figures arrived at in the statistical analyses of the data are brought together. 36 Illinois Natural Histoky Survey Buixetin notes or data of any kind could be secured to show either its relative prev- alence or destructiveness. As more than 1,.'?00 acres of oats were examined without a single stem rust infection being found, the indexes of prevalence and destructiveness may be considered as zero for the year. In contrast with the preceding year, in 1924 oats throughout the State were abundantly infected. The infection was particularly heavy in the northern half of the State. Fifty-eight fields distributed among the 34 counties shown in Figure 31 furnished data on prevalence and destructiveness. Again in 1925, the oats crop was infected throughout the northern half of the State. Data were taken from 8 fields in the seven counties shown in Figure 32. Although apparently inadequate in total acreage, the fields were widely distributed and were representa- tive, to that extent, of the entire oats acreage. Data were taken from 42 fields in 1926, these fields being scattered among the 29 counties shown in Figure 33. They were representative to a remarkable degree of the oat acreage of the State. Stem rust was very abundant, and quite destructive in the northern counties, as com- pared with previous years. The indexes of prevalence and destructiveness for the five years are brought together in the following table. Diseases of Grain Crops. 1922-1926 37 Figs. 30-33. Sovkces of stem bust data for oats The black areas on these maps show the regions from which the data dis- cussed in the text were taljen in 1922, 1924. 1925. and 1926. No infection was found in 1923. though more than 1.300 acres were examined. The fields exam- ined in the other tour years contained a total of 2.201 acres. (See page 36.) 38 Illinois Natural History Survey Bulletin The Stem Rust on Barley Very few records of the occurrence of stem rust on barley have been made in lUinois south of the north third of the State. So far as the annual epidemic on barley is concerned, only the northern third of the State needs to be considered, not only because of the rarity of the rust southward but also because of the scarcity of the crop there. Because of the limited range of the crop and the limited acreage devoted to it, records of its diseases are much fewer than those for more common and more widely grown cereals. In 1922 infection was found in the five counties shown in Figure 34. Data were obtained from eight fields which were representative of the barley district. Stem rust appeared to be completely absent in 1923, but in 1924 the infection became very general in the north. During the latter season, one barley field was examined in each of the 15 counties shown in Figure 35. In 1925, fields were examined in the four counties shown in Figure 36, and stem rust was found to be prevalent, though in a very light form, in each. By the time the observer reached the barley region in 1926 the crop had been cut and was standing in shocks. Although examinations made under these conditions are not wholly satisfactory, one field in each of the three counties, Boone, DeKalb, and McHenry, was given a very careful examination ; and the data taken from them may be considered typical of the stem rust infection in the barley region. The indexes of prevalence and destructiveness secured from the computations made with field data are as follows : The figures given above, which are reliable indexes of the serious- ness of stem rust on barley during each of the five years 1922-1926, indicate definitely the relative importance, in each year and throughout the period, of this rust and the leaf rust (see page 26) as agents in reducing the barley yield, stem rust being in the main much the less significant. Diseases of Grain Crops, 1922-1926 39 Fig. 36 192^ Figs. 34-36. Sources of steji hunt data koh liAiu-EY The region from which data were taken in the years 1922. 1924. and 1925 are shown in black on these maps. In 1923, there was no stem rust infection. In 1926, 3 fields, one in Boone, one in DeKalb, and one in McHenry County, were examined. (See page 38.) 40 Illinois N.\tir.\l Hi.stouy Sirvey Biij.hmK The Stem Rust ox Rye In spite of the wide distribution of rye in Illinois and the ex- tensive use of it along newly ])laced hard roads, which might be ex- pected to aid in distributing stem rust infection on the crop, it has been possible to secure but very little data on stem rust prevalence in rye fields. In 1922 only two fields, one in Mason and one in Grundy County, were found infected. No infection was recorded in 1933. In 1924 data were secured from 10 fields located in the seven counties shown in Figure 37. During the two subsequent years. 1925 and 1926, no instance of infection was seen in any of the fields examined, though Fig. 37. Socrcks of stkm ru.st data for rye in 1924 In the seven counties shown in black, 10 fields containing a total of 70 acres furnished data on the stem rust attack of 1924. infected rye plants were found occasionally in wheat fields and along roadsides. The data on the severity of stem rust on rye for the years covered by the survey are as follows : Diseases of Grain Ckoi's, 1922-1926 41 Comparison of stem rust infections on the cereal crops subject to its attack With stem rust, as with leaf rust, both the degree of prevalence and the degree of destructiveness have varied from year to year. These variations, which have not been of the same relative amounts for each of the cro]is, are best shown graphically. The indexes of prevalence on each of the cereals, for each year from ]!)a2 through lil2(i, are shown in Figure :iS. i\ common trend is siiown by all except wheat. In 1!)22 stem rust was moderately prev- alent on Ijarley, rye, and oats, but only mildly so on wheat. The fol- lowing year prevalence dropped i)ractically to zero on the first three crops, but on wheat it increased considerably. The crop season of 1924 saw a marked increase on all the cereals except wheat, while in 1925 prevalence was less on oats and rye, but more on wheat and barley, Fig. 38. Puevai.ecnce of stem uust on the cereai, cKors, 1922-1926 The indexes of prevalence computed from field data and given in tables in the text are shown graphically here. There is a greater uniformity of prev- alence trend from year to year on all crops than was true with the leaf rusts (compare with Fig. 23); but in contrast with the prevalence of the leaf rusts, stem rust reached its highest point in 1926. 42 Illinois Natural History Survey Buixetin than the year before. In 1926 it increased markedly on all except rye, reaching the greatest degree recorded during the survey. It is evident from an inspection of Figure 3S that the climatic conditions of any one year have not aflfected the prevalence of stem rust to the same extent, or in the same way, on every crop. The intensities of attack, as indicated by the indexes of destruc- tiveness calculated from the field data, are shown in Figure 39. Prac- tically the same tendencies are exhibited on all the crops, wheat alone showing a marked divergence. In 1924, when the trend of destruc- tiveness was upward on the three other crops, it declined very markedly on wheat. There has been a distinct parallel between prevalence and destruc- tiveness throughout the five seasons, the only break in it being the at- tack on wheat in 1923, in which year the increase in prevalence over the previous season was accompanied by a large decrease in destructiveness Fig. 39. Destkuctiveness of stem rust ox the cereal crops, 1922-1926 The indexes of destructiveness given in tables in the text are shown graph- ically here. The trend of destructiveness from year to year follows very closely the trend of prevalence (compare with Fig. 38), being greatest In 1926 and 1922, and generally very low in the other years. Diseases of Giwin Croi-s, 1922-1926 43 CEREAL SMUTS All of the cereal crops are subject to diseases classed under the general name, smut ; but the smuts, like rusts, are different on the dif- ferent crops, exhibiting different symptoms and being possessed of dif- ferent means of spread and infection. On wheat there are three; on barley, two ; on oats, two ; on rye, two ; and on corn, one. According to their appearance, they can be classed as "loose", "covered", and "stripe" smuts. In general, "covered" smuts are much less abundant than "loose" or "stripe" smuts. The covered smuts of oats and bar- ley, because of their rarity, have not furnished sufificient data to de- serve inclusion in this discussion ; and the limited range of the "stripe", or "flag", smuts on wheat and rye precludes their inclusion. The outstanding characteristic of all the smuts, except the one on corn, is that they nearly always destroy completely, either directly or indirectly, the head or the grain. The loose smuts replace the grain and the floral glumes with a mass of smutty powder, the covered smuts replace the grain and often the floral glumes as well, and the stripe smuts produce such malformations of the leaf and head that diseased stems rarely are productive. As a result, the prevalence of a smut is also a direct measure of its destructiveness. Wheat Loose Smut Caused by Ustilago tritici (Pers.) Rostr. This is the least important of the common cereal smuts. Though distributed throughout the wheat fields of the State and conspicuous enough to be classed among the common diseases, its means of ac- complishing infection, which is limited to the very short time the wheat flowers are open, prohibits it from becoming abtmdant, at least under Illinois conditions. For the same reason, the extreme fluctuations ex- hibited by other diseases will not be found with it. During 1922, loose smut was seen in wheat fields located in the 29 counties shown in Figure 40. Wheat harvest was already so far advanced when the examinations were made that only the northern fields furnished data. Data secured in 1923 from 81 fields located in the ''is counties shown in Figure 41 were representative of the loose smut infection of the season, not merely because of the large wheat-acreage from which they were taken but also because of their extensive distribution and their location in regions both of intensive and less intensive wheat culture. The most abvmdant data obtained in any of the five years of the survey were procured in 1924. In the 58 counties shown in Figure 42, 145 wheat fields aggregating 5,661 acres were examined. The wheat grown in the counties represented by the field records aggregates 1,7'54.S35 44 Illinois Natural Histoky Survey Bulletin Figs. 40-44. Sourcks of data for the loose smut of whkat, 1922-1926 The black areas on these maps show the regions from which data were taken In each year of the period. The fields examined in these five years contained a total of 16,425 acres. (See page 43.) Di.SK.\sKS OF Grain Cuoi-.s, 1922-1920 45 acres, which is shglitly mure than W) per cent of the State's acreage for that year. Data were obtained in 1!)2.5 from 51 fiehls located in the 2G counties shown in Figure 4.'!. Although the acreage they contained was only about one third as great as in the previous year, the total county acreage of which they were representative constituted nearly "ill i)er cent of the State's acreage. In l!)2(i. data were obtained by examining 4S wheat fields located in the 3!) counties shown in Figure 44. Though somewhat less repre- sentative with respect to size of acreage and wideness of distribution than the records of the previous two years, the increased care employed in obtaining them makes them the most dependable of all the e.timates obtained. During these five seasons, data have been taken from :)[U wheat fields containing a total of i;],42.5 acres. The indexes of jirevalence obtained from the analysis of the data compare as follows: 46 Illinois Natural History Survey Bulletin few days before harvest. The fields in which it is found are, there- fore, both small in number and not representative in distribution, and the data obtained from them are, very probably, understatements of the amount of infection. The only means of supplementing the field counts is to obtain reports from wheat dealers on the marketing of smut in- fested wheat. In analyzing data on bunt prevalence, the procedure is not quite the same as for other diseases. The smallness of the amount of data secured from fields, together with their limited distribution, renders them incapable of complete analysis. As a rule the only figure which can be computed is an index of prevalence in the acreage known to have been infested. It can be considered only as typifying the amount of in- fection prevailing in the crop found b^ other means to have been in- fected. The most useful data on bunt prevalence have been obtained in re- ports from grain dealers. Circular letters were sent annually to the dealers in the State, requesting reports of the number of bushels pur- chased and offered that were smutted ; and the replies received, which have been morr numerous than is usual with circulars, are believed to be especially dependable because the dealers were asked to reply only when such lots had been offered them. The data secured in this way, being in terms of bushels rather than acres, require a treatment somewhat different from that accorded acreage statistics. The following short section taken from the analysis of the 1922 reports shows the method. Sheet County Bushels reported as infested County yield Percentage of the county yield infested lAdams I [Champaign iChampaign Champaign IChampaign jChampaign I I I IDouglas JDouglas 3;P IPiatt 31 IPiatt I 1 I 32 IPutnam 2.000 622 2,622 1,500 1,500 2,000 500 2,500 2,000 1,011,030 518,315 392,632 581,600 203,381 .506 .382 .430 .983 I Total 8,622 2,706,958 .318 The dealers' reports are assigned places in the permanent file with other disease records. These for each year are kept in separate folders. Diseases of Grain Crops, 1922-1926 47 the sheets being assorted by counties as are the other reports, and the sheets of each folder numbered serially. In the calculations made with the data, only one tabulation is made. In the first column, the number of the report sheet is entered ; in the second column, the coimty from which it came: and in the third column, the number of bushels of smut- infested wheat. When this is comi)leted, the number of bushels in- fested is obtained for each county and for all the reports by summing the items in column 3. In column 4, o])posite each county total, the county yield given in the agricultural statistical reports already mentioned is entered. The total yield of all the counties concerned in the report is the sum of these entries. By dividing the county yields by the county totals of infested wheat, a percentage, shown in column 5, is obtained which re])resents the proportion of smut infestation for each county ; and by dividing the sum for column .'i by that for column 4 a per- centage is obtained which represents the average i)roportion of smut infestation for all the counties in the tabulation. The percentages are carried to the third place beyond the decimal for the reason that the per- centages dealt with in this disease are always very small and the ma.ximum of accuracy is desirable. No further manipulation of these data is permissible. They must be considered as representative of the entire territory from which they came, and the complete lack of data from other territories prohibits further calculations. The figures secured by this method are empirical in that they rep- resent prevalence less directly than definite field data. A more definite conception of prevalence can be obtained if the results of field examina- tions and of dealers' reports are taken together. For example, dealers' reports show that, in 1!)32, .3T1 per cent of the State's yield was in- fested, the fields in which the infestation occurred having, according to the results of field examinations, 1.9 per cent of the heads bunt-in- fected. During 1922 data on held prevalence were secured in six counties. These, together with dealers' reports, gave information from the 17 counties shown in Figure 45. The infestation present in 30. S per cent of the State's yield of .55,432,000 bushels is represented. The field data secured in 1923 are the most extensive obtained dur- ing the survey. Definite records were taken in 40 fields located in 24 widely distributed counties: but the acreage of these fields—2123 acres in all—is less than .2 per cent of the total acreage in the counties in which they were located, emphasizing again the lack of general prev- alence. Grain dealers' reports received from 48 counties, give data on 64.7 per cent of the yield of the State. The counties in which bunt was shown to have occurred, by field examinations and dealers' reports, number 35 and have the distribution shown in Figure 46. From the two sets of data, it may be inferred that fields furnishing .317 per cent of the State's yield were bunt-infested to the average extent of 2.5 per cent. Field examinations made in 1924 revealed the presence of bunt in 14 widely scattered counties Reports for this year were received from 48 Illinois Natural History Sirvky Bili.ktin Fig. 48 7925 Fig. 49 7926 Pigs. 45-49. Sources of wheat bunt data, 1922-1926 The black areas on these maps show the territory from which data were talien in each year of the period. The data for this disease are obtained in part by direct examinations of fields, and in part from the reports supplied by grain dealers. During the five years of survey, data were taken from fields aggregating 2,785 acres. The dealers' reports applied to from 30.8 per cent to 64.7 per cent of the State's annual yields. (See pp. 45-47.) J BiSEASES UK Grain Crops, 1922-1926 49 grain dealers in 46 counties. The total distribution of hunt included the 34 counties shown in Figure 47. The lack of adequate assistance in the survey during the summer of 1925 made it inadvisable to circularize the grain dealers, but data were secured by carefully examining seven bunt-infested fields in the seven widely scattered counties indicated in Figure 48. Although the number of fields is small and their acreage almost negligible, their wide distril)ution and the variation in degree of infestation which they exhibited justify considering the data as representative. Field examinations showing bunt prevalence in 1926 were limited, including five fields, all of which were in four centrally located counties. However, grain dealers' reports were abundant. Responses to inquiries were received from 193 purchasing stations, the crop directly repre- sented by these reports constituting a very large percentage of the State's yield,. The range of bunt infection for the year, compiled from dealers' reports and field examinations, is indicated in Figure 49, which shows this disease to have occurred in 44 counties. Summaries of the data showing field prevalence and the infes- tation of harvested grain are given in the table below. 50 Illinois Natural Histokv Survky Billktin Fig. 53 192S Fig. 54 1926 Figs. 50-54. Sources of data for oats loose smut, 1922-1926 The black areas ou these maps show the territory from which data were taken in each year. In all. fields totaling 7,667 acres were examined, the small- est acreage (723) being examined in 1922, and the largest (2,412 acres) in 1924. (See page 49.) Diseases of Grain Crops, 1922-1926 51 shown in Figure 51. The extent of the regions from which the records were taken make them much more representative than those of 1922. An acreage still larger than that of 1923 furnished data on oats smut prevalence in 1924. In all, 135 fields containing 2,412 acres were examined. They had the distribution shown in Figure 52, which in- cludes ")? counties so widely scattered as to be typic:d of the entire state. The acreage in these cotmties constitutes 04. T per cent of the State's acreage. The acreage examined for smut in llt25, though less than that of 1924, was close to 2000. In the 4.S counties shown in Figure 53. data were taken from 104 fields, which may be considered typical of the in- fection throughout the state even though a considerable territory in the northwest is not represented directly. In 1926, 41 oats fields were examined for smut prevalence. Though their total acreage was small when compared with previous years, they were widelv distributed and were directly representative of the oats crop grown in the 32 counties shown in Figure 54. These counties are so distributed as to be typical of the entire state. The indexes of prevalence, which in these cases are also indexes of destruct.iveness, obtained from the field data are given below for each of the 5 years. Year .\creage examined Prevalence; Calculated percentage of smut in- fected oats heads For the acreage examined For the county For the acreage represented State's acreage 1922 1 52 Illinois Natural History Survey Bulletin Fig. 55 1923 Fig. 56 1924 Fig. 57 1925 Fig. 58 7926 FiGE. 55-58. Sources of data fob barley loose smut, 192S-1926 The black areas on these maps show the territory from which data were taken In each of the four years. The acreage furnishing the records was not large, in comparison with other cereals, but the concentration of barley in the northern part of the State makes small acreages quite representative. (See page 16.) Di.SEA.SES OF Gkain Ckop.s, 1922-1926 53 More abundant data were secured in 1924. Barley fields—32 in number—scattered among the 17 counties shown in Figure 56 were given careful examination. All but two were located in the northern barley district, their wide distribution among the northern counties re- sulting in a typical sampling of smut prevalence. In li)2.3, records were taken from only ."i fields. They were located in the five counties shown in Figure 57. The seven fields examined in ]y2() were located in the seven counties shown in Figure 58. Al- though not so widely scattered as the fields examined in previous years, the uniformity of the data secured from them indicates a typical repre- sentation of the northern infection. The ])revalence of barley loose smut, as shown by calculations made from field data, is shown below, year by year. 54 Illinois Natural History Survey Bulletin Fig. 62 1925 Fig. 63 1926 Figs. 59-63. Sources op corn smut data, 1922-1926 The black areas on these maps show the territory from which data were taken in each of tlie five years. The prevalence of smut was recorded, during these years, in 244 fields wh.ch contained a total of 8,866 acres. (See page 53.) Diseases of Grain Crops, 19.';2-1926 55 they contained represented directly 6.16-1.400 acres of corn—68.9 per cent of the State's acreage. Twenty-four fields, distributed among the 21 counties shown in Fig- ure 62. were examined in 1925. The data secured from them are not as typical as might be desired, for the reason that the large, centrally located corn region including McLean County is not represented. However, the universal occurrence of smut and the evident similarity of infections in widely separated regions both lend authenticity to the data. Records of prevalence were taken in 1926 from 84 fields distributed among the 70 counties shown in Figure 63. The data secured are repre- sentative, not merely because of the extensiveness of the fields, which contained 1.5T4 acres, nor because of their very wide distribution, but es- pecially because greater care was used than in previous years. The indexes of prevalence for corn smut, determined from the field data, are as follows : Year Acreage examined Prevalence: Calculated percentage of infected stalks For the acreag examined For the county ] For the acreage represented ! State's acreage 1922 56 Illinois Natural History Sirvky Bullktin prevalence of the two first named is in very decided contrast with the low prevalence of the last two. A similar distinct contrast is exhibited in the years 1925 and 1926, the loose smuts of both wheat and barley showing decided increases in prevalence, the loose smut of oats and wheat bunt showing decreases. The curve for com smut conforms to no other. Fig. 64. Prevalence op cereal smuts, 1922-1926 The year-to-year trend of prevalence for the entire period covered by the data is not the same for any ivro of these smuts. The closest similarity is exhibited by oats smut and wheat bunt, which followed the same trend in 1924, 1925, and 1926. Another similarity is shown by the trends of the loose smuts of wheat and barley from 1923 to 1926, but the large variations of the barley smut contrast sharply with the small variations of the wheat smut. Corn smut prevalence appears to be completely independent of the others. SCAB OF CEREALS Caused by GibbcrcUa saubiiictii (Mont.) Sacc. As considered here, the disease commonly known as scab will be lim- ited entirely to the effects its causal fungus produces upon the heads or spikes of the crops which it attacks. Measurement of its prevalence is accomplished in the same way as for other diseases of cereals, and its destructiveness, which might be termed severity with more propriety, is determined by exact counts of the diseased spikelets occurring on dis- eased heads ; these heads, selected at random as the prevalence counts are made, are examined and counted later with care. Diseases or Grain Crops. 1922-1926 57 ScAii o.\ Wheat Though probably more widely distributed than our records show, scab infection of wheat had such limited prevalence in 1!)"^"2 that records were taken from only 9 fields. Located in the 4 counties shown in Figure 65, they contained a total of 193 acres and, as sam])le fields, represented directly the scab prevalence in 212,1)00 acres of wheat. The data for this year show ])revalence only, no special counts having been made to show destructiveness. Records were taken from a much larger number of fields in 1923, and the territory represented by them was also much more extensive. In this year, as in 1923, the data obtained showed prevalence only. In all, .5(5 fields located in the 26 counties shown in Figure (56 and including a total of 2,1 (if acres were subjected to examination. The total county wheat acreage, of which they were samples, constituted 35. :! ])er cent of the State's acreage. The season of 192-1 gave the most abundant data, 146 fields located in the 59 counties shown in Figure 67 and containing a total of 3,166 acres being examined. It is unfortunate that, in this year of unusual prev- alence, data bearing upon the destructiveness of the disease could not be secured. The 1,260,400 acres of wheat grown in the area shown in black in I'igure 66, of which the sample fields were typical, constituted 54.6 per cent of the State's wheat acreage. A very much smaller amount of data was secured in 1925. ]irimarily because of the very late apj^earance of infections but also because the additional task of taking data on destructiveness required more time in each field. Only seven fields yielded data. They contained a total of 125 acres and were located in the ~ counties shown in Figure 68. A total of 173.010 acres of wheat— 1. 7 per cent of the State's acreage—was grown in these counties. There was an increase in scalj infection on wheat in 1926 which re- sulted in data being taken from 19 fields located in and representing the wheat acreage of the 14 counties shown in Figure 69. The data show both the prevalence of the disease and the destructiveness of the attack in a total of 435 acres. The data taken from wheat fields to show the ijrevalence and de- structiveness of scab, are compared year by year in the following table: 58 Illinois Natural History Survey Bulletin Figs. 65-69. Source.s of data for scajb on whf_\t, 1922-1926 The black areas on these maps show the regions from which scab data were taken by direct examination of wheat fields in each year. A very large number of fields were examined, data being taken directly from a total of 6,083 acres (see page 57.) Diseases of Grain Crops. 1922-1926 59 Scab ox Oats Although large acreages of oats have been examined in each of the years of the survey, but few data have been accumulated on scab in- fection. During 1922 one infested field was found in Stephenson County, and in ]!I23 an infested field was found in Logan Countv. In 1924 scab was sufficiently prevalent to show, on the basis of the data taken from 5 fields located in the four counties shown in Figure TO. an average preva- lence of .To per cent and a destructiveness too small to be detemiined. The fields furnishing these estimates contained an even hundred acres. In both of the two seasons. 1925 and 1926. scab was so rare on oats that it was neither collected nor recorded. The prevalence may be stated for the five years as follows : Tear 1922 1923 1924 1925 1926 .\cres examined 723 1,695 100 1,967 870 Percentage of panicles infected Trace Trace .75 Pig. 70. Soirck of the i)at.\ for SCAB ox OATS, 1924 The black areas on this map show the regions from which scab data were taken by direct examina- tion of oats fields in 1924. In the 5.355 acres of oats examined during the period 1922-1926. scab was prac- tically absent, except in 1924. Fig. 71. Soirck of the uata for SCAB ox BARLEV. 1924 Although a considerable acre- age of barley was examined in each year of the period 1922-1926. data on scab were obtained only in 1924. In the other years infection was practically absent. Scab ox Barlev As was the case with oats, infection has been very light on barley during the years of the survey. Only in 1924 was it possible to secure positive data on its prevalence. In this season. 30 tields were examined. Thev were located in the 20 counties shown in Figure '1 and contained 60 Illinois Natural History Survky Bulletin a total of 392 acres. Their location in the northern counties of the State makes them typical of the barley region. In fact, the 209,780 acres of barley which they represented directly, were slightly more than 93 per cent of the State's acreage. In these fields, the data indicate a prevalence equal to 13.43 per cent ; in the acreage directly represented, they indicate an average prevalence of 12.77 per cent; and in the State, of 12.86 per cent. During 1925 and 1926 scab was so rare in barley fields that no rec- ords of it could be taken. Scab on Rye In contrast with the infections on oats and barley, scab on rye was most abundant in 1923, though the attack was too mild to measure. In the seasons of 1922, 1924, 1925, and 1926, infected rye was found only as mixtures in wheat fields. Summary of data on cereal scab The data obtained in field examinations, though much less voluminous than those for other cereal diseases, indicate the variations in prevalence among the individual crops from 1922 through 1926 and the intensity of the attacks in 1925 and 1926. The summaries of prevalence only, arrived at by methods previously outlined and given above for each of the crops, are compared in Figure 72. Besides indicating that, of the four cereals commonly attacked by scab, wheat and barley alone have had serious epidemics. Figure 72 Fig. 72. Prevalence of scab on cereals, 1922-1926 The prevalence indexes for scab calculated from field data and given in the text are shown graphically here. Wheat and barley have had serious infec- tions, and wheat alone is diseased every year. Diseases of Grain Crops, 1922-1926 61 shows that scab, like other diseases, finds conditions in a given year more favorable for spread on one crop than on another. Thus, in one season it may be very prevalent on wheat but practically absent from the other crops. Wheat appears to be the only crop consistently susceptible. There was an increase in prevalence on oats and barley in 1!)54; but. in contrast, the prevalence on wheat diminished. The slight rise in prevalence on rye in 11123 was accompanied by an increase on wheat. Except for these two coincidences, the tendency toward an increase in scab prevalence on one crop appears entirely independent of that tendency on another. LEAF SPOT AND LEAF STRIPE DISEASES Besides the common and conspicuous diseases just discussed, there is in Illinois at least one disease of each of the cereals, excepting corn. which attacks its host by entering the leaves, damaging the plants, and reducing yields by killing i)arts of the leaves. Barley is subject to a leaf stripe which not only results in a reduction in the amount of leaf but also generally kills the stem it attacks. Wheat is affected by "speckled" leaf blotch, which shows first as small spots but later kills large parts of the leaf. Rye is subject to a leaf spot much like the wheat leaf spot, but its occurrence in Illinois is very rare. B.\ULEV Stripe Caused by Hclminthosporitim graiiiiiu-iiiii Rabh. Stripe ]irevalence ajid stripe destructiveness are jiractically equivalent, for diseased stems die early, rarely maturing any grain. In obtaining data for this disease it is necessary, therefore, only to determine preva- lence. No data were taken in 192"^ and 192;); but the lack of data from the many fields examined during these years should not be construed as showing the disease to have been absent. In the summer of 1924 data were taken from 43 fields containing (!83 acres. They were distributed among the 21 counties included in the black area of Figure 73 and, as samples of the 215,275 acres of bar- ley grown in the area shown, were directly representative of 95.7 per cent of the State's acreage. The records for 1925 were taken from five fields only, one in each of the five counties shown on Figure 74. Though containing a total of only SO acres, they were representative of 5(5.320 acres of barley—16.2 per cent of the State's acreage. In ]92() data were taken from fi fields located in the fi counties shown in Figure 75. These fields, which contained 120 acres, were located mainly in the northeast corner of the state, in a region devoted to barley production, and may be considered as typifying the prevalence of stripe infection. 62 Illinois Natural History Survey Bulletin Fig. 75 1926 FiGE. 73-75. Sources of barley stripe data, 1924-1926 The black areas on these maps show the regions from which data were taken by direct examination of barley fields in each year. No data were taken in 1922 and 1923. but the fields examined in 1924 were representative of 95 per cent, and in 1925 of 46 per cent of the State's acreage (see page 61). Diseases of Grain Crops. 1922-1926 63 The data accumulated show, when subjected to anal\-sis, the follow- ing degrees of prevalence and destructiveness. 64 Illinois Natitral History Sirvev Buixktin to do so involves personal judgment to an extent certain to result in gross error. Instead, each part of the sample is considered as falling in the class represented by the scale diagram having the lower value of the two between which it falls. A record is made on the field data sheet of the number of parts of the sample falling in each class, and the average amount of leaf spot injury shown by the sample is determined in the manner described on page 1 9. An analysis of the data on prevalence and destructiveness then is made in the manner described in the first part of this paper ( see pages 8-U). Fig. 76. Standard for mea.suring the destructiv'eness of the speckled leaf spot op wheat The black areas on each diagrammatic leaf represent the tissue killed by the leaf spot fungus. The area of the black spots has been measured and is given as a percentage of the entire leaf area for each diagram. Random samples taken in each field are compared with this scale, and the destructiveness of the disease is computed in terms of leaf area destroyed (see page 63). DisKASKS OF Grain Crops, 1922-1926 65 Figs. 77-79. Siiritn-:: Il.^lA FOR rilK Sl'KCKI.Kl) I.EAF SI'OT OK WHEAT IN 1923. 1924, AND 1926 The black areas on these maps show the regions from which data were taken by direct examination of wlieat fields in each year. During tlie period 1922-1926 fields containing a total of 12.7S7 acres were examined for this dis- ease, 862 acres showing none of it in 1922 and 3,225 acres having only a trace in 1925, Illinois Natural Histokv Survey Buhetin Due in a large measure, perhaps, to the exceedingly heavy attack of leaf rust in 1!)22, leaf spot appeared practically absent from wheat fields. No positive data on its prevalence were secured that year. In 1923, however, a considerable quantity was obtained from 41 fields located in the 24 counties shown in Figure 77. The fields examined contained 2,428 acres and were representative of 40.3 per cent the State's wheat acreage. The season of 1924 furnished the largest accumulation of leaf spot data of any of the seasons covered by the survey. Examinations were made in 149 fields distributed in somewhat varying numbers among 53 counties. They furnished a representative sampling of leaf spot preva- lence and intensity for the region shown in black in Figure 78. The fields themselves contained 5,667 acres, while the region which they represent grew 1,389,975 acres—60.3 per cent of the State's 2,307,000 acres of wheat. Speckled leaf spot was rare in 1925. In 3,225 acres examined, only four instances of infection were found, one each in Clark, Crawford, Pulaski, and St. Clair counties. Though sufficient to give samples for laboratory diagnosis, they were so light that they could not be estimated, either in prevalence or severity. The infection for 1925 is to be con- sidered, therefore, as only a trace. In 1926, leaf spot was prevalent in fields throughout the central and southern parts of the state, but was not found in the northern third. Rec- ords of prevalence and severity were taken from 23 fields, which contained 605 acres and were located in the 14 counties shown in Figure 79. The data obtained were representative only of the infection in the southern half of the state, and the prevalence and intensity of the attack, expressed respectively by indexes of 11.0 and .39 calculated from these data, will be reduced considerably when 1926 acreage statistics become available for further calculations. The fluctuations in prevalence and intensity of attack, as shown by the field data when subjected to analysis, are as follows : Diseases of Grain Croi's. 1922-1926 67 Oats Halo Blight Caused by Psciidoiuoiias coronafiicicns ( Elliott) Stev. Though the syniptonis presented by the halo blight of oats are quite (litTerent from those shown by the leaf spot of wheat discussed above, the method of estimating its destructiveness is much the same. Its preva- lence is determined, of course, as the percentage ratio of disease-bearing stalks. For estimating the intensity of the attack, a special scale, shown with considerable reduction in Figure SO, has been devised. The eftect produced by halo blight in destroying leaf tissue appears to be consider- ably less than the similar elTect of the wheat leaf spot. The degrees of destructiveness are more readily distingnishalile. which ijerniits the use Pig. 80. Standard for .mi:asi:ring the destructiveness of oats iiai.o blight The black spots on each diagrammatic leaf represent the tissue killed by the halo blight bacterium. The area of the spots has been measured and is given as a percentage of the entire leaf area for each diagram. Random sam- ples taken in each field are compared with this scale, and the destructiveness of the disease is computed in terms of leaf area destroyed (see page 69). of a larger number of individual standards in the scale and results in a consideiably greater degree of accuracy in the estimates. Otherwise, the methods of obtaining field data are the same as those previously discussed. During l'J22, with halo blight of oats as with several other cereal diseases, data were recorded exclusively in the northern half of the State. Definite records were taken in ") fields which contained a total of ilT acres and were located in the 4 counties shown in Figure 81. Al- though their combined acreage is not impressively large, their wide 68 Illinois Nattrai. History Survkv BrLiETix Fic.Sl 1922 Fig. 82 1<>24 Fig. 83 192S Fig. 84 1926 Figs. 81-84. Sol-rces of data on the halo blight op o.\ts, 1922 and 1924-1926 The black areas on these maps show the regions from which data were secured by direct examination of oats fields in each year. Fields containing a total of 4,198 were examined for this disease during the period 1922-1926, the 1,306 acres seen in 1923 showing only a trace of infection. Diseases of Grai.v Crop.s, 1922-1926 69 separation together with the constancy of the data endows them with a certain degree of dependabiHty. In 1!>23 only two records of halo blight were obtained. -\ small field of o acres in Macoupin County, examined June 20, yielded the informa- tion that 65 per cent of the stems bore leaves having an average of 50 per cent of their area destroyed. In another small field of ;i acres in Fayette County, examined June 12. o per cent of the stalks bore leaves having 17.5 per cent of their area destroyed. In view of these facts, the prevalence and destructiveness of halo blight in 1!)2'3 were too light to be estimated. In TJSi data were taken from T.S fields which contained a total of 1,21(1 acres. They were distributed among the 43 counties which make up the black areas of Figure 82. The oats grown in these areas, 1,906,500 acres, make up 43.5 jier cent of the State's acreage. Agam in 1 1 abundant data were obtained. The fields examined, 56 in numljer, contained 1,1SI) acres and represented the infection pre- vailing in the :>1 counties shown in Figure S3. These counties contained 2,237,510 of the State's 4,'; 24,000 acres devoted to oats production. The data, therefore, apply directly to 47.4 per cent of the State's acreage. Much less al)undant in 1!)2() than in the two preceding seasons, halo blight was, nevertheless, widely jirevalent. Records were taken in the 9 counties included in the black areas of Figure 84. Prevalence and destructiveness. as shown by data obtained by the survey, are given in the following table : 70 Illinois Natihal Histouy Survey Bullktin 1922 Fig. 85. Prevalence of the leap spot and stripe diseases. 1922-1926 1 Diseases of Grain Crop.s, 1922-1926 71 U)2;3, their positions were reversed, the former being low and the latter very high. The summer of 1934 was marked by a sharp decline of speckled leaf spot and a sharp rise of halo blight, the two being very nearly equal in ])revalence and in destructiveness. In 11125 both diseases stood below their marks of the previous years, but the decline of halo blight was slight and that of leaf spot great, leaving the former still serious and the latter insignificant. In contrast to these two, there was a slight rise in the prevalence of barley stripe. Finally, in 1926, halo blight and barley stripe exhibited declines in both prevalence and destructiveness, which contrasted sharply with the increase shown by the s]ieckled leaf spot of wh-eat. SUMMARY OF ALL DATA SHOWING PREVALENCE AND DESTRUCTIVENESS All of the diseases considered in the previous pages have been of a kind which attacks and destroys above-ground parts of the plant. When considered in relation to their etTects, they are of three types: destroy- ers of stem tissue, of leaf tissue, and of heads. One only, the smut of maize, commonly attacks all parts of the plant. Data on at least one of each of these types of disease have been given for each of the cereal crops for nearly every one of the five seasons. In any season an individual plant of any of these crops may bear one, and is likely to bear two or even more, types of disease. This fact is not particularly significant with respect to the prevalence of the several diseases, but it is very important in determining the destructiveness resulting from disease attack. The prevalence indexes previously given for the several types of disease are brought together in the following tables in closer relation to the individual crops. For each crop, the indexes of the several dis- eases are added together to give an index of total prevalence for each year ; and averages are obtained for the entire priod of the survey. This gives a fairly exact comparison of the five seasons. 72 Illinoi.s Natur.vl Histoky Survey Bullktin of 1925 ranks lowest, 1922 next, then 1926 and 1924, and finally 192:i at the peak with a total prevalence of 227.31. The prevalence of oats diseases is given in the following table: Diseases of Grain Crops, 1922-1926 73 Although only three diseases are considered, instead of the five or six diseases of the crops just discussed, similar seasonal variations in prevalence are evident. For corn, the ntunber of diseases treated is of necessity small, and the information secured concerning them, due to conditions imposed, has often been inadequate. Yet the indexes of prevalence obtained show the following comparisons : 74 Illinois Natural History Survey Bulletin heavy or light on all of the five cereals. The closest approach to such a condition occurred in 1925, the total prevalence of disease in that year being, for each of the five crops, less than the year before. Begin- ning with the wide range of total prevalences displayed in the season of 1922, the trend was downward for rye, corn, and oats diseases, biit upward for those of barley and wheat in 1923; down for wheat and rye diseases but up for those of oats, barley, and corn in 1924 ; down for the diseases of all five crops in 1925 ; and in 1926 down for corn and rye, but up to relatively high points for wheat, oats, and barley diseases. The average total prevalence over the period of years covered by the survey is remarkably characteristic for each of the five crops. This fact is illustrated in Figure 88, in which the five-year index averages given in the foregoing five tables are shown graphically. The high prev- alence of infection on wheat and oats is indicative of the fact that these crops suffer most from diseases of their above-ground parts. The Oa/j iVheat 40 60 80 /oo ao Preya/ence /nde^ Fig. 88. Average disease prevai,ence for each crop The indexes of prevalence given in the text for the diseases of each crop have been added and averaged in order to show the normal prevalence of dis- eases on each crop in Illinois. lesser prevalence shown for barley and rye result, in part at least, from the limited range of culture of the first crop and from the scattering cultivation accorded the latter. In the case of corn, however, the very low average prevalence may be considered as directly correlated with the relatively minor ranks held by smut and rust' among corn diseases. It is not possible to gain a clear-cut conception of the destructive- ness of diseases through the simple process of adding the indexes, as has been done for prevalence. The types of injury caused by the various Diseases of Grain Citors. 1922-1926 75 diseases are dissimilar and. therefore, re(|uire a more complicated treat- ment. From the data previously given to show the comparative intensity of attack, those relating to wheat are selected and shown together in the following table : 76 Illinois Natural History Survey Bulletin Fig. 89. Destructiveness of wheat diseases, 1922-1926 The effect of disease attack upon the stems, leaves, and heads of wheat, as indicated by the indexes of destructiveness given in the text, is shown for each year of the period 1922-1926 in comparison with an average healthy stalk. Standard 1921 1925 1924 I92'5 1926 Fig. 90. Destructiveness of oats diseases, 1922-1926 The effect of disease attack upon the stems, leaves, and panicles of oats, as indicated by the indexes of destructiveness given in the text, is shown in com- parison with an average healthy stalk. See the text on page 75 for further explanation. Diseases of Grain Crops, 1922-1926 77 The destructiveness of the diseases of oats, barley, and rye may be regarded in the same way as the diseases of wheat. A statement of their destructiveness may be made, consequently, by omitting summarizing tab- ulations like that given above for wheat diseases and using, in their stead, illustrations similar to F"igure SI). In Figure HO, the total effect of disease attack on oats, as indi- cated by the indexes of destructiveness given in previous pages, is illustrated for each year in comjiarison with an undiseased stalk of aver- age jiroportions. The stature of the stalk is shown reduced 29 per cent by stem rust attack in 1922, not at all in I'.ri'A. 1:3.8 per cent in 1924, 4.5 per cent in 1925, and 42.? per cent in 1926; the normal leaf complement is illustrated as being reduced 40.9 per cent by the com- bined attacks of crown rust and halo blight in 1922, 30.6 per cent in 192;i, 19.2 per cent in 1924, 6.4 per cent in 1925, and 9.9 per cent in 1926; and the sjjikelets are depicted as reduced 5.62 per cent by scab and loose smut in 1922, ;i.arent agreement often is that weather has had a dominating influence. This a])i)lies ])articularly to diseases which attack ])lants above the ground ; for a number of diseases that are soil borne or attack the parts of plants that are in the soil have been subjected to extensive investigation in the laboratories of the Wisconsin Agricultural Experiment Station*, where many nf their relations to temperature and soil moisture have been as- certained. lUit, with diseases attacking as well as spreading to and from aerial plant parts, the conditions which govern infection are chiefly those of the atmosphere, though it nuist be recognized that concurrent soil conditions may so hasten or delay the plant's growth that even un- der ideal atmosjiheric conditions heavy infection does not nccur. The difflculties which attend the artificial duplication of out-of- door weather in the limited space of a laboratory demand that the main facts of the disease-and-weather relation shall be determined by observ- ing the two as they occur naturally. The conclusions arrived at may be subjected later to experimental verification and refinement ; but the ]5rinciples worthy of practical application to the jiroblems of the farm certainly will be those derived by observing natural phenomena. It is not within the scope of this jxqx'r to define exacth' anv of these principles, for the chief intention has been to describe ways of measuring disease phenomena; but a small contribution mav be luade by showing some of the ways in which disease data mav be correlated * Jones, L. R., ,Janies Juhiisun, .ind James G. Dit-kson. Wist-onsin studies upon the relation of soil temperature to plant disease. Wis. Agr. E.\p. Sta. Research Bull. 71. 1926. 80 Illinois Nati'ual Hihtouy Sikvky ErLL:.;T:N with weather. xA.s this material is presented mainly by way of example, it will be limited to the leaf rust of wheat, leaving the fuller treatment of it and other diseases to later papers. RELATION TO MEAN ANNUAL TEMPERATURE AND TOTAL ANNUAL RAINFALL The yearly history of wheat leaf rust may be considered to extend from about the first of July to the end of the following June. During this time, there are various periods of rest and active growth, the sum of which is expressed in the degree both of prevalence and destructive- ness reached by the rust at harvest. There is, first, a period of meagre existence in shocks, on stubble, in straw stacks, and on volunteer plants. Following it there often is a short period of very rapid multiplication u])on the young, fall-sown wheat, which is terminated by the arrival of the winter months with their low temperatures. Finally, there is the spring and early summer, when the rust that was overwintered, aided by an accretion of wind-blown infection from the south, develops with exceeding swiftness, step by step with the wheat, to harvest. The success with which the rust-producing organism passes these various periods is determined by many factors, including temperature, rainfall, humidity, wind, and light, each of which is changing continually and independently. While varying greatly in individual seasons, the most important of these factors—temperature and rainfall—vary remark- ably little, as a rule, from year to year. The normal yearly mean tem- perature for Illinois, according to weather records extending back to 1878, is 52°, and the average yearly rainfall is 36.42 inches. During the years covered by the disease survey reported here (1922-1926), the mean temperature of a year has never been more than 2.3° above nor more than 1.4° below the normal mean, and the rainfall of a year has never been more than 1.57 inches above nor more than 3.59 inches below the normal total. Striking correlations ought to exist, therefore, between very small annual departures and the large annual variations in rust attack. The indexes of wheat leaf rust are given in comparison with the mean tem- perature and total rainfall for each rust year, that is for the period July through June rather than January through December, in the fol- lowing table : Year DiSEASKS OF GuAi.N- Ciioi's, 1922-1926 81 No complete parallels exist between either i)hase of the rust at- tack and mean temperature or total rainfall for year-long periods, but they are observable when temperature and rainfall are considered to- j^^ether. In 1923, when the rust indexes were highest, both mean tem- perature and total rainfall were highest: while in li)23 the rust indexes, the mean temperature, and the total rainfall were distinctly lower. The destructiveness index was greater in li)2.i than in 1!>2G, and the combina- tions of mean temperature and total rainfall for these two years stood in the same relation. These facts, and the deductions to be drawn from them, are shown more clearly by diagram in Figure !io. The horizontal scale represents inches of total annual rainfall; the vertical scale, degrees of mean annual tem]ierature. The heavily-lined axes are drawn from the points of aver- age mean annual temperature and average total annual rainfall for Illinois. The points labeled A. B, C, D, and E are the combinations of temperature and rainfall listed in the preceding table for the consectitive years ]!t22-I!)2(). In the U]iper half of the diagram, a trend of rust increase from points B and D to point ,\ is indicated bv a dotted line : and in the lower half a similar trend is indicated by the dotted line drawn through points E and C. These two lines are nearly parallel. Pa/nfa// 'n /nches 30 3S 40 45 54 53 sz 82 Iii.ixois Naukai. History Sxrvf.y Bvi.i.i.tin running upward at an angle very close to 45°. The trends which they indicate suggest strongly that rust destructiveness increases when both the mean annual temperature and the total annual rainfall increase. Although these lines are determined by a very small number of points, the fact that they run ujnvard at an angle so close to -io° is strongly indicative of a certainty of correlation between increasingly heavy rust attacks and combinations of increasingly high mean annual temperatures and totals of annual rainfall. If it is assumed that this correlation is perfect, the line F-G may be drawn at a 45° angle through the intersection of the temperature and rainfall axes. Starting at the lower end, it should pass through points representing combinations of temperature and rainfall which would result in rust attacks with indexes rising steadily from to lOO. Certain of the. indexes theoretically included in this line are shown, however, actually to have been produced by temperature-rainfall combina- tions not crossed by the line, thus suggesting that rust attacks measured by a given index may result from an infinite number of such combina- tions. To determine them exactly requires data for a period of years much greater than is now available ; but a method for determining them approximately is suggested in the following : Point D is so situated that a line drawn from it to the intersection of the axes stands nearly at a right angle to the line F-G and, when continued, cuts the dotted line E-C at H. The point H lies about at the place between E and C where an index of 17 A would be expected from in- spection, and it may be given that value. By rotating either point D or point H about the intersection of the axes in conformity with the known values indicated by points already placed, a hyperbolic line may be drawn through the points D, K, and H, which theoretically runs through all the possible combinations of temi)erature and rainfall resulting in a rtist attack having an index of 17.1. A line drawn from point B to the axis intersection departs at an angle of 55° from the temperature axis. If a line is drawn from the in- tersection, departing from the line F-G at t'-e same angle, to cross the line E-C, it will determine the point L, to which may be assigned the in- dex of point B, 31.3. By rotating either point B or point L as directed above for D, the hyperbola L, M, R is obtained, which runs through all of the combinations of temperature and rainfall capable of producing a rust attack of the index 31.3. From so small a numlaer of points, the proper location of the hyper- bolic lines can be only suggested ; but with the accumulation of data year after year, it will be possible to determine them with great pre- cision for every magnitude of rust attack. Practically, the clelineation of a s^t of hyperbolic bands for predetermined ranges of rust indexes will be more serviceable than hyperbolic lines in predicting intensity of rust attack, for the absolute regularity of the line, which is determined Diseases of Grain Ckoi's, 1922-1926 83 only by temi)erature-rainfall combinations, will be shattered by the oc- currence of varied atmospheric saturation deficits, light intensities, and other minor factors not readily subjected to quantitative measurement. RELATIONS OF DISEASE TO SEASONAL TEMPERATURE AND RAINFALL ll has been i)ointed out ihat the leaf rust of wheat exists under <|uite diverse conditions during different periods of its year. Since these periods correspond to a certain extent, but not wholly, with the seasons of the year, it is necessary to visualize the points of difference and similitude among various years in order to determine the extent and characteristics of the rust seasons. Making use of diagrams constructed according to the method sug- gested by Taylor', the mean temperature and total rainfall for each month of the years included in the survey may be so represented as to be com- pared easily, year with year and month with month. This is done in Figures !il-'.)!), in which the vertical scale represents Fahrenheit degrees of mean monthl\' temi)erature and the horizontal scale, inches of total monthlv rainfall. The combination of mean temperature and total rain- fall occurring in any month is shown by a point placed at the intersec- tion of the [iroper temperature and rainfall lines, and the month is desig- nated by an arable numeral (1=; January, etc.). Figure !I4 shows the progress of a "normal" year in Illinois, beginning with July and ending with the following June, the normal mean annual temperature and the normal total monthly rainfall for the state being indicated by the heavy, dotted lines. Figures !).")-!)!• are diagrams of the years 1!)22-1!)26, in which, to facilitate comparison, the mean annual temperature and total monthly rainfall of a normal year are indicated as in Figure !)4. These diagrams show that no one of the _\ears with which we are dealing ap])roaches a normal year very closely. The outstanding fact illustrated by all of them is that the rust year is divided roughlv into halves, the first of which is characterized by falling, and the second by rising, temperatures. This we may take to be a characteristic of every rust year, to which we should attach only this very broad and obvious significance: From July through December the rust organism is being subjected to increasingly rigorous temperatures and is decidedly on the down-grade, while from January through June the general u]3ward trend of temjierature has an increasingly favorable influence u])On its rate of propagation. Lacking definite names, the terms '"fall trend" and "spring trend" may be given to these ])eriods. When a fall trend or a spring trend in any one of these years is com- pared by means of these diagrams with its counterpart in any other vear, distinct differences are evident. Keeping in mind the variations in rust attack stated previously, the observable differences suggest at once that ' Tine settlement (if trnpieal Australia. By G. Tavlor in Oeog. Rev. Vol. S, pp. .S4-11.5. nils. 84 Illinois Natuhal History Survey Bulletin r ao Disease.^ of Grain Crops. 1922-1926 85 rust abuiKlance is related to the mean temperature of the fall trend and the total rainfall of the spring; trend rather than to the rainfall of the fall trend and the temperature of the spring trend or combinations of the two. The fdllowiiig table emjihasizes these points. 86 Illinois N.vtvr.^l History SriiVEY Buixetin Practically, we are lead to believe that, whenever a fall trend has rain exceeding 15 inches in amount, abundant rust is probable the follow- ing spring, if the mean temperature of the fall is 59°F. or more; but if the temperature is around 56°, light rust is to be expected. A spring trend with a mean temperature of 45° or more may be expected to be productive of a heavy rust epidemic when it follows a fall trend of 50°, if its rainfall is in excess of 17.5 inches, and of a light rust attack when it follows a fall period of 56°, if its rainfall is less than 16.2 inches. Records of the weather in Illinois are available since 1878, making it possible to estimate the frequency with which years favorable and unfavorable to leaf rust may be expected to occur in the future, on the basis of their occurrence during the past 48 years. Favorable, intermed- iate, and unfavorable fall and spring trends, alone and in various com- binations, have occurred the number of times shown in the following table : Combinations of temperature and rainfall, classified witli I'espect to rust development DiSE.\SES OF Grain Crops. 1922-1926 87 s];rin,L;.s iccurring together is approximately 4 in .). Stated in terms of leaf rust attack, it is probable that on an average only one year in twelve will be productive of a severe epidemic and only one year in eight of a very light epidemic, while four year.s in every five should have moderate rust attacks. This statement, however, should not be expected to apply rigidly to any small number of years: it is significant only for a period of half a century or more. Referring again to Figures !)4-i)i), three divisions of the year may be distinguished, including first the months July, August, and September, next tlie months October through March, and finally the months April, May, and June. For convenience, these periods may be designated re- spectively as aestival, hibernal, and vernal, these terms applying not only to periods of the year but also to distinct phases in the rust organism's history. The mean temperatures and total precipitation of each of these periods for the years during which rust observations were made are com- pared in the following table, with the indexes of rust destructiveness. 88 Il.l.INOI.'i N.\TfH.\l. HlSTOIiY SriiVF.Y Bri.I.KTIX to furnish satisfactory conclusions. We might assume, for example, that the influences of temperature and rainfall are cumulative and that their total effect is expressed in the amount of rust eventually produced. On this basis, we could develop a linear equation for each year and, by comparing the equations, arrive at average effects to be assigned to each degree of mean temperature and each inch of rainfall for each of these seasons. Actual tests of this process with the data at hand have yielded results which, though very suggestive, are still too indefinite to be presented, ex- cept in generalizations. They suggest that there is a well defined mini- mum temperature below which rust infection can not occur, and that above this threshold the amount of rust produced increases more and more rapidly as the temperature increases. The months of May and June should be regarded as the most criti- cal of the year in the development of a leaf rust epidemic, for at this time the infections that have survived the rigors of the previous months begin to grow anew, producing new spores and starting the new epidemic. As the days pass, infection is increased not only from the local sources of spore supply but also by spores blown in on the wind from fields to the southward. Even under the most adverse circumstances, sufficient infec- tive material is present in the fields of this State to produce a very de- structive epidemic. Whether it remains mild, as it dicl in 1925 and 1926, or becomes serious, as in 1922, 1923, and 1924, depends very largely up- on the spring temperature and rainfall. A comparison of rust indexes and the weather of May and June is given below. Diseases of Grain Chops, 1922-1926 89 be made clear. The normal mean tenijierature for the May-June period in Illinois, 67.1°, is shown by the heavy horizontal line, and the normal total rainfall, S.ii.s inches, by the heavy vertical line. The point at which these lines cross rejiresents the normal mean temperature and total rain- fall condition for the May-June |ieriod. The actual combinations for the years in question are shown on the diagram, circled dots marking years of heavy rust and circled crosses years of light rust. It is remarkable that the points re])resenting heavy rust lie in a nearlv straight, diagonal line running from the u]iper left to the lower right corner of the diagram, while those for li.ght rust lie close together near the middle of the u])per part of the cold, dry cjuarter. The number of seasons for which there are records are far too few to permit complete acceptance of the relations the diagram suggests ; but in view of the principles underlying mathematical correlation, it is quite probable that the general trend, at least, of the May-June mean temperature-total rainfall combinations likely to produce heavv rust at- tacks is indicated. 90 Illinois Natlrai. Hiktoisv Svrvey Bullkti-x The geographical occurrence of dates of first spring infection, and the temperature relations governing the subsequent development of an epidemic It has been said that there is always sufficient infective material pres- ent to produce a destructive epidemic if conditions are favorable. Al- though, as has just been shown, several distinguishal^le parts of the year, as well as the year as a whole, influence very definitely the intensity of disease attack, the fact remains that each year's epidemic is inde- pendent, in a measure, of all the seasons of the year except the spring, for then spores of the rust organism blown from distant places by strong winds serve to introduce new infection. For this reason, it is worth while to inquire into the time when spring infection is likely to occur and the conditions under which various grades of intensity are developed. The earliest appearance of rust in Illinois, as shown by our survey records, varies considerably according to locality. The late start on field work in 1922 failed to furnish any such records for that year, but in the years 1923, 1924, and 1925 a number of dates of earliest infections were secured. In 1923, five such observations were made; in 1924, 8; and in 1925, 11. Arranged roughly from south to north, they are given by coimties in the following table under the "observed date." DiSKASKS OF Gkaix Chofs, 1922-1926 91 states, essentially, that "nther conditions lieing equal, the variation in the time of occurrence of a given periodical event ... is at the general aver- age rate of 4 days to each 1 degree of latitude. ."> degrees of longitude and 401) feet of altitude, later northward, eastward, and upward in the spring and earlv summer, and the reverse in late summer and autumn." In accordance with this view, the State of Illinois has l)een divided into the sections shown in Figure 102, which represent theoretical time variates of periodical events in harmony with adjacent states. The slant- ing longitudinal lines represent a difference of approximately four-fifths Fio. 102. Theorkticai. ti.me-vakiate lines for Illinoi.s, aimouding to Hopkix's Bioci.iM.\Tic Law The earliest wheat leaf rust infection occurs in the south, and at a given latitude infection occur.s earlier in the west than in the east. The seasonal progress of infection is in the northeasterly direction taken by the longitudinal lines. The horizontal lines represent, from south to north, a difference of one day, and the longitudinal lines a difference of four-fifths of a day, from west to east, in the appearance of first rust infections. When the first infection is observed in the south, the date of its appearance in any more northern region can be predicted accurately. 92 Illinois Natuhai. Histoky Sirvey Bulletin of a day, and the somewhat slanted horizontal lines a difference of ap- proximately one day in the occurrence of a given event. Selecting the first date given in the foregoing table for each vear, it is possible to compute from Figure 102 a theoretical date of first in- fection for the other counties listed. As an example, in 1!)23 rust v/as first seen in Bond County June fifth, but it was not seen in Coles County until June eighth. Referring to F'igure 102, it is readily determined that, by the provisions of the law just stated, rust should have appeared through the southern part of Coles County June seventh and through the northern part June eighth. There is a practical coincidence of ob- served date (June 8) and theoretical date (June 7 and 8). Theoretical dates so determined are given in the preceding table for each of the observed dates. A comparison of the observed and theoretical dates shows in all cases a very close agreement. The statement seems warranted that, when the appearance of rust has been observed in the southern part of the State, an accurate prediction can be made of the date on which initial infections will occur in any part further north. The dates of spring infection influence the total amoimt of disease subsequently produced. If they are early, there are longer periods of multiplication ; but if they are late, the periods of multiplication are shorter. This, however, is probably much less important than the tem- perature and rainfall of the multiplication period, for it is commonly observed that low temperatures inhibit the development of infection, though prolonging somewhat the growing period of wheat, while high temperatures not only inhibit infection but greatly curtail the growing season of the crop and consequently the multiplication period of the dis- ease. As temperature appears to be the controlling factor, it is to be expected that certain average temperatures occurring through periods of given extent should result, other conditions being equal, in definite amounts of infection. These other conditions will not be equal, of course ; but when a large territory is included, the variations they cause should fluctuate about the general trend of the tetnperature effects. The amount of rust observed at a certain time in a certain field may be considered, therefore, to be in the main an expression, somewhat modified by other factors, of the effect of the temperatures accumulated from the time of the initial infection to the time of examination. A statement of the accumulated temperatures may be obtained by adding the daily mean temperatures recorded by the nearest Weather Bureau station. In the following table, these totals as well as the average mean daily temperature during rust development are given in connection with amounts of rust observed in l!)2(i. It should be stated that the first infection was observed June 2 at Paris. From it, theoretical dates of first infection were determined, by the prediction method just described, for instances not definitely observed. Diseases of Gr-^in Crops. 1922-1926 93 OBSKRVKI) Rl ST INFKCTIONS, COIXCIOF.NTALLY ACClMri.ATEI) TOTAL MflVX DALLY Degrees, a.xd Average Meax Daily Temperatires from First I.NFECTiox to Date of Examixatiox 94 Illinois Natukal Histohy Sirvey BrLLETiN of the state mean daily temperatures are close to the optimum. For con- venience, these parts of the state may be designated as lying south of the line numbered 41 in Figure 102, between lines 41 and 43, and north of line 43. The relation between the rapidity of rust development and accumulated temperatures in these districts is illustrated in Figure 103. These differences in rapidity of rust development may be explained further, in view of their geographic locations, on the seasonal character- istics of the region in which they occur. Spring and summer come earlier and endure longer in the south. There, optimum temperatures for rust development may occur before the wheat can be infected readily and may be succeeded by temperatures above the optimum before the epi- demic has progressed appreciably. In the central section, optimum tem- peratures appear somewhat later and endure through much of the grow- ing season. Northward, unfavorably low temperatures, characteristic of Diseases of Graix Crops. 1922-1926 95 the early growing season, are succeeded tor a short period by optimum temperatures as the crop nears maturity, and iiiimediately are followed by the high temperatures of summer. The curves shown in Figure 10-3 suggest that a more complete analy- sis (not possible with the data we now possess) would show that tem- peratures could be classed as effective or non-eftective in the development of an epidemic and that e.xact values could be assigned, in terms of ve- locitv of disease development, to effective temperatures, which would make it possible to predict with a high degree of accuracy the ultimate intensity of attack in any season. This problem, however, requires data much more extensive, detailed, and com])lete than have been obtained thus far. SUMMARY AND CONCLUSION With strikingly few exceptions, attempts made heretofore to esti- mate variations in disease attack have resulted, with respect lo com- mercial crops, in statements so patently generalized, both as to prevalence and damage, as to be far from convincing. Based on inexact observa- tions and evaluated from memory or from too brief and inadequate notes, they have dealt too much with reductions in yield and money losses, neither of which can be estimated fairly with our present knowledge. While such estimates have been, as a rule, far too conservative, placing losses inordinately low. the fact remains that an abundant yield is de- termined chiefly by soil, weather, and proper cultivation. It is only in the exceptional season that yield is determined by disease attack alone. The methods for collecting data on disease outlined in the preceding pages take into account the effect of disease upon the individual plant and its ])arts rather than ujion yield. Disease prevalence and disease destructiveness have been observed directly, in accordance with definite rules, in fields selected at random, in order that representative samples might be secured ; statistical methods of a relatively simple nature have been devised and used in evaluating the data thus secured ; and indexes have been computed which represent the disease attack as it actually occurred. Because they do not involve the necessity of comparing or evaluating indirectly estimated reductions in yield or money losses, these indexes furnish the most satisfactory means now available for comparing disease attacks, year with year, or region with region. They also provide a means of visualizing the effect of disease upon the crop subjected to attack, as was illustrated in Figures 89-92. The use that is to be made of exact data collected by studying and measuring epidemics under natural conditions has been exemplified, though briefly, with some of the wheat leaf rust data. It has been shown that there is a well defined relation between intensity of attack and annual mean tem])eratures and yearly totals of rainfall ; that in the July-Decem- ber period of the vear temperature has a greater influence than rainfall. 96 Illinois Natural History Sl'rvey Bullktin while in the January-June period rain is the more important, in deterniin ing the destructiveness of a year's rust attack; that 'b;re is a possibiHt} of predicting with reasonable accuracy the date up^n which the first spring infection will occur in any part of the State in any year ; that dur- ing the months of May and June the correlation between disease develop- ment and weather conditions lies mainly with temperature ; and that the degree of destructiveness which an epidemic, once started, is likely to attain may be predicted from the rapidity with which degrees of mean daily temperature accumulate. Such results as these outline only the broadest of the principles which underlie the yearly development of disease epidemics on the crops grown on the farms of Illinois. Although they have no immediate application to the problems of the farm, they are the general laws from which the rules of practice must be drawn. What has been learned in the case of wheat leaf rust must be learned also for the other diseases of wheat, and for the diseases of other crops ; the laws pertaining to all must be refined by further observations and analyses ; and the practical rules at last determined must be subjected to final tests, before the task is finished.