Bulletin 3 Z aturail Nov w fi J985 125 Years of Biological Research 1858-1983: A Symposium STATE OF ILLINOIS DEPARTMENT OF ENERGY AND NATURAL RESOURCES NATURAL HISTORY SURVEY DIVISION CHAMPAIGN, ILLINOIS VOLUME 33, ARTICLE 3 SEPTEMBER 1985 STATE OF ILLINOIS DEPARTMENT OF ENERGY AND NATURAL RESOURCES BOARD OF NATURAL RESOURCES AND CONSERVATION n„n Fl.hi^nn Ph D Chairman- H S Gutowsky. Ph.D., Secretary. Chemistry: Robert L, Metcalf. PhD B.otogy: Walter E, Hanson MS i^» r^mr LirinrNXg Jr., Ph.D., Forestty: L.L, Sloss, Ph.D., Geology, Theodon. L. Bro^n. Ph.D., Represent.ng the PresuUn, T>heuZZ7''oTnUnot.. and John C. Guyon. PhD, Representing the President of Southern llUriots Unuersity. NATURAL HISTORY SURVEY DIVISION, Champaign, Illinois SCIENTIFIC AND TECHNICAL STAFF Paul G. Risser, Ph.D., Chief Alice K, Adams, Secretary to the Chief SECTION OF AQUATIC BIOLOGY Gorden, Robert W., Ph.D., Aquatic Biologist and Head Buck, D. Homer, Ph.D., Aquatic Biologist Larimore, R. Weldon, Ph.D., Aquatic Biologist Sparks, Richard E. Ph.D., Aquatic Biologist Herendeen, Robert A., Ph.D., Associate Aquatic Biologist Philipp, David R, Ph.D., Associate Aquatic Biologist Reinbold, Keturah A., Ph.D.. Associate Aquatic Biologist Ross, Philippe, Ph.D., Associate Aijuadc Biologist Bayley, Peter B., Ph.D., Assis(an( Aquatic Biologist Henebi-y, Michael S., Ph.D., Assistant Aquatic Biologist Hicks, Randall, Ph.D., Assis(an( Aquatic Biologist Horns, William H., Ph.D., Assislanf Aquatic Biologist Lubinski, Kenneth S, Ph.D., Assistant Aquatic Biologist Starnes, Wayne, Ph.D., Assistant Aquatic Biologist Storck. ltd W., Ph.D., Assistant Aquatic Biologist Wiley, Michael J., Ph.D., Assistant Aquatic Biologist Sandberger, Jens-Dieter, M.S., Assis(an( Supporlwe Scientist Waite, Jana L., M.S., Assistant Supportive Scientist Warren, Gary L., B.S., Assistant Supportive Scientist Austen, Douglas, M.S., Junior Professional Scientist Hooe, Michael. M.S., Junior Professional Scientist Hubert, Beverley, Junior Professional Scientist Kwak, Thomas J., M.S., Junior Professional Scientist ftrry. Lance, M.S., Junior Professional Scientist Sobaski, Stephen T, B.S., Junior Professional Scientist Sons, Eugene, Junior Professional Scientist Tkzik, Pamela P, M.S., Junior Professional Scientist Deviling, David, M.S., Technical Assistant Ewing, Katharyn L., B,&, Technical Assistant Koppelman, Jeffrey, M.S., Technical Assistant Lyon, Dale, M.S., Technical Assistant Newman, Barry E., B.S., Technical Assistant Schmittler, Craig D., M.S, Technical Assistant Wagner, Ruth, Technical Assistant Claussen, Julie, B.S., Junior Technical Assistant Crawford, LouAnn, B.S., Junior Technical Assistant Crossett, Lorrie, M.S.. Junior Technical Assistant DeQuenne. Barbara, Junior Technical Assistant Day. David. B.S., Junior Technical Assistant Fields, Robert, B.S., Junior Technical Assistant Grubaugh, Jack, M.S., Junior Technical Assistant Magnelia, Steve. B.S.. Junior Technical Assistant Maimer, Susan, B.S., Junior Technical Assistant Mayer, Christine, B.S., Junior Technical Assistant McDowell, Brian D. M.B.A., Junior Technical Assistant McLuckie. Alan D., B.S., Junior Technical Assistant Miller, Suzanne M., M.S.. Junior Technical Assistant Wike. Lynn. M.S.. Junior Technical Assistant SECTION OF BOTANY AND PLANT PATHOLOGY Grunwald, Claus, Ph.D., Botanist and Head Crane, J., Leland. Ph.D., Mycologist Himelick, E.B., Ph.D.. Plant Pathologist Neeley. Dan, Ph.D., Plant Pathologist Schoeneweiss. D.F, Ph.D., Plant Pathologist Robertson, Kenneth R., Ph.D., Botanist Endress. Anton. G.. Ph.D., Associate Bo(anis( Iverson, Louis R., Ph.D.. Assistant botanist Sergent. James E., Associate Supportive Scienlisl Nelson. Betty. Assistant Supportive Scientist McKnight, Bill N.. MS, Junior Professional Scientist Burton, Philip, M.S., Technical Assistant Harrison, Robert A., Technical Assistant Huang, Li-Shar, Ph.D., Technical Assistant Karnes. Jean E.. M.S. Technical Assistant Moran. Robbin C. M.S.. Technical Assistant Nolte. David, B.S., Technical Assistant SECTION OF ECONOMIC ENTOMOLOGY Ruesink, William G,, Ph.D., Associate Entomologist and Head Appleby, James E., Ph D , Entomologist Armbrust, Edward J , Ph D., Emtomologist Kogan, Marcos, Ph.D.. Entomologist Kuhlman, Donald E , Ph D.. Entomologist, Extension Maddox, Joseph V.. Ph D . Entomologist Randell, Roscoe. Ph.D.. Entomologist. Extension Bouseman. John K.. M.S.. Associate Entomologist Eastman. Catherine E.. Ph.D.. Associate Enlomologitt Felsot, Allan S. Ph.D.. Associate Entomologist Irwin. Michael E.. Ph D.. Associate Entomologist Levine. Eli. Ph D.. Associate Entomologist Steffey. Kevin, Ph D . Associate Entomologist, Extension Briggs, Stephen P. M.S. Assistant Specialist. Extension Fischer. Daniel. Ph.D.. Assistant Professional Scientist Jeffords, Michael, Ph.D., Assistant Entomologist Killion, Eugene E , B.S, Assistant Entomologist extension Lamp, William O , PhD , Assistant Professional Scientist Onstad, David, PhD , Assis(an( Professional Scientist Miller, Frederic D.. Ph.D.. Assistant Entomologist, Extension Wei, Lester, Ph.D., Assistant Professional Scientist Weinzierl, Richard A., Ph.D., Assistant Entomologist. Extension Zavaleta, Luis. Ph.D., Assistant Professional Scientist Helm. Charles G.. M.S.. Associate Supportive Scientist Kogan. Jenny, M.S.. Associate Supportive Scientist Wilson. Jean G.. B.A.. Associate Supportive Scientist Auble. Jo Ann. Assistant Supportive Scientist Brewer. Ellen. MS., Assistant Supportive Scientist Chu, TzuSuan, MS, Assistant Supportive Scientist Guse, Charles. B.S. Assist ElL Spec. Computer Applications Kampmeier, Gail. M.S.. Assistant Supportive Scientist LePar, Gerald, Assistant Supportive Scientist McGiffen. Milton E,. Jr.. M.S., Assistant Supportive Scientist O'Hayer, Karen, MS, Assistant Supportive Scientist Roberts. Stephen J . B.S.. Assistant Supportive Scientist SwofTord. David, Ph.D., Assistant Supportive Scientist Walsh. Laura. M.S. Assistant Supportive Scientist Dazey, Doyle, Junior Professional Scientist Post, Susan, Junior Supportive Scientist Carpenter, Sheria L.. A. A.. Technical Assistant Case. Laurie, B.S., Technical Assistant Peters, Brenda M , Technical Assistant Sherman, Renee. B.S.. Technical Assistant Wissmann. Janet L.. B.S. Technical Assistant Withrow. Jane. rcr*nica( Assistant Morris. Marilyn. M.S., Junior Technical Assistant McGuire. Michael, M.S.. Junior Technical Assistant Grider, Dorothea, Transcribing Secretary. Extension Erdman, Diane. CJrr*-7\pis/ /// Tiaub. Joan. Clerk-Typist 111 Szkodzinski, Alice. Clerk-Typist II Baughman. Thomas A.. Graduate Research .Assistant Arif. S.ved Mohammad. M.S. Graduate Research Assistant Day, Eric. MS. Graduate Research Assistant Schroeder. Alan. MS. Graduate Research Assistant Smith. Lane, MS, Graduate Research Assisfanl Fielding, Dennis J,. Grnrfiiate Teaching Assistant SECTION OF FAUNISTIC SURVEYS AND INSECT IDENTIFICATION LaBerge. Wallace E . Ph D . /nsi-c/ Ta-vonomist and Head Brigham, Warren U . Ph D . Insect Th.ronomist Page. Lawrence M . Ph D, Fish Taxonomisl Webb, Donald W, Ph D , Insect Taxonomist Brigham. Allison R , Ph.D., Associate Professional Scientist Godfre.v. George L , Ph D . As-wciote Insect Thxonomisl Unzicker, John D, Ph D , A.'i.^ociote Insect ThxonomiH Voegtlin, David J , Ph D . Assistant Insect Ttixonomisl Treworg>-. Colin G , B.S.. Associate Supportive Scientist Butcher" Matthew K, MS, A.Siiociate Supportive Scientist (Continued to Back Cover) us ISSN 0073-4918 ILLINOIS Natural History Survey BULLETIN 125 Years of Biological Research 1858-1983: A Symposium STATE OF ILLINOIS DEPARTMENT OF ENERGY AND NATURAL RESOURCES NATURAL HISTORY SURVEY DIVISION VOLUME 33, ARTICLE 3 CHAMPAIGN, ILLINOIS SEPTEMBER 1985 I WHEREAS, the lllinots Stiile Natural History Society was organized in Bloom- ington in IS5S. and has evolved into the highly respected Illinois State Natural His- tory SuT\-ey of the Department of Energy and Natural Resources, now located in Urbana. and ii'HEREAS. Dr. Paul G. Rtsser currently heads the staff of 200 which includes nationally acclaimed scientists; and WHEREAS, the Natural History Survey's work has received national recog- nition in the areas of insect control, fish and pond management, floricultural disease control, game species, waterfowl, and periodic reports on endangered animal spe- cies, and WHEREAS, the Natural History Survey has brought world-uide recognition to the State of Illinois and information of immeasurable value to lllinoisans on both the history and the future of the flora and fauna in the state. THEREFORE. I. James R. Thompson. Governor of the State of Illinois, pro- cLim September 24. 1983. as NATUR.'iL HISTORY SURVEY DAY in Illinois, in celebration of its 125th anniversary. \._^^^/ife',Ae^'eii^i/o^^,/m^yAix/n€^ /;£'A. seventh vj^y.^o/^ SEPTEMBER .j/n.A^ ^^^. EIGHTY-THREE , and 1.000- 2,000 mm of rain falls during the re- mainder of the year. Portions of the park were an opei'ating cattle ranch from no later than 1710 to 1978. The vegetation contains at least 680 species broadleafed plants (at least 400 species of perennial woody plants) and sup- ports at least 3,000 species of cater- pillars plus several hundred species of other animals that eat living plant parts. There are checklists of plants (Janzen & Liesner 1980K birds (Stiles 1983), reptiles and amphibians (Scott jj et al. 1983), mammals (Wilson 1983), and butterflies (DeVries 1983) for the park. The plant distributions within this vegetation range from nearly monospecific stands of very large trees (e.g., 10-20-m-tall stands of Qiiercus oleoides Cham, and Schlecht., Hymen- aea coui'baril L., Ateleia hcrbert-smithii (Pittier) to highly mixed vegetation where as many as 200 species of woody plants may occur in 100 ha and adult conspecific crowns are usually sepa- September 1985 125 Years of Biological Research 143 rated by one to many allospecific crowns. At Santa Rosa, herbivory by caterpillars, the focus of this essay, is chai-acteristically highly heterogene- ous among years and among individ- uals, species, and age classes of plants (e.g., Janzen 1981). ANATOMY OF A DEFOLIATION EVENT The event During the 1983 rainy season, a representative defoliation event oc- curred at Santa Rosa. The impact of the herbivores was highly heterogeneous. 1 briefly describe the ecology of this im- pact as an example of a pattern that could have been generated either by the heterogeneity of internal plant chemistry or by mortality factors exter- nal to the plant (or both). In fact, the pattern seems to have been generated by external factors and therefore pro- vides an example for the introductory comments of this essay. It is described in more detail elsewhere (Janzen un- published manuscript). While the study was conducted in a patch of Santa Rosa forest of approximately 2 km- (Bosque San Emilio, approxi- mately 2 km northeast of the park ad- ministration area), the 11,000-ha park contains tens of square kilometers of this type of forest. Cursory examina- tion of other forest patches showed that the events described here occurred at those sites as well. The first generation From 28 May to 4 June 1983, when 50-90-year-old secondary successional deciduous forest at Santa Rosa was just beginning budbreak following the first significant rains of the rainy season, the forest was .sprinkled with oviposit- ing Aellopos titan (Cram.). One or more of these small diurnal sphingid moths (Haber & Frankie 19831 could be seen by simply standing and looking through the forest for 5 minutes or less at any time during daylight hours. The moths darted among the shrubs and treelets at a height of about 1-3 m. They touched branchlets, twigs, and buds with legs and the tip of the abdo- men. Upon encountering a plant of Randia karstenii Polak or R. subcordata (Stand.) Standley, the moth hesitated a moment longer and .sometimes laid a single spherical pale gi'een egg on the bud, newly expanding leaf, thorn, or twig it contacted (Fig. 1). It then flew to other branches of the same plant or, about equally frequently, off to neigh- boring plants. Both species of Randia were beginning budbi'eak; a few indi- viduals were covered with a thin layer of newly expanding leaves, while others had only swelling buds. The outcome of this oviposition, by what must have been several thousand moths in the study area, was the deposition of tens of eggs to a thousand or more eggs on each Randia in the forest (of 214 plants briefly examined, all had some eggs). For example, on 5 June, all or a ma- jor portion of the potential oviposition sites on 14, 1.5-5-m-tall R. subcordata were searched thoroughly for A. titan eggs. As many as 10 eggs had accumu- lated on some branch ends (Fig. IC), but 1 or 2 per branch end or bud was more usual (Fig. lA, B). The estimated or actual numbers of eggs on these plants ranged from 21 to 1,212, with an average of 197 (SD = 316). Smaller plants (0.7-1.4 m in height) had 1-19 eggs on them (x = 6.0, SD = 5.1, n — 10). The small plants were not only shorter, but also had only one or a few stems and branchlets. While I did not count them, the numbers of A. titan eggs on the R. karstenii appeared to be about the same. Each hectare of this 2-km- patch of forest contains at least four individuals of these two Randia species that are 1.5 m or more in height, which allows a rough estimate of 800 oviposition plants carrying a minimum of about 160,000 A. titan eggs. If I assume that each female can lay 100 eggs, these eggs represent the oviposition by 1,600 moths. 144 Illinois Natural History Survey Bulletin Vol, 33. Art. 3 Fig. 1. — (A) Egg of Aellopos titan laid on a spine of Randia subcordata at ttie time of budbreak. (B) Accumulated eggs of A. titan following multiple ovipositions over several days on R. subcordata (C) Egg of A. titan laid on tfie underside of an expanding new leaf of R. subcordata- 30 May 1983, Santa Rosa National Park, Costa Rica. From the outset, the number of eggs present on a given plant was the prod- uct of several processes. Even while females were ovipositing, eggs were be- ing carried off by Pseudomyrmex, Azteca, Crematogaster, and other ants, picked off by warbler-sized and -shaped birds, and fed on by anthocorid bugs. The eggs hatched 5-6 days after ovi- position; even as the first were hatch- ing, more were still being laid. While minute parasitic Hymenoptera were observed ovipositing in the eggs in the field, no parasitoids were reared from a sample of 682 eggs collected from 24 different R. subcordata plants, and no parasitoids appeared when 100 of these eggs were reared to the adult stage. The first-instar larvae are pale green, match well the color of the new- ly expanding Randia leaves, and wander widely through the foliage of the plant on which they hatch. They appear to be highly edible to the ants and birds mentioned above, reduviid bugs, Fblistes wasps, other wasps, small spiders, and carabid beetles; all of these animals were observed to capture and eat them or carry them off on numer- ous occasions. By 4-7 June, tens to hundreds of first- or second-instar A. titan larvae could be found on any Ran- dia more than 1 m tall, and some lar- vae were present even on plants as small as 30 cm in height. On all plants the larvae appeared to be healthy and were feeding heavily. I was absent from the site 8-13 June, and upon my return, it was evi- dent that the outbreak had developed to the extent that three levels of defoli- ation could be recognized among both species of Randia. There were plants with only a few larvae (1-10 per plant by quick inspection), plants with mod- erate numbers of larvae (10-30 per plant), and plants with hundi-eds of lai'- vae. On all plants, the lai-vae appeared I September 1985 125 Years of Biological Research 145 to be healthy and to have about the same size distribution. On 18-20 June, 49 R. subcordata plants were thorough- ly searched at night for larvae. By searching at night with a strong flash- light, at least 99 percent of the larvae present on a plant were located. By these dates, virtually the entire popu- lation was in the ultimate (Fig. 2) or penultimate instar. A few larvae had already left their host plants to pupate or, if the host was defoliated, to search for more food. No larvae were on 43 per- cent of the plants, and all of these plants showed minor defoliation; how- ever, there was enough defoliation to make it clear that some A. titan cater- pillars had developed to moderate size on these plants before leaving or being preyed upon. The plants (33 percent) with 1-10 larvae had either moderate defoliation (10-50 percent of their leaf Fig. 2. — (A) Ultimate-instar caterpillar of Aellopos titan; this green morpti tias a white lateral posterior diagonal stripe and light lateral diagonal white and magenta side stripes; it is the most common morph. (B) Ultimate-instar caterpillar of A. titan, this dark morph is dorsally lavender and ventrally black, with white lateral markings. It is the rare morph except during crowding, heavy shading, or total defoliation. The previous two instars of A titan are extremely similar to these caterpillars, 20 June 1983, Santa Rosa National Park, Costa Rica, 146 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 surface area) or total leaf loss. In the latter case, the larvae present were wandering over the surface of the plant eating off petioles, leaf blade frag- ments, and buds. They were obviously a small remnant ofmuch greater num- bers of larvae on the tree. Finally, there were trees (25 percent) with 11 to sev- eral hundred larvae. For example, I counted, by removal sampling, 246 last- instar caterpillars from one R. subcor- data, and there were at least another 200 on the tree (Fig. 3). At the time of the census, these densely populated trees still had 10-30 percent of their leaf surface remaining; however, within 5 days, when all larvae had been preyed upon or had left the tree to pupate, all of these leaves had been eaten. Such trees appeared to have had caterpillars that were slightly delayed in their development or initially to have had fewer caterpillars on them than those already stripped of their leaves had had. It was evident that complete defol- iation (Fig. 3, 4) was a function of the number of caterpillars on a plant and the size of the plant. Part of the hetero- geneity of defoliation was generated purely by this interaction. The large ultimate- and penulti- mate-instar lai-vae (Fig. 2 ), were eaten by small to medium-sized birds, redu- viid bugs, scorpions, and unidentified mammals. One mammal scat, of a size that could have been produced by a gray fox, small cat, or procyonid, con- tained 26 head capsules of penultimate and ultimate instar A. titan larvae. While such an enormous population of moth larvae would seem to be an easy substrate for intense parasitization, of 617 ultimate and penultimate instar larvae collected and reared, only seven caterpillars were parasitized by tachi- nids, and none contained parasitic Hymenoptera (Table 1). The prepupal wandering stage ofA titan walks or di-ops off a plant and bui-- rows into the litter to pupate. Of those which pupated in captivity, all living individuals emerged from the pupal Fig. 3. — Randia subcordata adult tree defoliated by Aellopos titan: 3 days before this photograph was taken, this tree had at least 446 ultimate-instar A. titan caterpillars on it. The top of the parachute background is 3,8 m above the ground at each end, 23 June 1983, Santa Rosa National Park, Costa Rica. September 1985 125 Years of Biological Research 147 Fig. 4 — Randia karstenii subadult treelet defoliated by Aellopos titan- 23 June 1983, Santa Rosa National Park, Costa Rica. stage about 13-15 days after pupation (between 6 and 20 July). Pupal-stage duration was not influenced by wet or extremely dry conditions. The newly emerging adults simply left the site where they had developed; they were not observed at flowers that are stand- ard nectar hosts for Aellopos (e.g., Cedrela odorata L., which were visited by hundreds of female moths at the time of oviposition in early June). Despite extensive and intensive noctur- nal searches of Randia shrubs from mid-July to the end of August, only three A. titan larvae were located. There was no evidence of an attempted or realized second generation of the moth at this site or in other parts of the Santa Rosa forest. The pupae in the litter were subject to moderate predation by vertebrates. Collared peccaries [Tayassu tajacu (L.)] and nine-banded armadillos (Dasypus novemcinctus L.) snuffled intensively through the litter below the Randia that had had large numbers of cater- pillars, and I presume they were har- vesting A. titan pupae. Spiny pocket mice [Liomys salvini (Thomas)] ate the pupae readily in the laboratory and foraged incessantly at night in the forest litter for seeds and insect pupae; on one occasion a mouse brought an A. titan pupa into a live trap and ate part of it there. However, living A. titan pupae could easily be found by sorting through litter until the time of adult eclosion in mid-July, and numerous newly eclosed adults were encountered on the foliage at that time. There is no doubt that despite the various sources of mortality mentioned above and later, a large number of adults eclosed in Bosque San Emilio and in other patch- es of deciduous forest in Santa Rosa. Two weeks after the A. titan larvae had disappeared from the Randia, I walked a line transect through the Bosque San Emilio and estimated the intensity of defoliation of all Randia encountered (14 July 1983). Of 173 R. Table 1. — Fate of 617 penultimate and ultimate instar Aellopos titan larvae collected from Randia subcordata and R. karstenii in Bosque San Emilio, Santa Rosa National Park, Costa Rica (1983). 148 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 subcordata, 41 percent were leafless (Fig. 3), 28 percent were moderately but conspicuously defoliated (often with one part of the crown more severely defoliated than another), and the re- mainder (31 percent) showed only traces of the feeding damage characteristic ofA titan larvae. Of 181 R. karstenii encountered, 63 percent were essentially leafless (Fig. 4), 30 per- cent were moderately defoliated, and the remainder appeared intact but had probably been fed on. Since R. karstenii leaves are very small and have a weak midrib, the larvae generally ate the en- tire leaf; small amounts of damage were therefore harder to recognize than they were on R. cordata. By the last half of August, all of the defoliated plants were putting out a new leaf crop and appeared approximately as they did during the first week of June. On 15 June 1983, a transect of forest along Quebrada Costa Rica (about 5 km SW of Bosque San Emilio and slightly drier) located 51 R. karstenii. All of these plants had lost all of their leaves to A. titan. During the 1979-1982 gi'owing seasons there was no defoliation of Randia by A titan (or by any other in- sect). Frequent searches of numerous Randia plants during these 4 years yielded eight AeUopos larvae from Bosque San Emilio forest, six of which were parasitized by tachinids or Hymenoptera. The enormous number of adult female A. titan which ap- peared, as if by magic, in the Santa Rosa deciduous forest 25-28 May 1983 must have come from elsewhere, and when their offspring eclosed, they went elsewhere. Whether they arrived as a consequence of having concentrated in a large down-wind area in response to the odor of newly foliating Randia, or whether some more complex congi-ega- tion event occurred, is beyond the scope of this essay. My field notes from the 1978 rainy season suggest that there was a high density of A. titan on Randia in the Bosque San Emilio in late May and early June. While I was not collecting caterpillars in the field at that time, other biologists brought me 23 casually encountered A. titan larvae from "Ran- dia" in this forest. This account suggests several ques- tions pertinent to this essay. Why did the newly eclosed adults not attempt a second generation at Santa Rosa by ovipositing on the foliated Randia? Why were predation and parasitization rates not high enough to eliminate the first generation or at least depress it to the levels of the 1979-1982 rainy seasons? Why was the final result of this "outbreak" a very heterogeneously defoliated array of Randia trees? No second generation At the time ofA titan eclosion (eai-- ly to mid-July), the undefoliated Ra/i- dia had at least 5, perhaps 6, months of leafiness ahead of them. Since the egg- to-adult time for A titan is about 35 days, there is ample time in the rainy season for at least three generations. Throughout the 1979-1982 rainy sea- sons, there were such generations at extremely low densities. In 1981, I recorded A. titan oviposition on R. sub- cordata as late in the rainy season as 10 January. In mid-August 1983, 1 en- countered two last instar larva on R. subcordata and a single last instar larva on R. karstenii, but it was clear that no large pulse of eggs and larvae appeared on Randia foliage in July as it did in late May and early June. Therefore it is clear that the newly eclosing A. titan females were con- fronted with a moderate number of leafy Randia, but either they rejected them as oviposition sites and went else- where or were progi-ammed to leave the site irrespective of available food. Ironi- cally, a leafy Randia that has escaped a defoliation event is probably an un- suitable oviposition site. If Randia are foliated because of leaf chemistry (un- likely; see below), then the female should not oviposit on leafy Randia. In contrast, until the second half of September 1985 125 Years of Biological Research 149 August, defoliated individuals did not have enough leaves to support even a few caterpillars. They were leafless at the time of eclosion of the first genera- tion. If, as I suspect, the foliated Ran- dia is an indicator of high carnivory risk to eggs and caterpillars, then again, it is not a good oviposition site; the leafless Randia suggests a low risk site but offers no food. Selection should favor females that will ignore foliated Randia at the site of their birth and search in space or time for sites not subjected to recent defoliation. It is appropriate to add here that I have observed A. titan females to ovi- posit on the new, partly expanded leaves of R. subcordata and R. karstenii during all months of the rainy season at Santa Rosa. Furthermore, I have reared their caterpillars on leaves of all ages in all months of the rainy season. While newly foliating Randia in late May and early June have more newly expanding leaves than at any other time, the plants also have some newly elongating branchlets at all times of the gi'owing season. There are always some oviposition sites for A. titan dur- ing the rainy season in the Santa Rosa deciduous forests, even if they will only oviposit on very young foliage. Relatively low carnivory rates While eggs and larvae of A. titan suffered conspicuous and intense pred- ation, it was not sufficient to eliminate a highly evident array of caterpillars. As a Randia was becoming fully defol- iated, for the last 2-8 days virtually no search was necessary to see tens of A. titan caterpillars on its ever more naked branches and petioles. However, these plants were not being subjected to hoards of insectivorous birds, mam- mals, or carnivorous insects. During several hours of observing Randia heavily infested with ultimate and penultimate instar caterpillars, it was customary to encounter 5-10 reduviid bugs, each killing one caterpillar; sev- eral caterpillars being stung or carried off by ants; one or two being carried off by small to medium-sized birds; and several apparently starving to death because they had fallen or wandered off their host plant. Some plants clearly had all of their caterpillars removed by carnivores, while others retained large numbers even after many had been removed; however, carnivore density was simply not high enough nor their searching thorough enough to depress the A. titan caterpillar density on all trees to a level even approximating the low level "normally" observed. Either the Santa Rosa carnivore density was exceptionally low in 1983 or exceptionally large numbers of A. titan eggs were laid there in early 1983. I have no formal census data on carni- vore density at Santa Rosa in 1979- 1982, but certainly the 1983 density of birds and other predators did not seem to be any lower than were the 1979- 1982 levels. I should add, incidentally, that these levels are much lower than those readily visible in, for example, a Min- nesota or Michigan woodlot in late May. However, I can state that there was no massive input of A. titan eggs and larvae in the 1979-1982 rainy seasons.' The startlingly low numbers of parasitized A. titan larvae (Table 1) are consistent with records for other species of caterpillars at Santa Rosa that have suddenly increased greatly in density (unpublished field records). A hypothesis consistent with this low rate of parasitization is that the enor- mous number of larvae satiated the ovipositional capacity of the parasi- toids present. It was striking that four of the seven parasitized larvae had dif- ferent species of parasitoids in them and that none of the parasitoids were species that I had reared from A. titan in the past. While I could not monitor disease levels in the field, the results of rearing 'Note added in press: Likewise, in the 1984 and 1985 rainy seasons at Santa Rosa there was no massive input of A. titan. 150 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 penultimate and ultimate instar wild- caught caterpillars show clearly that disease was taking a severe toll of the caterpillars. The 617 caterpillars col- lected on 18-24 June (Table 1) were placed in individual plastic bags and were fed with foliage from plants light- ly to moderately defoliated by A titan (no plants were available with no A. titan damage). The bags were new or had been used (less than 1 percent) to house pupae of undiseased Saturniidae. Overall 39 percent of the caterpillars died of an unknown disease as larvae and as pupae of all ages. By way of con- trast, Aellopos penultimate or ultimate instar larvae collected during 1979- 1982 at Santa Rosa (three species on four host plant species, n = 53) invari- ably produced either adults or para- sitoids, even though they were reared in exactly the same manner. The data in Table 1 suggest that the larger the collection of caterpillars, the fewer are killed by disease. However, the relation- ship actually lies with the age of the caterpillars at the time of collection. The 19 June sample contained the largest number of penultimate instars and the 23 and 24 June samples con- tained only ultimate instars with 1-3 days remaining before pupation. In the earliest samples, there were infected caterpillars that were fated to die in nature as larvae and that then died in my rearing bags. In the latest samples, those caterpillars that were fated to die in nature as caterpillars had already done so. It is unlikely that the A. titan "out- break" escaped annihilation because the predators and parasitoids were oc- cupied elsewhere with some other spe- cies of prey. The overall caterpillar density in Bosque San Emilia in 1983 was conspicuously lower than in 1979- 1982. Furthermore, 5 weeks, one A. titan generation, is insufficient time for a functional response by the predator and parasite array. A numerical re- sponse depends on both the distance over which the carnivores must move and the areal extent of the high density of prey. Little can be said about either in this tropical habitat, but it was clear that the high density ofA titan larvae covered hundi-eds of hectares. Also, this part of Costa Rica is not replete with large numbers of, for example, small in- sectivorous birds that could or would move to an "outbreak" of this sort. In short, it appears that the local carnivores did not eliminate the A titan outbreak because there were not enough of them to do so. I suspect that this was the case because ( 1 ) the out- break occurred during the first weeks of the rainy season before whatever seasonal increase in carnivores that might occur had occurred, and (2) the habitat is not rich enough in prey to maintain a sufficiently high level of carnivores that they could consume a prey pulse with the properties of the A titan caterpillars. Such an answer con- tains the implicit assumption that A titan caterpillars are not suitable or available prey to a moderately large portion of the diverse carnivore array at Santa Rosa. The reasons range from physiological incompatibility to size and behavior. For example, the Enicos- pilus (Ichneumonidae) parasitoids that were heavily parasitizing large Roths- childia lebeau (Guer.-Meneville) (Sat- urniidae) larvae at the same time in the same forest, have never been en- countered in over a thousand rearings of Aellopos and other Sphingidae cater- pillars in Santa Rosa. The Pseudomyr- mex and Azteca ants that so eagerly kill and carry off first and second instar A. titan larvae show no attack response to penultimate and ultimate instar A. titan larvae. Coatis [Nasua nasua (L.)] readily eat the larvae but do not readi- ly climb the very thin and spiny bran- ches of Randia. White-faced monkeys [Cebus capucinus (L.)] generally do not forage as low as 1-3 m above the gi-ound in the forest interior during the rainy season and might well have never noticed the outbreak going on below them. The few species of North Ameri- can migrant birds that visit Guana- caste during the northern winters had long since departed (see Janzen 1980). September 1985 125 Years of Biological Research 151 Heterogeneous defoliation The array of predators in a tropical forest possesses sufficient patterned variation in its foraging to generate heterogeneous defoliation from an in- itial high level and thoroughly spread layer of eggs and caterpillars. One Ran- dia may be close to and foraged exten- sively on by a colony of Azteca or Camponotus ants, and another only a few meters away may only be passed through by an occasional far-wander- ing Pseudomyrmex ant. A Campylo- rhynchiis wren family may have only a single Randia in its foraging territory and may therefore thoroughly glean it of its caterpillars; another wren family may have 20 Randia in its territory and therefore only moderately glean several and ignore the remainder (F. Joyce, Division of Ecology and System- atics, Cornell University, personal com- munication). A reduviid bug that set- tles on a small Randia with 10 half- gi'own caterpillars may well kill all of them before they have done serious defoliation; the same bug on a neigh- boring tree with 30 caterpillars will at best only slow the time to total defolia- tion by a few days, since the surviving 20 caterpillars will be quite enough to eat all the leaves off the tree. Had only a few A. titan eggs been laid in the for- est (as in 1979-1982), heterogeneity of predation would still have occurred, but the consequence would have been invisible against the backgr-ound of the initial heterogeneity of oviposition. Had there been no carnivoi-y in 1983, all the leaves would have been eaten off all the Randia trees, and there would have been major A. titan death by star- vation. It can be argued that the failure of females to avoid previously laid eggs when ovipositing may well be due to the absence of selection for such behavior if it is usual for the number of surviving offspring to be unrelated to the number of offspring initially present on the plant. Her best chance is to distribute her eggs thinly over the Randia population, thereby maximiz- ing the chance of placing some eggs on those Randia where there is low risk from predators. The strong heterogeneity of defolia- tion could, at least in theory, also have been generated by heterogeneity in the suitability of the Randia foliage in terms of nutrients and secondary chemical defenses. Numerous natural history facts and suppositions argue against this possibility: 1. When second- to fifth-instar lar- vae were removed from Randia that they had defoliated and then were placed on Randia that had lost their in- itial caterpillars, they fed, developed, and pupated normally whether this was done in the field (under a protec- tive net) or in plastic bags in the laboratory. In the laboratory, all larvae produced adults that were smaller than wild adults. However, they were equal- ly dwarfed if they were fed leaves from their partly defoliated tree or leaves from trees quite free from damage (i.e., living in a plastic bag is suboptimal). 2. Defoliation was a function of tree size as well as of caterpillar numbers. A representative 2-m tall tree with 30 developing middle-sized larvae on it was defoliated by the time they left the tree to pupate; a 4-m tall tree with the same number of larvae suffered only minor damage, since it had many more leaves than did the smaller tree. 3. At least 40 different and foliated Randia subcordata trees were defoli- ated by harvesting their leaves to feed larvae in the laboratory. These larvae all developed normally with virtually no death except by disease, parasitoids, and laboratory accidents; their develop- ment time was not longer than that of their siblings left on Randia in the forest. 4. All foliated trees were fed on to some degi-ee, and a careful search of many trees with no conspicuous dam- age revealed that one to five large and healthy caterpillars were feeding on them. When the foliage from these 152 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 trees was offered to second-instar lar- vae in the laboratory, these larvae fed readily and developed into normal adults. It is tempting to wish that one could a priori know which trees were to be defoliated and therefore could ob- tain a leaf sample whose chemistry could be compared with those of leaves from trees not defoliated. However, this procedure was, and is, impossible when the moth has only one defoliating gen- eration at a site in a season. Further- more, all Randia individuals were fed on by at least one or more caterpillars; defoliation was a matter of degree. Finally, there is the problem of know- ing what chemicals to search for in such a comparison. It might well turn out that the eventually defoliated trees would be found to differ chemically from those that were not defoliated. However, there would be no reason to believe that these particular chemical differences, rather than some other covarying plant traits, were responsi- ble for the final levels of defoliation. HETEROGENEITY AMONG LIFE FORMS AND HABITATS The risk of carnivory is important to caterpillars (e.g., Heinrich & Collins 1983; Holmes et al. 1979). This risk is not uniformly distributed among plant life forms (e.g., Thompson 1983; Roy- ama 1970; Niemela et al. 1982) nor among the habitat types associated with different life forms. For example, a single large caterpillar in the crown of a large tree that projects above the vegetation should have a quite differ- ent probability of being located by a bird than if it were in the crown of a small sapling of the same species of tree projecting from young secondary succession. Likewise, a caterpillar in prey-rich, forest-edge herbaceous vege- tation is more likely to be encountered by a Polistes wasp than if it were in the prey-poor canopy of a patch of ever- gi-een forest. At Santa Rosa, the dis- tribution of saturniid and sphingid caterpillars among their host plants provides an example. At Santa Rosa, there are 30 breed- ing populations of Saturniidae ( Janzen 1982) and at least 63 breeding popula- tions of Sphingidae (Janzen 1984b). In an ongoing study, I have located at least one native host plant for 80 per- cent of the saturniids and 77 percent of the sphingids known or strongly sus- pected to breed at Santa Rosa (Tables 2 and 3). After examining only the host Latin binomials and caterpillar fami- ly names in Tables 2 and 3, it is evident that the huge caterpillars of sphingids and saturniids have little in common in their use of food plants (Janzen 1984b). Furthermore, at least half of the overlap in the two lists is a more apparent than a real similarity of food choice. In keeping with current emphases on the causes of host specificity, several years ago a slightly reduced subset of the plants in Tables 2 and 3 was ana- lyzed for phenolics and alkaloids (Jan- zen & Waterman 1983). The foliage of the saturniid hosts was found to con- tain significantly more phenolics and less water than did that of the sphingid hosts, while the sphingid hosts much more frequently contained alkaloids than did the saturniid hosts. These results are in strong agi-eement with two general impressions held by lepi- dopterists: (1) Most saturniids feed on tree foliage (traditionally viewed as rich in tannins), and (2) most sphingids feed on plants in families famous for toxic small molecules, resins, and cop- ious latex flow. While not meaning to denigi-ate the reality of these patterns, there is obviously a quite different way to view the same set of host records. It is evident that, as found in na- ture, the saturniid caterpillars are gen- erally in the crowns of adult trees, while the sphingid caterpillars are found in plants ranging from small herbs to vines to saplings to crowns of large trees CKibles 2 and 3). The most September 1985 125 Years of Biological Research 153 Table 2. — Larval hosts of Saturniidae in nature in Santa Rosa National Park, Costa Rica (as of Decennber 1983). Larva Host Microhabitat Arsenurinae Arsenura armida (Cram.) Caio championi (Drc.) Copiopteryx semiramis (Cram.) Dysdaemonia boreas (Cram.) Titaea tamerlan (Maassen) Ceratocampinae Adeloneivaia isara (Dognin) atheroma bellavista (Draudt) atheroma lobesis (W. Rothschild) Eacles imperialis (Dry.) Othorene purpurascens (Schaus) Othorene verana (Schaus) Ptiloscola dargei Lemaire Schai/siella santarnsensis Lemaire Syssphinx colla (Dyar) Syssphinx mexicana (Bdv.) Bombacopsis quinatum (Jacq.) Dugand (Bombacaceae) Bombacopsis quinatum (Jacq.) Dugand (Bombacaceae) Manilkara chicle (Pittier) Gilly (Sapotaceae) Ceiba pentandra (L.) Gaerth. (Bombacaceae) ^ Bombacopsis quinatum (Jacq.) Dugand (Bombacaceae) Lysiloma dwariciata (Jacq.) Macbride (Leguminosael Lysiloma auritum (Schl.) Benth. (Leguminosae) Phoradendron quadrangulare (HBK) Krug & Urb. (Loranthaceae) Cochlosperum vitifolium (Willd.) Spreng. (Cochlospermaceae) Bursera simaruba (L.) Sarg. (Burseraceae) Spondias mombin L. (Anacardiaceae) Psidium guajava L. (Myrtaceae)*^ Calycophyllum candidissim u m (Vahl.)'DC. (Rubiaceae) Phoradendron quadrangulare (HBK.) Krug & Urb. (Loranthaceae) Cochlospermum vitifolium (Willd.) Spreng. (Cochlospermaceae) Bursera tomentosa (Jacq.) Ti'iana & Planch. (Burseraceae) Astronium graveolens Jacq. (Anacardiaceae) Cedrela ndorata L. (Meliaceae) Manilkara chicle Pittier (Gilly) Sapotaceae Quercus oleoides Cham. & Schlecht. (Fagaceae) Acacia tenuifolia (L.) Willd. (Leguminosae) Hymenaea courbaril L. (Leguminosael Pithecellobium soman (Jacq.) Benth. (Leguminosael Acacia cnllmsii Safford (Leguminosae) Acacia cornigera L. (Leguminosae) large tree crown large tree crown large tree crown large tree crown large tree crown large tree crown large tree crown parasite in crown of large tree large saplings large tree crown large saplings shrubby treelet large tree crown parasite in crown of large tree large saplings and large tree crown large tree crown large tree crown large tree crown large tree crown large tree crown saplings and large vine crown large tree crown large tree crown sapling to adult treelet crown sapling to adult treelet ci'own 154 Illinois Natural History Survey Bulletin Table 2. — Continued Vol. 33, Art. 3 Larva Host Microhabitat Syssphinx molina (Cram.) Syssphinx quadrilineata (G. & R.) Hemileucinae Automeris io (F.) Automeris rubrescens (Wlk.) Automeris zugana Drc. Pitheceltobium soman (Jacq.) Benth. (Leguminosae) Cassia grandis L. (Leguminosael '^ Atbizzia adinocephala (Donn. Sm.) Britt. & Ftose (Leguminosae) Pithecellobium soman (Jacq.) Benth. (Leguminosae) Crescentia olato HBK. (Bignoniaceae) '^ Mimosa pigra L. (Leguminosael Cassia biflora L. (Leguminosae) Rhynchosio reticulata (Swartz) DC. (Leguminosae) Gliricidia sepium (Jacq.) Walp. (Leguminosae) Inga vera Willd. (Leguminosae) Rourea glabra HBK. (Connaraceae) Guazuma ulmifolia Lam. (Sterculiaceae) Cassia biflora L. (Leguminosae) Querus oleoides Cham. & Schlecht. (Fagaceae) Cordia alliodora (R. & P.) Oken (Boraginaceae) Lonchocarpus minimiflorus Donn. Smith (Leguminosae) Calycophyllum candidissimum (Vahl.) DC. (Rubiaceae) DHJ 12175 (Bignoniaceae) Zuelania guidonia (SW.) Britt. & Millsp. (Flacourtiaceae) Crescentia alata HBK. (Bignoniaceae) '^ Cassia grandis L. (Leguminosae) Annona purpurea Moc. & Sesse (Annonaceae) Lonchocarpus costaricensis Pittier (Leguminosae) Querus oleoides Cham. & Schlecht. (Fagaceae) Cydista heterophylla Seib. (Bignoniaceae) Calycophyllum candidissimum (Vahl.) DC. (Rubiaceae) Hymcnaea courbaril L. (Leguminosae) Solanuni hazenii Britt. (Solanaceae) Lantana camara L. (Verbenaceae) '' Lonchocarpus criocarinalis Micheli (Leguminosae) Cenlroscma pubescens Benth. (Leguminosae) large tree crown large tree crown large tree crown large tree crown large tree crown shrub shrub herbaceous vine sapling sapling scandent shurb large tree crown shrub sapling sapling sapling large tree crown sapling vine large tree crown large tree crown large tree crown large tree crown large tree crown large tree crown large woody vine sapling to large tree crown sapling large herb large herb shrub large tree crown heib vine September 1985 125 Years of Biological Research Table 2, — Continued 155 Larva Host Microhabitat Dirphia m'ia (Stoll) Hylesia dalina Schaus Hvlesia lineata Drc. Cassia hayesiana (B. & R.) 156 Illinois Natural History Survey Bulletin Table 2. — Continued Vol. 33, Art. 3 Larva Host September 1985 125 Years of Biological Research Table 2. — Continued 157 Larva Host Microhabitat Saturniinae Copaxa moinieri Lemaire Rothschildia ervcina (Shaw) Rothschitdia lebeau (GueivMeneville) Cassia alata L. (Leguminosae) '^ Inga vera Willd. (Leguminosae) Ardisia revoluta HBK. (Myrsinaceae) Astronium graueolens Jacq. (Anacardiaceae) Hymenaea courharil L. (Leguminosae) Quercus oleoides Schlecht. & Cham. (Fagaceae) Miconia argentea (Swartz) DC. ( Melastomataceae ) Ocotea veraguensis (Meisn.) Vlez (Lauraceae) Exostema mexicanum (Jacq.) Roem. & Schult. Coutarea hexandra (Jacq.) Schum. (Rubiaceae) Exostema mexicanum Jacq. Roem. & Schult. (Rubiaceae) Spondias mombin L. (Anacardiaceae) Spondias purpurea L. (Anacardiaceae) Casearia corymbosa HBK. (Flacourtiaceae) Zuelania guidonia (Sw.) Britt. & Rose (Flacourtiaceae) Xanthoxylum setutosum P. Wilson (Rutaceae) large tree crown medium tree treelet sapling sapling large tree crown sapling saplings and lower branches of treelet large tree crown treelet crown large tree crown large tree crown treelet treelet large tree crown large tree crown ^ Not yet found in nature but accepts readily and dies on other Santa Rosa Bombacaceae. '' Not yet found in nature but accepts readily and has the appropriate color and behavior to use this host. ^ Plant introduced to Santa Rosa within past several hundred years. ••**••*•***• Table 3. — Larval hosts of Sphingidae in nature in Santa Rosa National Park. Cost Rica (as of December 1983). Larva Aellopos clavipes (R. & J.) Aellopos fadus (Cram.) Aellopos lilan (Cram.) Agrius cingulalus (F.) Host Randia karslenii Polak (Rubiaceae) Genipa americana L. (Rubiaceae) Aliherlia edulis A. Rich. (Rubiaceae I Randia karslenii Polak (Rubiaceae) Rnndia subeiirduta (Stand.) Standley (Rubiaceae! Merremia timbellata (L.I Hall (Convolvulaceae) Microhabitat sapling to troclct sapling to large tree shrub sapling to treelet .sapling to treelet herb vine 158 Illinois Natural History Survey Bulletin Table 3. — Continued Vol. 33, Art. 3 Larva Host Microhabitat Aleuron carinata Wlk. Aleuron iphis Wlk. Amplypterus gannascus (StoUl Amplypterus ypsilon R. & J. Callionima falcifera (Gehlen) Cautethia spuria (Bdv.) Cautethia yucatana B. P. Clark Cocytius dupnnchel (Poey) Enyo ocypete (L.) Erinnvis ello (L.) Erinnyis lasauxii (Bdv.) Erinnvis oenotrus (Cram.) Eiimorpha anchemola (Gram.) Eumorpha satellitia (L.) Eupyrrhoghssum sagra (Poey) Isognathus rimosus (Grt.) Manduca barnesi (Clark) Manduca coraUina (Drc.) Manduca dilucida (Hy. Edw.) Manduca florestan (Cram.) DHJ 12071 (Convolvulaceae) Doliocarpus dentatus (Aubl.) Stand. (Dilleniaceae) Tetracera volubilis L. (Dilleniaceae) Ocotea veraguensis (Meisn.) Vlez (Lauraceae) Ocotea veraguensis (Meisn.) Vlez (Lauraceae) Stemmadenia obovala (Hook & Arn.) K. Schum. (Apocynaceae) Exoslema mexicanum A. Gray (Rubiaceae) Coutarea hexandra (Jacq.) Schum. (Rubiaceae) Exostema mexicanum A. Gray (Rubiaceae) Annona purpurea Moc. & Sesse (Annonaceae) Annona reticulata L. (Annonaceae) Tetracera volubilis L. (Dilleniaceael Cissus rhombifoUa Vahl. (Vitaceael Sebastiana confusa Lundell (Euphorbiaceae) Sapium thelocarpum Schm. & Pitt. (Euphorbiaceae) Manilkara chicle (Pittier) Gilly (Sapotaceae) Sarcostemma glauca HBK. (Asclepiadaceae) Fosteronia spicata (Jacq.) G. Mey (Apocynaceae) Cissus rhombifoUa Vahl. (Vitaceae) Cissus sicyoides L. (Vitaceae) Cissus rhombifoUa Vahl. (Vitaceae) Cissus sicyoides L. (Vitaceae) Chomelia spmosa Jacq. (Rubiaceae) Guetlarda macrosperma D. Sm. (Rubiaceae) Plumeria rubra L. (Apocynaceae) Godmania aesculifolia (HBK.) Standi. (Bignoniaceae) Cordia allwdora (R. & P) Oken (Boraginaceae) Sapranthus patanga Fries (Annonaceae! Annona reticulata L. (Annonaceae) Pithecoctinium crucigerum (L.) A. Gentry (Bignoniaceae) Cydista hetcrophylla Seib. (Bignoniaceae) Tabcbuia ochracca (Cham.) Standi. (Bignoniaceae) Callichlamys latifoUa (L. Rich) K. Schum. (Bignoniaceae) Arrabidaea chica (H. & B.) Verl. (Bignoniaceae) Cornutia grandifolia iSchlecht. & Cham.) Schau. (Verbenaceae) Ceratophytum tetragonolobum (Jacq.) Sprague & Sandw. herb vine low perennial vine low perennial vine sapling to treelet sapling to treelet sapling to treelet sapling to large tree treelet treelet sapling to treelet sapling to treelet low perennial vine herb vine sapling to treelet sapling large tree low vine low perennial vine low perennial vine low perennial vine low perennial vine low perennial vine sapling to treelet sappling to treelet large tree sapling sapling to lai'ge tree sapling to treelet sapling to treelet low' perennial vine low perennial vine sapling low perennial vine low perennial vine shrub low perennial vine i September 1985 125 Years of Biological Research Table 3. — Continued 159 Larva Host Microhabitat Manduca lefeburei (Guer.) Manduca muscosa (R. & J.) Manduca occulta (R. & J.) Manduca rustica (F.) Manduca sexta (L.) Neococytius cluenlius (Cram.) Nyceryx coffeae (Wlk.) Pachygonia drucei R. & J. Pachylia ficus (L.I Plenontoma variabilis (Jacq.) Miers 160 Illinois Natural History Survey Bulletin Table 3, — Continued Vol. 33, Art. 3 Larva Host Microhabitat Pachylia syces (Hbn.) Pachylioides resumens (Wlk.) Perigonia lusca (F.) Prolambulyx strigilis (L.) Pseudosphinx tetrio (L.) Sphinx merops Bdv. Unzela pronoe (Drc.) Xylophones anubus (Cram.) Xylophones ceratnmioides (G. & R.) Xylophones chiron Dry. Xylophones juonito R. Xylophones moculator (Bdv.) Xylophones pluto (F. ) Xylophones parens (Hbn.) Xylophones turbala (Hy. Edw.) Xylophones tyndariis (Bdv.) Chlorophora tinctorio (L.) Gaud. (Moraceae) Castillo etastico Cerv. (Moraceae) Ficus ovalis (Liebm.) Miq. (Moraceae) Forsteronia spicoto (Jacq.) Mull (Apocynaceae) Colcophyllum candidissimum (Vahl.i DC. (Rubiaceae) Guellarda macrosperma D. Sm. (Rubiaceae) Astromum graueotens Jacq. sapling to large tree sapling large tree low perennial vine sapling to large tree sapling to treelet (Anacardiaceae) September 1985 125 Years of Biological Research 161 revealing addendum to this pattei-n is that if those same saturniid caterpil- lars are placed on small saplings inside screen nets, they develop into quite nor- mal adults. This is true whether they are transferred as first-instar or any later-instar caterpillars. I have done this with Rothschildia lebeau on Spon- dias mombin L. and Xanthoxylum setu- losum P. Wilson; R. erycina (Shaw) on Exostema mexicanum A. Gray and Cou- tarea hexandra (Jacq.) Schum.; Eacles imperialis (Dry.) on Cochlospermum vitifolium (Willd.) Spreng.; atheroma lobesis W. Rothschild on Calycophyllum candidissimum (Vahl.) DC, Spondias mombin, and Cochlosperum vitifolium; Arsenura armida (Cram.), Caio cham- pioni (Drc), and Titaea tamerlan (Maas- sen) on Bombacopsis quinatum (Jacq.) Dugand; Othorene purpurascens (Schaus) on Manilkara chicle (Pittier) Gilly; Othorene verana (Schaus) on Quercus oleoides; Syssphinx molina (Cram.) on Pithecellobium saman (Jacq.) Benth., Cassia grandis L., and Albizzia adinocephala (Donn. Sm.) Britt. & Rose; Syssphinx colla (Dyar) on Pithecellobium saman; Adeloneivaia isara (Dognin) on Lysiloma divaricata (Jacq.) Macbride and L. auritum (Schl.) Benth.; and Dirphia avia (Stoll) and Schausiella saniarosensis Lemaire on Hymenaea courbaril. The net protects the caterpillars from carnivores but does not, I assume, seriously modify the microclimate of the plant's physi- ology. The opposite experiment, putting these species of saturniid larvae on a host plant sapling in nature without a protective net, i-esults in rapid removal of the catei^pillars by ants, wasps, birds, scorpions, mammals, and spiders. The few survivors are very often parasi- tized. I am in the process of conducting large-scale experiments which will yield an understanding of the variation and how it compares with the mean and variance of the same experiments in adult host plant crowns. However, a few preliminary results are instructive. On 18 July 1983, 1 placed 340 first- instar, newly hatched Rothschildia lebeau caterpillars on a 4-m tall bushy Spondias mombin sapling growing in a tangle of secondary succession. The lower trunk was ringed with Tangle- foot, and the surrounding vegetation was trimmed so that the host plant did not contact any other plant. By 8 August, 21 days later, only 13 cater- pillars remained, all late penultimate instar. On 15 August, 3-5 days before they were due to wander off the plant to pupate, I removed the 12 remaining survivors, all of which spun cocoons. Of these caterpillars, 5 contained a large ichneumonid wasp larva (Enicospilus), 3 contained larvae of tachinid flies, 3 died of disease inside their cocoons be- fore they could pupate, and 1 produced a healthy pupa. On 30 June, I placed 120 third- and early fourth-instar larvae of Syssphinx molina on 60 sapling Pithecellobium saman growing in roadside secondary succession within 100 m of adult P saman. Shortly before these larvae (Fig. 5A) were due to leave the plant to pupate in the litter and only 11 days after they had been placed out, I col- lected the 14 survivors. Three of these contained larvae of a species of Thyre- odon, an ichneumonid wasp apparently specific to this caterpillar and its close relatives, and six produced healthy pupae; the remainder died of disease. As yet I have no extensive data on the survivorship o{ Rothschildia lebeau or Syssphinx molina caterpillars in the crowns of their large tree host plants, but numerous anecdotal observations suggest that it will be gi-eater than the low levels of survival mentioned above. There is a very conspicuous excep- tion to the generality that saturniids are concentrated in the crowns of adult trees. Hemileucine saturniids, the io moth to extratropical New World read- ers, are found both on juveniles of their hosts and on species of plants that are normally small as adults and are there- fore imbedded in low secondary succes- sion (these caterpillars are found as well in the crowns of large trees; see Table 2). Hemileucine caterpillars (Fig. 5B, C) are the most severe urticators of 162 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 M X. £**!: 'isej- FiQ 5 - (A) Ultimate-instar larva of Syssph/nx moban: this green morph has a greenish-white lateral s,r,pe ?eL ,Sac"ns (scol,), and po^er^r dorsal gold too,hin.e sp.nes, (B) ,^1— "-^-^^jf-| larva o Peuphoba arcaer. this green caterpillar has a pale whitish-green lateral line with a lavender ^ne usfa™! (C) UKimate'instar urticating larva of Automens lo. this green caterpillar has a white latera !'ne wra dark red lateral line ,ust above i? and lower lateral dark red patches with white dots, JulyAugust 1983, Santa Rosa National Park. Costa Rica, September 1985 125 Years of Biological Research 163 the saturniid world. As such they are largely invulnerable to a major subset of the vertebrate carnivores that prob- ably find it more profitable to search for food in secondary succession than in the huge monospecific expanses of leaves in the forest canopy. It is too early in this investigation to be able to offer a clean hypothesis about what allows sphingid caterpil- lars to feed on such a diversity of plant life forms (Table 3), especially on the small ones. However, some possibilities come to mind (cf. Janzen 1984b). 1. Sphingid adults live for weeks to months (as compared with 3-10 days for saturniid females) and therefore have more time to seek out widely scat- tered small individual hosts on which to lay only one or very few eggs at a time; with only a few nights to search for a host, a saturniid female may find it more profitable to search for a few large crowns and lay many eggs on each than to have to locate many saplings. 2. By being initially widely scat- tered, sphingid caterpillars avoid dens- ity-dependent effects, such as ai'e likely to occur when, for example, a Polistes wasp finds one of many saturniid cater- pillars on a plant and repeatedly re- turns to harvest them. 3. Because they are extremely host- specific, sphingid caterpillars are on average more cryptic than are satur- niid caterpillars and therefore have a gi-eater chance of surviving the visu- ally oriented component of the more in- tense carnivory that, I suspect, occurs in the low vegetation of secondary succession. 4. Sphingid caterpillars grow more rapidly to a given size than do satur- niids of the same body weight; they are therefore exposed to agents of larval mortality for a shorter period than are saturniid caterpillars. 5. Sphingid caterpillars may have chemical or physical defenses against invertebrate carnivores, defenses that saturniids lack. I have seen the same Polistes that aggressively harvests young Caio championi caterpillars from sapling Bombacopsis land on Manduca sexta (L.) caterpillars and simply walk on in search of food, appar- ently treating the caterpillars as though they were part of the plant. Three species of "primitive" sphin- gids (Smerinthinae) at Santa Rosa offer instructive exceptions to the generality that sphingids and saturniids have dif- ferent hosts at Santa Rosa. Protambu- lyx strigilis (L.) caterpillars are com- monplace on Spondias mombin and Astronium graveolens Jacq., ranging from 1.5-m saplings to the crowns of 20-m, full-gi'own trees; they are found on all sizes of their hosts from forest understory to tree falls to large ex- panses of recent secondary succession along roads and in old fields. They are heavily parasitized by tachinids on all of these plant life forms and in all habi- tats. In forests, and on forest edges, Rothschildia lebeau (Saturniidae) cater- pillars are found only in the crowns of medium-sized to large Spondias mom- bin; it is extremely difficult to imagine what besides predator pressure could have been responsible for the evolution or ecology of this restriction to adult S mombin. However, since female R. lebeau do occasionally oviposit on sap- ling S mombin, and I have on seven oc- casions found well-developed R. lebeau larvae on S. mombin saplings in secon- dary succession, it may be that there is little evolution involved. Rather, the R. lebeau females may be ovipositing throughout the S. mombin population, but carnivores may be eliminating those on the saplings. If this is the case, were S mombin the only R. lebeau host plant in the park and were S. mombin a small treelet only found in early secondary succession, R. lebeau might well be absent from the park. It is also evident that P strigilis is substantially more cryptic than is R. lebeau when on •S mombin. Amplyterus gannascus (Stoll) and A. ypsilon R. & J., the other two smer- inthine sphingids at Santa Rosa, feed exclusively on Ocotea veraguensis 164 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 (Meisn.) Vlez. O. veraguensis is a shrub- by subcanopy treelet found throughout the understory of several forest types at Santa Rosa. It is the only lauraceous plant in the habitat. A. gannascus and A. ypsilon are encountered feeding on the foliage of O. veraguensis at all heights, from 1-m saplings to the tops of the crowns of 8-m tall large adult plants. The saturniid Copaxa moinieri Lemaire has 0. veraguensis as its sole host at Santa Rosa. In contrast to other saturniids, C. moinieri is conspicuous in feeding entirely on the saplings and lower branches of its host and is virtua- ly never found more than 3 m above the gi'ound. O. veraguensis may range from forest edge to heavily shaded forest understory, but the vast majority are in the forest understory. However, if the caterpillars are placed on foliage in crowns of adult 0. veraguensis under protective nets or on similar foliage in plastic bags in the laboratory, they develop quite normally. Here again, the sphingid feeds throughout the life forms of its host plant, while the satur- niid is restricted to a particular subset of its host population. While as yet untested, one selective pressure for this behavior is easy to postulate. Since the host is a forest understory plant, feed- ing solely on the foliage within about 3 m of the ground places caterpillars in a zone quite free of foraging by white-faced monkeys. White-faced mon- keys are eager to eat large caterpillars, and in the last instars, the catei'pillars of C. moinieri are quite easy to find because of the characteristic damage that they do when feeding. This is so when there is a defoliation event, at which time the leafless shrubs are cleaned of caterpillars by the human collector, and presumably would be as well by insectivorous monkeys if they foraged close to the gi-ound. There are at least six other species of Copaxa in Costa Rica. A summai-y of Copaxa host records (Janzen unpub- lished manuscript) suggests that they are specialists on Lauraceae and may be able eat the foliage of any species of Ocotea, Nectandra, or Persea. Why then don't they occur at Santa Rosa, where O. veraguensis is common? At least three species of Copaxa occur only 10- 13 km to the east in a more evergreen forest that is continuous with the Santa Rosa forest. Could it be that these other species of Copaxa feed higher in the foliage of their host plant and therefore don't survive at Santa Rosa because this zone is lethal on O. veraguensis in the dry forest of lower stature, as found at Santa Rosa? In short, when a sphingid hops to a new host plant, it gets more than the plant's chemisti-y. The plant's nutrient value may be high and its defenses may be easily bypassed by existing chemical pathways, but if a gi'een catei-pillar has to sit on a brown leaf all day, its fitness on that plant may well be zero. Like- wise, if a species of saturniid caterpil- lar is essentially immune to most ver- tebrate carnivores because it has severely urticating spines, it has a major trait that predisposes it to the evolution of the biochemical ability to feed on a variety of host species. This appears to be the case with the hemi- leucine saturniids. They, severe urtica- tors, are found on many kinds of leafy backgi-ounds, as indicated by their long host lists. On the other hand, I do not wish to denigi-ate the impoilance of plant chemisti-y in all of this. If a caterpillar lineage comes up with a high quality general-purpose detoxifying system, as the ancestral species to the hemileu- cines must have done, then the mutant that subsequently invented urticating hairs had the companion trait neces- sary to allow the explosive radiation that produced and maintained the most species-rich and widespread of the four New World saturniid subfamilies. Over half the New World saturnids are hemileucines; the most species-rich genus, Hylesia, with over 200 species (C. Lemaire personal communication), has not only severely urticating cater- pillars (see Janzen 1984a for a review), but aggi'egating caterpillars as well. September 1985 125 Years of Biological Research 165 CAN A CATERPILLAR GET A FREE RIDE? A major aspect of the evolutionary accumulation of a plant species' her- bivore load is the intensity of fitness depression of the plant by an evolution- arily incoming caterpillar. Put most simply, the more severe is the caterpil- lar's impact, the more likely is the newcomer to depress the density or local distribution of the plant to a level at which it will not support the cater- pillar population or will select for a defense trait that will evict a portion of the herbivore load, or both. In short, one expects that the smaller the bite the herbivore takes, the more likely it is to be able to persist on a new host plant. But bites are measured not in gi-ams of seeds or area of leaves, but in units of population change or fitness depression. There is, however, one rarely dis- cussed way in which a herbivore can take large bites without depressing the fitness of the host plant. If the her- bivore feeds on a subset of the host population that is fated to die prior to reproduction irrespective of the her- bivore's impact, it will have no fitness- depressive effect whatsoever. Addition- ally, if it feeds on a plant that will live to I'eproduce, then the fitness-depres- sing effect of its herbivory should be directly proportional to the fitness of that plant relative to its conspecifics. In other words, the worst way to move evolutionarily among host species or increase your host list is to take big bites out of the individuals of the population with the highest potential fitness. At the opposite extreme are the herbivores that feed on plants that are clearly genetically dead though physio- logically living - the gi-een aborted leaf or fruit on the forest floor, the branch newly broken out of the tree crown, the seed defecated on the floor of a dry bat cave, the annual herb that has matured and shed its seeds, and so on. A detritivore is defined as an organ- ism that feeds on a dead plant (or animal), and from the viewpoint of evolutionary ecology, dead is defined as having zero potential fitness. While no living wild plant has zero probability of surviving to first reproduction, it is certain that the probability of attain- ing any given fitness value is neither distributed evenly among the members of the population nor at random with respect to a multitude of environmen- tal variables. The questions become whether sufficient cues exist by which herbivory can be evolutionarily or ecologically directed so that its impact is reduced, and whether individual selection can generate the necessary traits. To ask the question in a different way, are there caterpillars with little or no impact on their host populations simply because they feed (for whatever reason) on that subset of the plant population that has severely reduced potential fitness? Such a set of questions calls atten- tion to a major difference between plants and animals as hosts or prey. Since animals move around, it is harder to identify individuals, cohorts, or arrays that have a high likelihood of not attaining full adult status than is the case with plants. Despite the fact that we all know about this difference, ecologists have not developed either a terminology or conceptual base with which to discuss easily its potential significance to competition, herbivory, edaphic responses, etc. Before examining some examples at Santa Rosa, a hypothetical example is in order. First, consider the case of a monophagous caterpillar species that is distributed over its host plants in a manner unrelated to the probability of any particular plant's attaining its potential fitness. Assume the cater- pillar to be on 20 percent of the host individuals and that each caterpillar set (on each plant) eats half the leaves off its plant before pupating. The 10 percent of the plant population with a moderate chance of becoming adult (e.g., individuals growing in tree falls) will have 20 percent of its individuals 166 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 damaged to this degi'ee. The remaining 90 percent of the plant population will be damaged likewise, but assume that only 1 percent of these plants have a chance of surviving (i.e., those growing where a tree will fall). With respect to natural selection, the selective pres- sure for a defense trait that will repel the caterpillar is generated by 50 per- cent herbivory of 20 percent of 10.9 percent of the individuals, while the amount of leaf eaten by the caterpillar population is 50 percent of 20 percent of 100 percent of the individuals. In short, enough leaf material to sustain a species of caterpillar may be harves- ted with only the amount of impact that would have been generated by a much less voracious species of caterpil- lar feeding on a plant population with a high proportion of individuals with a high chance of attaining their poten- tial fitness. Alternatively, if the incom- ing caterpillar fed only on individuals with a high chance of surviving, it would have the same impact with much less herbivory that it would have if it also included ill-fated conspecifics in its diet. This scenario may also be stated as, "The more the caterpillar population is focused on ill-fated plants, the more likely it is to persist." At the limit, the caterpillar feeding on shed leaves is totally free of evolu- tionary responses by the plant to its feeding. Likewise, as the percentage of the plant population made up of ill- fated plants rises, there is a rise in the amount of herbivore load that will be sustained free of selective chai'ge if the herbivores are distributed at random or focused on the ill-fated plants. Two kinds of ecological complexity should be added to this hypothesized process. First, if the herbivory raises the probability that a plant will die before reproduction or that it will be a less prolific reproducer, there is the complication that the same amount of herbivory is likely to lower the realized fitness of a weakened plant more than that of a healthy one. This differential response may alleviate some of the ill- fated plant effect (eliminate some of the free ride) by rendering the herbivore less of a detritivore and more of a predator or parasite. Second, while the herbivore sub- sisting largely or totally on ill-fated plants may not be selecting for traits to repel itself even if this leads to earlier plant death than would other- wise have occurred, it will still be ecologically thinning out its host plants. Furthermore, it will have to sur- vive despite the defenses selected for by other more fitness-reducing species of herbivores, just as must the more con- ventional detritivores consuming litter. It is critical to recognize that the selective process that may result in a larger herbivore load on ill-fated in- dividuals or in larger herbivore loads on species that normally have a high proportion of ill-fated individuals does not favor local mutant herbivores for this trait per se. The individual hei-- bivore that feeds on an ill-fated plant in preference to a plant with high po- tential fitness does not raise its own fitness. Rather, the selective process is that of the plant suffering or repelling herbivores that are serendipitously arriving through the evolution of other traits or by immigi-ation. What fraction of the leaf-eating herbivores in Santa Rosa habitats are largely supported by ill-fated plants with little or no potential fitness? Fur- thermore, what proportion of the her- bivores subsists on ill-fated plants because those plants were actively sought by ovipositing females or be- cause their habitats are so sought? In both cases, I assume that the selection for this choice was not driven by a selective value to the individual female ovipositing on the plant with a low potential fitness. It is easy to imagine that many of the Santa Rosa seed predators subsist on seeds that have zero fitness because of their site of deposition following dispersal (or lack thereof). It is much more difficult to divine the potential fitness of the young gi-owing plants whose leaves are September 1985 125 Years of Biological Research 167 important food for so many caterpillars at Santa Rosa. There certainly are many species of caterpillars at Santa Rosa that wholly or largely feed on saplings of forest canopy trees in the deeply shaded understory where the host has virtually no chance of sm'vival to reproduction unless a tree fall occurs at that site. I cannot yet offer data on the influence of this herbivory on how long a plant can wait before finally dy- ing of starvation, distorted stems, disease, etc. CLIMATE IS A HOST-PLANT TRAIT The physical environment undoubt- edly influences caterpillar fitness. We tend to forget that each local popula- tion of plants, and even distinctive subsets of the population, is associated with a distinctive microclimate. The caterpillar has to contend with this as much as with the directly genetically progi-ammed traits of the plant and the plant's distinctive carnivory I'egime. Variation in this climatic regime over a plant's distribution may as well as responsible for the absence of a cater- pillar in parts of the plant's range as is geographic variation in host plant chemistry. That the saturniid moth Schausiella santarosensis is apparently found only in northwestern Guana- caste Province, Costa Rica, while its host tree, Hymenaea courbaril, ranges from southern Mexico to central South America, could have many causes be- sides local uniqueness of H. courbaril chemistry. Among these causes may be the Santa Rosa climate. At Santa Rosa, there are many ex- amples of climatic components of what might be termed the suitability profile of caterpillar host plants. One that is annoyingly omnipresent is the failure of many species of saturniids to have at least the minimum of three genera- tions per year that could occur if host foliage properties were the only challenge to the caterpillar. While in theory such failures may be equally attributable to seasonal change in carnivory risk or leaf chemistry, there are times when climate seems to be the likely primary cause. The four saturniids that feed on the foliage of H. courbaril offer an example. This tree bears resin-rich leaves (e.g., Langenheim et al. 1982) throughout the year, except for approximately 2 weeks in December or January (early dry season) when the old leaf crop is dropped and a new one is put out immediately thereafter. All four species of saturniids that eat H. cour- baril leaves can be reared readily under nets in the field or in plastic bags in the laboratory on foliage collected at any time of year (unpublished field notes). H. courbaril foliage of all ages is eaten by all four species in nature and in the laboratory. Throughout the dry season, from at least January until shortly before the first rains in late April or early May, three of the four saturniids are dormant in cocoons or as pupae in the litter below the fully leafed adult H. courbaril, and Hylesia lineata Drc. is dormant as eggs in an arboreal felt nest constructed by the female (Janzen 1984a). Dirphia auia anticipates the rains and times its emergence so that ovipositing females are in the forest as much as a week before the rains. The eggs require 2-3 weeks to hatch, with the consequence that the caterpillars begin feeding during the rains. Schausiella santaro- sensis, Periphoba arcaei (Drc), and Automeris zugana Drc. adults emerge within 3 weeks after the first heavy rains, and their first-generation cater- pillars are present for the first 2-3 months after the rains begin. H. lineata eggs hatch shortly after the rains begin as well, with the same consequence. Most of the pupae of the first genera- tion of S. santarosensis remain dormant in the variably wet and dry litter below the H. courbaril for 9-10 months before eclosing to repeat the process the next year; however, a small fraction of these 168 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 pupae eclose about 3 weeks after pupa- tion and constitute a second generation during the second half of the rainy season. This latter generation again demonstrates that the leaves of H. courbaril are chemically quite suitable as food for S. santarosensis at this time of year. The pupae of this second generation then join those of the first generation in remaining dormant below the leafy trees throughout the dry season. The other three saturniids regularly have two generations on the rainy season H. courbaril leaf crop (as well as on other species of hosts) and then become dormant at the end of the rainy season. It is hard to avoid the working hypothesis that the reason that these moths ignore a food souixe on which they could pass at least two more generations is because the dry season heat and wind, and attendant desiccation, are inimical to their development in the crown of a large tree. The first-instar larvae are prob- ably the most susceptible, but desicca- tion during molting and even during pupal formation may also be a problem. Larval mortality from desiccation will probably turn out to be a major reason for delaying oviposition until the rains start, even when the host plant is in full leaf. For example, at the beginning of the 1983 rainy season (late May), there was sufficient rain to cause some Cochlospermum vitifolium to leaf out. This plant is a natural host of Eacles imperialis (Table 2). I placed hundreds of 1 -day-old first-instar E. imperialis larvae on these leafy C. vitifolium. The plants were then naturally subject to winds, high tem- peratures, direct insolation, and dry air as severe as was characteristic of mid- day a few weeks earlier during the dry season. It was dry enough to stop fur- ther Cochlospermum branch elonga- tion. While the caterpillars did feed, few obtained enough water to replace that lost by desiccation, and I watched them quite literally dry up and blow off during the following 2 days. Two weeks later, after rainy season humidity, cloudiness, and still air had set in, similar inoculations of the same plants resulted in nearly 100-percent larval survival (the larvae were protected from carnivores with netting in these experiments). The same change occur- red with Rothschildia lebeau first- instar larvae put on insolated sapling Spondias mombin during the same days. For both species, the foliage of their host plants during the dry weather was excellent fodder when given to siblings of the dead cater- pillars in plastic bags in the labora- tory. In short, at the beginning of the rather erratic 1983 rainy season, the leaves of major host plants of E. imperialis and R. lebeau were present and edible but, in effect, unavailable. In addition, there was spatial hetero geneity to their availability. Inocula- tion experiments, using siblings of the victims at the beginning of the raim* season, were quite successful when the young caterpillars were placed on host plants that happened to be gi'owing in the shade and wind-buffered air of a patch of evergi-een trees. Had these shaded plants been of a different species, one of the other host species of these two moths, it would be easy to conclude that these ecologically available species of host plants had dif- ferent weather regimes from those of C. vitifolium and S. mombin. thereby giving them a longer period of suscep- tibility to the caterpillars during the year It is likewise easy to imagine that this different regime might well make them an acceptable host for a species of caterpillar that has no other species of host in the habitat. The weather becomes a trait of the host plant in more ways than just its direct effect on the caterpillar. By determining the duration of leafiness, the weather determines such things as the number of caterpillar generations possible. This may in turn determine the length of a generation, because the September 1985 125 Years of Biological Research 169 number of generations per rainy season must be an integi'al number. The length of a generation may in turn determine which hosts are too good and which are inadequate. An inadequate host may be one that the caterpillar can eat with impunity, yet lacks suffi- cient nutrients for the caterpillar to attain a reasonable size by the time it must pupate if it is to remain in syn- chrony with the remainder of its population. A too-good host might be one that makes the caterpillar gi-ow too fast; however, in this case, I would expect quick selection for a fixed larval period coupled with heavier cater- pillars on better host species or indi- viduals. Viewed the other way around, a given nutrient and defensive chem- istry may generate a moth egg-to-adult cycle that is incompatible with the seasonality of the site even if the resultant adult moths are quite normal in weight and other physiological para- menters. DISCUSSION A host plant is not the focus of the ecological and selective regime of a moth and its caterpillars. Rather, the focus is a space bounded by parameters involving carnivores, climate, host chemistry, host fitness depression through herbivory, and a variety of other traits. In a certain sense I am defining a Hutchinsonian niche for a caterpillar and arguing that this niche is not centered on any one of these traits. This notion is emphasized by recognizing that in any particular point in season, space, or ontogeny there are many individuals of the caterpillar's host plant population that are not being fed on by that species of caterpillar; this distribution is not necessarily generated by randomly placed misses, but because of predict- able caterpillar failure or disinterest under this or that circumstance. A par- ticular host is necessary, but not suffi- cient, for caterpillar presence. But is it even necessary? Why not eat just any plant? The current answer, and I am sure the largely correct one, is that in any given snapshot in evolu- tionary time, a caterpillar species is genetically (and subsequently, ecolog- ically as well) programmed so that it feeds on only a few species of plants in the habitat; female oviposition be- havior is here viewed as merely an ex- tension of caterpillar behavior How- ever, there are at least two ways in which this answer is probably incom- plete. First, and I think this is gen- erally accepted by students of animal- plant interactions, there are probably more species of plants in the habitat that can be physiologically processed by the caterpillar species than the caterpillar species actually eats in nature. I have discussed here some of the reasons why these plants may not be used as hosts; the physical climate, predators, etc., may stop the evolution of the choice of that plant, and current ecological processes may prevent its use even if it is chosen by the ovi- positing female. Second, given the right collection of ecological circumstances, perhaps just about any plant would be a suitable host in ecological time and certainly in evolutionary time. In other words, is it unthinkable that, if a species of moth were con- fronted with a habitat rich in plant species but quite lacking in carnivores (including diseases), inclement wea- ther, competitors, and the ability of plants to evolve further, it might well evolve a digestive physiology such that any individual caterpillar could feed on many or even all plant species? We can imagine that the genetic progi-amming might be too complex and that the biochemical machinery somewhat in- compatible for a caterpillar to contain all the collective food-processing abilities of the several thousand species of caterpillars now feeding on the plants of Santa Rosa. But what if, 170 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 instead, the caterpillar simply devel- oped a series of biochemical filters that, albeit slowly and incompletely, simply extracted the easily removed nutrients from the food and passed the rest on out? At Santa Rosa, Hypercompe icasia Cram. (Arctiidae) has over 60 species of broadleafed host plants recorded for it to date (in half as many plant families) and is one of the most slow- growing caterpillars I have ever reared. The sympatric H. suffitsa Schaus has more than 50 recorded hosts, and there is almost no overlap of its host list with that of//, icasia; it not only gi'ows very slowly, but if the food dries up, it spins a flimsy silk nest and waits until more green leaves appear. It can wait as long as 3 months at ambient temperatures. Why do most other Santa Rosa cater- pillars appear to lack such digestive abilities? I doubt that it is because Hypercompe has invented some miraculous digestive system never hit upon by other Lepidoptera. I view Hypercompe not as the epitome of anything, but simply one of many ways to cope with the wide variety of selec- tive and ecological pressures that con- front a caterpillar. I have mentioned quite enough complications in this essay that it seems truly a miracle for there to be any patterns at all. However, patterns do exist. The large moths at Santa Rosa — saturniids and sphingids - can be partitioned with respect to many of the traits mentioned here. I have already mentioned that sphingids are generally found in carnivore-rich mixes of small plants in low vegetation as well as in the crowns of lai'ge trees, while saturniids are much more cater- pillars of the crowns of large trees. Sphingids rely heavily on crypsis specific to the host plant (or even leave the host plant during the day), while saturniids have both ostentatious real defenses and mimicry in addition to a more generalized crypsis. Closely linked to this contrast, sphingid cater- pillars are more host-specific than are saturniids and grow faster than do saturniids. On the one hand, it is nothing new to stress that a host plant is moi-e than just its chemistry. But the topic needs stress now for two reasons. First, most analyses of nonchemical traits (e.g., plant apparency, size of the plant, or plant population as an island) have examined them as lai'ge-scale attri- butes, with the stress being on how these traits correlate with some caricature of host plant chemistry or numbers of insect species (e.g., Southwood et al. 1982). Such analyses are an essential part of the story but bypass the fine-scale interactions between species and individuals (e.g., Karban and Ricklefs 1983; Lawlon 1982; Rausher and Papaj 1983: Connor et al. 1983) and the small-scale non- chemical traits. The nonchemical traits desei've the same detailed attention as specificity determinants that secon- dary compound chemistry is now receiving. Second, the topic needs stress because of the inherent difficul- ty in the technology' of the subject. One cannot determine a caterpillar"s risk of being eaten, for example, from any intrinsic trait of the plant (or the cater- pillar) that can be obtained from laboratory analyses of collected materials. Equally, the depression of host plant fitness by a caterpillar population cannot be determined from any compilation of facts about plant species, including vegetation analyses, percentage of herbivory, percentage of cover, and all those other things tradi- tionally measured. These traits are situation dependent. They have mean- ing only in context. Here it is appropriate to digi'ess on the subject of the concept of average or baseline herbivory. Plants live a long time. Grant proposals are for 3 or fewer years, and field studies are rarely more than one to two times this duration. A commonplace herbivory regime is small amounts for a run of years, with severe defoliation occurring at long and September 1985 125 Years of Biological Research 171 irregular intervals. Since severe defoliation events are sufficiently far apart that any one study usually sees none or only one of them, there is a strong temptation to view them as "unusual" and therefore not to be considered in determining the intensi- ty of herbivory. While it does not make biological sense to average the years of peak herbivory with the many more years of much lower herbivory, somehow the impact of the plant's life- long herbivory regime needs to be taken to be the same norm. In characterizing the entire herb- ivory regime with its causes for varia- tion, one finds that many of the nonchemical traits of the plant take on great importance. Leaving aside herbivore-induced changes in chemis- try (e.g., Edwards & Wratten 1983), the defense chemistry and nutrients in a particular set of leaves are relatively invariant as causes of herbivory fluc- tuation when compared with other traits. Most of the individual cater- pillars that actually attempt to feed on a given species of plant either can or cannot do it. And if they can, they can do it over much or all of a plant's lifespan. It is easy to forget that much of a plant's herbivoi'e defenses are against those animals that, because of these defenses, are virtually never responsible for any leaf loss. We are in the silly situation that much of the study of secondary compound chemical ecology occurs in the vacuum of having no idea what herbivores selected for the defenses, and the intensity of much of the herbivory we measure is deter- mined and patterned by events gener- ally not measured by ecologists that study herbivoiy. The ultimate irony is that just as in the defense budgets of nations, the largest cost levied by herbivores is probably the cost of the defenses rather than the amount of material removed during herbivory (Janzen 1983b). I have argued that a multispecies mix of small plants has a very different carnivory regime than does an equal- sized large tree crown, which is essen- tially a monospecific stand. For almost all biologists, the closest approxima- tion of the multispecies mix of small plants is roadside and old field secon- dary succession. Such habitats are often analogized with natural disturb- ance sites such as tree falls, river banks, landslides, etc. I would only emphasize that the analogy is not a close one and likely to be particularly deceptive with respect to the case at hand. For example, the old field does not have nearby intact forest as an herbivore reservoir, refuge, and barrier between small portions of the old field. Likewise, ti'opical roadsides are, for example, generally continually bathed in dust, which is one of the best con- tact pesticides known (especially for small insects like parasitoids). For example, the only reason that the Syssphinx molina experiments men- tioned earlier could be conducted was that they were conducted along a rarely traveled paved road with no shoulders. It is clear that any herbivore enter- ing a new habitat is stepping into a battlefield rich in plant defenses se- lected for by other herbivores and by the constraints of defense economics. This situation applies even more to the nonchemical traits of a host plant, the traits I have been discussing; many of the traits of importance were not even selected for by biotic agents, to say noth- ing of herbivores. The opportunities for coevolution have been minimal. The traits that render a Hymenaea cour- baril crown a low-yield area for a carni- vore were certainly not selected for by the four saturniid caterpillars that feed with relative impunity on the leaves in large H. courbaril crowns. It is unlikely that host leaf colors, shapes, and sizes, are selected for through their render- ing caterpillars of a particular species easily found by carnivores. Unfortunately for evolutionary ecology, it is at least theoretically pos- 172 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 sible for natural processes to produce a habitat full of interacting caterpillars and their hosts with no coevolution and even very little (if any) evolution. This is particularly true for the nonchemical traits of plants. One can imagine an arriving caterpillar finding a number of edible species of plants but ending up using only a small fraction of the species and biomass because of the negative effects of processes like those mentioned in this essay. Whether the caterpillar persists until it is evolu- tionarily modified by the many new selective pressures put upon it should depend largely upon nonchemical fac- tors, if there is even one species of plant present that the caterpillar can eat. The immigi'ation placement of a caterpillar in a new habitat, and its ecological adjustment to the plants there, has much in common with the process of the introduction of new pollinators, new dispersal agents, animal-dispersed plants, etc. At first, the persistence of the invader (or a resi- dent newly deprived of its interactant) depends on other organisms with traits similar to those of the organisms it left behind, even if the invader cannot achieve the same level of fitness that it had originally. At least theoretically, the interactions could then continue indefinitely with no evolution - though the population structure of the incoming species might be quite dif- ferent than it was where it came from (as might also be those of its new associates). In fact, it is easy to imagine a species being passed from interactant to interactant over long evolutionary time, with a truly new phenotype appearing only rarely but with largely the same phenotypes of organisms con- tinuing to interact in this or that habitat (e.g., Janzen 1983c). Perhaps evolutionary rates proceed most rapid- ly where species richness is lowest, since in such a situation the probabili- ty that alternate or substitute interact- ants ai'e present is lowest. Yet, there has to be enough diversity of substrate for the immigi'ant at least to get established. At the other extreme, if there are many similar species in a habitat, the addition of one more may result in virtually no discernible evolu- tionary change, because each of the ecologically similar species adjusts just a small amount. SUMMARY To the herbivore evolutionarily moving onto a host, or surviving there, a host plant has, in addition to its chemistry and other personal traits, a carnivore regime, weather regime, and fitness regime that must be overcome by the herbivore. The interplant heterogeneity of a Costa Rican defolia- tion of Randia treelets by Aellopos titan sphingid caterpillars is described as due to heterogeneity in the carnivory regime rather than in the interplant chemistry. That Costa Rican saturniid caterpillai-s feed primarily in the crowns of large trees while in the same habitat sphingid caterpillars are distributed over many more plant life forms and sizes is attributed to satur- niids (except the severely urticating species) being comparatively incompe- tent at contending with the more intense carnivory regime of dense, low, and plant-species-rich vegetation rather than to saplings having dif- ferent foliar chemistry than have their large parents. The saturniid phenology of there not being caterpillars on evergi-een trees during the dry season in the same forest is ascribed to desic- cation during this time of year rather than to different leaf chemistry in the dry season. Finally. I argue that the impact of herbivory, and hence the like- lihood that it will select for defense traits against a herbivore, is related to the potential fitness of the actual plants fed upon. A plant that is slated September 1985 125 Years of Biological Research 173 to die for other reasons is evolution- arily dead even if physiologically liv- ing; the accumulation, over evolution- ary time, of an herbivore load by ecolog- ically distinctive conspecifics of dif- ferent potential fitness should differ with that potential fitness. Further- more, the accumulation of an herbivore load by a species of plant may well be influenced by the proportion of the population of gi-owing juveniles that are in the category of "evolutionarily dead." ACKNOWLEDGEMENTS This study was supported by NSF DEB 77-04889. DEB 80-11558, and BSR 83-08388 and by the Servicio de Parques Nacionales de Costa Rica. W. Hallwachs and anonymous reviewers provided constructive criticism of the manuscript. Dr. Claude Lemaire, France, confirmed the saturniid iden- tifications, and J. M. Cadiou confirm- ed the sphingid identifications. Plant determinations were done by the author in compliance with Janzen & Liesner (1980) and more recent deci- sions by appropriate authorities; vouchers are deposited in the Missouri Botanical Garden, St. Louis. I especial- ly wish to thank J. Zarucchi and C. Hamilton for Rhynchosia and Randia determination. LITERATURE CITED Abrahamson, W. G., p. O. Armbruster, and G. D. Maddox. 1983. Numerical relationships of the Solidago altissima stem gall insect- parasitoid guild food chain. Oecologia 58:351-357. Ahmad, S. 1983. Mixed-function oxidase activity in a generalist herbivore in relation to its biology, food plants, and feeding history. Ecology 64:235-243. BoZA, M. and R. Mendoza. 1981. The national parks of Costa Rica, Institute de la Caza Fotogi-afica y Ciencias de la Naturaleza (INCAFO), Madrid, Spain. CoLEY, R D. 1982. Rates of herbivory on different tropical trees. Pages 123-132 in E. G. Leigh, A. S. Rand, and D. M. Windsor, eds., The ecology of a tropical forest. Seasonal rhythms and long-term changes. Smithsonian Institu- tion Press, Washington, D.C. 1983a. Herbivory and defensive charac- teristics of tree species in a lowland tropical forest. Ecological Monographs 53:209-233. 1983b. Intraspecific variation in herbi- vory on two tropical tree species. Ecology 64:426-432, Connor, E. F.. S. H. Faeth, and D. Simberloff. 1983. Leafminers on oak: the role of immigi'a- tion and in situ reproduction recruitment. Ecology 64:191-204. DeVries, P J. 1983. Checklist of butterflies. Pages 654-678 in D. H. Janzen, ed., Costa Rican natural history. University of Chicago Press, Chicago, IL. Edwards, P J., and S. D. Wratten. 1983. Wound induced defences in plants and their conse- quences for patterns of insect grazing. Oecologia 59:88-93. Farantinos, R. C, P. J. Capretta, R. E. Kepner, and V. M. Littlefield. 1981. Selective herbi- vory in tassel-eared squirrels: role of mono- terpenes in ponderosa pines chosen as feeding trees. Science 213:1273-1275. Futuyma, D. J., and M. Slatkin. 1983. Coevolu- tion Sinauer Associates, Sunderland, MA. Haber, W. a., and G. W. Frankie. 1983. Aellopos titan (Cinta blanca, white-banded sphinxlet). Pages 680-681 in D. H. Janzen, ed., Costa Rican natural history. University of Chicago Press, Chicago, IL. Hairston, N. G., F E. Smith, and L. B. Slobodkin. 1960. Community structure, population con- trol, and competition. American Naturalist 94:421-425. Hare, J. D. 1983. Seasonal variation in plant- insect associations: utilization of Solanum dulcamara by Leptinotarsa decemlineata. Ecology 64:345-361. Haukioja, E. 1980. On the role of plant defenses in the fluctuation of herbivore populations. Oikos 35:202-213. , and P. NiEMELA. 1979. Birch leaves as a resource for herbivores: seasonal occurrence of increased resistance in foliage after mechanical damage of adjacent leaves. Oecologia 39:151-159. Heinrich, B., and S. L. Collins. 1983. Caterpillar leaf damage, and the game of hideand-seek with birds. Ecology 64:592-602. HoLME.s, R. T, J. C. ScHULTZ, and P. Nothnagle. 1979. Bird predation on forest insects: an enclosure experiment. Science 206:462-463. Janzen, D. H. 1980. Heterogeneity of potential food abundance for tropical small land birds. Pages 545-552 m A. Keast and E. S. Morton, eds., Migrant birds in the neotropics. Smith- sonian Institution Press, Washington, D.C. 174 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 1981. Patterns of herbivory in a tropical deciduous forest. Biotropica 13:271-282. . 1982. Guia para la identification de mariposas nocturnas de la familia Saturn- iidae del Pai-que Nacional Santa PJosa, Guana- caste, Costa Rica. Brenesia 19/20:255-299. , ed. 1983a. Costa Rican natural history. University of Chicago Press, Chicago, IL. . 1983b. Food webs: who eats what, why. how, and with what effects in a tropical forest? Pages 163-182 in F. B. Golley, ed., Ti'opical rainforest ecosystem: structure and function. Ecosystems of the World Series 14A. Elsevier Scientific Publishing Co., New York. . 1983c. Dispersal of seeds by vertebrate guts. Pages 232-262 (;? D. J. Futuyma and M. Slatkin, eds., Coevolution. Sinauer Associates, Sunderland, MA. 1984a. Natural history ofHylesia lineata (Saturniidae: Hemileucinae) in Santa Rosa National Park, Costa Rica. Journal of the Kansas Entomological Society 57:490-514. 1984b. How to be a tropical big moth: Santa Rosa saturniids and sphingids. Oxford Surveys in Evolutionary Biology 1:85-140. , and R. Liesner. 1980. Annotated check- list of plants of lowland Guanacaste Province, Costa Rica, exclusive of gi'asses and non-vas- cular cryptograms. Brenesia 18:15-90. , and P. S. Waterman. 1983. A seasonal census of digestion-inhibitors and alkaloids in the leaves of tropical deciduous forest trees in Costa Rica. Biological Journal of Linnean Society 21:439-454. Karban, R. and R. E. Ricklefs. 1983. Host char- acteristics, sampling intensity, and species richness of Lepidoptera larvae on broad- leaved trees in southern Ontario. Ecology 64:636-641. Kareiva, p. 1982. Experimental and mathemati- cal analyses of herbivore movement: quantify- ing the influence of plant spacing and quality on foraging discrimination. Ecological Mono- gi-aphs 52:261-282. KoGAN, M. 1977. The role of chemical factors in insect plant relationships. Pages 211-227 in XV International Congress of Entomology Proceedings. Washington, D.C. Langenheim, J.H., D. E. Lincoln, W. H. Stubble- bine, and A. C. Gabrielli. 1982. Evolutionary implications of leaf resin pocket patterns in the tropical tree Hymenaea (Caesalpinioideae: Leguminosae). American Journal of Botany 69:595-607. Lawton, J. H. 1982. Vacant niches and unsatur- ated communities: a comparison of bracken herbivores at sites on two continents. Journal of Animal Ecology 51:573-595. 1983. Plant architecture and the diver- sity of phytophagous insects. Annual Review of Ecology and Systematics 28:23-39. McClure, M. S. 1983. Population dynamics of a pernicious parasite: density-dependent vital- ity of red pine scale. Ecology 64:710-718. NiEMELA, P., J. TaH\'ANALVEN, J. SoR.JL-NEN. T. HOK- kanen, and S. Xelvon-en. 1982. The influence of host plant gi-owth form and phenology on the life strategies of Finnish macro- lepidopterous larvae. Oikos 39:164-170. NiTECKi. M. H. 1983. Coevolution. University of Chicago Press, Chicago, IL. Price, R W, C. E. Bouton, R Gross, B. A. McPheron, J. N. Thompson, and A. E. Weis. 1980. Interactions among three tropic levels; influence of plants on interactions between insect herbivores and natural enemies. An- nual Review of Ecology and Systematics 11:41-65, Rausher, M. D., and D. R. Papa.j. 1983. Demo- graphic consequences of discrimination among conspecific host plants by Battus phi/enor butterflies. Ecologv' 64:1402-1410. Royama. T. 1970. Factors governing the hunting behavior and selection of food by the gi"eat tit [Parus major L.). Journal of Animal Ecology 39:619-668. ScHULTZ. J. C. 1983. Habitat selection and forag- ing tactics of caterpillars in heterogeneous trees. Pages 61-90 in R. F Denno and M. S. McClure, eds.. Variable plants and herbivores in natural and managed systems. Academic Press, New York. P. J. NoTHNAGLE, and I. T Baldwin. 1982. Seasonal and individual variation in leaf quality of two northern hardwoods tree species. Amer- ican Journal of Botany 69:753-759. ScoTT, N. J., J. M. Savage, and D. C. Robison. 1983. Checklist of reptiles and amphibians. Pages 367-374 in D. H. Janzen, ed.. Costa Rican natural history. University of Chicago Press. Chicago. IL. SoLiTHwooD, T. R. E., V. C. MoR.\.v. and C. E, J. Kexnxdy. 1982. The richness, abundance and biomass of the arthropod communities on trees. Jour- nal of Animal Ecology 51:635-649. Stanton. M. L. 1982. Searching in a patchy en- vironment: food plant selection by Colias p epiphyle butterflies. Ecology 63:839-853. Stiles, F. G. 1983. Checklist of birds. Pages 530-544 (/) D. H. Janzen. ed.. Costa Rican natural history. University of Chicago Press. Chicago. IL. Thompson, J. N. 1983. Selection pressures on phytophagous insects feeding on small host plants. O^kos 40:438-444. Washburn. J. O.. and H. J. Cornell. 1981. Para- sitoids. patches, and phenolog>-: their possi- ble role in the local extinction of a cjnipid gall wasp population. Ecology 62:1597-1607. Wilson. D. E. 1983. Checklist of mammals. Rages 443-447 in D. H. Janzen, ed.. Costa Rican natural history. University of Chicago Pi"ess, Chicago, IL. WiNT. W. 1983. The role of alternative host-plant species in the life of a polyphagovis moth. Operophtcra hrumala (Lepidoptera: Geometridae). Journal of Animal Ecology 52:439-450. Plant Kairomones and Insect Pest Control Robert L. Metcalf "There can be hardly an entomologist today who does not know that host selection is the very heart of agricultural entomology and that secondary plant substances are the clues to the problem." — G. Fraenkel (1969) "The extraordinary variety of insect plant relationships is likely to be based on more diverse mechanisms than can be foreseen at present The promise ofrewards is rich for investigators who approach the problem free ofdeeply rooted bias." — A. J. Thorsteinson (1960) Estimates suggest that there are at least 100,000 chemical compounds pro- duced during the gi-owth and develop- ment of the more than 200,000 species of flowering (vascular) plants (Har- borne 1982). The vast majority of these chemicals are the so-called "secondary plant compounds," not essential for the normal physiology of the plants, but rather the seemingly capricious out- poui'ings of nature's chemical factories, the alcohols, esters, ketones, terpen- oids, steroids, alkaloids, flavonoids, phenylpropanoids. It is only when we examine carefully the evolutionary processes that have brought about the diversification and speciation of plants that this enormous array of organic chemicals takes on purposeful and orderly significance. The external plant environment is pervaded by these compounds of second- ary metabolism that ooze from leaves, blossoms, and fruits. While humans are well aware ofmany ofthem because they define the colors, odors, and tastes that characterize our interactions with plants, there are countless others, less conspicuous in either quality or quan- tity, that dominate the lives of the 500,000 or so species of insects that have coevolved with the flowering plants. Many of these allelochemicals generate olfactory or gustatory stimuli that convey behavioral messages to species involved in the coevolutionary Dr. Robert L. Metcalf is a Professor of Biology and Research Professor of Entomology, Univer- sity of Illinois, Urbana-Champaign, and Principal Scientist, Illinois Natural History Survey. and ecological interrelations of food webs. From an arthropogenic point of view these allelochemics are perceived by insect sensory receptors as attract- ants, repellents, or arrestants, and in a more complex way, as antibiotics that disrupt normal growth, development, fecundity, and longevity or as anti- xenotics that disrupt normal host selec- tion processes. From the viewpoint of the processes of coevolution between plants and insects, these allelochemics are classified as allomones if they con- vey adaptive advantage to the plant producer and as kairomones if they con- vey adaptive advantage to the insect receiver (Kogan 1983). Fraenkel (1959) reviewed much of the early history of our understanding of the reasons for the existence of plant allelochemics and suggested the criter- ia for their positive identification as isolation and identification of the chemical, initiation of the kairomone or allomone response when applied to a neutral surface, and demonstration of a quantitative relationship between the concentration of the allelochemic and the insect response. The same chemical compound may act as an allo- mone, protecting the plant against some herbivores, and as a kairomone, stimulating the feeding of other herbi- vores, depending upon the vagaries of mutations and coevolution. Specific kairomones may be formed by a wide variety of plant families, and thus, these kairomones may be attractive to particular insect pests that exhibit a wide host range. 175 176 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 Coevolution of Plants and Insects The coevolutionary history of the plant kingdom with the insect world is an important area of investigation for both fundamental and applied studies in entomology. The relatively enor- mous numbers of species involved, some 500,000 insects and 200,000 plants, and the vast expanse of more than 300 million years of evolutionary time concerned, serve not only to pro- vide almost unlimited examples of plant-insect interactions for study, but also have obscured many of the key interrelationships in the mists of time. Some of the salient history, as record- ed in the fossil records, is portrayed in Fig. 1. The first land plants are thought to have appeared about 420 million years ago in the Devonian Period, and the first flowering plants occurred about 225 million years ago in the Triassic Period. These preceded the appearance of fruits in the Cretaceous Period, little more than 100 million years ago. This segment of evolutionary time was overlapped by the evolution of the Insecta, which evolved about 300 million years ago in the Carboniferous Period and were well diversified into modern orders by the Permian Period 270 million years ago (Riek 1970). The first fossil records of insect-damaged leaves are found in the Permian Pferiod. These relationships between plants, chemicals, and insects are sem- inal in understanding speciation, co- evolution, zoogeogi-aphy, ecologj', and insect behavior, and they are of pri- mary importance in the applied ecol- ogy of host-plant susceptibility and resistance to insect attack. In this review we will consider two gi'oups of insects that have exploited plant kairo- mones so that some of their species have become dominant pests of agi'icul- ture. The principal thesis explored is that insect response to plant-produced kairomones is largely a function of specific kairomone receptor organs on antennae, palpi, or tarsi that have active sites structurally complemen- September 1985 125 Years of Biological Research 177 tary in conformation to the structures of the kairomones. The specific kairo- mone I'eceptor is the product of millions of years of coevolution between host plant and insect pest and has been gi-adually modified by successive muta- tions that have provided the insect with adaptive advantages in terms of previously inaccessible ecological niches. As plant allelochemics have become more and more diversified, in- sect kairomone receptors have become shaped to accommodate them. PHENYLPROPANOIDS IN COEVOLUTION OF THE DACINI The fruitflies of the tribe Dacini (order Diptera, family Ifephritidae) con- stitute a closely related and rapidly evolving gi'oup of more than 1,000 species found in the Old-World tropics (D. E. Hardy, University of Hawaii, per- sonal communication). The female flies have sharp piercing ovipositors and typically insert their eggs into a wide variety of fruits and vegetables in which the larvae develop. Many of the Dacinae are remarkable for their wide host range: Dacus dorsalis Hendel, the oriental fruitfly, has been reared from more than 150 plants, and D. tyroni (Froggatt), the Queensland fruitfly, from more than 106 hosts (May 1953). D. cucurbitae (Coquillett) the melon fly, has been recorded as attacking more than 40 different kinds of plants in 12 families, including melons, cucumbers, squash, gourds, tomatoes, and peppers (Hardy 1949). Many other species of this subfamily are much more restrict- ed in host range (AUwood & Angeles 1979), and the host plant preferences for the majority of species seem not to be recorded. Hewlett (1912, 1915) first reported that citronella oil was attractive to the male fruitflies, Dacus diversus Coquil- lett and D. zonatus (Saunders), in India and subsequently demonstrated that 3,4-dimethoxyallylbenzene, or methyl eugenol, a constituent of citronella oil, was highly attractive to males of these flies and of D. dorsalis {=D. ferrugineus (Fabricius). D. cucurbitae males were shown to be highly attracted to 4-(p-methoxy- phenyl)-2-butanone, or anisyl acetone (Barthel et al. 1957), and 4-(p-acetoxy- phenyl)-2-butanone, or cue-lure, was developed subsequently as a more effec- tive lure (Beroza et al. 1960). Cue-lure is intrinsically less attractive than its hydrolysis product, 4-(p-hydroxyphe- nyl)-2-butanone, raspberry ketone or Willison's lure (Kaiser et al. 1973; Drew 1974), a natural product first isolated from the raspberry (Schinz & Seidel 1961). Drew and his coworkers, in a mon- umental series of trapping experiments in Australia and New Guinea, have surveyed the response of more than 150 species of the tribe Dacini to methyl eugenol and to cue-lure or raspberry ketone (Drew 1974; Drew et al. 1981; Drew & Hooper 1981). From their data and those of others (see Review of Ap- plied Entomology), it appears that ap- proximately 90 species of Dacini have been shown to respond to cue-lure or raspberry ketone and 40 species to methyl eugenol. No species has been shown to respond to both types of lure, and all species within each complex of closely related species responded to the same lure, indicating that the olfactory response has profound evolutionary im- plications. Hardy (1979) estimates that 90 percent of all Dacini respond to one or the other of these lures. Both methyl eugenol and raspbeny ketone are phenylpropanoid secondary plant compounds derived during plant evolution from phenylalanine through a common precursor, p-hydi'oxycinnamic acid or p-coumaric acid (Geissman & Crout 1969; Friederich 1976), as sug- gested in Figure 2. It appears that an ancestral Dacini in Southeast Asia must have developed an association with rotting fruits containing coumaric acid and that this compound became a kairomone regulating early Dacini behavior. The Tephritidae are believed to have arisen in the Paleocene era about 65 million years ago (Fig. 1). 178 Illinois Natural History- Survey Bulletin Vol, 33. Art. 3 COH O^JHp phenyl^ ^-^ *- ninnir. 'COH alanine (gr^ - cmnpjnic / V 'COH p-coumanc acid chovicol HOerT'COH ,^r-e. HO'l^J raspberry ketone Wf^ sugenol OCH, cHsoO;;^-.^^ Fig. 2. - Plant evolution of phenylpropanoids from phenyl alanine. Arrows indicate divergence leading to the development of kairomones for two distinct groups of Dacini. Both methyl eugenol for D. dorsalis and raspberry ketone for D. cucurbitae are attractive in nanogram quantities on filter paper, and D. dorsalis has been shown to respond by attraction, orien- tation, and feeding to as little as 100 pg of ^H methyl eugenol present on the antennae where the olfactory re- ceptors are located (Metcalf et al. 1981). Thus, the sensitivity of the responses of these fruitflies to these two kairo- mones approaches that of various male lepidoptera to the female sex phero- mones. However, in contrast to the very precise stereochemical conformity be- tween sex pheromone and receptor (Blum et al. 1971; Priesner et al. 1975), the kairomone receptors of both D. dor- salis and D. cucurbitae ai'e responsive over a range of about 10" in concentra- tion and to a wide variety of organic chemicals with structural resem- blances to methyl eugenol or to rasp- berry ketone (Metcalf et al. 1975, 1979, 1981, 1983b) (Tkble 1). The male Dacini respond to odorant compounds by a characteristic behavioral sequence of a) orientation, b) searching, c) arrest, and d) compulsive feeding fMetcalf et al. 1979) that is highly reproducible. The degree of complementarity' or "fit" of individual odorants to the kairomone receptors on the male Dacini antennae can be judged by determining the least amount of odorant, applied to filter paper under standardized conditions and exposed to fruitfly populations of uniform age, that produces the behav- ioral sequence. This value is termed the limit of response (LR) (Metcalf et al. 1979, 1983). Methyl eugenol is widely distrib- uted in plants, and there are numerous recorded observations of male Dacini aggregating and feeding on plant sour- ces, as shown in Table 1. Phloretic acid has recently been found in the blos- soms of Cucurbita pepo Linnaeus (Itok- awa et al. 1983), and its presence may account for the association of at least 10 species of Dacini, including the melon fly Dacus cucurbitae, with vari- ous Cucurbitaceae. KAIROMONE RECEPTOR INTERACTIONS IN THE DACINI The sensory receptor organs that trigger the responses of male Dacini to the phenylpropanoids are located on the scape of the antenna, and bilateral antennectomy abolishes the response (Metcalf et al. 1975). The antennal receptors, as revealed by scanning elec- tron microscopy, are typical sensilla basiconica. The most notable feature of the Dacini response to phenylpro- panoid kairomones is the sharp demar- cation of the tribe Dacini into species responding to 4-(p-hydroxyphenyl)-2- butanone, or raspberry ketone, and its close analogues, e.g., Dacus cucurbitae, the melon fly, and species responding to 3,4-dimethoxyallylbenzene, or meth- yl eugenol, and its close analogues, eg., Dacus dorsalis (Drew 1974; Drew & Hooper 1981). The data in T^ble 2 show the sensi- tivity of response (LR) of these two key September 1985 125 Years of Biological Research 179 Table 1. — Plants containing methyl eugenol to which male Dacini are attracted. ^ Family Plant Species Dacus Species Anacardiaceae Araceae Bromeliaceae Caricaceae Labiatae Lecythidaceae Leguminosae Myrtaceae Piutaceae Saxifragaceae Mangifera indica Linnaeus (mango flower) Colocasia antiquorum Schott Vriesea heliconioides (Humboldt, Bonpland & Kunth) Hooker ex Walpers (blossom) Carica papaya Linnaeus (papaya flower) Ocimum basilicum Linnaeus O. sanctum Linnaeus (tulsi) Couroupita guianeitsis Aublet (cannon-ball tree, blossom) Cassia fistula Linnaeus (golden shower tree, blossom) Pimenta racemosa (Miller) J. W. Moore Syzygium cumini (Linnaeus) Skeels (= Eugenia jambolana) (jamum tree) S. aromaticum (Linnaeus) Merrill & L. M. Perry (= Eugenia aromaticum) (clove) Pelea anisata Mann (leaf, twig) Zieria smithii Andrews (leaf) Brexia madagascariensis (Lamarck) Noronha (blossom) D. diversus Coquillett, D. correctus (Bezzi) D. dorsalis Hendel, D. zonatus (Saunders) D. dorsalis D. dorsalis, D. diuersus D. dorsalis D. correctus D. dorsalis D. dorsalis D. diversus, D. dorsalis, D. zonatus D. caudatus, D. diversus, D. dorsalis, D. zonatus D. zonatus D. dorsalis D. cacuminatus (Hering) D. dorsalis ^ Howlett (1915), Kawano et al. (1968), Metcalf et al. (1975), Fletcher et al. (1975), (1976). Shah & Patel species to a variety of phenylpropanoids and related compounds (Metcalf et al. 1983b). It is evident that D. cucurbitae shows maximum response to phenyl- propanoids with a p-OH group on the phenyl ring and a C = group about 2 atomic diameters removed from the phenyl ring, as typified by raspberry ketone and methyl phloretate (see Fig. 2). D. dorsalis shows maximum re- sponse to phenylpropanoids with m- and P-CH3O groups on the phenyl ring and a side chain with an unsaturated double bond, as typified by methyl eugenol and methyl isoeugenol. Neither species responded appreciably to the most effective odorants for the other species. The antennal receptors of D. cucurbitae and D. dorsalis must provide maximum structural complementarity to the kairomones which promote max- imum attraction and feeding stimula- tion, raspberry ketone for D. cucurbitae and methyl eugenol with D, dorsalis. The degree of structural overlap be- tween the two types of receptors is minimal, as shown in Table 2. Present-day understanding of the evolution of phenylpropanoid synthesis in the plant kingdom indicates phenyl- alanine as a precursor and p-hydroxy- cinnamic acid (p-coumaric acid) as the common ancestral substance from which there was divergence to the eugenol-methyl eugenol and to the raspberry ketone-phloretic acid phenyl- propanoids pathways (see Fig. 2) (Fried- rich 1976; Metcalf et al. 1979, 1983b). 180 Illinois Natural History Survey Bulletin Vol, 33, Art. 3 Table 2. — Limit of response (LR) of Dacus cucurbitae and D. dorsalis to kairomone analogues. September 1985 125 Years of Biological Research 181 dents of a mutant form whose antennal receptors developed complementarity to the 3,4-dimethoxyphenylpropanoids, thus opening up new ecological niches. There is good evidence that the primary attractive site on the male D. dorsalis antennal receptor is comple- mentary to the P-CH3O group of methyl eugenol (Metcalf et al. 1983). For this series of compounds there is a high degree of negative correlation between LR values for D. dorsalis and the tt value for octanol/water partition of the substituent side chain, demonstrating the lipophilic nature of the receptor (Metcalf et al. 1981). Chemical Ecology of Host Selection in the Dacini The role of phenylpropanoid kairo- mones in the regulation of the behav- ioral ecology of the Dacini is complex. The lack of responsiveness of the fe- males to these substances is puzzling. However, it appears that in the pres- ence of male Dacini the kairomones can act as short-range ovipositional stimulants. The sex pheromones of the Dacini are produced in rectal glands of the males and attract virgin females. Thus, the plant kairomones may pro- mote host selection by producing male aggregation and consequent sex phero- mone release to bring females to suit- able sites, where short-range oviposi- tional stimulation occurs (Metcalf et al. 1983b). CUCURBITACINS IN COEVOLUTION OF THE LUPERINI The more than 300,000 described species of beetles of the order Coleop- tera comprise nearly one-third of all insects and form the largest order of living organisms. The preponderance of species of Coleoptera feed on plants, and representatives first appeared about 260 million years ago in the Permian era (Fig. 1; Riek 1970). The Chrysomelidae, or leaf beetles, (over 20,000 described species) feed almost exclusively on plants, and the coevolutionary association between plants of the family Cucurbitaceae and beetles of the tribe Luperini provides the most comprehensive example of the role of allelochemics acting as kairo- mones to promote host selection and feeding by phytophagous insects. The family Cucurbitaceae contains some 900 species of plants in about 100 genera, many familiar as the gourds, squash, cucumbers, and melons of Cucurbita, Cucumis, Citrullus, Lagen- aria, Marah, Sicyos, Echinocystis, Ecbal- lium, and Bryonia. At least 100 species in 30 genera of Cucurbitaceae have been shown to contain a group of more than 20 oxygenated tetracyclic triter- penoids, the cucurbitacins (Cues) (Fig. 3). The Cues are responsible for the characteristic bitter taste of most wild Cucurbitaceae (Rehm 1960; Lavie & Glotter 1971). Current thinking holds that the Cues were selected by coevolu- tionary processes to protect the Cucur- bitaceae against attack by a wide variety of both invertebrate and verte- brate herbivores. The Cues are the most intensely bitter substances yet characterized and can be detected in aqueous solution at concentrations as low as 1 ppb (Metcalf et al. 1980). They are also extremely toxic, with intra- peritoneal LD50 values of Cue A mouse 1.2, rat 2.0, Cue B mouse 1.0 mg per kg (David & Vallance 1955). There have been instances of severe poisoning and death in sheep and cattle that con- sumed bitter Cucumis and Cucurbita fruits during drought conditions (Watt & Breyer-Brandwijk 1962). Very re- cently an epidemic of human illness in Australia resulted from the eating of zucchini-type Cucurbita pepo fruit that contained about 1 mg per g of Cues (fresh weight), apparently as the result of a genetic reversion (Ferguson et al. 1983a). A number of insect pests, in- cluding the leaf beetles Phyllotreta nemorum (Linnaeus), P. undulata (Kutschera), P tetrastigma (Comolli), 182 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Fig. 3. - Suggested plant evolution of cucurbitacins B, E, D, I, F, G, and L. Numbers refer to specific enzymes involved: 1) cucurbitacin A' dehydrogenase, 2) cucurbitacin acetylesterase. 3) cucurbitacin A^' reductase, 4) cucurbitacin CS-hydroxylase, and 51 cucurbitacin C24-hydi-oxylase (Lavie & Glotter 1971). Phaedon cochleariae (Fabricius), P cru- ciferae (Goeze), and Cerotoma trifurcata (Forester), are strongly deterred from feeding by the presence of Cues (Niel- son et al. 1977; Metcalf et al. 1980). In contrast, the most important in- sect pests of Cucurbitaceae worldwide are found in a large group of beetles of the tribe Luperini, comprising 1,528 species of Old-World Aulacophorina (535 species) and the New-World Dia- broticina (993 species) (Wilcox 1972). Luperini contains such destructive pests of cultivated crops as Diabrotica undecimpunctata howardi Barber, the spotted cucumber beetle or southern corn rootworm; D. undecimpunctata undecimpunctata Mannerheim, the western spotted cucumber beetle; D. balteata LeConte, the belted or banded cucumber beetle; Acalymma vittatum (Fabricius) and A. trivittatum (Man- nerheim), the striped cucumber bee- tles; D. barberi (Smith and Lawrence), the northern corn rootworm; D. vir- gifera virgifera LeConte, the western corn rootworm; D. virgifera zea Kryson and Smith; D. speciosa Germar of South America; Aulacophora foveicollis of Asia, Africa, and Europe; and A. femoralis of Siberia, China, and South- east Asia. The larvae of these beetles are rootworms, and the adults are found feeding on cucurbits, corn, beans, peppers, and a variety of other plants. Host plant records are sorely lack- ing, and Wilcox (1972) in the authori- tative Coleopterorum Catalogus lists only 29 for 1,528 species of Luperini; 72 percent of the recoi'ds are for Cucur- bitaceae. A literature search has dis- closed 49 species, listed in Tkble 3, as feeding on Cucurbitaceae (e.g., Tkki- zawa 1978). These 49 species represent more than 80 percent of the published host i-ecords for the Luperini and por- tray a relationship between these bee- tles and the Cucurbitaceae that is widely distributed between the Old- World Aulacophorina and the New- World Diabroticina. Compulsive feed- ing of the Luperini beetles on the leaves and fruits of wild, bitter Cucur- bitaceae has been described for many of these species, and Diabrotica balteata, D. barberi (as D. longicornis). September 1985 125 Years of Biological Research 183 Table 3. — Chrysomelidae: Galeruclnae; Luperini Associated with Cucurbitaceae ^ Species Plant Hosts Locations Aulacorphorina Aulacophora A. abdominalis (Fabricius) A. atripennis (Fabricius) A. bicolor (Weber) A. cincta (Fabricius) A. coffeae (Hornstedt) A. excavata Baly A. femoralis (Motschulsky) A. fbveicollis (Lucas) A. hilaris (Boisduval) A. lewisii Baly A. loochooensis Chujo A. nigripennis nigrippennis Motschulsky A. n. nitidipennis Chujo A. olivieri Baly A. quadrimaculata (Fabricius) A similis (Olivier) Paridea P angulicollis (Motschulsky) P. costata (Chujo) P sauteri (Chujo) P sexmaculala (Laboissiere) P testacea Gressit & Kimoto Agetocera A. discedens Weise A. taiwana Chujo Lamprocopa L delata (Erichson) Paragetocera P ini'oluta Laboissiere Diabrotica D. balteata Leconte cucumber, melons, pumpkin gourds, pumpkin, muskmelon wild, cultivated Cucurbitaceae snake gourd, bitter gourd, bottle gourd melons, pumpkin CitruUus. Cucumis, Cucurbita, Lufta wild, cultivated Cucurbitaceae pumpkin, squash, muskmelon pumpkin, marrow wild, cultivated Cucurbitaceae Cucurbitaceae squash, cucumber Cucurbitaceae melons, cucumber, pumpkins, squash CitruUus lanatus (Thunberg) Matsumura & Naki, Cucurbita pepo Linnaeus CitruUus lanatus. Cucumis melo Linnaeus, C. satiuus Linnaeus, Cucurbita pepo Cucurbitaceae wild Cucurbitaceae Cucurbitaceae cultivars Indonesia, Australia Indonesia Taiwan, Indonesia, China, India, Japan, Phillipines India, Ceylon Indonesia, Fiji India Siberia, Japan, Phillipines, Viet Nam Asia, Africa, 8. Europe Australia, Micronesia S.E. Asia, Pacific Islands Ryukyu Islands Siberia, China, Japan, Korea, Taiwan Ryukyu Islands Australia Pacific Islands, Australia S.E. Asia, Samoa, Fiji Gymnostemma pentaphyUum 184 Illinois Natural History Survey Bulletin Table 3. — Continued Vol. 33, Art. 3 Species Plant Hosts Locations D. cristata (Harris) D. longicornis (Say) D. graminea Baly D. speciosa (Germar) D. tibialis Jacoby D. virgifera Leconte D. undecimpunctata howardi Barber D. undecimpunctata undecimpunctata Mannerheim Acalymma Cucurbitaceae Cucurbita foetidissima Humboldt, Bonpland & Kunth Cucurbita andreana Naudin Cucurbitaceae Cucurbita foetidissima Cucurbitaceae cultivars Cucurbitaceae cultivars Central USA Central USA Costa Rica, West Indies, Cuba, Puerto Rico Mexico to Argentina S. Texas, Mexico Central USA Central & S. USA W. USA A. bivittatum (Fabricius) September 1985 125 Years of Biological Research 185 developed detoxication and excretory systems for Cues (Ferguson & Fischer 1985). These species of Diabroticina are able to detect nanogram quantities of Cues by means of well-developed chemosensory organs located on the maxillary palpi (Metcalf et al. 1980), and the limit of response to pure cucur- bitacins on silica gel thin layer plates varies with the individual species and the chemical nature of the Cue (Table 4). The high degree of sensitivity and specificity of response of the beetles are demonstrated by the "beetle prints" resulting from the feeding of Diabro- ticina on Cues extracted from Cucur- bitaceae and separated by thin-layer chromatogi-aphy (TLC) (Fig. 4). Using this technique, we have studied the Table 4, — Limit of response (LR) of Diabroticina beetles to pure cucurbitacins. Species LR in Micrograms of Cucurbitacin E I Egly. Diabrotica balteata 186 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 quantitative distribution of the Cues in 18 species of Cucurbita, as shown in Table 5 (Metcalf et al. 1982). No Cues were detectable in leaves, fruits, or roots of the domesticated species C. ficifolia, C. maxima, C. mixta, C. mos- chata, and C. pepo (zucchini) down to a limit of 0.02 mg per g of fresh weight. Indeed, fruits of these species would be inedible if Cues were present. The Cues content of the fruits of wild species ranged from 0.53 mg per g in C. pedatifolia to 3.2 mg per g in C. andreana. The nature of the Cues pres- ent and the beetle feeding responses observed agree reasonably well with evolutionary groupings based on numer- ical taxonomy (Rhodes et al. 1968), cross compatibilities (Bemis et al. 1970), and isozyme analysis (Puchalski & Robinson 1978). Group (1) included the Cue B- and Cue D- forming species found in subgroup (a) C. andreana and C. ecuadorensis and subgroup (b) C. gracilior, C. palmeri, and C. sororia. Group (2) included the Cue E- and I- forming species found in subgroup (a) C. martinezii and C. okeechobeensis and subgroup (b) C. cylindrata. C. palmata, C. foetidissima, and C. texana formed a preponderance of Cue E glycoside (Metcalf et al. 1980, 1982). The role of the Cues as feeding stimulants was demonstrated conclu- sively by painting microgram quan- tities of Cue B on soybean leaves, not eaten normally by Diabrotica barberi, D. u. howardi, and D. v. virgifera. All three species fed heavily upon leaf areas treated with Cue B. In field ex- periments, these beetles ate a variety ofweeds treated with Cue extracts and homogenates of bitter Cucurbita. The plant Iberis umbellata Linnaeus, or candy tuft, (Cruciferae) is one of the few species outside the Cucurbitaeeae known to produce Cues. Plants of this species interspersed with bitter Cucur- bita were massively attacked by D. u. howardi and D. v. virgifera and were completely defoliated and killed within 2 weeks. Thin-layer chromatogi-aphy of Iberis extracts isolated Cues E and I. which were readily fed upon by Dia- brotica beetles (Metcalf et al. 1980). Evaluation of Diabroticina feeding patterns on TLC chromatogi'ams from extracts of Cucurbita andreana eontain- Table 5, — Cucurbitacin content of fruits of Cucurbita spp. ^ Cucurbita spp. September 1985 125 Years of Biological Research 187 ing CuCucs B and D, C. okeechobeensis containing Cues E and I, and C. texana containing Cue E glycoside showed an almost identical qualitative response for Diabrotica balteata, D. cristata, D. barberi, D. u. howardi. D. v. virgifera, and Acalymma vittatum (Metcalf et al. 1980). There is substantial evidence of the lengthy coevolutionary association of Cucui'bitaceae plants and Luperini beetles. The genus Cucurblta is indig- enous to the Americas, where it has existed since pre-Columbian times. It is considered to have its center of origin in the tropical or semitropical region of southern Mexico, from which the 27 species (22 wild, 5 cultivated) have radiated to North and South America (Whitaker & Bemis 1964, 1975). The present distribution of species of Cucur- blta is (Whitaker & Bemis 1964): Southern Mexico and Central America, 14 species; Northern Mexico, 6 species; Northern South America, 4 species; Southern North America, 8 species. The Diabrotica beetles have re- markably similar geographic distribu- tion, as shown in Fig. 5. The putative area of origin of this genus is in nor- thern South America or Central America, from which a few species have radiated into both southern South America and North America (data from Wilcox 1972). Fig. 5 also shows the species distri- bution of the Aulacophora (Maulik 1936), the Old-World counterparts of the Diabrotica, which also must have coevolved with Cucurbitaceae. This genus appears to have evolved in Indo- nesia and to have radiated into south- east Asia, extending north to China and Siberia and south to Australia. The very close systematic relationships of the Aulacophorina and Diabroticina, the intimate association of both sub- tribes with Cucurbitaceae, and their common response to cucurbitacins strongly suggest a common ancestral co-evolution with an early cucurbi- taceous species during a geologic per- ?*«51i^ m^i V -/>^ 53 Fig. 5. - Distribution and numbers of described species of Diabroticina in tlie New World and Aulacopliorina in the Old World. (Data from Wilcox 19721. 188 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 iod when continental land bridges were present (Metcalf 1979). KAIROMONE RECEPTOR INTERACTIONS IN THE DIABROTICINA Among the Diabroticina beetles the primary sensory receptor organs that trigger the compulsive feeding response to the cucurbitacins are lo- cated on the maxillary palpi of both sexes. Surgical amputation of the max- illary palpi abolishes the compulsive feeding response, but this response is not altered by antennectomy (Metcalf et al. 1980). Scanning electronmicro- scopy of these organs in Diabrotica balteata, D. cristata, D. barberi, D. u. howardi, D. v. virgifera, ariA Acalymma vittatum has demonstrated similar morphology of the sensory receptors in all species, and their ultra structure is under study (J. R. Larsen, University of Illinois, unpublished research). These species of Diabi'oticina respond qualitatively to the spectrum of Cues found in choloroform extracts oiCucur- bita andreana, C. okeechobeensis, and C. texana, as separated by TLC on silica gel thin layer plates, in an almost iden- tical pattern (see Fig. 4). Thus, there is no evidence of any significant change in the spectrum of response to the Cues by the various species of Diabroticina over an evolutionary period of more than 40 million years, as estimated from isozyme studies (Harvey et al. 1983). This evolutionary stability of receptor response is remarkable in view of the substantial differences in the present host preferences of these species of Diabroticina. The adults of a number of North American species have a proclivity for feeding on cucur- bits and have been given trivial names suggesting this: Diabrotica balteata, the belted or banded cucumber beetle; D. u. howardi, the spotted cucumber bee- tle; D. II. undecimpunctata Manner- heim, the western spotted cucumber beetle; D. picticornis Horn, the painted cucumber beetle; D. (Paranapiacaba) connexa (LeConte), the saddled cucum- ber beetle; Acalymma trivittatum, the western striped cucumber beetle, and Acalymma vittatum, the striped cucum- ber beetle (Chittenden 1910 (. The most generalized feeders are probably D. balteata, whose adults also feed on cucumber, squash, melons, beans, soy- beans, eggplant, and vetch, and D. u. howardi and D. u. undecimpunctata, whose adults feed on cucumbers, melons, squash, beans, soybeans, peas, cabbage, peppers, and corn (Chittenden 1910). Acalymma trivittatum and A vit- tatum are more specialized, and the adults generally restrict their attacks to cucumber, squash, and melons. The northern corn rootworms, Diabrotica barberi Smith and Lawrence, and the western corn rootworm, D. v. virgifera, were originally described fi'om adults collected on blossoms of Cucurbita foetidissima (Smith & Lawrence 1967) and D. v. virgifera adults readily attack cucumber, squash, and melons (Howe et al. 1976). The larvae of both species apparently develop only in the roots of gi'asses, especially corn, and the adults feed avidly on corn silks and pollen. Branson & Krysan (1981) suggest that both the northern and western corn rootworms have become pests of corn only relatively recently through con- vergent evolution. D. cristata is found almost exclusively in relict prairies and may develop only on the roots of prairie gi-asses, such as Andropogou, but the adults have been collected from squash blossoms (personal observation). The demonstration that functional Cue receptors are present in these species suggests not only that they originally coevolved with the Cucurbitaceae, but also that the development of other host preferences may have been relatively recent (Branson & Kiysan 1981). The data in Table 4 indicating the limits of response (LR) for the various species of Diabroticina exposed to a variety of pure Cues indicate that Cue B was consistently detected at lower levels than was any of the other Cues. September 1985 125 Years of Biological Research 189 Therefore, it appears that Cue B has maximum complementarity to the Cue receptors on the maxillary palps and is likely to be the primitive Cue to which the sensory receptor of an ancestral Luperini must have become evolution- arily tuned (Metcalf et al. 1980, 1982). This possibility is supported by the chemical resemblance ofCue B to bryo- genin, considered the parent tetracyclic triterpenoid formed from mevalonic acid (Geissman & Crout 1969). Bryo- genin and Cue B both have the eyclo- hexane moiety in ring A, differing from Cue E, which has a pai'tially aroma- tized ring with a C=C bond (Fig. 3). Cue B is the dominant form found in the Cucurbitaceae and was detected in 91 percent of 46 species examined by Rehm et al. (1957); Cue E was detected in 42 percent of the species. Further- more, Cue B was the dominant form found in the cotyledons of 19 species of wild Cucurbita, Cucumis, and Citrullus and in 46 commercial cultivars of these genera, even when the mature leaves and fruits contained Cue E (Ferguson et al. 1983c). Cue E is formed from Cue B by an enzyme, cucurbitacin A^ dehydrogenase; Cue D is formed from Cue B and Cue I from Cue E by cucur- bitacin acetyl esterase. The chemical relationships between these and other Cues having alterations in the degi'ee of oxygenation of ring A or in the unsaturation of the C23=C24 side chain are shown in Fig. 3. It appears that all of the common Cues can be formed from Cue B. The sensitivity of response of the Diabroticina examined to the various pure Cues (Table 4) was D. u. howardi = D. u. undecimpunctata >D. balteata >D virgifera >D. barberi >D. cristata, Acalymma vittatum is considerably less responsive. Cue B was consistently detected at levels 0.1 - 0.3 of that of Cue E, and the deacetoxy Cues D and I were substantially less effective in triggering the compulsive feeding response than were Cues B and E. The changes in maxillary receptor sensitivity to the various Cues demon- strated by all the species of Diabro- ticina examined impel speculation about the nature of the Cue receptor. It seems likely that receptor depolari- zation follows allosteric changes in the receptor protein resulting from inter- actions of the free paired electrons associated with the several oxygen atoms in the Cue molecules (Metcalf et al. 1980). The structural change in Cue B by introduction of a single double bond at ring A to form Cue E seems trivial (Fig. 3), yet this change produces a tenfold decrease in receptor affinity (Table 4). The introduction of the C=C into ring A substantially changes the orientation of the three contiguous O atoms (C3=0, C2-0H, and C11 = 0) from a staggered configuration in the eyclohexyl moiety of Cue B to a planar configuration in Cue E. This change seems ample, from my observations of molecular models, to decrease receptor affinity and depolarization (Metcalf et al. 1980). Cues D and I exhibit about a tenfold decrease in receptor affinity compared with their C25 acetoxy deriv- atives, Cues B and E (Fig. 3). This fact suggests that the acetoxy-C = must also be involved in complete binding to the receptor. PLANT KAIROMONES AS ATTRACTANTS AND ARRESTANTS FOR INSECT PEST CONTROL The intrinsic nature of the coevolu- tionary process that produced the specific interactions between plant kairomones and insect sensory organs that lead to profound behavioral chang- es suggests that kairomones can be employed for insect control in a variety of baits and traps. This use of the in- stinctual behavior of the insect pest in the presence of the kairomone chemical is analogous to the use of insect sex pheromones for population monitoring, removal trapping, and mating con- fusion (Shorey & McKelvey 1977; Mitchell 1981; Nordlund et al. 1981). 190 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 There has been surprisingly little application of this use of kairomones for insect control, and several of the most successful examples were devel- oped without appreciation of the role of the kairomones in the chemical ecology of the pests (Mitchell 1981). Kairomone Baits for Monitoring and Controlling Dacini Fruitflies Howlett (1915) first demonstrated that methyl eugenol was a specific attractant to males of several species of Dacini, including the oriental fruit- fly Dacus dorsalis. During the outbreak of D. dorsalis in Hawaii following its discovery there in May 1946, oil of citronella was evaluated by the USDA as an attractant and was so effective that it was used as the standard sub- stance for monitoring oriental fruitfly populations (Steiner 1952). Oil of citro- nella contains about 8 percent methyl eugenol, and it was soon demonstrated that the purified chemical was far superior in attracting male D. dorsalis, for which it is also an arrestant and compulsive feeding stimulant (Steiner 1952). Methyl eugenol was shown to attract male D. dorsalis upwind from as far as 0.5 mile away and to stimulate male fruitflies to compulsive feeding that can kill them from overindulg- ence. The use of methyl eugenol for population monitoring of D. dorsalis was rapidly adapted to a variety of invaginated glass and plastic ti-aps con- taining water to drown the flies. Ti-aps containing about 1 g of methyl eugenol trapped as many as 2,600 - 7,300 male D. dorsalis in a single day. The simple 8-oz (230-ml) bottle trap baited with methyl eugenol on a cotton wick has become a standard monitoring device to detect incipient oriental fruitfly invasions (Steiner 1957). A simple box trap for area-wide control of D. dorsalis was developed by treating the inside of 3- x 12- x 16-inch (75- x 300- x 400-mm) wick boxes with 0.5 g of parathion insecti- cide as a wettable powder and over- spraying with 2 ml of methyl eugenol (Steiner 1952). Such traps, open on one side, attracted and killed 13,000 - 15,000 fruitflies per trap. These traps were used to control D. dorsalis through male annihilation in pine- apple fields in Hawaii, and it was shown that marked flies were attracted from as far as 1-1.5 miles (1.6- 2.4 km) away. In a 125-acre (50-ha) pineapple field, 45 box traps killed thousands of male flies and substan- tially reduced the male fly population over an area of at least 4 square miles (10.4 km^). The male annihilation method has been refined by employing cane fiber blocks 2.5 inches (62.5 mm) square and 0.37 inch (9.25 mm) thick saturated with a bait mixture of 97 percent methyl eugenol and 3 percent naled in- secticide (dimethyl l,2-dibromo-2,2- di- chloroethyl phosphate), so that each block contained about 23.3 g of methyl eugenol and 0.7 g of insecticide. These were dropped from aircraft at the rate of 125 per square mile (2.6 km^) over the island of Rota in the Mariannas at about 2-week intervals for 8 months. The wild D. dorsalis population was monitored by methyl eugenol trapping and declined from a pretreatment count of 262 male D. dorsalis per trap to 18.4 males per trap after the first treatment and to 0.028 male per trap after the fourth treatment. No flies were caught after the 7th month, and the oriental fruitfly population was re- duced by at least 99.6 percent (Steiner et al. 1965). This extremely efficient control effoi't used only 3.5 g of insecti- cide per acre (0.4 ha) per application and remains a classic demonstration of the efficiency and effectiveness of kairomone lures for insect pest control. Raspberry ketone (Willison's lure) and its p-acetoxy-derivative, cue-lure, have been widely used for monitoring populations of male Dacini, such as the Queensland fruitfly, Dacus tyrorti, and the melon fly, D. cucurbitac. that re- spond to this kairomone (Drew 1974; Drew & Hooper 1981), Fiberboard September 1985 125 Years of Biological Research 191 blocks 2.5 inches (62.5 mm) square and 0.5 inch (12.5 mm) thick were treated with a bait mixture of 95 percent cue- lure and 5 percent naled insecticide so that each block contained about 23.75 g of cue-lure and 1.25 g of insecticide. These blocks were tied 2-5 feet (0.6-1.5 m) above the ground on trees or stakes at the rate of about 585 blocks per square mile (2.6 km^) over an isolated area of the island of Hawaii and were replaced with freshly treated blocks each month for 4 months (Cunningham & Steiner 1972). The wild D. cucurbitae population was monitored by cue-lure trapping and declined from a pretreat- ment count of 169 male D. cucurbitae per trap to 62 males per trap after the first treatment (96.3 percent decline) to 0.22 per trap after the fourth treatment (99.9 percent decline). This level of population reduction was maintained for over 60 days. This extremely effi- cient control effort used only 1.2 g of insecticide per acre (0.4 ha) per treat- ment and illustrates again the very high efficiency of kairomone baiting and the safety and specificity of this method of insect pest control. The male annihilation technique using poisoned kairomone baits has produced high levels of control of sev- eral species of Dacini and, when applied to island populations, has resulted in the eradication of the Queensland fruitfly, D. tyroni, from Easter Island (Bateman et al. 1973). This method should be used with great caution against indigenous infestations because, as Hardy (1979) has emphasized, only a fraction of the 1,000 species of Dacws and related genera of Tephritidae are of commercial importance as pests in- festing agricultural crops. Overenthus- iastic efforts to eradicate pest species by the use of kairomone lures can lead to the devastation and even extinction of the endemic fauna in the Oriental, Australian, and Ethiopian zoogeo- graphic regions. Thus no eradication progi-am should be undertaken until the probable environmental impacts have been fully assessed. Kairomone Baits for Monitoring and Controlling Diabroticina Beetles The tetracyclic triterpenoid cucur- bitacins B and E are of high molecular weights and of very low volatility in contrast to methyl eugenol or rasp- berry ketone (the estimated v.p. of methyl eugenol is 0.01 mm Hg at 25 °C) and do not have any long-range vapor attractancy for Diabroticina beetles. The Cues, however, resemble methyl eugenol and raspberry ketone in that they are powerful arrestants and com- pulsive feeding stimulants, detectable by Diabroticina beetles on inert sur- faces at concentrations as low as 1 ng (Metcalf et al. 1980). In contrast to the phenylpropanoid kairomones for the Dacini, Cues kairomones appear to be effective with both male and female Diabroticina. Despite their lack of vapor attrac- tancy, the Cues can be used success- fully to monitor Diabroticina beetle populations for integi'ated pest man- agement (IPM) progi-ams, and their kairomonal activity can be exploited for use in poison baits in a way that is analogous to the successful use of methyl eugenol and raspberry ketone for control of the Dacini. Whereas methyl eugenol and raspberry ketone and its analogue cue-lure are relative- ly simple organic chemicals that can be made synthetically, the Cues are very complex chemicals that have not been synthesized. Nevertheless, the relative- ly high Cue content of wild bitter Cucurbita provides a useful source of kairomones for study. Early experi- ment with Diabroticina beetles ex- posed equally to sliced bitter and sweet Cucurbita fruit demonstrated conclu- sively that the great preponderance of the Diabroticina beetles (95-98 per- cent) remained feeding on bitter fruits, such as C. andreana (Contardi 1939), C. foetidissima (Sharma & Hall 1973), and C. pepo (Howe et al. 1976). In a comparison of the arrestant powers of fresh, sliced fruits of 11 species of Cucurbita with total Cue contents ranging from <0.02 mg per g (C. max- 192 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 ima, C. mixta, C. moschata, and C. pepo) to 3.20 mg per g (C andreana), the correlation coefficient between average numbers of beetles feeding and Cue content was 0.70 for Diabrotica u. howardi and 0.58 for D. v. virgifera (Metcalf et al. 1982). These bitter Cucurbita baits are rapidly consumed by Diabroticina beetle feeding, and the arrestant effect is lost within a day or two. The effectiveness can be prolonged for several weeks by sprinkling sliced bitter Cucurbita fruit with about 0.1 g of rapidly acting contact insecticide, such as trichlorfon or methomyl. Such treated bitter Cucurbita fruits remained effective in killing Diabroticina beetles for 3 weeks, even after heavy rains, and single fruit halves killed in excess of 2,000 beetles (Fig. 6) (Rhodes et al. 1980). The prolonged arrestant power of dried bitter Cucurbita fruit for Diabro- ticina beetles is remarkable and illus- trates the profound effect of the Cue kairomones on beetle behavior. Ground or pelleted bitter Cue fruits from C. andreana and C. texana or roots of C. foetidissima retain arrestant power for more than 3 years of storage indoors, and when impregnated with 0.1 per- cent w/w of a variety of carbamate and organophosphorus insecticides, are effective in arresting and killing Dia- broticina beetles for 3 weeks or more under summer field conditions (Metcalf et al. 1983). These dry poisoned kairo- mone baits are being used to monitor Diabroticina populations to determine economic thresholds for IPM progi'ams. Shaw et al. (1984) developed a 4-oz (120-ml) plastic vial, with holes to admit Diabroticina, baited with about 0.1 g of dried bitter Cucurbita bait con- taining carbaryl insecticide that is efficient and effective. To use the Cue kairomones for control of Diabroticina beetles, it is necessary to have an abundant source of bitter Cucurbita fruit, and wild Cucurbita spp. do not provide depend- able sources of Cues. They are more dif- ficult to grow in temperate regions and yield less than domesticated species, and in some cases fruiting is dependent upon the photo-period. Genes eonroll- ing the formation of Cues were trans- ferred to domesticated cultivars by crossing C andreana x C. maxima to produce long-vined plants with large fruits, averaging 3.90 kg and contain- ing a total of 1.26 mg of Cue B and D per gram of fresh weight. The dried bait contained 5.0 mg of Cues per gram. A hybrid of C. texana x C. pepo produced bushy plants with fruits aver- aging 0.73 kg and containing a total of 0.48 mg of Cues E, I, and E-glycoside per gram of fresh weight. The di'ied bait contained 6.1 mg of Cues (F, fruit) and 3.0 mg of Cues (Fj fruit) per gi'am (Rhodes et al. 1980; Metcalf et al. 1983). Another source of Cue kairo- mones is in the roots of C. foetidissima, which contain about 3.1 mg of Cues E, I, and E-glycoside per gi-am of fresh weight (Metcalf et al. 1982). Dried C foetidissima roots, gi-own as a semi- commercial crop for starch production in Arizona, contained 4.0 mg of Cues per gi-am (Berry et al. 1978). These air-dried and gi-ound Cucur- bit baits were impregnated with a vari- ety of insecticides, including the carba- mates, carbaryl, carbofui'an, bendicai-b, and methomyl; the organophosphates, malathion, dimethoate, terbufos, and isofenphos; and the pyrethroids, fenval- erate, permethrin, and decamethrin. It was determined that the most effective concentrations of insecticides were 0. 1 pei'cent w/w for carbamates and organ- ophosphates and 0.01 percent w/w for pyrethroids. Such baits have been eval- uated for Diabroticina beetle control in cucurbits, sweet corn, and dent corn by broadcasting them at rates of 10-100 lb (4.5-45 kg) of bait per acre (0.4 ha). Applications of these baits at 10-30 lb (4.5-13.5 kg) per acre produced reduc- tions of adult Diabrotica u. howardi, D. virgifera, and Acalymma vittatum rang- ing from 75 to 99-f percent within 1 to 3 days. Baits poisoned with methomyl and isofenphos were effective at appli- cation rates of 4.5-13.5 g of insecticide per acre (0.4 ha), and the decamethrin bait at rates as low as 0.45 g per acre September 1985 125 Years of Biological Research 193 Fig, 6.-Diabrotica beetles killed after 5 days of feeding on cut fruit of C. pepo x C. te.xana hybrid dusted with 0. 1 g of methomyl insecticide (Rhodes et al. 1980). (0.4 ha). (Metcalf et al. 1982, 1983a). In experiments in sweet and dent corn such applications killed from 160,000 to 230,000 Diabroticina beetles per acre (0.4 ha) and remained effective for at least 2 weeks. A notable feature of the Cuc-kairomone insecticide baits is that the quantity of insecticide re- quired per unit of area is only about 1 percent of that required for conven- tional spray applications for Diabro- ticina beetle control. The applications are highly selective because of the kairomonal effect on the Diabroticina, and other insects, including beneficial insects, are not appreciably affected. HOST-PLANT RESISTANCE BY ANTIXENOSIS (NONPREFERENCE) Antixenosis is a major type of host- plant resistance to insect attack in which the plant lacks the characteris- tics desired by insect pests and is an unsuitable host. Antixenosis, therefore, is the major resistance factor limiting most phytophagous insects to a mono- phagous or oligophagous host range. Increasing appreciation of the role of plant allelochemics acting as kairo- mones, suggests that genetic manipu- lations to remove them from cultivars is a logical approach to host-plant resistance. Indeed, this removal has oc- curred inadvertently in the Cucur- bitaceae from primitive man's efforts to find palatable squash, melons, and cucumbers free of the bitter cucur- bitacins. However, antixenosis has had only limited exploitation in the de- velopment of pest-resistant cultivars compared with the development of an- tibiosis, involving genetic manipula- tion to produce or increase allelochem- ics or other factors adverse to host selection and pest development and reproduction (Kogan 1983). The importance of antixenosis as a mechanism for host-plant resistance in the Cucurbitaceae to cucumber beetles (Diabroticina) has been demonstrated for the squash Cucurbita pepo (Nath & Hall 1965), the watermelon CitruUus lanatus (Chambliss & Jones 1966b), and the cucumber Cucumis sativus (DaCosta & Jones 1971b). In the Cucurbitaceae, cucurbitacin synthesis is initiated by a single dominant "bit- ter gene," Bi, and antixenotic resistance is associated with the recessive pheno- type, bi bi (DaCosta & Jones 1971a; Robinson et al. 1976). Nonbitter fruit may develop from bitter seedlings in the presence of a modifier suppressing synthesis in the fruit. Quantitative 194 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 estimations of the total Cue content in Cucurbita have shown a partial domin- ance for low Cue content with a multi- genic additive type of inheritance (Nath & Hall 1965). The Cue content of the cotyledon leaves appears to be substantially independent of that in mature leaves and fruits (Ferguson et al. 1983b). Protection of the early growth stages of Cucurbitaeeae cultivars from attack by Diabroticina beetles is a major necessity for the successful sur- vival of these plants. Study of 46 cultivars of Cucurbita, Cucumis, and Citrullus showed that substantial amounts of Cues were present in the seedling stages of 29 of these cultivars at a limit of detection of about 0.1 ppm of fresh weight. These Cues were demonstrated by extraction and thin- layer chromatography, followed by the feeding of the spotted cucumber beetle, Diabrotica u. howardi, and the banded cucumber beetle, D. balteata (Ferguson et al. 1983b). Field experiments showed that thei-e was almost complete correlation be- tween Cue content of the cotyledon leaves of 25 Cucurbita cultivars and the extent of the destruction of seedling plants by early spring and summer feeding of Diabrotica u. howardi and Acalymma vittatum beetles (Ferguson et al. 1983b). This demonstration of the value of antixenosis provides a useful example of means for the selection of Cucurbita cultivars with substantial antixenotic host-plant resistance to Diabroticina pests. Such antixenotic host-plant resistant varieties can play an important role in IPM progi-ams. It should be emphasized that the cucur- bitacins also function as allomones to confer host-plant protection against a variety of herbivores apart from the Diabroticina. Therefore, the ultimate effect of antixenosis in crop protection is one of balance. For the Cucurbita- eeae, the depradations of the Diabro- ticina and the wilt diseases that they innoculate are so severe that the balance appears to be tilted strongly in favor of antixenotic protection. KAIROMONES IN TRAP CROPS From evolutionary considerations, it appears that the use of kairomones in situ in plants attractive to insects should provide an optimum way to con- trol insects (Martin 1940:317). The principle of insect control by "trap crop- ping" has its origin in folklore but can be developed on a rational basis through knowledge of the qualitative and quantitative nature of the kairo- mone contents responsible for insect attraction, arrest, compulsive feeding, and oviposition. Curtis (I860) recom- mended controling the parsnip web- worms, Depressaria depressella (Hiib- ner) and D. dauceUa (Denis & Schiffer- muUer), (Oechophoridae) by setting parsnip plants (Heracleum sphondyl- ium Linnaeus) 6-10 feet (1.8-3.0 m) apart among carrots (Daucus carota Linnaeus) (Umbelliferae). The web- worm moths prefer to lay their eggs upon the parsnips, and the larvae de- velop in the parsnip heads and can be destroyed by cutting and burning or by drowning. A modern example of successful trap cropping is Newsoms (1978) sug- gestion that lima beans iPhaseolus lunatus Linnaeus) planted contigu- ously with soybeans (Glycine max (L.) Merrill) will trap the Mexican bean beetle, Epilachna varivestis Mulsant. The principle of trap cropping is stated to be an important component of IPM for soybean pests, reflecting economy, minimum adverse effects on beneficial insects, reduction of environmental pollution by pesticides, and minimal selective pressure to delay insecticide resistance. Detailed knowledge of the kairo- mones present in plant species and varieties now affords a firm scientific basis for the application of the trap- crop principle. Contardi (1939) specu- lated about the use of bitter Cucurbita andreana as a trap crop to protect cucurbit cultivars from attack by Diabrotica speciosa. A single vine of Cucurbita ecuadorensis, whose leaves September 1985 125 Years of Biological Research 195 contain a total of 0.41 mg of Cue B per gi-am of fresh weight, planted at the periphery of a 0.1-acre (0.04-ha) melon patch of Cucumis melo cultivar was observed to have arrested hundreds of spotted cucumber beetles, while only a single beetle was found in the musk- melon. Experiments using a design of hybrid Cucurbita andrena x C. maxima plants, one on each side of plots of 20 muskmelon plants, Cucumis melo, showed that 82 percent of Diabrotica u. howardi, D. v. virgifera, and Acalymma vittatum were found on the trap crop (average 18 per plant) compared with those found on the melons (average 4.2 per plant) (unpublished data). Such trap crops can be used in conjunction with occasional applications of contact insecticides to destroy the pests on the trap crop, or the trap crop can be planted over an application of gi-anular systemic insecticide. Scientifically de- signed trap cropping, employing spe- cies or cultivars with high kairomone levels that do not hybridize with culti- vars, offers a technological challenge to the organic farmer and an economical method for substantially decreasing the use of insecticides. ACKNOWLEDGMENTS Research from the writer's labora- tory summarized here was supported in part by grants from the SEA, U.S. Department of Agriculture, Com- petitive Research Grants Office, No. 5901-0410-8-00C7, "Coevolutionary Behavior of Corn Rootworms and Cu- cumber Beetles Attacking Corn and Cucurbits," and 59-2171-1-1-1-659, "Kairomones in Corn and Cucurbits for Monitoring and Controlling Corn Root- worms." Any opinions, findings, conclu- sions, or recommendations are those of the author and do not necessarily reflect the view of the U.S. Department of Agriculture. A portion of the re- search was also supported by the U.S. National Science Foundation, Grant No. BNS-8305833, "Chemical Ecology and Coevolution of Dacini." The research contributions ofnum- erous colleagues, including A.M. Rhodes, W.C. Mitchell, R.A. Metcalf, Esther R. Metcalf, Jane Ferguson, W. Howe, and Po-Yung Lu, are gratefully acknowledged. Thanks are due to D. Lavie and P. Enslin for supplying samples of pure cucurbitacins. LITERATURE CITED Allwood, a. J., and T. S. Angeles. 1979. Host records of fruitflies (family Tephritidae) in tlie Northern Territory. Queensland Journal of Agi'icultural and Animal Sciences 36:105-113. Barthel, W. F., N. Green, I. Keiser, and L. F. Steiner. 1957. Anisyl acetone, synthetic at- tractant for male melon fly. Science 126:654. Bateman, M. a., V. Insuza, and P. Arretz. 1973. The eradication of the Queensland fruitfly from Easter Island. FAO Plant Protection Bulletin 21(5):111. Bemis, W. p., a. M. Rhodes, T. W. Whitaker, and S. G. Carmer. 1970. Numerical taxonomy applied to Cucurbita relationships. American Journal of Botany 57:404-412. Beroza, M.. B. H. Ale.xander, L. ¥. Steiner, W. C. Mitchell, and D. H. Miyashita. 1960. New synthetic lui'es for the male melon fly. Science 131:1044-1045. Berry, J. W., J. C. Scheerens, and W. P. Bemis. 1978. Buffalo gourd roots, chemical composi- tion and seasonal changes in steroid content. Journal of Agi'icultural and Food Chemistry 26:345-356. Blum, M. S., R. Boch, R. E. Doolittle, M. T Trib- ble, and J. G. Traynham. 1971. Honey bee sex attractant: conformational analysis, struc- tural specificity, and lack of masking activity of cogener. Journal of Insect Physiology 17:.349- 364. Bogawat, J. K., and S. N. Pandey 1967. Food preference in Aulacophora spp. Indian Jour- nal of Entomology 29:349-352. Branson, T F., and J. L. Krysan. 1981. Feeding and oviposition behavior and life cycle strat- egies o{ Diahrotica: an evolutionary view with implications for pest management. Environ- mental Entomology 10:826-831. Chambliss, O. L., and C. M. Jones. 1966a. Cucur- bitacins: specific insect attractants in Cucur- bitaceae. Science 153:1392-1393. , and . 1966b. Chemical and genetic basis for insect resistance in cucurbits. Amer- ican Society for Horticultural Science Pro- ceedings 89:394-405. 196 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Chittenden, F. H. 1910. Some insects injurious to truck crops: notes on tlie cucumber beetles. U.S. Department of Agi-iculture Entomology Bulletin 82:(5):67-75. CoNTARDi, H. 1939. Estudios genticos en "Cucur- bita" y consideraciones agronomicas. Physis 18:331-347. Cunningham, R. T, and L. F. Steiner. 1972. Field trials of cue-lure + naled on saturated fiber- board block for control of the melon fly by the male-annihilation technique. Journal of Economic Entomology 65:505-509. Curtis, J. 1860. Farm insects. Blackie and Son, Glasgow, Scotland. DaCosta, C. p., and C. M. Jones. 1971a. Cucum- ber beetle resistance and mite susceptibility controlled by the bitter gene in Cucumis sativus L. Science 172:1145-1146. , and . 1971b. Resistance in cucum- ber, Cucumis sativus to three species of cucumber beetles. HortScience 6:340-342. David, A., and D. K. Vallance. 1955. Bitter prin- ciples of cucurbitaceae. Journal of Pharmacy and Pharmacology 7:295-296. Drew, R. A. I. 1974. The responses of fruitfly species (Diptera:Tephritidae) in the South Pacific area to male attractants. Australian Entomological Society Journal 13:267-270. , D. L. Hancock, and M. C. Romig. 1981. Australian Dacinae (Diptera:Tephritidae); new species from Cape York Peninsula, a dis- cussion of species complexes and key to species. Australian Journal of Zoology 29:49-91. , and G. H. S. Hooper. 1981. The responses of fruit fly species (Diptera:Tephritidae) in Australia to various attractants. Australian Entomological Society Journal 20:201-205. Ferguson, J. E., D. C. Fischer, and R. L. Met- calf. 1983a. A report of cucurbitacin poison- ing in humans. Cucurbit Genetics 6:73-74. , R. L. Metcalf, and D. C. Fischer. 1985. Disposition and fate of Cucurbitacin B in five species of Diabroticina. Journal of Chemical Ecology. (In press.) , E. R. Metcalf, R. L. Metcalf, and A. M. Rhodes. 1983b. Influence of cucurbitacin con- tent in cotyledons of Cucurbitaceae cultivars upon feeding behavior of Diabroticina beetles (Coleoptera:Chrysomelidae). Journal of Economic Entomology 76:47-51. Fletcher, B. S., M. A. Bateman, N. K. Hart, and J. A. Lamberton. 1975. Identification of a fruitfly attractant in the Australian plant Zieria snulhii as 0-methyl eugenol. Journal of Economic Entomology 68:815-816. Fraenkel, G. 1959. The raison d'etre of secondary plant substances. Science 129:1466-1470. . 1969. Evaluation of our thoughts on secondai-y plant .substances. Entomologia Ex- perimcntalis et Applicata 12:473-486. Friedrich. H. 1976. Phenylpropanoid consti- tuents of essential oils, Lloydia 39:1-7. Geissman, T A., and D. H. G. Crout. 1969. Organic chemistrj' of secondary plant metab- olism. Freeman. Cooper & Company, San Francisco, CA. Hanson, K. R., and E. A. Havir. 1979. An intro- duction to the enzymology of phenylpropanoid biosynthesis. Pages 91-137 in T. Swain. J. B. Harborne, and C. Van Summers, eds.. Recent advances in phytochemistiy. Vol. 12. Appleton- Century-Crofts, New York. Harborne, J. B. 1982. Introduction to ecological chemistry. Academic Pi-ess, New York. Hardy, D E. 1949. Studies in Hawaiian fruitflies. Entomological Society of Washington Pro- ceedings 51:181-205. . 1979. Book review. Economic fruitflies of the South Pacific region by R. A. I. Drew, G. H. S. Hooper, and M. A. Bateman. Pacific Insects 20:429-432. Harvey, J., R. A. Metcalf, and R. L. Metcalf. 1983. Evolutionary relationships in Diabro- ticina. Ms. in preparation. Howe, W. L., J. R. Sanborn, and A. M. Rhodes. 1976. Western corn rootworm adult and spotted cucumber beetle associations with Cucurbita and cucurbitacins. Environmental Entomol- ogy 5:1043-1048. HowLETT, F M. 1912. The effect of oil of citronella on two species of Dacus. Entomological Socie- ty of London Ti-ansactions 60:412-418. . 1915. Chemical reactions of fruitflies. Bul- letin of Entomological Research 6:297-305. Itokawa, H., Y. Oshida. A. Iklta, H. Inatond. and T Adacm. 1983. Phenolic plant growth inhibi- tors from the flowei-s of Cucurbita pepa Ph>-to- chemistry 21:1935-1937. Kaiser, I., S Nak.xgawa. R. M. Kob.JlYashi. D. L. Chambers, T. Urago, and R. E. DooLnrLE. 1973. Attractiveness of cue-lure and the degi-adation product 4- (/>-hydroxyphenyl)-2- butanone to male melon fiies in the field in Hawaii. Journal of Economic Entomology' 66:112-114. Kawano, Y, W. C. MrrcHELL, and H. Matsumoto. 1968. Identification of the male oriental fruit- fly attractant in the golden shower blossom. Joui-nal of Economic Entomologj' 61:986-988. KoGAN. M. 1983. Plant resistance in pest manage- ment. Pages 93-134 i;j R. L. Metcalf and W H. Luckmann, eds.. Itroduction to insect pest management. 2nd ed. John Wiley & Sons. New York. Lavie. D.. and E. Glotter. 1971. The cucur- bitacins. a gioup of tetracyclic triterpenes Fortschritte der Chemie Organischer Natur- stoffe 29:307-362. Martin, H. 1940. The scientific principles of plant protection. 3rd ed. E. Arnold. London. Maulik, S. 1936. The fauna of British India In- cluding Ceylon and Burma. Coleoptera. Chrj- somelidae, Galerucini. Taylor and Francis, London. May. a. \V. S. 1953. Queensland host records for the Dacinae (fam. TephritidaeK Queensland Journal of Agricultural Science 10:36-79. September 1985 125 Years of Biological Research 197 Metcalf, R. L. 1979. Plants, chemicals, and in- sects: some aspects of coevolution. Ento- mological Society of American Bulletin 25(l):30-35. , J. E. Ferguson, D. Fischer, R. Lampman, and J. Andersen. 1983a. Controlling cucum- ber beetles and corn rootworms with baits of bitter cucurbit fruits and root. Cucurbit Genetics 6:79-81. , E. R. Metcalf, and W. C. Mitchell. 1981. Molecular parameters and olfaction in the oriental fruitfly Dacus dorsalis. National Academy of Science Proceedings USA 78:4007-4010. , , and L. W. Y. Lee. 1979. Evolution of olfactory receptor in oriental fruitfly Dacus dorsalis. National Academy of Science Proceedings USA 76:1561-1565. , R. A. Metcalf, and A. M. Rhodes. 1980. Cucurbitacins as kairomones for diabroticite beetles. National Academy of Science Proceed- ings USA 77:3769-3772. W. C. Mitchell, T. R. Fukuto, and E. R. Metcalf. 1975. Attraction of the oriental fruitfly Dacus dorsalis to methyl eugenol and related olfactory stimulants. National Acad- emy of Science Pi-oceedings USA 72:2501-2505. , and E. R. Metcalf. 1983b Olfac- tory receptors in the melon fly Dacus cucur- bitae and in the oriental fruitfly Dacus dorsalis. National Academy of Science Pro- ceedings USA 80:3143-3147. , A. M. Rhodes, R. A. Metcalf, J. Fergu- son, E. R. Metcalf, and Po-yung Lu. 1982, Cucurbitacin contents and Diabroticite feed- ing upon Cucurbita spp. Environmental Ento- mology 11:931-937. Mitcihell, E. R., ed. 1981. Management of insect pests with semiochemicals: concepts and prac- tice. Plenum Press, New York. Nath, R, and C. V. Hall. 1965. The genetic basis for cucumber beetle resistance in Cucurbita pepo L. American Society for Horticultural Science Journal 86:442-445. Newsom, L. D. 1978. Progi-ess in integrated pest management of soybean pests. Page 171 in E. H. Smith and D. Pimentel, eds.. Pest control strategies. Academic Press, New York. Nielson, J. K., L. M. Larsen, and H. J. Soren- son. 1977. Cucurbitacins E and I in Iheris amara, feeding inhibitors for Phyllotrela nemorum. Phytochemistry 16:1519-1522. Nordlund, D. a., R. L. Jones, and W. J. Lewis. 1981. Semiochemicals: their role in pest con- trol. John Wiley & Sons, Inc., New York. Priesner, E., M. Jacobsen, and H. J. Bestmann. 1975. Structure-response relationships in noc- tuid sex pheromone reception. Zeitschrift fuer Naturforschung 30:283-293. Puchalski, J. T, and R. W. Robinson. 1978. Com- parative electrophoretic analysis of isozymes in Cucurbita species. Cucurbit Genetics 1:28. Rehm, S. 1960. Die Bitterstoffe der Cucurbitaceae Ergebnisse der Biologie 22:106-136. , P. A. Enslin, a. D. J. Neeuse, and J. H. Wessels. 1957. Bitter principles of the Cucur- bitaceae VII. The distribution of bitter princi- ples in the plant family. Journal of the Science of Food and Agriculture 8:679-686. Rhodes, A. M., W. P. Bemis, T W. Whitaker, and S. E. Carmer. 1968. A numerical taxonomic study of Cucurbita. Brittonia 20:251-266. , R. L. Metcalf, and E. R. Metcalf. 1980. Diabroticite beetle response to cucurbitacin kairomones. American Society for Horticul- tural Science Journal 105:838-842. RiEK, E. F. 1970. Fossil history. Pages 168-186 in Insects of Australia. Melbourne University Press, Melbourne, Australia. Robinson, R. W., H. M. Munger, T. W. Whitaker, and G. W. Bohn. 1976. Genes of the Cucur- bitaceae. HortScience 11:554-568. ScHiNZ, H., and C. F. Seidel. 1961. Nachtrag zu der Arbeit Nr., 194 im Helvetica Chimica Acta 40:1829 (1937). Helvetica Chimica Acta 44:278. Shah, A. H., and R. C. Patel. 1976. Role of the tulse plant Ocimum sanctum in control of the mango fly Dacus correctus Bezzi. Current Science (India) 45:313-314. Sharma, G. C, and C. V. Hall. 1971. Influence of cucurbitacins, sugars, and fatty acids on cucurbit susceptibility to spotted cucumber beetle. American Society for Horticultural Science Journal 96:675-680. , and 1973. Relative attractance of spotted cucumber beetles to fruits of fifteen species of Cucurbitaceae. Environmental Entomology 2:154-156. Shaw, J, T, W. G. Ruesink, S. R Briggs, and W. H. LucKMANN. 1984. Monitoring populations of corn rootworm beetles with a trap baited with cucurbitacin. Journal of Economic En- tomology 77:1495-1499. Shorey, H. H., and J. J. McKelvey, Jr., eds. 1977. Chemical control of insect behavior. John Wiley & Sons, New York. Sinha, a. K., and S. S. Krishna. 1969. Feeding of Aulocophora foveicollis on cucurbitacin. Journal of Economic Entomology 62:512-513. , and . 1970. Further studies on the feeding behavior oi Aulacophora foveicollis on cucurbitacin. Journal of Economic Ento- mology 63:333-334. Smart, J., and N. F Hughes. 1973. The insect and the plant progressive paleological integration. Pages 143-155 in H. F. van Emden, ed., Insect/plant interrelationships. Entomological Society of London Symposium 6. Smith, R. F 1966. Distributional patterns of selected western North American insects. Entomological Society of American Bulletin 17:108-110. , and J. F Lawrence. 1967. Clarification of the status of the type specimen.s of Dia- broticites (Coleoptera, Chrysomelidae, Galerucini). University of California Press, Berkeley. Steiner, L. F 1952. Methyl eugenol as an attrac- tant for oriental fruitfly. Journal of Economic Entomology 45:241-248. 198 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 1957. Low cost plastic fruitfly trap. Jour- nal of Economic Entomology 50:508-509. _, W. C. Mitchell, E. J. Harris, T. T. KozuMA, and M. S. Fujimoto. 1965. Oriental fruitfly eradication by male annihilation. Journal of Economic Entomology 58:961-964. Takizawa, H. 1978. Notes on Taiwanese Chryso- melidae I. Kontyu (Tokyo) 46:123-134. Thorsteinson, a. J. 1960. Host selection in phytophagous insects. Annual Review of Ento- mology 5:193-218. Watt, J. M., and M. G. Breyer-Brandwijk. 1962. The medicinal and poisonous plants of south- ern and eastern Africa. 2nd ed. E. & Sl Living- stone, Edinburgh. Whitaker, T W., and W. P. Bemis. 1964. Evolu- tion in the genus Cucurbita. Evolution 18:553-559. , and . 1975. Origin and evolution of the cultivated Cucurbita. Torrey Botanical Club Bulletin 102:362-368. Wilcox, J. A. 1972. Chi-ysomelidae: Galerucinae, Luperini, Coleopterorum. Catalogus Supple- menta. Pars 78, Fasc. 3. 2nd ed. Dr. W. Junk, s^Gravenhagen, Netherlands. Early Contributions of Insect Toxicology To the Evolution of Environmental Toxicology Allan S. Felsot The celebration of an anniversary connotes the remembrance of the past from the perspective of the present with anticipation for the future. Sci- ence, like law, is built upon precedent. The continued advancement of any scientific endeavor must entail a peri- odic assessment of where that disci- pline has been. Historical observations inevitably lead to a better understand- ing of the present state of the art and of future needs. It is appropriate on the occasion of the 125th anniversary of the Illinois Natural History Survey to examine the historical roots of one of its disciplines. The study of environ- mental toxicology is relatively new, but it has a long tradition of support within both the Survey's Section of Economic Entomology and the University of Illi- nois' Department of Entomology. Environmental toxicology is a broad- ly based discipline that involves the in- tegrated efforts of scientists from many other disciplines. It is currently enjoy- ing unprecedented public attention in the media. Not since the publication of Silent Spring (Carson 1962) has more public attention been focused on envi- ronmental contamination from chemi- cals and the potential biological effects. Amidst all the controversy surround- ing the manufacture, use, and disposal of synthetic organic chemicals, it is easy to forget how this environmental awareness came about. More impor- tantly, the evolution of the science that provided the fundamental basis for this awareness has become obscured. Dr. Allan S. Felsot is an Associate Ento- mologist, Section of Economic Entomology, Illi- nois Natural History Survey. It is hypothesized that the field of environmental toxicology began as a subdiscipline of applied entomology. Early environmental toxicological i"e- search was mainly conducted by eco- nomic entomologists who needed to know the effective dose required to kill pest insects and the environmental fac- tors that would affect the efficacy of the pest control operation. Today, the Ento- mological Society of America recog- nizes pesticide chemistry and toxicol- ogy as research areas within the scope of two of its disciplinary sections. The objective of this paper is to document the early contributions of economic en- tomologists (especially insect toxicolo- gists) to the development of environ- mental toxicology by examining the en- tomological literature prior to 1962 (the year of publication of Carson's Silent Spring). This objective was ac- complished by scanning the Journal of Economic Entomology (JEE), volume 1 (1908) through volume 55 (1962), for papers that dealt with environmental toxicology. In addition, papers from other journals were cited if they seemed to have had a significant influence on the content and direction of the insect toxicological research reported in JEE. Seven areas of research that form the essence of modern environmental toxicology could be seen in the applied entomological literature. These includ- ed the measurement of toxicity; symp- tomology, mode of action, and metabo- lism; insecticide resistance; pesticide selectivity and comparative toxicology; insecticide residues and analytical methods development; hazard evalua- tion; and environmental chemodynam- ics. The contributions of entomologists to each of these areas were reviewed. 199 200 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 MEASUREMENT OF TOXICITY Chemicals have long been used to help control insect pests. Early ac- counts of insect toxicology were anec- dotal with little quantitative data from which to determine effective dosages even though the number of chemicals available was very limited by today's standards. Economic entomologists recognized the need to develop a stan- dardized system for comparing the tox- icity of different chemicals and for determining the efficacy of control. Early studies were devoted to find- ing appropriate bioassays for the deter- mination of toxicity of cyanide fumi- gants. Coleman (1911) developed a bioassay to determine the most effec- tive concentration of cyanide gas against scale insects. He observed mor- tality of insects in the laboratory at one-sixteenth the dosage that had been prevalent in fumigation work. Hartzell (1924) was the first to use correlation analysis to study the relation of control efficiency to dosage, temperature, and percentage of open space. Dosage was found to exert the greatest influence on control. Abbott (1925) presented a tech- nique that is still used routinely for correcting the observed mortality that may occur in the untreated population of a bioassay. A technique for comparing the rela- tive toxicities of several chemicals to a variety of invertebrate and vertebrate species was first reported by Marco- vitch (1928). He adapted a mathemati- cal formula, using time and concentra- tion as variables, to determine the numerical toxicities of arsenical and fluorine compounds. He suggested that the mosquito be used as a standard test organism in insect toxicology as the white rat was for pharmacological studies. Comparative toxicity tests of different compounds were also studied by Campbell (1930). Campbell used the concept of median lethal dose, the dose which kills 50 percent of a large gi'oup of animals, to express the toxicity of dif- ferent compounds to the silkworm and mosquito. The concept of median lethal dose had been proposed earlier by Ti-evan (1927) to replace the prevalent expression of minimal lethal dose, which was regarded as highly variable. Campbell further showed the impor- tance of the route of exposure to the toxicity of a compound. Shepard & Richardson (1931) also used Trevan's (1927) ideas to draw "toxicity curves," described as being characteristic of a specific insecticide regardless of how the dosage was measured. After Trevan's concept of median lethal dose was published, many toxi- cologists observed that the response of animal populations to increasing dosages of a toxicant followed a sig- moidal pattern. Bliss (1934) used this characteristic phenomenon to develop a standardized technique for determin- ing mortality at a specified dosage. He transformed dosage to logarithmic units and percent mortality to proba- bility units that he termed probits. The dosage-mortality curve became linear and thus amenable to regi'ession analy- sis. In later work, Bliss (1935a) dev^el- oped statistical procedures for calculat- ing dosage-mortality curves. In a re- lated paper Bliss (1935b) introduced the term "LD50" to replace median lethal dose in comparisons of dosage- mortality data. Later, Bliss developed techniques for expressing mortality relative to time of exposure and the relationship between exposure time and concentration (Bliss 1937, 19401 Although researchers have studied the intricacies of the dosage-mortality curve and have suggested improve- ments (e.g., Sun & Shepard 1947; Wadley & Sullivan 1943; Beard 1949; Lanchester 1951), Bliss' techniques have remained standard practice in all fields of toxicology. SYMTOMOLOGY, MODE OF ACTION AND METABOLISM Observations on pesticide sympto- mology, mode of action, and metabo- September 1985 125 Years of Biological Research 201 lism are classified today as pharmaco- kinetic or toxicodynamic studies. Early observations of symptoms exhibited by poisoned insects gave clues to the mode of insecticide action (Shafer 191 la,b; Tischler 1935). Mode of action studies gradually became more sophisticated with the evolution and advancements in biochemistry. For example, Fink (1927) compared glutathione levels in normal insects and in insects treated with arsenicals. His observations were in agreement with the published find- ings on arsenic toxicity in vertebrates (Voegtlin et al. 1925), that arsenic in- terfered with the normal functioning of glutathione. Metcalf& March (1949) produced a highly cited study on the mode of action of parathion in insects that influenced the course of insect tox- icology research. These authors con- ducted pioneer studies on the relation of the chemical structure of a series of organophosphorus insecticides to tox- icity and cholinesterase inhibition. They were the first to show that there was a direct relationship between in vitro cholinesterase inhibition and in vivo toxicity. The importance of metabolism in explaining differential toxicity among insect species was recognized early in the entomological literature (Fernald 1908). Prior to World War II few metab- olism studies were reported because analytical methods were too insensi- tive for the determination of toxicants and metabolites in small tissue sam- ples. However, the value of radiotracer methodology as a sensitive detection technique was shown in studies on arsenic disposition in the silkworm (Campbell & l.ukens 1931; Norton & Hansberry 1941). After the introduc- tion of DDT, the metabolism of chlorin- ated hydrocarbons in insects was studied by using only wet chemical methods of analysis (Ferguson & Kearns 1949; Sternburg et al. 1950; Sternburg & Kearns 1956). Most studies on insect metabolism after 1950 employed radioisotopes of the chlorinated hydrocarbon and organo- phosphorus insecticides (e.g., Roan et al. 1950; Lindquist et at. 1951a,b). In the 1950's and thei'eafter radioisotopes were increasingly used to study the metabolism and distribution of insecti- cides in nontarget vertebrates (e.g., Pankaskie et al. 1952; Robbins et al. 1956; March et al. 1956a,b; Lindquist et al. 1958; Kaplanis et al. 1959). Another area of early toxicodynam- ic research in the entomological liter- ature is the study of the effect of chem- ical interactions on toxicity and metab- olism. Specifically, synergistic interac- tions between insecticides and nontoxic compounds, such as sesamin, were found to be important in increasing the toxicity of the unstable pyrethrums (Haller et al. 1942; Kerr 1951). Robbins et al. (1959) first reported the syner- gism of organophosphates by piperonyl butoxide in mammals. The interactions between synthetic organic insecticides and botanical insecticides were also studied (Sun 1948). INSECTICIDE RESISTANCE The study of insecticide resistance has received a great deal of attention by insect toxicologists. Although many reviews of this subject have appeared since the 1940's (e.g., Quayle 1943; Babers 1953; Brown 1968; Georghiou & Taylor 1976), several points perti- nent to the development of environ- mental toxicology should be noted. First, insect resistance was originally reported in 1914 by Melander and in- volved the decreased effectiveness of lime-sulphur against the San Jose scale. Prior to 1940 most reports in- volved resistance of scale insects or the codling moth to various fumigants (Quayle 1943). After the introduction and widespread use of DDT in the mid- 1940's, reports of resistance grew ex- ponentially. March & Metcalf (1949a,b) were probably the first in the United States to report both DDT resistance and possible cross resistance to other chlorinated hydrocarbon insecticides in houseflies. Until that time most resis- tance studies involved anecdotal re- 202 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 ports of decreased control of pests in the field or the comparison of insecti- cide toxicity against laboratory strains and field-collected populations. A breakthrough in understanding the biochemical basis of resistance came with the reports of differences in DDT metabolism between susceptible and resistant houseflies (Sternburg et al. 1950; Perry & Hoskins 1951). When Sternburg et al. (1953) reported that DDT was enzymatically dehydrochlor- inated by resistant houseflies, the stage was set for sophisticated studies on detoxication mechanisms and the bio- chemical genetics of I'esistance. PESTICIDE SELECTIVITY AND COMPARATIVE TOXICOLOGY Pesticide selectivity refers to the determination of the relationship be- tween chemical structure and toxicity in various target and nontarget spe- cies. Comparative toxicology includes studies of differences among inverte- brates and vertebrates in detoxication mechanisms and pesticide interactions with biochemical targets. Kearns and his coworkers (e.g., Kearns & Flint 1937; Metcalf & Kearns 1941; Dahm & Kearns 1941) pioneered the concept of testing a large number of compounds in a homologous series against different insects to deter- mine the effect of chemical structure on toxicity. Early studies focused on nitro- gen-containing synthetic organics, such as cyclohexylamine, picramic acid, toluenesulfonyl chloride, and alkyl secondary amines. Other structure- activity studies followed with investi- gations of phosphorous esters (Ludvik & Decker 1947, 1951), N-heterocyclics (King & Frear 1943), DDT analogs (Metcalf 1948), and N-methyl carba- mates (Metcalf et al. 1960, 1962). Some studies were qualitative in the sense that long lists of compounds and per- centages of mortality were presented. Other studies became more quantita- tive by gi-aphing the relationship be- tween mortality and specific chemical properties (e.g., Dahm & Kearns 1941; Fukuto et al. 1961; Metcalf et al. 1962). Pesticide selectivity studies were important in reconciling the role of biological control of insect pests with that of chemical control. For example, Ripper et al. ( 1951 ) published a compre- hensive study on the comparative tox- icity of insecticides between pests and their natui-al enemies. Today, quantita- tive structui'e-activity coiTelations play an important role in the discovery of new pest control agents and in the development of com.pounds that are selective for target organisms. Although research had been con- ducted on the toxic effects of pesticides on nontarget vertebrates, few studies directly compared responses of insects and vertebrates before the advent of the synthetic organic insecticides. Mar- covitch ( 1928) was perhaps the first en- tomologist to compare the toxicity of inorganic insecticides between mosqui- toes, eai'thworms, and rabbits. Goldfish and cockroaches were compared in an investigation of the toxic effects of nic- otine (EUisor 1936). The first compre- hensive comparative toxicological study was reported by Metcalf& March (1950). These workers compared the propeities of acetylcholinesterase from the bee, the housefly, and the mouse to study the relationship between para- thion derivatives and insecticidal ac- tion. The motivation behind studies of this nature is clear from the authors' conclusion, "A detailed knowledge of the properties of various cholinesterases and correlation of the structure of various organic phosphate anticholin- esterases with their specific action may result in the development of insecti- cides with a gi-eater margin of safety to warm-blooded animals than those currently in use. and may enable the entomologist to select compounds toxic to certain insect pests, but relatively harmless to beneficial parasites and predators." Metcalf & March (1950) gi'eatly influenced the course of toxi- cology as evidenced by the volume of September 1985 125 Years of Biological Research 203 research devoted to comparative toxi- cological studies in the 1950's and 1960's (e.g., Johnson et al. 1952; Casida & Stahmann 1953; March et al. 1955, 1956a; O'Brien 1956, 1957a,b; O'Brien et al. 1958; Krueger & Casida 1957; Krueger & O'Brien 1959; Krueger et al. 1960; Casida et al. 1960; Afshar- pour & O'Brien 1962). INSECTICIDE RESIDUES AND DEVELOPMENT OF ANALYTICAL METHODS The problem of pesticide residues on food and in the environment has concerned entomologists since the early 1900's. JEE has long served as a forum for the reporting of residues. This func- tion was especially important when there was essentially no other outlet for this kind of investigation. Prior to World War II and the introduction of DDT, most investigations centered on the occurrence of arsenic and lead resi- dues owing to the heavy spraying of lead arsenate for the control of orchai'd pests. Early studies were concerned with arsenic residues in soils because of the potential for phytotoxicity to trees in alkaline soils (Ball et al. 1910; Headden 1910). Later studies measured the arsenic and lead content of sprayed apples (O'Kane 1913; MacLeod et al. 1927; Hartzell & Wilcoxon 1927, 1928; McLean & Weber 1928). Attention was also given to residues of the organic in- secticides, nicotine, and derris (Norton & Billings 1941; Cassil 1941). The introduction of DDT into gen- eral use in agriculture greatly in- creased the amount of insecticides used. Consequently, the attention devoted to pesticide residues greatly in- creased. The prolonged persistence of DDT was quickly recognized. Almost every crop that had been sprayed with DDT was studied for residues (e.g., Wilson et al. 1946b; Borden 1947; Smith et al. 1948a; Eden & Arant 1948). Residues of other synthetic organic insecticides, especially the organophosphates, were investigated almost as soon as they were introduced (e.g., Ginsburg et al. 1949, 1950; Hos- kins 1949; Smith et al. 1952; Gunther & Jeppson 1954; Gunther et al. 1954; Brett & Bowery 1958). The tremendous interest in pesti- cide residues after the introduction of the synthetic organics was more than mere concern about the possibility of exceeding established tolerances. Shoi't- ly after the general introduction of DDT it was discovered that this chem- ical could be bioconcentrated in milk and passed through the food chain (Telford & Guthrie 1945; Woodward et al. 1945; Wilson et al. 1946a). Many reports on the occurrence of DDT and other chlorinated hydrocarbon insecti- cides in milk and meat products ap- peared in JEE after 1946 (e.g.. Smith et al. 1948b; Carter et al. 1949a,b; Claborn et al. 1950a,b; Frear et al. 1950; Bushland et al. 1950; Fahey et al. 1955; Ely et al. 1957). These studies generally showed that DDT could bio- concentrate in milk regardless of whether the insecticide was applied directly to cows, barns, or feed. Studies of organophosphate and carbamate in- secticide residues in milk soon followed the DDT studies, but these indicated no significant bioconcentration (Goulding & Terriere 1959; Eheart et al. 1962). Pesticide residues were also studied in poultry and eggs. Lindane, a chlor- inated hydrocarbon, was found at high levels in fat and eggs several months after poultry houses had been sprayed, and recommendations for its use were cancelled (Ivey et al. 1961). Residues of Co-Ral, an organophosphate, were ex- tensively studied by Dorough et al. (1961a,b). These authors emphasized the need also to study metabolite resi- dues. Bioconcentration of carbamate insecticides was shown to be insignifi- cant in a study of carbaryl residues in poultry products (McCay & Arthur 1962). It should be noted that residue studies of organophosphate and carba- 204 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 mate insecticides during the 1950's and 1960's were using detection limits of approximately 0.1 ppm. Today, detec- tion limits easily reach the 0.01 or 0.001 ppm level, and the probability of finding "biodegradable" pesticide resi- dues in food has increased accordingly. Food processing, such as the mill- ing of wheat, was found to aid in resi- due reduction (Schesser et al. 1958). Siakotos (1956b) observed that vapor- ized lindane could enter stored pack- aged foods regardless of the container. Since it was observed that pesticide residues on food were generally inevi- table, some attention was devoted to techniques for residue removal. A num- ber of studies were conducted on the removal of lead arsenate residues from fruit (Robinson 1929; Fisher 1931; Weber & McLean 1933); later, the effect of washing on organophosphates was studied (Smith et al. 1955). Concomitant with the need for ac- curate pesticide residue determina- tions was the development of sensitive analytical methods. Although other journals published analytical methods for pesticide residues (e.g.. Journal of the Association of Agricultural Chem- ists, published prior to World War II, and the Journal of Agricultural and Food Chemistry, initiated in 1953), JEE had been a reliable forum for this kind of information since its inception. Ear- ly reports included methods for the analysis of nicotine (Safro 1917), lead arsenate (Hamilton & Smith 1925; Ginsburg 1928), oil (Ebeling 1940; McCall & Kagy 1940), and rotenone (Gunther 1942). The widespread use of the synthet- ic organic pesticides stimulated investi- gation of more sophisticated and diverse techniques that were required for the analysis of the gi'owing number of sampling matrices (Carter 1955). Originally the chlorinated hydrocar- bons were determined by total chlorine analysis (Ginsburg 1946). Eventually, chromatogi'aphic separation methods and, later, colorimetric analysis were adopted for the analysis of many organ- ochlorine insecticides (e.g., DDT, Schechter & Haller 1945; aldrin and dieldrin, O'Donnell et al. 1954, 1955; heptachlor and chlordane, Polen & Silverman 1952; lindane, Schechter & Hornstein, 1952; EDB, Sinclair & Crandall 1952). Radiolabelled pesticides were useful for testing the extraction of residues from complex matrices, such as milk (Timmerman et al. 1961). A cholinesterase-inhibition technique, useful for detecting organo- phosphate insecticides (Giang & Hall 1951), was modified for analyzing weak inhibitors, such as the organophosphor- odithioates (Miskus et al. 1959; Miskus & Hassan 1959). The bioassay of ex- tracts, using mosquito larvae, was employed as a sensitive technique for the detection of toxic compounds and metabolites (Bushland 1951; Hartzell 1952). Sampling techniques to ensure reliable residue data wei'e also studied (Cassil et al. 1943; Anderson & Gun- ther 1951; Van Middelem et al. 1956; Huddleston et al. 1960b; Lichtenstein et al. 1960). In sum, the literature on insecti- cide residues shows a long-standing and deep concern among entomologists about health hazards in food supplies. Perhaps the prevailing philosophy was expressed best by White (1934), who stated at a meeting of the American Association of Economic Entomolo- gists, "Let us not again go along for years putting on other poisons in total ignorance of the amounts going to the consumer and of the effects upon his health." This theme was acted upon in 1951 when the chairmen of most mid- western departments of entomologA- met in Chicago and prepared the outline of a research project entitled, "Hazards Resulting from the Use and Misuse of Pesticides and Means for their Elimination" (E.P. Lichtenstein personal communication). This project was activated in 1954 and represented one of the first regional projects (NC-19) established in entomology-. The objective of this cooperative venture was to isolate, define, minimize, or eliminate hazards connected with the use of pesticides. September 1985 125 Years of Biological Research 205 HAZARD EVALUATION Hazard evaluation involves the determination of acute and chronic effects on nontarget organisms and on the environment from direct or indirect exposure to pesticides or their metab- olites. Four main areas of hazard eval- uation research that have appeared in the entomological literature are haz- ards to livestock and other mammals; hazards to nontarget insects, fish, and birds; effects on soil biota and plants; and effects on human health. The use of insecticides to protect livestock from insect infestations that might reduce production efficiency had been investigated since the late 1920's (Melvin 1932). Naturally, the safety of the animals was a major concern. Early studies focused on the physiological and toxic effects of oil sprays on cattle (Melvin 1932; Freeborn et al. 1934; Atkeson et al. 1944). No acute effects owing to the treatment of forage with pyrethrum (De Ong 1937) or cryolite (Wilford & Mott 1944) were observed in pastured livestock. In other tests it was observed that the injection of pyreth- rum extract into rats produced severe adverse reactions (Leonard 1942), but a similar mode of exposure to cryolite failed to produce acute toxic symptoms (Sweetman & Bourne 1944). Feeding studies with chickens showed no ad- verse effects with low doses of arsenic bait, but it reduced egg laying at high doses (Wilson & Holmes 1936). The first widespread use of DDT after World War II was for the control of livestock pests. Immediately after the introduction of DDT there was a proliferation of reports on its effects on a wide variety of animals. Much of this research attempted to etablish safe levels of insecticide exposure resulting from single or repeated adminstrations by direct spraying or by feeding (e.g., Orr & Mott 1945; Telford & Guthrie 1946; Wilson et al. 1946a; Bushland et al. 1948; Batte & Turk 1948; Welch 1948; Radeleff 1950). Both acute toxic symptoms and histopathological changes were recorded in response to extremely high dosages and dosages likely to be received under normal use conditions. In general, low dosages of many of the chlorinated hydrocarbons had little ef- fect, but moderate to severe histopatho- logical effects were noted at the highest dosages. Ingle (1947) used white rats to com- pare the acute and chronic toxicities of chlordane and DDT. From 1950 to 1955 there were a number of studies on the toxicity of the new chlorinated cyclo- diene insecticides to poultry (Eden 1951; Turner & Eden 1952; Arant 1952; Sherman & Rosenberg 1953, 1954). Moore (1952) studied the toxic ef- fect of lindane vapors on canaries and pigeons. After the early 1950's, research was directed to the acute effects of exposure to the organophosphate insecticides. Studies were conducted on the toxicity of direct application (Furman & Wein- mann 1956), exposure through treated forage (Dobson et al. 1960), and expo- sure to vapors (Ti-acy et al. 1960). Re- search was also directed toward the toxicity of systemic organophosphate insecticides intentionally added to feeds (Radeleff & Woodard 1956, 1957). Methods for measuring cholinesterase activity in erythrocytes wei'e developed to determine the degree of exposure of livestock to organophosphate residues (Robbins et al. 1958; Hermenze & Goodwin 1959). Effects of organochlor- ine and organophosphate insecticides on the metabolism and motility of mammalian spermatozoa were studied by Beck (1953). The use of DDT and organophos- phates as mosquito larvicides stimu- lated research on the acute toxicity of these pesticides to fish and aquatic invertebrates (Ginsburg 1945, 1947; Eide et al. 1945; Darsie & Corriden 1959; Mulla & Isaak 1961). Some studies determined the acute toxicity to fish of insecticides that could reach water through indirect routes (Schulze et al. 1952; Haynes et al. 1958). Other studies investigated the ecological ef- 206 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 fects of insecticides on aquatic com- munites (Tiller & Cory 1947; Hitchcock 1960; Webb 1960). Arsenic baits used in grasshopper control were found nontoxic to pheas- ants (Lilly 1940). DDT-treated insects were fed to nestling birds in one study, and toxicity was only manifested when the birds' food supply was limited (George & Mitchell 1947). Although the effects of insecticides on bees had received much attention before 1900 (Shaw 1941), little research had been conducted on the hazards of insecticides to other beneficial insects, including predators and parasites, be- fore the introduction of DDT (Haug & Peterson 1938). Both arsenicals and DDT were the focus of much of the re- search on pollinators before 1950 (Doane 1923; Webster & Crews 1934; Linsley & MacSwain 1947). Observa- tions by many workers of the resur- gence of pest populations and the emer- gence of secondary pests after crops were sprayed with DDT (DeBach 1947; Newson & Smith 1949; Griffiths 1951) stimulated further research on the acute toxicity of various insecticides to beneficial insects (Ripper et al. 1951). Concern for the potential hazards associated with pesticide accumulation in soil was manifested long before the introduction of the persistent chlorin- ated hydrocarbons. Hyslop (1914) ex- pressed concern about the effect of cyanide on beneficial soil microbiota. Others studied the effect of the accum- ulation of lead and calcium arsenate on their phytoxicity (Scott & Karr 1942; Fleming et al. 1943). The concern with adverse effects on microbiota and with phytoxicity continued after the intro- duction of DDT (Wilson & Choudhri 1946; Morrison et al. 1948; Cullinan 1949; Stitt & Evanson 1949; Gould & Hamstead 1951; Simkover & Shenefelt 1951) even though the accumulation of DDT residues from earlier applications had not been recognized until the 1950's (Chisholm et al. 1950; Ginsburg & Reed 1954; Lichtenstein 1957). Most studies did show some phytotoxic ef- fects from chlorinated hydrocarbons, such as lindane (Morrison et al. 1948). In general, few effects on soil micro- organisms or microarthropods were observed after insecticide treatments (Bollen et al. 1954a,b; Hartenstein 1960). Since the early 1970's gi-eat empha- sis has been placed on the assessment of human exposure to pesticides. Re- cently a research conference and work- shop on minimizing occupational expo- sure to pesticides were conducted (Gun- ther 1980). Entomologists have always been interested in this topic because of their own potential exposure. The first study on human hazard assessment published in JEE concerned the effects of fumigation with hydrocyanic gas (Yothers 1910). A number of studies before World War II commented on the hazards to human health from various pesticides, including arsenic (O'Kane 1916; Huckett 1934;), derris (Wells et al. 1922), and cryolite (Marcovitch & Stanley 1938). Wilson & Holmes (1936) assessed the hazards of eating arsenic- tainted poultry by comparing the resi- due values obtained from arsenic-fed chickens with levels known to be toxic to humans. Marcovitch & Stanley (1938) also used cryolite residue values on food and the median lethal dose to humans to determine the "factor of safety" for this pesticide. Chlorinated hydrocarbon insecti- cides (especially DDT) were initially thought to be safe for humans because large differences in acute toxicities of these compounds were observed be- tween insects and mammals. Human hazard assessments of these com- pounds were absent from the early entomological literature. On the other hand, several entomologists conducted detailed hazard evaluations of the organophosphate insecticide, parathion (Griffiths et al. 1951; Ashdown et al. 1952; Braid & Dustan 1955; Quinby et al. 1958). Due to the extremely high acute toxicity of parathion and its lack of selectivity, it was deemed respons- ible for a number of severe injuries and deaths resulting from occupational ex- September 1985 125 Years of Biological Research 207 posures. Studies showed that absorp- tion of parathion through the skin over several exposure periods created the greatest hazard to appUcators and field hands. Metcalf (1951) developed a tech- nique for estimating blood cholinester- ase that was useful in the detection of possible exposure to organophosphorus insecticides. Fulton et al. (1955) evalu- ated the efficiency of respiratory protec- tive devices for agi-icultural use. Protec- tive clothing, gloves, and respirators had long been known to affoi'd protec- tion against overexposure to pesticides (Metcalf 1951). Entomologists have consistently recommended using pesti- cides only when necessary and taking proper safety precautions. ENVIRONMENTAL CHEMODYNAMICS The study of the fate and behavior of pesticides in the environment has emerged as a relatively new discipline called environmental chamodynamics (Haque & Freed 1974). Pesticide cham- odynamics essentially involves four areas of I'esearch: 1. Physicochemical properties of pesticides that influence their behavior in the environment; 2. Mechanisms of the partitioning of pesticides among environmental components (air, water, soil, biota); 3. Attenuation processes affecting pesticide fate (e.g., photodecomposition, chemical and microbial degradation, etc.); 4. Understanding and modeling of environmental transport processes. The ultimate objective of environ- mental chemodynamic research is to assess accurately the exposure of target or nontarget organisms to a pesticide or any contaminant. Since exposure to a chemical is largely dependent on its environmental distribution, the eco- logical impact of pesticides in any eco- system is directly related to the sum of the chemodynamic processes. Pesticide chemodynamic research appearing in JEE between 1908 and 1962 included: residue dynamics on plants; translocation into plants and systemic dynamics; metabolism in plants and transformation products; persistence and translocation in soil; volatilization; formulation chemistry; and bioconcentration. The determina- tion of pesticide residues on crops had received a gi'eat deal of attention owing to the establishment of tolerances by the Food and Drug Adminstration and the concern about hazards to human health. In contrast, the interest in residue dynamics seemed to be moti- vated by rapid decreases in insecticidal activity after application. Early studies investigated the effect of fruit growth and climatic conditions on residue losses (Hamilton 1929; Fahey & Rusk 1940). The effect of sunlight and oxy- gen on residues was also studied (Jones et al. 1933; Gunther 1943), and meth- ods were explored for the inhibition of insecticide decomposition (Gunther et al. 1948). Some studies were concerned only with the dispersion of residues on foliage and their subsequent rates of loss (MacLeod & Sherwood 1937; Gun- ther et al. 1946; Dawsey & Markwood 1940). In the 1950's the focus of pesticide research shifted to the study of residue dynamics with respect to the evalua- tion and elimination of hazards. Trans- location of foliar-applied materials into other parts of the plant, such as the fruit, and the actual location of the residues therein were studied (Smith & Clifford 1950; Blinn et al. 1959; Mat- sumura 1960). Other studies concen- trated on the effects of application rate, number of applications, weathering, and crop dehydration in reducing resi- dues (Stansbury & Dahm 1951; Sloan et al. 1951a,b; Dahm 1952; Hopkins et al. 1952a,b; Westlake & Butler 1953; Waites & Van Middelem 1958). Many researchers graphically displayed resi- due persistence data with the concen- tration recovered as a dependent vari- able and the time after application as an independent variable. The resulting curves were biphasic; an initial rapid loss of pesticide was followed by a much 208 Illinois Natural History Survey BuLLEmN Vol. 33. Art. 3 slower degradation rate (e.g., Decker et al. 1950; Hopkins et al. 1952a,b; Fahey et al. 1952). Gunther & Blinn (1955) discussed these curves as representing a two-step process in which the rapid initial loss of residue is related to weathering, and the slower loss is re- lated to metabolically induced alter- ations. The application of systemic insecti- cides to soil has been viewed as a rela- tively efficient and selective method for controlling pests. Questel & Connin (1947) were first to show that the treat- ment of soil with parathion produced plant tissues lethal to the European corn borer. Terriere & Ingalsbe (1953) showed that potato tubers absorbed chlorinated hydrocarbon insecticides from soil as long as 3 years after the initial treatment. Getzin & Chapman (1959) proved that soil type greatly influenced the amount of insecticide translocated from the soil. Reynolds & Metcalf (1962) showed that there was a direct relationship between the water solubility of a chemical and the amount taken up by a plant. The metabolism and systemic dy- namics of organophosphate insecticides were extensively studied by Metcalf, March, Casida, and their coworkers. Casida et al. (1952, 1954) showed that plants could metabolize octamethylpy- rophosphoramide to a toxic, anticholin- esterase product. Metcalf et al. (1955, 1957a) and Fukuto et al. (1956) showed that phorate and the isomers of Systox were oxidized in plants to toxic metabo- lites. Gannon & Decker (1958) demon- strated that aldrin, a chlorinated cyclo- diene, was oxidized in plants to the equally toxic epoxide, dieldrin. Metcalf & March (1952) and Metcalf et al. (1954, 1956, 1957b) conducted detailed, comprehensive studies on the systemic dynamics of radiolabeled organophos- phates. Radioautogi'aphs of whole leaves were used to visualize distribution of the insecticide. Throughout the 1950's studies pro- liferated on the persistence of the chlorinated hydrocarbon and cyclo- diene insecticides in soil (Fleming & Maines 1953, 1954; Kiigemagi et al. 1958; Young & Rawlins 1958; Lichten- stein & Polivka 1959; Lichtenstein et al. 1960). Little work had been pub- lished on the persistence of the organ- ophosphates at that time (Menn et al. 1960). The major concern seemed to be with the accumulation of residues in soil and their possible phytotoxic ef- fects. There was also a need to under- stand residue dynamics so that control of soil pests could be optimized. Lich- tenstein and his coworkers conducted extensive research on the effects of soil type, temperature, moisture, mode of application, and cover ci'op on pesticide persistence and translocation (Lichten- stein 1958; Lichtenstein & Schulz 1959b, 1961; Lichtenstein et al. 1962). Some work was reported on the adsorp- tion of insecticides by soils (Chisholm & Koblitsky 1943; Getzin & Chapman 1959; Weidhaas et al. 1961) and on volatilization (Harris & Lichtenstein 1961). In most of the insecticide-soil interaction studies, soil organic matter content appeared to be the most signifi- cant property affecting pesticide trans- location and persistence. Edwai-ds et al. (1957) showed an inverse relationship between soil organic matter content and toxicity. Several studies proved that insecti- cides could be oxidized in soil to biolog- ically active metabolites. Gannon & Bigger (1958) and Lichtenstein & Schulz (1959a) were the first to show that aldrin and heptachlor were trans- formed into the toxic epoxides, dieldi'in and heptachlor epoxide. Lichtenstein & Schulz (1960) showed that the transfor- mation was biological and did not take place in autoclaved soils. The organo- phosphate insecticide, phorate, was also shown to be oxidized in soil or by soil microorganisms to products simi- lar to those found in plants by Metcalf and his coworkers (Ahmed & Casida 1958; Getzin & Chapman I960). Several early experiments measm'ed the volatility of insecticide formula- tions containing nicotine, DDT, and some fumigants in relation to effects on insect control (DeOng 1923; Roark & September 1985 125 Years of Biological Research 209 Nelson 1929; Fleck 1944). One study reported the contamination of food and air by lindane applied as a surface spray (Siakotos 1956a). Other studies of formulations measured the efficiency of various materials as carriers for dif- ferent insecticides (Headlee & Rudolfs 1922; Eddy 1926; Weidhaas 1957) and the stability of various spray mixtures (Gunther 1947; Cutkomp 1947). No studies were found in JEE prior to 1962 that measured bioconcentra- tion factors, but the accumulation of insecticides in nontarget organisms after they had fed on treated forage or after area-wide sprays was studied (Wilson et al. 1946a). Early studies reported the absorption of arsenic by vegetables grown in soils treated with lead arsenate (McLean et al. 1944). Dairy cows pastured in fields sprayed with chlorinated hydrocarbon and cyclodiene insecticides secreted the pesticides in their milk (Entomology Research Division 1959; Gannon & Decker 1960). DDT residues were found in dairy and poultry products from New York farms following widespread aerial spraying for gypsy moth control (Huddleston et al. 1960a). One study showed that leaves from elm trees in- jected with the organophosphate, Tet- ram, were toxic to earthworms and rats (Al-Azawi et al. 1961). CONCLUSIONS The publication of Silent Spring (Carson 1962) has been credited with informing the general public of the hazards of pesticide use and creating an attitude of enviornmental aware- ness. The purpose of this historical review of environmental toxicology was to trace the evolution of this environ- mental awareness from the perspective of the entomological literature. It be- came clear after reviewing over 200 papers published primarily in the Jour- nal of Economic Entomology between 1908 (volume 1) and 1962 (volume 55) that the interest in envii-onmental toxi- cology evolved from the research efforts of insect toxicologists before and im- mediately after World War II. JEE, originally published as the proceedings of the American Association ofEconomic Entomologists, served as a continous forum for a multitude of insecticide tox- icology studies. The Journal of Agri- cultural and Food Chemistry, published by the American Chemical Society, also served as a forum after 1953. Today more than 10 journals regu- larly publish research on environmen- tal toxicology. Many of these journals have existed for less than 15 years. Toxicological research encompasses a gi'eat number of environmental con- taminants, and researchers are trained in a wide variety of fields. It is perti- nent to note that the curi'ent and past scientific advisory panels to the U. S. Environmental Protection Agency have had entomologists trained in in- sect toxicology as members. The com- bination of strong training in biology and chemistry has enabled insect toxi- cologists to adapt their research to a wide range of environmental problems that go well beyond the bounds of tradi- tional economic entomology. With con- siderable foresight Moore (1923) ob- served, "The economic entomologist should be well grounded in chemistry, especially organic and physical chem- istry for the purpose of enabling him to see and solve the problems in con- nection with the use of insecticides." In retrospect, it appears that eco- nomic entomologists have traditionally been concerned with both the benefits and the risks of pesticide use. The dilemma of the entomologist in balanc- ing crop protection with environmental protection was perhaps best expressed by Dr. George Decker of the Illinois Natural History Survey. In a presenta- tion about DDT residues before the In- sect Control Committee of the National Research Council, Decker (1946) said, "The entomologist frequently finds himself in difficult positions. He is ex- pected to assist the farmers by supply- ing information on the latest and best possible recommendations for the con- trol of insect pests. At the same time. 210 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 he is obligated to guard against recom- mending any pi'ocedure or treatment that might endanger the life or health of man or beast. In the absence of reli- able data, he must proceed with rea- sonable caution and must at all times avoid being carried away by current waves of public opinion. Today, he must balance the merits ofDDT against the hazards that might accompany its use and weigh carefully the evidence on both sides of the equation." It is incumbent upon insect toxi- cologists to understand and explain the diverse aspects of the fate and behavior of insecticides in the environment so that they may be used in the most effi- cient and safest manner possible. LITERATURE CITED Abbott, W. S. 1925. A method of computing the effectiveness of an insecticide. Journal of Economic Entomology 18:265-267. Afsharpour, F, and R. D. O'Brien. 1962. Hydro- lases in insects and mammals. Jom-nal of Eco- nomic Entomology 55:317-320. Ahmed, M. K., and J. E. Casida. 1958. Metabo- lism of some organophosphorus insecticides by microorganisms. Journal of Economic En- tomology 51:59-63. Al-Azawi, a., D. M. Norris, Jr., and J. E. Casida. 1961. Hazards associated with the implanta- tion of tetram into elm trees for Dutch elm disease control. Journal of Economic En- tomology 54:127-129. Anderson, L. D., and K A. Gunther. 1951. Samp- ling techniques for determination of DDT residue on sweet corn. Journal of Economic Entomology 44:1008-1010. Arant, F. S. 1952. Toxicity of aldrin to chickens. Journal of Economic Entomology 45:121. Ashdown, D., R. G. Dahms, W. O. Ridgway, and C. F. Stiles. 1952. Hazards in the use of para- thion for greenbug control. Journal of Eco- nomic Entomology 45:82-84. Atkeson, F. W., a. R. Borgmann, R. C. SMrrH, and A. O. Shaw. 1944. Relative effects of several base oils used in livestock sprays on the skin of cattle. Journal of Economic Entomologj' 37:419-428. Babers, F H. 1953. Chemical control and resis- tance to insecticides by agi-icultural pests. Journal of Economic Entomology 46:869-873. Ball, E. D., E. G. Titus, and J. E. Greaves. 1910. The season's work on arsenical poisoning of fruit trees. Journal of Economic Entomology 3:187-197. Batte, E. G., and R. D. Turk. 1948. Tbxicity of some synthetic insecticides to dogs. Journal of Economic Entomology 41:102-103. Beard, R. L. 1949. Time of evaluation and the dosage-response curve. Journal of Economic Entomology 42:579-585. Beck, S. D. 1953. Effect of insecticides on the metabolism and motility of mammalian sperm- atozoa. Journal of Economic Entomologj' 46:570-574. Blinn, R. C, G. E. Carman, W. H. Ewart. and F A. Gunther. 1959. Residual behavior of various insecticides on and in lemons and oranges. Journal of Economic Entomolog>' 52:42-44. Bliss, C. I. 1934. The method of probits. Science 79:38-39. 1935a. The calculation of the dosage- mortality curve. Annals of Applied Biology 22:134-167. 1935b. The comparison of dosage mortali- ty data. Annals of Applied Biology 22:307-333. . 1937. The calculation of the time-mortal- ity curve. Annals of Applied Biology 22:134-167. 1940. The relation between exposure time, concentration and toxicity in experi- ments on insecticides. Entomological Society of America Annals 33:721-766. BoLLEN, W. B., H. E. Morrison, and H. H. Crowell. 1954a. Effect of field treatments of insecticides on numbers of bacteria. Strep- tomyces, and molds in the soil. Journal of Economic Entomology 47:302-306. and L. 1954b. Effect of field and laboratory treatments with BHC and DDT on nitrogen transformations and soil respiration. Journal of Economic Entomology 47:307-312. Borden, A. D. 1947. DDT dust deposits on peai-s. Journal of Economic Entomolog\- 40:926-927. Braid. P. E., and G. G. Dustan. 1955. Parathion residues on imniatui-e peaches and the hazard in spraying and thinning operations. Journal of Economic Entomologj' 48:44-46. Brett. C. H.. and T G. Bowery. 1958. Insecticide residues on vegetables. Journal of Economic Entomolog>' 51:818-821. Brown. A. W. A. 1968. Insecticide resistance comes of age. Entomological Society of Amer- ica Bulletin 14:3-9. September 1985 125 Years op Biological Research 211 Bushland, R. C. 1951. Attempts to utilize mos- quito larvae in a bioassay method for insecti- cide residues in animal products. Journal of Economic Entomology 44:421-423. , H. V. Claborn, H, F. Beckman, R. D. Radeleff, and R. W. Wells. 1950. Contamina- tion of meat and milk by chlorinated hydro- cai-bon insecticides used for livestock pest control. Journal of Economic Entomology 43:649-652. R. W. Wells, and R. D. Radeleff. 1948. Effect on livestock of sprays and dips contain- ing new chlorinated insecticides. Journal of Economic Entomology 41:642-645. Campbell, F. L. 1930. A comparison of four methods for estimating the relative toxicity of stomach poison insecticides. Journal of Economic Entomology 23:357-370. , and C. Lukens. 1931. A radioactive indi- cator method for estimating the solubility of acid lead ai'senate within the alimentary tract of the silkworm. Journal of Economic Ento- mology 24:88-94. Carson, R. 1962. Silent Spring. Fawcett Publica- tions, Inc., Greenwich, CT. Carter, R. H. 1955. Status of analytical methods with respect to the determination of minimal quantities of insecticides. Joui'nal of Economic Entomology 48:424-425. P. E Hubanks, H. D. Mann, F F Smith, P. G. PiQUETT, J. C. Shaw, and L. P. Ditman. 1949a. DDT content of milk from cows fed pea vine silage containing DDT residues. Journal of Economic Entomology 42:119-122. , R. W Wells, R. D. Radeleff, C. L. Smith, R E. Hubanks, and H. D. Mann. 1949b. The chlorinated hydi-ocarbon content of milk from cattle sprayed for control of horn flies. Joui'nal of Economic Entomology 42:116-118. Casida, J. E., K. B. Augustinsson, and G. John- SSON. 1960. Stability, toxicity, and reaction mechanism with esterases of certain carba- mate insecticides. Journal of Economic Ento- mology 53:205-212. R. K. Chapman, and T C. Allen. 1952. Relation of absorption and metabolism of octa- methylpyrophosphoramide by pea plants to available phosphorus. Journal of Economic Entomology 45:568-578. , , M. A. Stahmann, and T. C. Allen. 1954. Metabolism of Schradan by plants and insects to a toxic phosphoramide oxide. Jour- nal of Economic Entomology 47:64-71. , and M. A. Stahmann. 1953. Metabolism and mode of action of Schradan. Journal of Agi-icultural and Food Chemistry 1:883-887. Cassil, C. C. 1941. Denis residue on marketable cabbage. Journal of Economic Entomology 34:72-74. F. M. Wadley, and F P Dean. 1943. Sampling studies on orchard .spray residues in the Pacific Northwest. Journal of Economic Entomology 36:227-231. Chisholm, R. D., and L. Koblitsky. 1943. Sorp- tion of methyl bromide by soil in a fumigation chamber. Journal of Economic Entomology 36:549-551. , , J. E. Fahey, and W E. Westlake. 1950. DDT residues in soil. Journal of Economic Entomology 43:941-942. Claborn, H. V., H. F. Beckman, and R. W. Wells. 1950a. Contamination of milk from DDT sprays applied to dairy barns. Journal of Economic Entomology 43:723-724. , , and . 1950b. Excretion of DDT and TDE in milk from cows treated with these insecticides. Journal of Economic Ento- mology 43:850-852. Coleman, G. A. 1911. Methods of determining the toxicity of hydrocyanic acid gas. Journal of Economic Entomology 4:528-531. CuLLiNAN, F P. 1949. Some new insecticides - their effect on plants and soils. Journal of Economic Entomology 42:387-391. Cutkomp, L. K. 1947. Thermal decomposition of DDT dispersed in water. Journal of Economic Entomology 40:444-445. Dahm, p. a. 1952. Effects of weathering and com- mercial dehydration upon residues of aldrin, chlordane, and toxaphene applied to alfalfa. Jouimal of Economic Entomology 45:763-766. , and C. W Kearns. 1941. The toxicity of alkyl secondary amines to the housefly. Jour- nal of Economic Entomology 34:462-466. Darsie, R. F, and F E. Corriden. 1959. The tox- icity of malathion to killfish. (Cyprinodon- tidae) in Delaware. Journal of Economic Ento- mology 52:696-700. Dawsey, L. H., and L. N. Markwood. 1940. Persis- tence of nicotine on rose bushes sprayed with nicotine sulfate combinations. Journal of Economic Entomology 33:722-723. DeBach, p. 1947. Predators, DDT, and citrus red mite populations. Journal ofEconomic Ento- mology 40:598-599. Decker, G. C. 1946. Argicultural applications of DDT, with special reference to the importance of residues. Journal of Economic Entomology 39:557-562. , C. J. Weinman, and J. M. Bann. 1950. A preliminary report on the rate of insecticide residue loss from treated plants. Journal of Economic Entomology 43:919-927. DeOng, E. R. 1923. The relation between the volatility and toxicity of nicotine in sprays and dusts. Journal of Economic Entomology 16:486-493. 1937. Tblei-ance of animals to pyrethrum extracts. Journal of Economic Entomology 30:921-927. DoANE, R. W. 1923. Bees vs. spraying. Journal of Economic Entomology 16:527-531. DoBSON, R. C, G. O. Throneberry, and T. E. Bell- ing. 1960. Residues in established alfalfa treated with gi'anulated phorate (Thimet) and their effect on cattle fed the hay. Journal of Economic Entomology 53:306-310. 212 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 DoROUGH, H. W., U. E. Brady, Jr., J. A. Timmer- MAN, Jr., and B. W. Arthur. 1961a. Residues in tissues and eggs of poultry dusted with Co- Ral (Bayer 21/199). Journal of Economic Ento- mology 54:25-30. , , , and . 1961b. Resi- dues in tissues and eggs of poultry receiving Co-Ral (Bayer 21/199) in the feed. Journal of Economic Entomology 54:97-100. Ebeling, W. 1940. A method for determination of oil-spray residue on citrus foliage. Journal of Economic Entomology 33:900-904. Eddy, C. 0. 1926. Studies of some adsorbed in- secticides. Journal of Economic Entomology 19:77-86. Eden, W. B. 1951. Tbxicity of dieldrin to chickens. Journal of Economic Entomology 44:1013. Eden, W. G., and F. S. Arant 1948. DDT residues on alfalfa. Journal of Economic Entomology 41:383-387. Edwards, C. A., S. D. Beck, and E. R Lichten- stein. 1957. Bioassay of aldrin and lindane in soil. Journal of Economic Entomololgy 50:622-626. Eheart, J. P., E. C. Turner, and J. Dickinson. 1962. Residues of Sevin in whole milk from sprayed and dusted cows. Journal of Economic Entomology 55:504-505. EiDE, P. M., C. C. Deonier, and R. W. Burrell. 1945. The toxicity of DDT to certain forms of aquatic life. Journal of Economic Entomology 38:492-493. Ellisor, L. D. 1936. A toxicological investigation of nicotine on the goldfish and the cockroach. Iowa State College Journal of Science 11:51-53. Ely, R. E., L. A. Moore, R. H. Carter, and B. A. Apr 1957. Excretion of endrin in the milk of cows fed endrin-sprayed alfalfa and technical endrin. Journal of Economic Entomology 50:348-349. Entomology Research Division. 1959. Residues in fatty tissues, brain, and milk of cattle from insecticides applied for gi-asshopper control on rangeland. Journal of Economic Entomology 52:1206-1210. Pahey, J. E., T. A. Brindley, and M. L. Spear. 1955. DDT residues in fat from steers pas- tured on corn stover in DDT-treated fields. Journal of Economic Entomology 48:606-607. , D. W. Hamilton, and R. W. Rings. 1952. Longevity of parathion and related insecti- cides in spray residues. Journal of Economic Entomology 45:700-703, _, and H. W. Rusk. 1940. Effect of fruit growth and weather on deposits of insecticides on apples in southern Indiana. Joui-nal of Eco- nomic Entomology 33:505-511. Ferguson, W. C, and C. W. Kearns. 1949. The metabolism of DDT in the large milkweed bug. Journal of Economic Entomologv' 42:810-817. Pernald, H. T 1908. The fundamental principles of spraying. Journal of Economic Entomology 1:265-267. Fink, D. E. 1927. Is glutathione the arsenic recep- tor in insects? Journal of Economic Ento- mology 20:794-801. Fisher, D. P. 1931. Recent developments in spray residue removal. Journal of Economic Ento- mology 24:526-531. Fleck, E. E. 1944. Rate of evaporation of DDT. Journal of Economic Entomology 37:853. Fleming, W. E., F. E. Baker, and L. Koblitsky. 1943. Effect of lead arsenate in soil on vege- tables. Journal of Economic Entomology 36:231-233. , and W. W. Maines. 1953. Persistence of DDT in soils of the area infested by the Japa- nese beetle. Journal of Economic Entomology 46:445-449. , and . 1954. Persistence of chlor- dane in soils of the area infested by the Japanese beetle. Journal of Economic Ento- mology 47:165-169. Prear, D. E. H., W. N. Bruce, and A. C. Ragsd.^le. 1950. DDT in milk following barn spraying. Joui-nal of Economic Entomology 43:656-657. Freeborn, S. B.. W. M. Regan, and L. J. Berry. 1934. The effect of petroleum oil fly sprays on dairy cattle. Jom-nal of Economic Entomology 27:382-388. PuKUTO, T. R., R. L. Metcalf, and M. Y. Winton. 1961. The insecticidal properties of esters of phosphonic and phosphinic acids. Journal of Economic Entomology 54:955-962. J, R Wolf, III, R." L. Metcalf, and R. B. March. 1956. Identification of the sulfoxide and sulfone plant metabolites of the thiol isomer of systox. Journal of Economic Ento- mology 49:"l47-151. Pulton, R. A., F. F. Smith, and R. P. Gelardo. 1955. Respiratory protective devices for agri- cultural use. Journal of Economic Entomology 48:457-458. Purman, D. R, and C. J. Weinmann. 1956. Tox- icity of malathion to poultry and their ecto- parasites. Journal of Economic Entomolog>' 49:447-450. Gannon, N., and J. H. Bigger. 1958. The conver- sion of aldrin and heptachlor to their epox- ides in soil. Journal of Economic Entomologj' 51:1-7. , and G. C. Decker. 1958. The conversion of aldrin to dieldrin on plants. Journal of Eco- nomic Entoniolog>' 51:8-11. , and . 1960. The excretion of diel- drin, DDT, and heptachlor epoxide in milk of dairy cows fed on pasture treated with diel- drin, DDT, and heptachlor Journal of Eco- nomic Entomolog\' 53:411-415. George, J. L.. and R. T. Mitchell. 1947. The ef- fects of feeding DDT-treated insects to nest- ling birds. Journal of Economic Entomologj' 40:782-789. Georghiou, G. P., and C. E. Taylor. 1976. Pesti- cide resistance as an evolutionary phenome- non. Pages 759-785 in XVth International Congress of EntomologN' Pi-oceedings. Getzin. L W.. and R. K. Chapman. 1959. Effect of soils upon the uptake of systemic insecti- cides by plants. Journal of Economic Ento- mology" 52:1160-1165. September 1985 125 Years of Biological Research 213 , and 1960. The fate of phorate in soils. Journal of Economic Entomology 53:47-51. GlANG, P. A., and S. A. Hall. 1951. Enzymatic determination of organic phosphorus insecti- cides. Analytical Chemistry 23:1830-1839. GiNSBURG, J. M. 1928. A comparison between complete and incomplete digestion of sprayed apple foliage in determining arsenic by the Gutzeit method. Journal of Economic Ento- mology 21:588-598. 1945. Tbxicity of DDT to fish. Journal of Economic Entomology 38:274-275. 1946. Chemical methods for analysis of dichlorodiphenyltrichloroethane (DDT). Jour- nal of Economic Entomology 39:174-177. . 1947. Results from feeding mosquito lar- vae, killed by DDT, to goldfish. Journal of Economic Entomology 40:275-276. , R. S. FiLMER, and J. R Reed. 1950. Longevity of parathion, DDT and dichlorodi- phenyl dichlorethane residues on field and vegetable crops. Journal of Economic Ento- mology 43:90-94. , , , and A. R. Paterson. 1949. Recovery of parathion, DDT and certain ana- logs of dichlorodiphenyl-dichlorethane from treated crops. Journal of Economic Ento- mology 42:602-611. and J. P Reed. 1954. A survey on DDT- accumulation in soils in relation to different crops. Journal of Economic Entomology 47:467-474. Gould, E., and E. O. Hamstead. 1951. The tox- icity of cumulative spray residues in soil. Journal of Economic Entomology 44:713-717. Goulding, R. L., and L. C. Terriere. 1959. Mala- thion residues in the milk of dairy cows treated for horn fiy control. Journal of Eco- nomic Entomology 52:341. Griffiths, J. T. 1951. Possibilities for better citrus insect control through the study of the ecolog- ical effects of spray programs. Journal of Eco- nomic Entomology 44:464-468. , C. R. Stearns, Jr., and W. L. Thompson. 1951. Parathion hazards encountered spray- ing citrus in Florida. Journal of Economic Entomology 44:160-163. GuNTHER, F. A. 1942. Analytical isolation of rotenone from a spray solution. Journal of Economic Entomology 35:458. 1943. Effects of oxygen and sunlight on decomposition of rotenone in spray mixtures. Journal of Economic Entomology 36:273-281. 1947. Thermal decomposition ofDDT and benzene hexachloride mixtures. Journal of Economic Entomology 40:874-877. , ED. 1980. Research conference and workshop on minimizing occupational expo- sure to pesticides. Residue Reviews 75:1-189. , J. H. Barkley, and W. H. Ewart. 1954. Harvest residues of apparent dieldrin in peel and juice of navel oranges. Journal of Eco- nomic Entomology 47:1033-1035. , and R. C. Blinn. 1955. Interpretation of data. Pages 131-148 in Analysis of insecti- cides and acaricides. Interscience Publishers, New York. , G. E. Carman, and M. I. Elliot. 1948. In- hibition of the field decomposition of DDT spray deposits on citrus; preliminary results. Journal of Economic Entomology 41:895-900. , and L. R. Jeppson. 1954. Residues of p-chlorophenyl-p-chlorobenzenesulfonate (com- pound K-6451) on and in lemons and oranges. Journal of Economic Entomology 47:1027-1032. , D. L. LiNDGREN, M. I. Elliot, and J. P. LaDue. 1946. Persistence of certain DDT deposits under field conditions. Journal of Economic Entomology 39:624-627. Haller, H. L., F. B. LaForge, and W. N. Sullivan. 1942. Effect of sesamin and related com- pounds on the insecticidal action of pyreth- rum on houseflies. Journal of Economic Entomology 35:247-248. Hamilton, C. C. 1929. The growth of the foliage and fruit of the apple in relation to the main- tenance of a spray coating. Journal of Eco- nomic Entomology 22:387-396. , and C. M. Smith. 1925. A colorimetric method for showing the distribution and quantity of lead arsenate upon sprayed and dusted surfaces. Journal of Economic Ento- mology 18:502-509. Haque, R., and V. H. Freed. 1974. Behavior of pesticides in the environment: "Environ- mental chemodynamics." Residue Reviews 52:89-116. Harris, C. R., and E. P. Lichtenstein. 1961. Fac- tors affecting the volatilization of insecticidal residues from soils. Journal of Economic Ento- mology 54:1038-1045. Hartenstein, R. C. 1960. The effects of DDT and malathion upon forest soil microarthropods. Joui'nal of Economic Entomology 53:357-362. Hartzell, a. 1952. Bioassay of organic insecti- cides in processed foods. Journal of Economic Entomology 45:1102. , and F. Wilcoxon. 1927. The arsenic con- tent of sprayed apples. Journal of Economic Entomology 20:204-213. , and 1928. Analyses of sprayed ap- ples for lead and arsenic. Journal of Economic Entomology 21:125-130. Hartzell, F. Z. 1924. The estimation of dosage for volatile dusts: an illustration of the value of correlation mathematics to entomology. Journal of Economic Entomology 17:279-285. Haug, G. W., and A. Peterson. 1938. The effect of insecticides on a beneficial coccinellid, Hip- podamia convergent Guer. Journal of Eco- nomic Entomology 31:87-92. Hayne.s, H. L., H. H. Moorefield, A. J. Borash, and J. W. Keays. 1958. The toxicity of Sevin to goldfish. Journal of Economic Entomology 51:540. Headden, W. p. 1910. Arsenical poisoning of fruit trees. Journal of Economic Entomology 3:32-36. 214 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Headlee, T. J., and W. Rudolfs. 1922. Some fur- thei- experience with contact dusts. Journal of Economic Entomology 15:75-82. Hermenze, F., and W. J. Goodwin. 1959. Normal cholinesterase activity in the erythrocytes of cattle. Journal of Economic Entomology 52:66-68. Hitchcock, S. W. 1960. Effects of an aerial DDT spray on aquatic insects in Connecticut. Jour- nal of Economic Entomology 53:608-611. Hopkins, L., G. G. Gyrisco, and L. B. Norton. 1952a. Effects of sun, wind and rain on DDT dust residue on forage crops. Journal of Eco- nomic Entomology 45:629-633. , L. B. Norton, and G. G. Gyrisco. 1952b. Persistence of insecticide residues on forage crops. Journal of Economic Entomology 45:213-218. Hoskins, W. M. 1949. Deposit and residue of re- cent insecticides resulting from various con- trol practices in California. Journal of Economic Entomology 42:966-973. HucKETT, H. C. 1934. The situation in the vege- table industry with respect to the use of ai-sen- icals and ai'senical substitutes. Joui'nal of Eco- nomic Entomology 27:156-167. HuDDLESTON, E. W., G. G. Gyrisco, and D. J. Lisk. 1960a. DDT residues on New York dairy farms following the gypsy moth eradication program. Journal of Economic Entomology 53:1019-1021. , K. H. Thompson, G. G. Gyrisco, D. J. LiSK, T. W. Kerr, Jr., and C. E. Olney. 1960b Sampling experimental plots of alfalfa for in- secticide residues. Journal of Economic Ento- mology 53:1078-1080. Hyslop, J. A. 1914. Soil fumigation. Journal of Economic Entomology 7:305-312. Ingle, L. 1947. Tbxicity of chlordane to white rats. Journal of Economic Entomology 40:264-268. IvEY, M. C, R. H. Roberts, H. D. Mann, and H. V. Claborn. 1961. Lindane residues in chick- ens and eggs following poultry house sprays. Joui'nal of Economic Entomology 54:487-488. Johnson, G. A., J. H. Fletcher, K. G. Nolan, and J. T Cassaday. 1952. Decreased toxicity and cholinesterase inhibition in a new series of dithiophosphates. Journal of Economic Ento- mology 45:279-283. Jones, H. A., W. A. Gersdorff, E. L. Gooden, F. L. Campbell, and W. N. Sullivan. 1933. Loss in toxicity of deposits of rotenone and related materials exposed to light. Journal of Economic Entomology 26:451-470. Kaplanis, J. N., W. E. RoBBTNS, D. I. Darow, D. E. Hopkins, R. E. Moore, and G. Treiber. 1959. The metabolism of dimethoate in cat- tle. Journal of Economic Entomologv 52:1190-1194. Kearns, C. W., W. R Flint. 1937. Contact insecti- cidal properties of various derivatives of cyclohexylamine. Journal of Economic Ento- mology 30:158-166. Kerr, R. W. 1951. Adjuvants for pyrethrins in fly sprays. Commonwealth Scientific and Indus- trial Research Organization Bulletin 261. Melbourne, Australia. KnoEMAGi, U., H. E. Morrison, J. E. Roberts, and W. B. BoLLEN. 1958. Biological and chemical studies on the decline of soil insecticides. Jour- nal of Economic Entomology 51:198-204. King, H. L., and D. E. H. Frear'. 1943. Relation of chemical constitution of some N-hetero- cyclic compounds to toxicity to Tetranychus telerius (L.I. Journal of Economic Entomology 36:263-265. Krueger, H. R., and J. E. Casida. 1957. Toxicity of fifteen organophosphorus insecticides to several insect species and to rats. Journal of Economic Entomology 50:356-358. , and R. D. O'Brien. 1959. Relationship between metabolism and differential toxicity of malathion in insects and mice. Journal of Economic Entomology 52:1063-1067. , , and W. C. Dauterman. 1960. Rela- tionship between metabolism and differential toxicity in insects and mice of diazinon, di- methoate, parathion, and acethion. Journal of Economic Entomology 53:25-31. Lanchester, H. P 1951. A modified probit method for comparison of insecticides for wire- worm control. Journal of Economic Ento- mology 44:779-784. Leonard, C. S. 1942. The effects of pyrethrins on certain mammals. Journal of Economic Ento- mology 35:261-264. Lichtenstein. E. p. 1957. DDT accumulation in mid-western orchard and crop soils treated since 1945. Journal of Economic Entomology 50:545-547. . 1958. Movement of insecticides in soils under leaching and nonleaching conditions. Journal of Economic Entomologj' 52:380-383. _, L. J. DePaw, E. L. Eshb.augh. and J. P. Sleesman. 1960. Persistence of DDT, aldrin, and lindane in some midwestern soils. Jour- nal of Economic Entomolog>' 53:136-142. , C. H. Mueller, G. R. Myrdal, and K. R. ScHULZ. 1962. Vertical distribution and pei-sis- tence of insecticidal residues in soils as influ- enced by mode of application and cover crop. Journal of Economic Entomology' 55:215-219. , and J. B. Polivka. 1959. Persistence of some chlorinated hydrocarbon insecticides in turf soils. Journal of Economic Entomology 52:293. , and K. R. Schulz. 1959a. Breakdown of lindane and aldrin in soils. Journal of Eco- nomic Entomolog>' 52:118-123. , and . 1959b. Persistence of some chlorinated hydi'ocarbon insecticides as influ- enced by soil types, rate of application and temperature. Jom'nal of Economic Entomolog>" 52:124-131. and . 1960. Epoxidation of aldrin and heptachlor in soils as infiuenced by auto- claving, moisture, and soil types. Journal of Economic Entomology' 53:192-197. and . 1961. EfTect of soil cultiva- tion, soil surface and water on the pei-sistence of insecticidal residues in soils. Joui'nal of Eco- nomic Entomology ,54:517-522. Lilly. J. H. 1940. The effect of arsenical gi'ass- hopper poisons upon pheasants. Journal of Economic Entomology 33:501-505. September 1985 125 Years of Biological Research 215 LiNDQUiST, A. W., A. R. Roth, R. A. Hoffman, and J. S. Butts. 1951a. The distribution of radio- active DDT in house flies. Journal of Eco- nomic Entomology 44:931-934. , , W. W. Yates, and R. A. Hoffman. 1951b. Use of radioactive tracers in studies of penetration and metabolism of DDT in house flies. Journal of Economic Entomology 44:167-172. LiNDQUiST, D. A., E. C. Burns, C. P. Pant, and P. A. Dahm. 1958. Fate of P32-labeled Bayer 21/199 in the white rat. Journal of Economic Entomology 51:204-206. LiNSLEY, E. G., and J. W. MacSwain. 1947. The effects of DDT and certain other insecticides on alfalfa pollinators. Journal of Economic Entomology 40:358-363. LuDviK, G. R, and G. C. Decker. 1947. Toxicity of certain esters of phosphorus acids to aphids. Journal of Economic Entomology 40:97-100. , and . 1951. The insecticidal proper- ties of some esters of phosphorus acids. Jour- nal of Economic Entomology 44:405-418. MacLeod, G. F., D. E. Haley, and R. H. Sudds. 1927. A study of arsenical residues on apples in Pennsylvania with respect to efficient spraying practices. Journal of Economic Ento- mology 20:606-614. , and H. F. Sherwood. 1937. Grenz radio- graphs of sulfur dispersion on foliage. Jour- nal of Economic Entomology 30:395-398. March, R. B., T. R. Fukuto, R. L. Metcalf, and M. G. Maxon. 1956a. Fate of P32-labeled malathion in the laying hen, white mouse, and American cockroach. Journal of Eco- nomic Entomology 49:185-195. , and R. L. Metcalf. 1949a. Laboratory and field studies on DDT-resistant houseflies in southern California. California State De- partment of Agi-iculture Bulletin 38:93-101. _, and . 1949b. Development of resis- tance to organic insecticides other than DDT by houseflies. Journal of Economic Ento- mology 42:990. , T R. Fukuto, and F. A. Gunther. 1956b. Fate of P32-labeled malathion sprayed on Jersey heifer calves. Journal of Economic Entomology 49:679-682. _, and M. G. Maxon. 1955. McCay, C. F, and B. W. Arthur. 1962. Sevin resi- dues in poultry products. Journal of Economic Entomology 55:936-939. McLean, H. C, and A. Weber. 1928. Influence of spray practices on arsenical residues. Jour- nal of Economic Entomology 21:921-928. , A. L. Weber, and J. S. Joffe. 1944. Arsenic content of vegetables grown in soils treated with lead arsenate. Journal of Eco- nomic Entomology 37:315-316. Melander, a. L. 1914. Can insects become resis- tant to sprays? Journal of Economic Ento- mology 7:167-174. Melvin, R. 1932. Physiological studies on the ef- fect of flies and fly sprays on cattle. Journal of Economic Entomology 25:1150-1164. Menn, J. J., G. G. Patchett, and G. H. Batch- elder. 1960. The persistence of trithion, an organophosphorus insecticide, in soil. Journal of Economic Entomology 53:1080-1082. Metcalf, R. L. 1948. Some insecticidal proper- ties of fluorine analogues of DDT. Journal of Economic Entomology 41:416-421. . 1951. The colorimetric microestimation of human blood cholinesterases and its appli- cation to poisoning by organic phosphate in- secticides. Journal of Economic Entomology 44:883-890. _, T. R. Fukuto, and R. B. March. 1957a. Metabolism of Systox in the white mouse and American cockroach. Journal of Economic Entomology 48:355-363. Marcovitch, S. 1928. The relative toxicities of arsenicals and fluorine compounds to various organisms. Journal of Economic Entomology 21:108-114. and W. W. Stanley. 1938. New facts con- cerning cryolite spray residues. Journal of Economic Entomology 31:480-482. Matsumura, F 1960. Malathion residues on and in the leaves of Phaneolus vulgaris. Journal of Economic Entomology 53:452-454. McCall, G. L., and J. F Kagy. 1940. A colori- metric method for the determination of oil deposit on citrus leaves. Journal of Economic Entomology 33:905-908. Plant metabolism of dithio-Systox and Thimet. Journal of Economic Entomology 50:338-345. , , , and E. M. Stafford. 1956. The systemic behavior of Systox thiol isomer sulfoxide and methosulfate in plants. Journal of Economic Entomology 49:738-741. „, and M. Y. Winton. 1960. Alkox- yphenyl N-methylcarbamates as insecticides. Journal of Economic Entomology 53:828-832. , , and . 1962. Insecticidal carbamates: position isomerism in relation to activity of substituted-phenyl N-methylcarba- mates. Journal of Economic Entomology 55:889-893. _, and C. W. Kearns. 1941. The toxicity and repellent action of some derivatives of pic- ramic acid and of toluenesulfonyl chloride to the greenhouse leaf tier. Journal of Economic Entomology 34:306-309. and R. B. March. 1949. Studies of the mode of action of pai-athion and its derivatives and their toxicity to insects. Journal of Eco- nomic Entomology 42:721-728. , and 1950. Properties of acetyl- choline esterases from the bee, the fly and the mouse and their relation to insecticide action. Journal of Economic Entomology 43:670-677. , and . 1952. Behavior of octamethyl pyrophosphoramide in citrus plants. Journal of Economic Entomology 45:988-997. , , T R. Fukuto, and M. Maxon. 1954. The behavior of systox-isomers in bean and citrus plants. Journal of Economic Ento- mology 47:1045-1055. 216 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 , , and M. G. Maxon. 1955. The nature and significance of Systox residues on plant materials. Journal of Economic Ento- mology 48:364-369. , E. M. Stafford, T. R. Fukuto, and R. B. March. 1957b. The systemic behavior of 0,0- diethyl S-2(diethylamino)ethyl phosphorothio- late and its salts. Journal of Economic Ento- mology 50:205-210. MisKus, R., and S. Hassan. 1959. Preassay activa- tion of Thimet. Journal of Economic Ento- mology 52:353-354. , M. E. T^anakakis, and S. M. Smth. 1959. Determination of Bayer 19639 residues in agricultural crops by cholinesterase inhibi- tion. Journal of Economic Entomology 52:76-78. Moore, S., III. 1952. The effect of lindane vapor on canaries and pigeons. Journal of Economic Entomology 45:1092-1093. Moore, W. 1923. The need of chemistry for the student of entomology. Journal of Economic Entomology 16:172-176. Morrison, H. E., H. H. Cropwell, S. E. Crumb, Jr., and R. W. Lauderdale. 1948. The effects of certain new soil insecticides on plants. Jour- nal of Economic Entomology 41:374-378. Mulla, M. S, and L. W. Isaak. 1961. Field studies on the toxicity of insecticides to the mosquito fish, Gambusia afpnis. Journal of Economic Entomology 54:1237-1242. Newson, L. D., and C. E. Smith. 1949. Destruc- tion of certain insect predators by applications of insecticides to control cotton pests. Journal of Economic Entomology 42:904-908. Norton, L. B., and O. B. Billings. 1941. Charac- teristics of different types of nicotine sprays. I. Nicotine residues. Journal of Economic En- tomology 34:630-633. , and R. Hansberry. 1941. Radioactive tracer methods for determination of the dis- position of arsenic in the silkworm. Journal of Economic Entomology 34:431-437. O'Brien, R. D. 1956. The inhibition of chloin- esterase and succinoxidase by malathion and its isomer. Journal of Economic Entomology 49:484-490. 1957a. The effect of malathion and its isomer on carbohydrate metabolism of the mouse, cockroach and house fly. Journal of Economic Entomology 50:70-83. 1957b. Properties and metabolism in the cockroach and mouse of malathion and mala- oxon. Journal of Economic Entomology 50:159-164. , G. D. Thorn, and R. W. Fisher. 1958. New organophosphate insecticides developed on ra- tional principles. Journal of Economic Ento- mology 51:714-718. O'Donnell, a. E., H. W. Johnson, Jr., and F. T Weiss. 1955. Chemical determination of diel- drin in crop materials. Journal of Agi'icultural and Food ChemLstrv 3:759-762. M. M. Neal, R T Weiss, J. M. Bonn, J. D. DeCino, and S. C. Lau. 1954. Chemical determination of aldrin in crop materials. Journal of Agricultural and Food Chemistry 2:573-580. O'Kane, W. C. 1913. Arsenical residues on fruit. Journal of Economic Entomology 6:159-160. 1916. Arsenic on fruit and forage follow- ing spraying. Journal of Economic Entomology 9:90-91. Orr, L. W, and L. O. Mott. 1945. The effects of DDT administered orally to cows, horses, and sheep. Journal of Economic Entomology 38:428-432. Pankaskie, J. E., F. C. Fountaine, and P. A. Dahm. 1952. The degi-adation and detoxication of parathion in dairy cows. Journal of Economic Entomology 45:51-60. Perry, A. S., andW M. Hoskins. 1951. Detoxifica- tion of DDT as a factor in the resistance of house flies. Journal of Economic Entomology 44:850-857. PoLEN, P. B., and P. Silverman. 1952. Detection and determination of 1,4,5,6,7,8,8,-heptachlor- 3a,4,7,7a-tetrahydi-o4,7-methanoindene Ana- lytical Chemistry 24:733-735. QuAYLE, H. J. 1943. The increase in resistance in insects to insecticides. Journal of Economic Entomology 36:493-500. QuESTEL, D. D, and R. V. Connin. 1947. A chemi- cal treatment of soil which produces plant tis- sue lethal to European corn borer Journal of Economic Entomology 40:914-915. Quinby, G. E., K. C. Walker, and W. F. Durham. 1958. Public health hazards involved in the use of organic phosphorus insecticides in cot- ton culture in the delta area of Mississippi. Journal of Economic Entomology 51:831-838. Radeleff, R. D. 1950. Acute toxicity of chlor- in- ated insecticides applied to livestock. Journal of Economic Entomologii- 43:358-364. , and G. T Woodard. 1956. Toxicological problems in the use of systemic insecticides for livestock. Journal of Economic Ento- mology 49:89-91. and . 1957. Toxicological studies of Dow ET-57 in cattle and sheep. Journal of Economic Entomolog>' 50:249-251. Reynolds, H. T, and R. L. Metcalf 1962. Effect of water solubility and soil moisture upon plant uptake of granulated systemic insecti- cides. Journal of Economic Entomology 55:2-5. Ripper, W. E., R. M. Greenslade. and G. S. Hartley. 1951. Selective insecticides and bio- logical control. Journal of Economic Ento- mology 44:448-459. Roan. C. C., H. E. Fernando, and C. W. Kearns. 1950. A radiobiological study of four organic phosphates. Journal of Economic Entomology 43:319-325. Roark. R. C. and D. A. Nelson. 1929. Maximum weights of various fumigants which can exist in vapor form in a 1.000 cubic foot fumigating chamber Journal of Economic Entomology- 22:381-387. September 1985 125 Years of Biological Research 217 RoBBiNS, W. E., T. L. Hopkins, and D. I. Darrow. 1959. Synergistic action of piperonyl butoxide with Bayer 21/199 and its corresponding phos- phate in mice. Journal of Economic Ento- mology 52:660-662. and G. W. Eddv. 1956. The metab- olism of P32-labeled Bayer L 13/59 in a cow. JoLU-nal of Economic Entomology 49:801-806. _, and A. R. Roth. 1958. Applica- tion of the colorimetric whole-blood method to the measurement of bovine red-blood-cell cholinesterase activity. Journal of Economic Entomology 51:326-329. Robinson, R. H. 1929. Facilitating the removal of the spray residue. Journal of Economic En- tomology 22:693-698. Safro, V. I. 1917. How to test for the presence of nicotine on sprayed plants. Journal of Eco- nomic Entomology 10:459-461. ScHECHTER, M. S., and H. L. Haller. 1945. Colori- metric determination of DDT. Industrial Engi- neering Chemistry 17:704-709. , and I. HORNSTEIN. 1952. Colorimetric determination of benzene hexachloride. Ana- lytical Chemistry 24:544. ScHESSER, J. H., W. E. Priddle, and E. P. Farrell. 1958. Insecticidal residues in milling fractions from wheat treated with methoxytchlor, mala- thion, and lindane. Jom'nal of Economic Ento- mology 52:516-518. Schulze, F., H. L. Sweetman, and P. J. Spear. 1952. Tbxicity of lindane vapor to tropical fish. Joui-nal of Economic Entomology 45:1094-1097. Scott, D. B., Jr., and E. H. Karr. 1942. The influ- ence of insecticides added to soils on gi-owth and yield of certain plants. Journal of Eco- nomic Entomology 35:702-708. Shafer, G. D. 1911a. The effect of certain gases and insecticides upon the activity and respira- tion of insects. Journal of Economic Ento- mology 4:47-50. 1911b. Some properties that make lime- sulphui' wash effective in killing scale-insects. Journal of Economic Entomology 4:50-53. Shaw, F. R. 1941. Bee poisoning: a review of the more important literature. Journal of Eco- nomic Entomology 34:16-21. Shepard, H. H., and C. H. Richardson. 1931. A method of determining the relative toxicity of contact insecticides with especial reference to the action of nicotine against Aphis rumicis. Jownal of Economic Entomology 24:905-914. Sherman, M., and M. M. Rosenberg. 1953. Acute toxicity of four chlorinated dimethanonapth- thalene insecticides to chicks. Journal of Eco- nomic Entomology 46:1067-1070. , and . 1954. Subchronic toxicity of four chlorinated dimethanonaphthalene in- secticides to chicks. Journal of Economic En- tomology 47:1082-1083. SlAKOTOs, A. N. 1956a. Contamination of food and air by lindane vapor. Journal of Economic En- tomology 49:212-214. 1956b. Package exposure to continuous- ly vaporized lindane. Journal of Economic Entomology 49:481-484. SiMKOVER, H. G., and R. D. Shenefelt. 1951. Ef- fect of benzene hexachloride and chlordane on certain soil organisms. Journal of Economic Entomology 44:426-427. Sinclair, W. B., and P. R. Crandall. 1952. Methods for determining ethylene chlorobro- mide and ethylene dibromide. Journal of Eco- nomic Entomology 45:882-887. Sloan, M. J., W. A. Rawlins, and L. B. Norton. 1951a. Residue studies on DDT and parathion applied to lettuce for control of the six-spotted leafhopper. Journal of Economic Entomology 44:691-701. , , and . 1951b. Factors affect- ing the loss of DDT and parathion residues on lettuce. Journal of Economic Entomology 44:701-709. Smith, F. F, and P. A. Clifford. 1950. Transloca- tion of parathion from foliage applications Journal of Economic Entomology 43:708-712 , F. I. Edwards, P. Giang, and R. Fulton 1952. Residues of organic phosphorus com pounds and DDT on gi'eenhouse vegetables Jom-nal of Economic Entomology 45:703-707 P Giang, and E. A. Taylor. 1955. Reduc tion of malathion residues on vegetables by washing. Journal of Economic Entomology 48:209-210. Smith, R. F., O. H. Fullmer, and P. S. Messenger. 1948a. DDT residues on alfalfa hay and seed chaff Journal of Economic Entomology 41:755-758. , W. M. HosKiNS, and O. H. Fullmer. 1948b. Secretion of DDT in milk of dairy cows fed low-residue alfalfa hay. Journal of Eco- nomic Entomology 41:759-763. Stansbury, R. E., and R A. Dahm. 1951. The ef- fect of alfalfa dehydration upon residues of aldrin, chlordane, parathion, and toxaphene. Journal of Economic Entomology 44:45-51. Sternburg, J., and C. W. Kearns. 1956. Penta- chlorocyclohexene, an intermediate in the metabolism of lindane by house flies. Journal of Economic Entomology 49:548-552. , , and W. N. Bruce. 1950. Absorp- tion and metabolism of DDT by resistant and susceptible house flies. Journal of Economic Entomology 43:214-219. E. B. Vinson, and C. W. Kearns. 1953. En- zymatic dehydrochlorination of DDT by resis- tant flies. Journal of Economic Entomology 46:513-515. Stitt, L. L., and J. Evanson. 1949. Phytotoxici- ty and off-quality vegetables gi-own in soil treated with insecticides. Journal of Economic Entomology 42:614-617. Sun, Y. -P. 1948. Synergistic action of chlordan in dusts containing nicotine or rotenone. Jour- nal of Economic Entomology 41:89-91. and H. H. Shepard. 1947. Methods of calculating and correcting the mortality of insects. Journal of Economic Entomology 40:710-715. 218 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 SwEETMAN, H. L., and A. I. Bourne. 1944. Tox- icity of natural and synthetic cryolites to rats. Journal of Economic Entomology 37:299-300. Telford, H. S., and J. E. Guthrie. 1945. Trans- mission of the toxicity of DDT through the milk of white rats and goats. Science 102:667. , and 1946. Effects of oral dosages ofDDT on certain vertebrates. Journal of Eco- nomic Entomology 39:413-414. Terriere, L. C, and D. W. Ingalsbe. 1953. Trans- location and residual action of soil insecti- cides. Journal of Economic Entomology 46:751-753. Tiller, R. E., and E. N. Cory. 1947. Effects of DDT on some tidewater aquatic animals. Journal of Economic Entomology 40:431-433. TiMMERMAN, J. A., Jr., H. W. Dorough, J. R. BuT- TRAM, and B. W. Arthur. 1961. In vitro stabili- ty and recovery of insecticides from milk. Journal of Economic Entomology 54:441-444. Tischler, N. 1935. Studies on how derris kills in- sects. Journal of Economic Entomology 24:905-914. Tracy, R. L., J. G. Woodcock, and S. Chodroff. 1960. Tbxicological aspects of 2,2-dichlorovinyl dimethyl phosphate (DDVP) in cows, horses, and white rats. Journal of Economic Ento- mology 53:593-601. Trevan, J. W. 1927. The error of determination of toxicity. Royal Society (London) Pi'oceedings (B) 101:483-514. Turner, H. R, and W. G. Eden. 1952. Tbxicity of chlordane to chickens. Journal of Economic Entomology 45:130. Van Middelem, C. H., J. W. Wilson, and W. D. Hanson. 1956. Sampling studies related to in- secticide residues on cabbage and celery. Joui'- nal of Economic Entomology 49:612-615. Voegtlin, C, H. a. Dyer, and C. S. Leonard. 1925. On the specificity of the so-called arsen- ic receptor in the higher animals. Journal of Pharmacology and Experimental Therapeu- tics 25:297. Wadley, F. M., and W. N. Sullivan. 1943. A study of the dosage-mortality curve. Journal of Economic Entomology 36:367-372. Waites, R. E., and C. H. Van Middelem. 1958. Residue studies of DDT and malathion on tui-- nip tops, collards, snap beans and lettuce. Journal of Economic Entomology 51:306-308 Webb, F. E. 1960. Aerial forest spraying against spruce budworm - a problem of mutual inter- est in Canada and the United States. Journal of Economic Entomology 53:631-633 Weber, A. L., and H. C. McLean. 1933. Removal of lead and aisenic spray residues from apples. Journal of Economic Entomology 26:727-730. Webster, R. L., and A. Crews. 1934. Spray poison in the Yakima Valley. Journal of Economic Entomology 27:614-617. Weidhaas, D. E. 1957. Adsorption of DDT, meth- oxychlor and some related compounds on in- secticide dust diluents and carriers. Journal of Economic Entomology 50:429-435. , M. C. Bowman, and C. H. ScHNaor. 1961. Loss of parathion and DDT to soil from aque- ous dispersions and vermiculite granules. Joui-nal of Economic Entomologj' 54:175-177. Welch, H. 1948. Tests of the toxicity to sheep and cattle of certain of the newer insecticides. Journal of Economic Entomologj' 41:36-39. Wells, R. W, F C. Bishipp, and E.' W. Laake. 1922. Derris as a promising insecticide. Jour- nal of Economic Entomology 15:90-97. Westlake, W E., and L. L Butler. 1953. Resi- dues of malathion on fruits and vegetables. Joui-nal of Economic Entomologj' 46:850-852. White, W. B. 1934. The current season's experi- ence in enforcing spray residue tolerances. Journal of Economic Entomolog>' 27:125-133. WiLFORD, B. H., and L. O. Mott. 1944. Results of tests with domestic animals confined on pas- tures sprayed with natural cryolite. Journal of Economic Entomology 37:291. Wilson, H. F, N. N. Allen, G. Bohstedt, J. Betheil, and H. A. Lardy. 1946a. Feeding ex- periments with DDT-treated pea vine silage with special reference to dairy cows, sheep, and laboratory animals. Journal of Economic Entomology 39:801-806. , and C. E. Holmes. 1936. Effect on chick- ens of arsenic in giasshopper bait. Little danger in eating arsenic-fed chickens. Jour- nal of Economic Entomology 29:1008-1014. , A. S. Srivastava, W B. Hull, J. Betheil, and H. A. Lardy. 1946b. DDT residues on pea vines and canned peas from fields treated with DDT dusts. Journal of Economic Ento- mology 39:806-809. Wilson, J. K., and R. S. Choudhri. 1946. Effects of DDT on certain microbiological processes in the soil. Journal of Economic Entomology 39:537-538. Woodward, G., R. R. Ofner, and C. M. Mont- gomery. 1945. Accumulation of DDT in the body fat and its appearance in the milk of dogs. Science 102:177-178. Yothers, W W. 1910. The effects of fumigation with hydrocyanic gas on the human system. Journal of Economic Entoniolog>' 3:317-319. Young, W. R., and W. A. Rawlins. 1958. The per- sistence of heptachlor in soils. Journal of Ek:o- nomic Entomology 51:11-18. Biotic and Abiotic Stresses as Primary and Predisposing Factors Affecting Illinois Trees D. F. Schoeneweiss, Dan Neely, and E. B. Himelick Illinois, the Prairie State, is in the heart of the nations breadbasket. Deep, flat, prairie soils and equitable climate are ideal for the modern, highly mech- anized production of row crops. With nearly 12 million aci'es of corn and 10 million acres of soybeans annually, Illi- nois is a major agricultural state and is known around the world for high yields. In the face of this intensive agricultui'e, it is small wonder that the forests and woodlots of Illinois have received only limted attention as val- ued natural resources. With less than 11 percent of its total acreage forested, Illinois ranks last among states east of the Mississippi in proportion of wooded area (Stewart 1980). Before 1800, over 40 percent of the state was covered by forest, particular- ly in southern, western, and northern Illinois and along major rivers (Heren- deen & Rolfe 1983). By 1858, the date from which the Illinois Natural His- tory Survey marks its beginning, much of the wooded land had been cleared for farming. As long ago as 1886, members of the State Laboratory of Natural His- tory, a forerunner of the Survey, were voicing concern over the rapid loss of Illinois' trees (Carter 1958). The acreage of forest has continued to de- cline to the present day due to logging, disease, clearing for cropping, and the development of urban and industrial complexes. Because of the relative scar- city of trees, their protection and pres- ervation in Illinois are of vital concern to its citizens. FOREST RESOURCES OF ILLINOIS The forest resource base in Illinois can be divided into two categories, rural and urban (Stewart 1980). The rural base consists of national, state, county, and private forests; tree planta- tions; farm woodlots; and windbreaks. The urban base is made up of trees on parkways and private property and trees in municipal and industrial plant- ings. In rural areas, windbreak trees, which were widely planted after the dust bowl years of the 1930's, have all but disappeared, as have many farm woodlots, due to the increased conver- sion of acreage to row crops. The result- ing loss of fertile topsoil by wind and water erosion on deforested land has become a vital issue facing the state. Of the 3.5 million acres of Illinois forest remaining in 1967, 96 percent was des- ignated as commercial forest, with 93 percent of this commercial portion in private ownership (Stewart 1980). The Shawnee National Forest in southern Illinois totals 277,000 acres, with coun- ty forests making up 126,000 acres and state forests only 11,500 acres. Nearly all of the rural forest is hardwoods, predominantly oak (42 percent) and hickory (11 percent). The extent and value of the urban forest resources of Illinois have not been adequately defined. Community- owned forests in the 55 largest urban areas total over 50,000 acres (Stewart 1980). It has been estimated (Himelick 1976) that there are over 25 million trees along streets and on private prop- erty in urban areas of the state. Some idea of the dollar value of urban shade trees can be gained by looking at the cost of the removal and replacement of Drs. D. F. Schoeneweiss, Dan Neely, and E. B. Himelick are Plant Pathologists, Section of Botany and Plant Pathology, Illinois Natural History Survey. 219 220 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 diseased or damaged trees. The mean in-house costs nationwide for munici- pal tree and stump removal are $140 per tree, with an additional $45 for replacement planting (Anonymous 1982). If the job is contracted, as is com- monly done in cities, the mean costs are $208 for removal and $86 for replant- ing. Removal and replacement take up 43 percent of municipal tree care budg- ets. It is estimated that in 1980, 57 communities in Illinois with over 25,000 population spent a total of $11,280,000 for municipal tree care (Stewart 1980). If the costs for privately owned trees were added to this figure, the estimate would likely be doubled. Tree care in Illinois supports sev- eral major horticultural industries. Illi- nois ranks among the top 12 states in the production of nursery stock, includ- ing trees, with an estimated wholesale value of over $100 million annually. Many growers import planting stock from other states and export high- quality specimen landscape trees throughout the eastern United States and Canada. Other sizable industries dealing with trees in the state are represented by the Illinois Landscape Contractors and Illinois Commercial Arborists associations. Also directly involved in tree planting and care are the Illinois Department of Ti'ansporta- tion, the Division of Forestry in the Department of Conservation, and the many county and municipal govern- ments in the state (Stewart 1980). Thus, trees are an important economic as well as aesthetic natural i-esource. BIOTIC AND ABIOTIC STRESSES Trees, like all living things, can be- come weak and sick and eventually can die. Even forest trees growing in a natural habitat are damaged by in- sects, diseases, and climatic stresses. During the rapid agricultural and in- dustrial expansion of the last century, the nation's forests were viewed by many as inexhaustible supplies of lum- ber and as natural barriers to agricul- ture to be cleared so that the land could be put to better use as farmland. This concept was challenged when the chest- nut blight fungus appeared in New York State in 1904 and within 30 years vii'tually eliminated one of the most valued tree species in North America. Since then, several droughts and such pests as the gypsy moth and the oak wilt and Dutch elm disease fungi, have destroyed vast numbers of trees, empha- sizing the vulnerability of our forest resources. Biotic Stresses Stresses caused by insects and dis- eases are classed as biotic (Fig. 1). Nearly all tree species are under cons- tant attack by pests, most of which are endemic and may be present wherever trees ai-e gi-owing. Disease stresses of trees are caused by fungi, bacteria, viruses, mycoplasma-like organisms (MLO), and nematodes (Hepting 1971). Except for crown gall and fire blight (bacteria), elm phloem necrosis (MLO). and pine wilt (nematode plus fungus), the major known biotic diseases of trees are caused by fungi (Carter 1975: Hept- ing 1971; Himelick 1976). Fungus species that require living plant tissue as a gi-owth substrate are obligate parasites, those that gi-ow on nonliving organic media as well as plant tissue are facultative parasites (Baker & Cook 1973; Wheeler 1975). Parasitic fungi vary considerably in their ability to attack higher plants. Obligate para- sites and virulent or aggi'essive faculta- tive parasites may attack otherwise healthy plants and act as primary stress factors causing disease damage. Weak or nonaggi-essive parasites cause damage only if plants are predisposed by other stresses (Schoeneweiss 1975b). Fungal pathogens enter suscept- ible and resistant host plants with equal frequency (Baker & Cook 1973; Wheeler 1975). Even wound pathogens that cannot penetrate host plant sur- faces directly find avenues of entrance through small wounds, cracks, and lenticels that are present on all plants. September 1985 125 Years of Biological Research 221 Fig. 1. - Biotic stresses as primary factors affecting Illi- nois trees. (A) Ver- ticillium wilt on green ash. (B) Nee- tria canker on thorn- less honey locust. (C)Sporulating cedar- apple rust galls on red cedar. (D) An- thracnose lesions on black walnut. Consequently, disease outbreaks and epidemics are influenced by environ- mental conditions of light, temperature, and humidity (Colhoun 1973). If a sus- ceptible host and a virulent pathogen are present and environmental condi- tions favor infection, disease damage will appear. Disease outbreaks reach epidemic proportions when pathogen inoculum is abundant and large num- bers of susceptible hosts are available, as is the case with street and parkway plantings in which species diversity is lacking. Some spore-bearing fungi and certain bacteria are spread from tree to tree by wind and rain, and others are transmitted by animals, primarily in- sects. Where pathogens and host trees have coexisted for many years, trees have usually developed resistance to disease through mutation and natural selection. Pathogens introduced into a population of trees that have not developed resistance can cause devas- tating disease epidemics. In 1934, the Dutch elm disease fungus came into this country from Europe. The Ameri- can elm, which was widely planted as a parkway tree throughout North America and was also a common woodland species, had virtually no resistance to the fungus. As popula- tions of the European elm bark beetle vector built up, an epidemic of Dutch elm disease began in the east and is still spreading toward the west coast. As already mentioned, the chestnut blight fungus, which spread rapidly by wind- and rain-carried spores, killed 222 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 nearly all of the American chestnuts within three decades, because the spe- cies was uniformly susceptible. This remains the only case on record of a plant pathogen virtually eliminating its host species (Hepting 1971). Abiotic Stresses Physical, chemical, and environ- mental factors that adversely affect tree health are called abiotic stresses, or physiological diseases (Parker 1965). The most common environmental stress factors are drought, flooding, and freez- ing temperatures. Man-induced stresses include wounds, toxic chemicals, pollu- tants, nutrient deficiencies, changes in soil grade and drainage, soil compac- tion, and loss of roots during trans- planting. Trees suffering from abiotic stress may exhibit a wide variety of symptoms (Fig. 2), from the dropping of older leaves to severe wilting, die- back, and eventual death. When dam- age occurs as a result of these factors without the involvement of a pathogen, the stress exerted is a primary abiotic stress. Nearly all trees suffer at one time or another from abiotic stress, but there are several critical periods when they are most susceptible to injury. A young seedling or a transplant strug- gling to establish a root system may be killed by even a short drought. After a tree becomes established, it is usually able to sm-vive all but unusually severe abiotic stresses for many years. Once the tree approaches maturity, however, the root system has ramified through- out the upper layers of soil and attained an increasingly delicate balance with the soil environment or rhizosphere. Abiotic factors, such as drought, flood- ing, soil compaction, or toxic chemicals, can stress mature and overmature trees and initiate a state of decline (Himelick 1976). Trees transplanted in- to a landscape setting and those in natural stands that are thinned by building contractors are frequently subjected to conditions that are likely to cause stress. In addition, landscape and parkway trees all too often are planted in disturbed soils where com- paction, alkalinity, and poor nutrient status are unfavorable for tree growth. The increased use of herbicides with lawn fertilizers has also damaged many trees in urban areas (Neely & Crowley 1974). The most critical period in the life of most urban trees is associated with transplanting, when abiotic stresses may become limiting factors in tree survival and performance. All tree planting stock, whether moved bare rooted or with a root ball, loses a major portion of its absorptive root system (Watson & Himelick 1982b). As a con- sequence, the uptake of water is re- duced, and transplants may suffer drought stress even though soil mois- ture content is adequate for establish- ed trees. To decrease weight and save shiping costs, more trees are being moved with tree spades or with small root balls; therefore, the amount of stress they are subjected to is increas- ing. Predisposing Stresses Although biotic and abiotic factors can, and often do, act as primary stresses affecting trees, it is common for several factors to interact in causing tree damage. When a tree is weakened by one stress so that it becomes more susceptible to another stress, the tree is said to be predisposed (Schoeneweiss 1975b). Ti'ees stressed by drought, flooding, freezing, or other factors often become predisposed to diseases caused by weak or nonaggi'essive pathogens that are not able to damage nonstressed trees (Schoeneweiss 1981). Most stem cankers, diebacks, and root rots occur on trees that have been predisposed by other stresses, especially recently trans- planted trees, which are often in a weakened condition and are predis- posed to biotic diseases (Himelick 1976). The interactions between abiotic and biotic stresses make the diagnosis of tree damage difficult, often requiring an examination of the tree by an expeit to determine the cause of the problem September 1985 125 Years of Biological Research 223 and to recommend the most effective prevention or treatment. Plant patholo- gists at the Illinois Natural History Sui'vey have provided such expertise as a service to Illinois citizens and institu- tions for over 50 years HISTORY OF TREE DISEASE RESEARCH IN ILLINOIS The involvement of the Illinois Natural History Survey with trees and their diseases is nearly as old as the Survey itself In 1881, Thomas J. Bur- rill, who was on the staff of the Illinois State Laboratory of Natural History as well as that of the state university, demonstrated for the first time that a plant disease, fire blight of pear, was caused by a bacterium. He later pub- lished articles on fungal and bacterial diseases of several forest, roadside, and street trees (Carter 1958). Stephen A. Forbes, director of the State Laboratoiy and guiding force in the evolution of the Natural History Survey, reported on a widespread dying of elms in Illi- nois in 1883 and again in 1912. Broad interest in the vegetational resources of Illinois led to the creation in 1921 of the Section of Botany, headed by Leo H. Tehon. Shortly thereafter, when elms began to die in many areas of the state, J. C. Carter joined the staff as a full-time plant pathologist to con- duct research on the causes and control of tree diseases. In 1935, the name of the section was changed to Applied Botany and Plant Pathology (Carter 1958) and was later shortened to Botany and Plant Pathology. In the 1930's research in plant pathology focused on disease problems of street and parkway elms. The rapid dying of elms was due to a disease called phloem necrosis, originally thought to be caused by a virus but recently shown to be caused by a mycoplasma- like organism (Carter 1958). Several other elm diseases were studied by Carter, including bacterial wetwood (Carter 1964). The effort devoted to tree disease research in Illinois was in- creased after the fungal disease known as oak wilt was found here in 1942. By 1950, when Dutch elm disease ap- peared in Illinois, Survey plant path- ologists were actively conducting re- search on tree diseases, and additional staff was added to meet this new threat to the rural and urban tree resources of the state. In 1955, plant pathologists scat- tered through several departments of the University of Illinois were brought together into a new Department of Plant Pathology. Since Survey patholo- gists were well established in tree dis- ease research at the time and the uni- versity pathologists were primarily in- terested in diseases of agronomic crops, the Survey accepted the responsibility for research on tree diseases, a commit- ment that continues today. Over the years. Survey plant path- ologists have cooperated in solving tree problems with many other state agen- cies, including the Departments of Transportation and Conservation and the Bureau of Plant and Apiary Protec- tion in the Department of Agriculture. As a state-supported institution, the Natural History Survey has provided diagnostic and counselling services to all Illinois residents. During the peak years of the Dutch elm disease epi- demic in the 1950's and 1960's, several thousand plant specimens were proc- essed annually in the pathology labor- atories. With the opening of the univer- sity's plant clinic in 1975, much of the specimen load has been reduced or diverted elsewhere; however, Survey pathologists continue to provide the public with diagnostic and counselling services. Nurserymen, landscapers, arborists, city foresters, pest control operators, Christmas tree growers, and private citizens receive willing coop- eration in solving their tree disease problems. In contrast to agi'onomic crops, in- volving only a relatively few species of hosts and pathogens, many abiotic and biotic factors affect dozens of species and cultivars of trees. Keeping up to date on tree disease problems requires 224 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 a continually evolving program of basic and applied research. As one problem is solved, new ones arise and become issues needing attention. RESEARCH ON BIOTIC STRESSES AS PRIMARY STRESS FACTORS Both the initiation and expansion of research projects on tree diseases at the Illinois Natural History Survey from the 1930's through the 1950's were motivated by the sudden appear- ance and rapid spread of vascular wilt diseases: phloem necrosis of elm in 1930, oak wilt in 1942, and Dutch elm disease in 1950 (Carter 1958). Elm Phloem Necrosis Little was known about phloem necrosis when it was identified in Illi- nois in the 1930's except that it was thought to be caused by a virus trans- mitted by the elm leafhopper Since virus diseases could not be controlled with chemicals and the application of foliar sprays to control leafhoppers on large trees is seldom practical, research efforts were mostly centered on disease diagnosis and the recording of disease spread and development. Accurate Sur- vey records of the urban epiphytotic of phloem necrosis and Dutch elm disease in the Champaign-Ui'bana communities (Carter & Carter 1974) are of value in predicting potential tree losses in other similar communities if control pro- grams are not initiated early enough. Oak Wilt The American oaks are the most important group of hardwoods in Illi- nois and furnish more native timber than any other group of bi'oad-leaved trees. Oak species have also been used extensively as shade and ornamental trees. By 1950, the fungus disease, oak wilt, first identified in Illinois in Win- nebago County in 1942, was destroying large areas of oak timber in northern Illinois. Intensive studies on the distri- bution of the fungus in Illinois, poten- tial vectors, and procedures to control its spread were begun in 1951 with a grant from the Cook County Forest Preserve District, later supplemented and eventually replaced with state and federal funds (Carter 1968; Himelick 1958). Oak wilt disease was found to be present in almost every Illinois county with large areas of native oaks (Hime- lick 1958). The sexual and asexual spore stages of the fungus and their role in the life cycle of the fungus were described (Himelick & Fox 1961). Squir- rels and various sap-feeding insects were shown to be potential vectors of the fungus (Himelick & Curl 1955; Himelick & Curl 1958) although none proved to be efficient enough to pose a threat of epidemic disease. Survey pathologists found that the most significant spread occurred through root gi-afts from diseased to adjacent healthy trees, and methods were tested for interrupting these root grafts. Both mechanical trenching and the use of the soil sterilant Vapam to form chem- ical barriers (see Dutch elm disease) prevented the spread of the fungus through root grafts (Himelick & Fox 1961). Poisoning trees with sodium arsenite or potassium iodide prevented the formation of fungus pads (Curl 1955) from which spores of the fungus could be carried by insect and animal vectors to healthy trees (Himelick & Fox 1961). The disease is now limited to small pockets oftimber in which con- trol measures have not been used. Oak wilt was, and still remains, a potential threat to the predominantly oak forest and woodlots of the state. If an efficient vector appears, Illinois forests would become highly vulner- able to oak wilt epidemics. Therefoi-e, Survey pathologists are keeping a close watch for any sudden increase in the incidence of oak wilt. ^, Dutch Elm Disease In contrast to the oak wilt fungus, the fungus that causes Dutch elm disease, and its bark beetle vector, found conditions in Illinois ideally pro- gi-ammed for disease epidemics. The streets and boulevards of most cities in September 1985 125 Years of Biological Research 225 the Midwest were planted almost exclu- sively to American elms, which are highly susceptible to the disease. The European elm bark beetle, an intro- duced species, breeds in dying elms. When the beetle arrived in advance of the fungus, beetle populations built up rapidly on elms dying from phloem necrosis. In 1950, the Dutch elm disease fungus appeared in Coles County and rapidly spread throughout the lower half of the state, where phloem necrosis was prevalent. By 1959, Dutch elm disease was found in all 102 counties of the state (Carter & Carter 1974). This phenomenal spread was much more rapid and extensive than that which had occurred earlier in the eastern states. Consequently, a program was initiated in the 1950's to attempt to find means of controlling the disease or at least slowing down its rapid spread so that communities could replace dying elms over an extended period. The beautiful colonnades of park- way elms in Illinois cities began dying rapidly in the 1950's from Dutch elm disease. Since the disease appeared to spread from infected trees to adjacent trees along parkways, Survey patholo- gists made extensive studies of disease progression and found that the fungus also moved from tree to tree through root gi-afts. Root gi-afts were so common among parkway trees that up to 100 percent of the trees within 20 feet and 90 percent of the trees within 35 feet of infected trees also became infected (Himelick & Neely 1962). To halt this type of spread, several methods were devised and evaluated for breaking or interrupting elm root grafts. Mechan- ical trenching proved to be effective but costly. Several soil fumigants were ap- plied by different methods to kill roots and form a barrier between adjacent trees (Neely & Himelick 1965). The most effective treatment found was the injecting of the soil sterilant Vapam, diluted 1:3 with water, into 3/4-inch diameter holes 15 inches deep and 6 inches apart in a line between adjacent elms. Vapam treatments reduced adja- cent tree infections by 60 percent in all Illinois cities where it was tested (Neely & Himelick 1965). This method became a standard recommendation for the control of Dutch elm disease in Illinois cities and was subsequently adopted as a control procedure in other states. Control strategies were also devised and tested to prevent or reduce the overland spread of the fungus by bark beetles emerging from brood galleries in diseased elms. Over 1 million beetles had been estimated emerging from a single dying elm; therefore, strict san- itation in the form of the prompt re- moval and destruction of dying trees and branches became an essential com- ponent of any effective control program. Unfortunately, few communities were able to organize and fund community- wide control programs until many trees had already been lost to the disease. The biggest problem for communities where elms were dying was that the prompt removal of dying trees was essential to contain the disease epi- demic, but sufficient funds and man- power for tree removal were simply not available. To help alleviate this situa- tion. Survey pathologists developed a technique based on previous oak wilt research in which either sodium arsenite at the rate of 100 g/1 of water or potassium iodide at the rate of 500 g/1 was placed in axe frills at the base of a tree (Himelick & Neely 1961). This treatment caused such rapid tree death that the tree did not become suitable for beetle breeding. The treatment gave almost complete prevention of beetle colonization and allowed communities to remove dead elms at convenient times while maintaining an effective control program. With the advice and recommenda- tions of Survey pathologists, 55 Illinois cities had begun control progi-ams for Dutch elm disease by 1957 (Neely 1967). The basic program included strict sanitation and the application of DDT sprays. When DDT was banned in 1968, more than half of the cities dropped their control programs, while the remainder substituted the less 226 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 effective insecticide methoxychlor for DDT. In cities with no control pro- grams, over 90 percent of the elm pop- ulations were lost to disease by 1966 (Neely 1967), and those that dropped their programs soon lost most of their remaining elms. The 21 cities that have continued control programs through 1982 have retained from 30 to 50 percent of their original elm popula- tion, allowing the gi-adual replacement of dying elms with other tree species over a 25-year period (Neely 1984). Con- trol strategies developed at the Survey have served as models for community- wide Dutch elm disease control pro- gi'ams in many parts of the country. Another phase of research which found practical application was a com- prehensive study on the spread of the Dutch elm disease pathogen within in- fected elms. When inoculations were made at the base of the trunk, the fungus was recovered at a height of 6 feet in 3 days, and at 6 days was present in the tops of trees 4 to 5 meters high. In contrast, twig inoculations, simulat- ing beetle feeding transmission, re- sulted in downward movement of only 2 cm per day, and it was weeks or months before the fungus reached the main stem (Neely 1968). Unless a main stem was reached in the first year of inoculation, the fungus became local- ized. Using this information, Survey pathologists found that if infected branches were surgically removed be- fore trees showed more than 5 percent wilt (flagging) symptoms, up to 65 per- cent of the infected trees could be saved (Himelick & Ceplecha 1976). Therefore, early detection and prompt pruning were included in control recommenda- tions. Many additional studies on Dutch elm disease were conducted, including a survey of resistant species and cul- tivars of elm. Of 13 species and sub- species of iilmus on the University of Illinois campus, only U. americana L., U. rubra Muhl., and U. thomasi Sarg. were lost to the disease (Neely & Carter 1965). Unfortunately, nearly all parkway trees on the campus and in the twin cities of Champaign-Urbana were U. americana. Of the original pop- ulation of 14,103 American elms, only 47 trees remained in 1972 (some of these have since succumbed to disease); the rest were killed by either phloem necrosis or Dutch elm disease (Carter & Cai-ter 1974). This loss resulted from the lack of any control program in the twin cities. Fortunately, other com- munities in the state that followed Survey recommendations were able to retain many of their elm trees. Verticillium Wilt Verticillium wilt is a vascular disease caused by a soil-borne fungus, Verticillium dahliae Kleb., and is the only vascular tree disease that affects a wide range of unrelated annual and perennial plant species (Himelick 1969). In Illinois, the disease is rare among forest trees, but it has become increasingly prevalent in ornamental plantings (Fig. lA). Presently, 60 species of trees and shrubs are known to be susceptible (Himelick 1969). There is evidence that the fungus varies in virulence, and the indiscrim- inate movement of nursery stock from one state to another has resulted in dis- tribution of the more virulent strains to areas of Illinois where they were not previously present. Sm-vey pathologists have reseai'ched the biology, control, and host range of Verticillium over many years. Diagnosis is difficult because trees weakened by root infection usually decline and be- come more susceptible to the effects of other stresses. Attempts at chemo- therapy with fungicides have been un- successful. Feeding and watering infected trees are the only treatments that have proved helpful in extending the longevity of trees having Verticil- lium wilt. OTHER BIOTIC DISEASES AS PRIMARY STRESS FACTORS Sycamore Anthracnose American sycamore, a common river-bottom species in the eastern half September 1985 125 Years of Biological Research 227 of the United States, has been widely planted as a shade tree on parkways and private property. It is a hardy, fast- growing species, but its value as an urban shade tree is limited by a fungal disease known as sycamore anthrac- nose (Himelick 1962; Neely 1976). Occasionally, sycamores appear quite healthy throughout the summer, but in most years anthracnose damage ranges from moderate to severe defoliation early in the growing season. In re- sponse to numerous requests from nurserymen, arborists, and home- owners for disease control recommen- dations, pathologists at the Natural History Survey conducted research on the biology and control of this disease. A taxonomic study of the pathogen revealed that it was distinct from the oak anthracnose fungus (Neely & Himelick 1967), and it was established as a separate species (Matteoni & Neely 1979). The disease cycle on syc- amore was found to be composed of several stages (Himelick 1962). The fungus overwinters in current-season twigs and causes bud, twig, and shoot blight symptoms following bud break in the spring. It also forms stem cankers in which spores are produced that serve as a source of new infections. During late spring and summer, spores infect leaves, causing the leaf-blight stage. The fungus invades leaf petioles and grows downward into the stem, where it remains over the winter Since the fungus spends most of its cycle within petiole and stem tissues, the disease is difficult to control with surface-active fungicides. Mercury fungicides, which penetrate plant tis- sues and act as localized systemics, gave good control of anthracnose, but mercuries have been banned from use because of their high mammalian tox- icity. No other foliar fungicides tested have been effective; however, pressure injection of trees with the fungicide Arbotect has given good control for as long as 3 years (Himelick & Duncan 1982). Research on the anthracnose cycle led to the development of a predictive model for disease outbreaks. Patholo- gists found that temperature controls the severity of shoot blight, the most damaging stage of the disease (Neely & Himelick 1963a). If mean daily tem- peratures during the 2 weeks following first leaf emergence average less than 13 °C (55 °F), severe damage occurs. When temperatures from 13° to 16 °C (55° to 60 °F) occur, damage is moder- ate, and at temperatures over 16 °C (60 °F) no shoot blight appears. Cool temperatures in both fall and spring favor fungal growth in twigs. After several years of testing and refinement, this model has proved to be quite accu- rate in predicting anthracnose severity and is useful in decision making as to if and when fungicides should be ap- plied for disease control. Walnut Anthracnose Black walnut {Juglans nigra L.) has long been a highly valued native tree in Illinois, both as a source of cabinet- grade and veneer lumber, and as a pro- ducer of edible nutmeats. Walnut plan- tations have sprung up throughout the southern part of the state as farmers and other landowners have rushed to take advantage of the increasing de- mand and rising prices for walnut prod- ucts. Because of nationwide interest in walnut, the U.S. Forest Service made funds available for research on walnut culture and pest control. A Survey plant pathologist received a grant to study control measures for walnut an- thracnose (Fig. ID), the most serious disease of walnut (Neely 1979). Since effective control measures for a plant disease depend upon a thor- ough knowledge of the growth habits and life cycle of the pathogen, studies were conducted on the anthracnose fungus, using artificial media in the laboratory and host plants in the field and gi'eenhouse. Detailed information was obtained on pathogen gi'owth and sporulation (Matteoni & Neely 1979) and on host susceptibility at different growth stages (Black & Neely 1978a). 228 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Wide variation was recorded in resis- tance to anthracnose among various species and hybrids within the genus Juglans, with a tendency toward greater resistance when trees are pro- ducing rapid growth (Black & Neely 1978b). Results of fungicide trials showed that several compounds gave good con- trol of walnut anthracnose, the best being benomyl (Benlate) applied as a foliar spray. Two applications are re- quired, the first in late May to control ascospore infection and the second in early July for conidial infection (Neely 1979). Since rapidly growing trees ap- peared to be more pathogen resistant, various fertilizer combinations were applied to plantation walnuts. The ap- plication of nitrogen, at the rate of 3 kg N/ha of soil surface, resulted in a 60-percent decrease in defoliation by the end of the season (Neelyl981). Con- sequently, Survey plant pathologists now recommend a combination of fun- gicide sprays and fertilization for max- imum control of walnut anthracnose. TREE SELECTION FOR DISEASE RESISTANCE Control of biotic tree diseases may involve several methods or approaches. These include chemotherapy, surface and soil applications of fungicides, sanitation, fertilization to promote tree vigor, and selection and breeding (Carter 1975). Since the breeding of trees requires many years, selection among available types is far more com- mon than breeding in the development of disease-resistant trees (Carter 1966). In most cases, species and varieties within a genus vary in resistance to biotic pathogens (Bingham et al. 1971); however, accurate information on gen- etic resistance in trees is often lacking due to the considerable time and effort needed to obtain this information. Sur- vey pathologists working with tree diseases over the years have made sur- veys and kept records of host resistance to several of the common tree diseases. Most of these studies have been con- ducted in cooperation with the Morton Arboretum at Lisle, Illinois, where a wide range of tree species and varieties is located. Lists have been published of junipers resistant to twig blight (Schoeneweiss 1969) and cedar rusts (Himelick & Neely 1960) and of varie- ties of horse chestnut and buckeye resistant to leaf blotch (Neely & Hime- lick 1963b). Results of sm-veys and inoc- ulation experiments have been com- bined into a list of woody hosts suscep- tible to the Verticillium wilt fungus (Himelick 1969). This information is valuable to growers and propagators in selecting and selling trees that are re- sistant to diseases. DISEASE CONTROL WITH FUNGICIDES With the development of organic pesticides after World War II, the most common method of disease control was the application of fungicides. However, the overuse and abuse of such pesti- cides, combined with increased envi- ronmental awareness, has led in i-ecent years to the banning or restriction of pesticide usage. The amount of scien- tific data now required before a fungi- cide can be labeled for a specific use imposes a particularly severe con- straint on the control of tree and shrub diseases, which are considered minor uses. With funds provided by the fed- eral progi'am for minor-use pesticides. Survey pathologists have tested many fungicides on ornamental plants and have submitted data on phytotoxicity to aid in obtaining label clearance so that these compounds can be legally used for disease control on nurseiy and landscape plants. Fungicide Bioassay Procedures To select a suitable fungicide to control a plant disease and to deter- mine the proper timing and number of September 1985 125 Years of Biological Research 229 applications needed, it is necessary to know at what rate of application the fungicide is effective, whether it will re- main active on or in the host, and whether the activity will persist or be lost through weathering. These data are usually obtained through the use of bioassays in fungicide screening tests. Because most bioassays gave in- consistent results. Survey plant path- ologists developed a cellophane disc technique, which can detect small amounts of fungicide (as low as 0.01 ppm) and may be used in the field as well as in the laboratory (Himelick & Neely 1965). This technique is simple, fast, and economical, and it is now used to determine whether a fungicide kills or merely inhibits disease organisms and whether it will persist on or in plant tissues long enough to give good disease control. Chemotherapy The term chemotherapy refers to the internal treatment of infected plants with chemicals to cure disease or arrest further infection. The field of chemotherapy for tree diseases is in its infancy. Injection into the soil and direct injection of systemic fungicides into the vascular system of trees have been used in attempts to control vas- cular wilt diseases, but results have often been either negative or inconclu- sive. Survey pathologists developed a pressure-injection apparatus with which they have injected as much as 110-150 1 of dilute systemic fungicides into several tree species (Himelick 1972). American elms in the early stages of Dutch elm disease have been saved by pressure injections of the fun- gicides Benlate and Arbotect (Himelick 1972), and good control of sycamore anthracnose for up to 3 years has been achieved with the injection of Arbotect (Himelick & Duncan 1982). We expect that other serious diseases of shade and ornamental trees will be controlled by pressure injection in the future with new systemic fungicides. RESEARCH ON ABIOTIC STRESSES AS PRIMARY STRESS FACTORS Although damage symptoms on trees are often attributed to infectious diseases, the fact is that much damage to trees and other plants is caused by abiotic stresses (Levitt 1980a, 1980b; Parker 1965). Since Survey tree path- ologists are concerned with the protec- tion and preservation of trees in gen- eral, many research projects have been conducted on abiotic stresses. Herbicide Damage The increased use of herbicides for weed control in both rural and urban areas of Illinois is reflected by a higher incidence of damage on sensitive tree species. Rural windbreaks and woodlot trees are being injured by herbicides used on crops, while urban trees are showing the effects of weed-and-feed lawn fertilizers which have become quite popular for the control of weeds in turf grasses (Fig. 2B). A recent study at the Survey revealed that the herbi- cide dicamba, commonly incorporated with lawn fertilizers, consistently caused significant damage to trees (Neely & Crowley 1974). Tree species differed in sensitivity to this com- pound, and the extent ofdamage varied with soil type and the amount of spring rainfall. Pathologists at the Survey have cautioned against the use of dicamba in the vicinity of trees and other woody plants. Iron Chlorosis Yellow foliage and slow growth are typical of many tree species planted along streets and around homes where the original topsoil has been removed or mixed with subsoil. The average urban soil has relatively poor physical, chemical, and biological characteris- tics. In Illinois, chlorosis (yellow foli- age) is a common physiological stress on pin oak (Fig. 2D) and to a lesser 230 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Fig. 2. - Abiotic stresses as primeiry factors affecting Dli- nois trees. (AlOak decline in northern Illinois. (B) Injury from lawn herbicide on yew. (C I Loss of roots during trans- planting. (D) Iron chlorosis symptoms on pin oak leaves. extent on sweet gum, white oak, bald cypress, river birch, red maple, and hackberry. Several Stirvey pathologists have investigated the cause and control of the chlorosis problem. Soil tests from areas of chlorotic trees have consistently shown that chlorosis is generally associated with pockets of alkaline clay soils, quite common in urban areas in the north- ern part of the state (Schoeneweiss 1973). The yellowing of leaves, with veins remaining green, is a typical symptom of iron deficiency and can be corrected by treating plants with cer- tain iron-containing compounds if the iron is in the proper form (Wallace 1971). Field research has centered around both soil injection and direct trunk implantation of iron compounds. Trunk implantations with ferric citrate and ferric ammonium citrate were found to be effective on pin oak, red oak, and swamp white oak; cypress; and sweet gum (Neely 1973; Himelick & Himelick 1980). These compounds are now commercially available in Medicaps"^ and are being used nation- ally to correct chlorosis by commercial arborists, nm-serymen, and operators of fruit orchards. Applications of inor- ganic iron compounds to the soil, prior to the 1960's a standard recommenda- tion, was totally ineffective in the alka- line clay soils of Illinois. However, pres- sure injection of liquid organic com- pounds, known as iron chelates, cor- rected even advanced stages of chlor- osis in highly alkaline soils for 3 or more years (Schoeneweiss 1973). Either treatment may be used for iron chlor- osis in Illinois. September 1985 125 Years of Biological Research 231 Nutrient Deficiencies Many urban trees planted in dis- turbed soils have poor growth charac- teristics due to deficiencies of nutri- ents. Although the fertilizing of park- way, street, and private-property trees in cities is a common practice, most fer- tilization recommendations are based on studies of crop plants or turf grasses and have not been evaluated on trees. To find out what nutrient materials and methods of application are best for promoting tree growth. Survey patholo- gists applied several different fertilizer combinations by several methods to experimental blocks of shade tree species at five locations throughout Illi- nois. They found that trees responded to soil applications of nitrogen fertili- zers and that the response was directly related to nitrogen content regardless of how the fertilizer was applied or other nutrients present (Neely & Himelick 1966). Ammonia, ammonium sulphate, urea, and ureaform were equally effective, based on equal nitrogen content. Optimum growth response was obtained with young, newly established trees when fertilized in April at the rate of 3 Kg N/ha of soil surface (Neely & Himelick 1966; Neely at al. 1970). Growth response varied with tree species, but all treated trees were healthier than untreated controls. Proper fertilizing of urban trees en- hances tree vigor and helps trees resist attack by many stress-related disease and insect pests. Transplanting Stress The bulk of the root systems of trees are lost when they are dug and moved from the nursery to a planting site (Watson & Himelick 1982a). Be- cause of high labor costs for hand ball- ing nursery stock, more trees are now being moved with large mechanical tree spades, which cut off a larger por- tion of the root system than does hand balling (Fig. 2C). Consequently, losses of transplanted trees from damage caused by abiotic stresses (Himelick 1976) and predisposition to biotic stresses (Schoeneweiss 1965, 1966) have increased. Survey plant pathologists recently conducted a 4-year study on the pat- terns of root distribution and regenera- tion of seven species of trees trans- planted with a commercial tree spade (Watson & Himelick 1983). The study was conducted in the Survey arbore- tum at Urbana, where core samples from root balls and the soil could be taken to the laboratory for detailed examination. Data were collected on root carbohydrate content, vertical and horizontal distribution of the original root system at the time of digging, and regeneration of fibrous feeder roots after transplanting. They found that as much as 98 per- cent of the original root system was lost during digging, yet the survival rate after 1 year was 97 percent, due to careful maintenance of the trees after transplanting (Watson & Hime- lick 1982b). Because the tremendous loss of roots caused many trees to become severely water stressed in mid- summer, frequent irrigation was essen- tial for survival. New fibrous roots formed at the severed ends of the roots at the edge of the root ball and were most numerous in the uper 10 cm of soil. Root regeneration was propor- tional to carbohydrate content, which was lowest during the period of bud break and new shoot growth in early spring, indicating that trees should not be moved until new growth has ma- tured or, preferably, after the onset of dormancy in the fall. Results of this study were compiled in a transplanting manual, published by the International Society of Arbori- culture (Himelick 1981), for the pur- pose of developing uniform transplant- ing procedures. Use of this information should help to reduce losses of urban shade and ornamental trees after trans- planting. Oak Decline Several thousand large, established oaks have died or are in various stages of decline on private and public prop- 232 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 erty in cities adjoining Lake Michigan (Fig. 2A). Aerial reconnaissance, using high-resolution photography, has shown extensive loss of valuable trees in many of the urban areas north of Chicago, where oak is the predominant shade tree (Himelick 1976). Although declin- ing trees have been attacked by disease and insect pests, the death of high pro- portions of the root systems, caused by alternating periods of excess soil mois- ture and drought, is responsible for the decline. Through field research it was demonstrated that a 5- to 10-cm layer of wood chips placed over the root area and the addition of surface-applied fer- tilizer were highly beneficial in restor- ing the vigor of delcining trees. Fine root development more than doubled in the top 15 cm of soil among trees that were mulched in this manner. Large, 75- to 100-year-old oaks also showed measurable response in greener foliage compared with that of untreated trees. stresses (Fig. 3) (Schoeneweiss 1981). Others, such as Hypoxylon canker on aspen and Armillaria root rot of maple and oak, that cause minor injury to vigorous trees may become much more damaging to trees under stress (Scho- enweiss 1978). Interactions between abiotic and biotic stresses have long been recognized, yet few studies had been conducted until recently on sti-ess predisposition in woody species. Envi- ronmental stresses that occur under field conditions are both highly vari- able and unpredictable, making it diffi- cult to obtain reliable data on stress- disease interactions from field studies. Most of the useful information has come from studies on controlled stress research at the gi-eenhouse and gro%vth chamber facilities of the Illinois Nat- ural History Survey. Although many different stresses may act or interact in predisposing trees to diseases, research thus far has been concentrated on drought, freezing, and defoliation stresses. STRESS PREDISPOSITION Many of the biotic diseases of trees that cause significant damage, such as the stem cankers, diebacks, and root rots, only appear on trees that have been predisposed to disease by abiotic Drought Stress The influence of drought, or water stress, as a predisposing factor in stem canker diseases of trees was investi- gated by placing potted tree seedlings in various stages of wilt into a specially designed humidity cabinet under equi- LOW Predisposition to Biotic Stress MODERATE Primary Abiotic Stress SEVERE LEVEL OF ABIOTIC STRESS Fig. 3. - The ex- tent of primary dam- age caused by abio- tic stress compared with the extent of biotic disease dam- age resulting from sti-ess pi-edisposition at increasing levels of abiotic stress se\-er- ity. Tives may become predisposed to infec- tious disease organ- isms at moderate lev- els of stress, while direct or primary stress damage usu- ally appears only on severely stressed trees. September 1985 125 Years of Biological Research 233 librium conditions of high humidity, constant temperature, and reduced light (Schoeneweiss 1975a). Under these conditions, plant water potentials became stable, and the level of water stress was correlated with the develop- ment of stem cankers caused by weak or nonaggressive pathogens (Crist & Schoeneweiss 1975; Schoeneweiss 1975b, 1978, 1983). These studies re- vealed that woody stems became pre- disposed to disease when plant water potentials fell below a threshold level of -12 to - 13 bars (1 bar = 0.97 atmos- phere) and that disease development increased with increasing stress severi- ty (Crist & Schoeneweiss 1975). The predisposing level would be considered moderate stress, as most tree species do not show signs of wilting at this level. In the field, water potentials of trees may fall well below the threshold level on a hot, windy day yet recover at night without causing predisposition. For predisposition to occur, plant water potentials must remain below the threshold level for several days before trees appreciably lose disease resis- tance (Wene 1979). Trees regain resis- tance several days after the stress is relieved. Most trees are under pro- longed water stress during droughts and after transplanting and should be irrigated regularly at these times to prevent predisposition to biotic disease. Freezing Stress The predisposing effect of freezing stress was studied by placing dormant potted tree seedlings in a program- mable walk-in freezer and lowering the chamber temperature to below freezing (Schoeneweiss 1974, 1977). Stems inoc- ulated with canker fungi became pre- disposed to disease when stem temper- atures exceeded a threshold degree, usually around -20°C (-13°F). Pre- disposed plants regained resistance within 1 to 2 weeks after stems were thawed (Wene 1979). Again, the level of stress resulting in predisposition was considered moderate, since no direct in- jury was apparent, and frozen plants that were not inoculated were indistin- guishable from unfrozen controls after growth resumed (Schoeneweiss 1981). Defoliation Stress Loss of foliage during the growing season reduces the photosynthetic capacity of trees and may weaken them sufficiently to predispose them to biotic diseases (Schoeneweiss 1981). Defoliat- ing insects, like the gypsy moth and cankerworms, have been followed by the decline and death of trees due to canker, dieback, and root rot fungi. Controlled studies at the Survey, where tree seedlings were manually defoli- ated, resulted in predisposition to canker fungi when plants were defoli- ated beyond a threshold period of sev- eral weeks (Crist & Schoeneweiss 1975; Schoeneweiss 1967). If allowed to refol- iate, seedlings recovered without ill effects in the absence of biotic path- ogens. Plant Tissues AffectecJ by Different Stresses Although many environmental stresses predispose trees to weak biotic pathogens and all seem to require some threshold level that must be exceeded, the effects of stress on plants may be quite different. Drought, defoliation, and most other stresses weaken the entire plant and may predispose any portion of the plant to attack (Schoene- weiss 1981). In contrast, freezing stress predisposes only that portion exposed to temperatures below the threshold degree (Wene & Schoeneweiss 1980). Freezing stress has a greater effect on older wood tissues, while other stresses predispose bark and sapwood (Schoene- weiss 1981). Pathogens also colonize the xylem of stems stressed by freezing to a much greater distance beyond canker margins than they do with other stresses. In pruning diseased stems to help stressed trees recover, it is helpful to know what environmental stress was involved. Stress-Related Pathogens The appearance of stem cankers, diebacks, and root rots is usually an in- 234 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 direct effect of predisposing abiotic stress. Some of the organisms that cause the diseases are present on plant- ing stock imported from other states, but most are common saprophytes of woody tissues and are widely dis- tributed. These pathogens are often re- ferred to as "secondary" organisms, and their role in causing damage has been disregarded; however, trees recover from stress without ill effects in the absence of these pathogens. Therefore, they should be thought of as essential components of the damage resulting from stress (Schoeneweiss 1981). As dif- ferent environmental stresses may dif- ferentially predispose tree species to disease, information derived from con- trolled studies with specific hosts and pathogens in combination is useful in formulating recommendations for dis- ease prevention or treatment. Out- breaks of diseases resulting from stress predisposition may not be noticed for days, months, or even years after the stress has been relieved. Climatic data and the cultural history of a tree may provide clues to the factors contribut- ing to disease damage. CURRENT RESEARCH ON BIOTIC AND ABIOTIC STRESSES Given the number and complexity of biotic and abiotic stresses and the wide range of species and cultivars of trees grown in Illinois, tree disease research is a continually changing and evolving process. Even in a paper of this size, many of the research projects conducted by Natural History Survey plant pathologists over the years must of necessity be omitted. In this section we briefly look at some of the major problems under investigation at the present time. Histology and Biochemistry of Predisposition When a tree is predisposed to dis- ease by stress it becomes more suscep- tible to another stress, in most cases a biotic pathogen (Schoeneweiss 1975b). Since nonstresed trees ai'e able to resist attack by most pathogens, a histolog- ical or biochemical change most likely occurs in the pi'edisposed host before infection (Bell 1981). Which mechan- isms change or break down dui'ing pre- disposition is a controversial subject about which little is known. Research is currently under way at the Survey to help answer these questions. Weak or nonaggressive stem-canker fungi are examples of biotic agents that only attack trees that have been pre- disposed to disease. When stems of nonstressed and stress-predisposed tree seedlings, inoculated with canker fun- gi, were examined with the light and scanning electron microscopes, no mor- phological barriers or deposits were found that could account for resistance to fungal growth in vascular tissues (Wene 1979). In resistant stems, the canker fungi appear to be inhibited biochemically or degi-aded by host en- zymes (McPartland & Schoeneweiss 1984). More research is needed to understand which mechanisms operate and how trees become predisposed. Once the resistance mechanisms are identified and characterized, it may be possible to select or breed trees less sensitive to predisposing stress or to apply treatments to enhance the ex- pression or genetic resistance to stress. Wound Healing Ti'ee wounds serve as avenues of en- trance for disease and decay organisms. Although much has been written about wound healing and wound treatment, few recommendations have been based on sound i-esearch data. In studies be- ing conducted on the healing of various sizes and shapes of wounds on different tree species, pathologists at the Natm-al History Survey have found that wound closure is directly related to radial stem gi'owth. Large wounds close more per unit of radial gi'owth than do small ones, but wounds on fast-gi'owing trees within a species close less per unit of gi'owth than those of slow-gi'Dwing September 1985 125 Years of Biological Research 235 trees (Neely 1983). Some tree species are more efficient in wound closure than others. If trees are maintained properly, wounds of less than 1.2 cm in diameter, such as those resulting from trunk injections, close within a year. Tests are under way on the closure and healing of branch pruning wounds, common on nursery and landscape trees. Data from this study will provide a sound basis for recommendations on wound treatment. Modeling of Cedar Rust Infections Cedar rusts are common and troub- lesome diseases of crabapple and haw- thorn. Using pathogen and environ- ment as the variable functions, we have developed models to predict the likeli- hood of infection by the cedar-apple and cedar-hawthorn rust fungi. These models will be tested against natural field infection to evaluate their accu- racy in forecasting disease epidemics. If accurate, they would be of consider- able value as a basis for effective fun- gicide recommendations. Tree-Grass Competition In most urban parkway and land- scape settings trees are surrounded by turf grasses. As a result, these trees are often stressed due to competition be- tween tree and grass roots for available water and nutrients. Pathologists at the Survey are collecting data on growth rates of trees in experimental plots with and without living tui'f cover and with and without supplemental fertilization. The aim is to determine whether killing the turf with herbi- cides will improve tree vigor. Hopefully, effective means of preventing or reduc- ing stresses in landscape trees will be derived from this study. Honey Locust Decline Of the many declines that occur on shade and ornamental trees, one of the most serious is the decline of the thorn- less honey locust. Because honey locust has been the preferred species used to replace elms lost to Dutch elm disease, it has been widely planted in urban areas of Illinois, and decline has be- come a prominent problem. Isolation and inoculation studies with fungi from declining trees indicate that Thyronectria astro-americana (Speg.) Seeler is a major cause of decline of older, established locust (Stim & Hime- lick 1981). Symptoms include cankers that girdle twigs and branches and, in advanced stages of decline, large, elon- gated cankers that develop on the trunk. An insect borer is associated with the canker and may be a vector of the fungus responsible for the dam- age. Symptoms appear particularly during periods of drought. Studies are in progress to determine if various cultivars of honey locust show resis- tance to the fungus. Pine Wilt For the past 5 years, a rapid decline of pines has been observed in both urban areas and pine plantations in Illinois and neighboring states. The incidence of this disease increased dramatically during 1980 and 1981, particularly among Scotch and Austrian pines in Illinois, and in some areas reached epidemic proportions. Many nurserymen and growers of Christmas trees have become concerned about the potential economic loss that could occur if the disease continues to spread. The pine wilt nematode, Bursephalen- chus xylophilus Mamiya & Kiyohara, and a blue-stain fungus, Ceratocysts ips (Rumb.) C. Moreau, appear to be closely associated with the disease (Himelick 1982). An international conference on pine wilt was held recently in Illinois, and a cooperative research effort in- volving entomologists, nematologists, and plant pathologists from the Natural History Survey and the University of Illinois has been initiated to investi- gate various aspects of the pine wilt syndrome. OUTLOOK FOR ILLINOIS TREES In spite of research efforts by Natural History Survey plant pathologists, damage caused by biotic and abiotic 236 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 stresses to Illinois trees is likely to in- crease through the foreseeable future. Urban trees, in particular, are being neglected because of high maintenance costs. Frequently, more nursery stock is moved with tree spades or grown in containers; in both cases, root systems are often inadequate to assure tree health after transplanting until new roots are established. Modern farming practices in Illi- nois, particularly the application of large quantities of herbicides and the plowing of every available square foot of land for row crops, are highly unfa- vorable to tree health. As a result, rural windbreak and woodlot trees are in poor condition, and many are show- ing signs of decline. While stresses are becoming more prevalent as factors affecting Illinois trees, the demand for trees in the state is increasing. Continuing urbanization is creating a greater demand for more shade and ornamental trees for beau- tification and as noise and pollution filters. The tremendous loss of topsoil to erosion on farm fields may stimulate new interest in windbreak plantings. Ti-ee plantations for the production of biomass as an energy source and for windbreaks are becoming popular. Re- forestation of mine spoils and marginal land is being funded at the state and federal levels. All of these factors are contributing to a greater demand for tree planting stock. The chances for the introduction of a disease organism that could result in a tree disease epidemic has increased in recent years. Many of the new culti- vars grown in Illinois nurseries are propagated in other regions of the country and are imported as planting stock along with the pathogens that occur in those i-egions. Under Illinois conditions, some of these pathogens will undoubtedly become serious threats. For these and many other reasons, biotic and abiotic stresses affecting Illi- nois trees will continue to be issues re- quiring attention from Natm-al Histoiy Survey plant pathologists. Information from both basic and applied research will be needed to meet these issues. LITERATURE CITED Anonymous. 1982. Municipal tree management. Urban Data Service Report 1(14):1-14. Inter- national City Management Association, Washington, D.C. Baker, K. F., and R. J. Cook. 1973. Biological Control of Plant Pathogens. W. H. Freeman, San Francisco, CA. Bell, A. A. 1981. Biochemical mechanisms of disease resistance. Annual Review of Plant Physiology 32:21-81. Bingham, R. T., R. J. Hoff, and G. I. McDonald. 1971. Disease resistance in forest trees. An- nual Review of Phytopathology 9:433-452. Black, W. M., and D. Neely. 1978a. Effects of temperature, free moisture, and relative humidity on the occurrence of walnut an- thracnose. Phytopathology 68(7): 1054- 1056. , and . 1978b. Relative resistance of Juglans species and hybrids to walnut anthrac- nose. Plant Disease Reporter 62(61:497-499. Carter, J. C. 1958. Applied botany and plant pathology. Pages 145-162 in A century of bio- logical research. Illinois Natural History Sur- vey Bulletin 27:84-234. 1964. The wetwood disease of elm. Illinois Natural History Survey Circular 50. 1966. Illinois trees: selection, planting, and care. Illinois Natural History Survey Circular 51. 1975. Diseases of midwest trees. Univer- sity of Illinois Special Publication 35. , and L. R. Carter. 1974. An uiban epiphy- totic of phloem necrosis and Dutch elm dis- ease, 1944-1972. Illinois Natural History Survey Bulletin 31:113-143. CoLHOUN, J. 1973. Effects of environmental fac- toids on plant disease. Annual Review of Phj-to- patholog^- 11:342-364. Crist. C. R.. and D. F. Schoeneweiss. 1975. The influence of controlled stresses on suscepti- bility of European white birch stems to attack bv Bolr\osphaena dothidea. Phvtopathology 65:369-373. Curl, E. A. 1955. Natural availability ofoak wilt inocula. Illinois Natural History Survev Bul- letin 27:277-323. Hefting, G. H. 1971. Diseases of forest and shade trees of the United States. U. S. Department of Agiiculture Forest Service Agriculture Handbook 386. September 1985 125 Years of Biological Research 237 Herendeen, R. a., and G. L. Rolfe. 1983. For- estry in Illinois. Illinois Research 25(l):3-5. University of Illinois Agricultural Experi- ment Station, Urbana. HiMELiCK, E. B. 1958. The oak wilt situation in Illinois - 1957. Proceedings of the Illinois Technical Forestry Association 1:5. 1962. Sycamore anthracnose. Pages 136-143 m R E. Tilford, ed., 37th National Shade Ti-ee Conference Proceedings. 1969. Ti'ee and shrub hosts of Verticillium albo-atrum. Illinois Natural History Survey Biological Notes 66. . 1972. High pressure injection of chemi- cals into trees. Arborist's News 37(9):97-104, 106-107. 1976. Disease stresses of urban trees. Pages 113-125 in Better Ti-ees for Metro- politan Landscapes Symposium Proceedings. U. S. Department of Agi-iculture, Forest Serv- ice General Tfechnical Report NE-22. National Arboretum, Washington, D.C. . 1981. Ti-ee and shrub transplanting manual. International Society of Arboricul- ture, Urbana, IL. . 1982. Pine blue-stain disease associated with the pine wilt syndrome. Journal of Arboriculture 8(8):212-216. , and D. W. Ceplecha. 1976. Dutch elm disease eradication by pruning. Journal of Arboriculture 2(5 ):8 1-84. and E. A. Curl. 1955. Experimental transmission of the oak wilt fungus by caged squirrels. Phytopathology 45(ll):581-584. , and , 1958. Transmission of Ceratocystis fagaceaarum by insects and mites. Plant Disease Reporter 42:538-545. , and D. R. Duncan. 1982. Control of syca- more anthracnose with injections of Arbotect 20-S. Results of 1981 Fungicide-Nematicide Tests, American Phytopathological Society 37:143. , and Howard Fox. 1961. Experimental studies on control of oak wilt disease. Univer- sity of Illinois Agricultural Experiment Sta- tion Bulletin 680. _, and Kirk Himelick. 1980. Systemic treat- ment for chlorotic trees. Journal of Arboricul- ture 6(71:192-196. and D. Neely. 1960. Juniper hosts of cedar-apple and cedar-hawthorne rust. Plant Disease Reporter 44(2):109-112. , and 1961. Pievention of bark bee- tle development in undesirable elms for the control of Dutch elm disease. Plant Disease Reporter 45(3): 180- 184. , and 1962. Root grafting of city- planted American elms. Plant Disease Reporter 46(2):86-87. , and 1965. Bioassay using cello- phane to detect fungistatic activity of com- pounds translocated through the vascular system of trees. Plant Disease Reporter 49(ll):949-953. Levitt, J. 1980a. Responses of Plants to Environ- mental Stresses. Vol. 1. Chilling, freezing and high temperature stresses. Academic Press, New York. . 1980b. Responses of Plants to Environ- mental Stresses. Vol. 2. Water, radiation, salt, and other stresses. Academic Pi-ess, New York. Matteoni, J. A., and D. Neely. 1979. Gnomonia leptostyla: growth, sporulation and hetero- thallism. Mycologia 71(5):1034-1042. McPartland, J. M., and D. F. Schoeneweiss. 1984. Hyphal morphology of Bolryosphaeria doth idea in vessels of unstressed and di-ought- stressed stems ofBetula alba. Phytopathology 74:358-362. Neely, D. 1967. Dutch elm disease in Illinois cities. Plant Disease Reporter 51(6):511-514. 1968. Twig inoculations on American elm with Ceratocystis iilmi. Phytopathology 58(11):1566-1570. 1973. Pin oak chlorosis - trunk implan- tations to correct iron deficiency. Journal of Forestry 71(6):340-342. . 1976. Sycamore anthracnose. Journal of Arboriculture 2(8):153-157. 1979. Etiology, epidemiology and control of black walnut anthracnose. Pages 58-61 in Walnut Insects and Diseases Workshop Pro- ceedings. U. S. Department of Agriculture, Forest Service General Technical Report NC-52. North Central Forest Experiment Station. 1981. Application of nitrogen fertilizers to control anthracnose of black walnut. Plant Disease 65(7):580-581. 1983. Tree trunk growth and wound closure. HortScience 18(1):99-100. 1984. Dutch elm disease control in Illi- nois municipalities. Plant Disease 68:302-303. _, and J. C. Carter. 1965. Species of elm on the University of Illinois campus resistant to Dutch elm disease. Plant Disease Reporter 49(6):552. , and W. R. Crowley, Jr. 1974. Toxicity of soil applied herbicides to shade trees. Hort- Science 9(2):147-149. _, and E. B. Himelick. 1963a. Tfemperature and sycamore anthracnose severity. Plant Disease Reporter 47(3):171-175. , and . 1963b. Aesculus species sus- ceptible to leaf blotch. Plant Disease Reporter 47(3):170. , and . 1965. Effectiveness ofVapam in preventing root gi'aft transmission of the Dutch elm disease fungus. Plant Disease Reporter 49(21:106-108. , and 1966. Fertilizing and watering trees. Illinois Natural History Survey Cir- cular 52. , and . 1967. Characteristics and nomenclature of the oak anthracnose fungus. Phytopathology 57(111:1230-1236. , , and W. R. Crowley, Jr. 1970. Fer- tilization of established trees: a report of field studies. Illinois Natural History Survey Bulle- tin 30(41:235-266. 238 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Parker, J. 196.5. Physiological diseases of trees and shrubs. Advancing Frontiers of Plant Sci- ence 12:97-248. Schoeneweiss, D. F. 1965. Fusicoccum canker of mountain ash in Illinois. Plant Disease Re- porter 49:251-252. . 1966. Cytospora canker on thornless honeylocust trees. Plant Disease Reporter 50:13-14. 1967. Susceptibility of weakened cotton- wood stems to fungi associated with black- stem. Plant Disease Reporter 51:933-935. 1969. Susceptibility of evergreen hosts to the juniper blight fungus under epidemic con- ditions. Journal of the American Society for Horticulture Science 94:609-611. 1973. Correction of lime-induced chlorosis of pin oak by liquid soil injection. HortScience 8:333-334." . 1974. Tubercularia ulmea canker of tallhedge: influence of freezing stress on dis- ease susceptibility. Plant Disease Reporter 58:937-941. 1975a. A method for controlling plant water potentials for studies on the influence of water stress on disease susceptibility. Can- adian Journal of Botany 53:647-653. . 1975b. Predisposition, stress, and plant disease. Annual Review of Phytopathology 13:193-211. 1977. Freezing stress predisposes Euony- mus alatus to attach by Nectria cinnabarina. Plant Disease Reporter 61:921-925. . 1978. Water stress as a predisposing fac- tor in plant disease. Pages 61-99 in T. T. Kozlowski, ed.. Water deficits and plant gi-owth. Vol. 5. Academic Press, New York. . 1981. The role of environmental stress . 1983. Drought predisposition to Cjlo- spora canker in blue spruce. Plant Disease 67:383-385. Stewart, C. A. 1980. Illinois forest resources op- portunity for total management. Illinois Institute of Natural Resoiirces Document 80/20A. Stim, J. A., and E. B. Himelick. 1981. Honey- locust decline in urban areas. Phytopathology 71:906. (Abstract) U. S. Department of Commerce, Bureau of the Census. 1981. 1978 Census of Agriculture. State and County Data. Illinois 1(13):6. Wallace, A. 1971. Regulation of the micro- nutrient status of plants by chelating agents and other factors. Arthur Wallace, Los Ange- les, CA. Watson, G. W, and E. B. Himelick. 1982a. Root distribution of nursei-y trees and its relation- ship to transplanting success. Joirrnal of Arboriculture 8(91:225-229. , and . 1982b. Seasonal variation in root regeneration of transplanted trees. Jour- nal of Arboriculture 8(121:305-310. , and . 1983. Root regeneration of shade trees following transplanting. Journal of Arboriculture l(2):50-52. Wene, E. G. 1979. Stress predisposition of woody plants to Botryosphaeria doth idea stem cankei-. Ph.D. Thesis. University of Illinois, Urbana-Champaign. , and D. F. Schoeneweiss. 1980. Localized freezing predisposition to Botryosphaeria canker in differentially frozen woody stems. Canadian Journal of Botany 58:1455-1458. Wheeler, Harry. 1975. Plant pathogenesis. Springer-Verlag, New York. in diseases of woody plants. Plant Disease 65:308-314. Morning Session Summary Catherine E. Eastman and Marcos Kogan The Illinois Natural History Sur- vey, as envisioned by its first chief, Dr. Stephen A. Forbes, was to concern itself not merely with the cataloging of organ- isms and their distribution in Illinois, but also with the study of the relation- ships of organisms with their environ- ment (Mills et al. 1958; Hays 1980). The presentations in this symposium today indicate that such studies are still a primary focus of Survey research in keeping with the finest work of our sister biological research institutions and universities. The environment within which or- ganisms must exist has changed con- siderably since the early days of the Survey. For example, one has only to look at the landscape of central Illinois to understand the impact that the needs of an increasing human popula- tion have had on our own environment and that of other organisms. With the change from marshy prairie and for- ested hills to drained, plowed farmland well under way in the 1850's, the di- verse vegetation well adapted to a prairie environment was replaced by an overwhelming dominance of two plant species — corn and wheat. Such is the case today, although soybeans have replaced wheat in acreage. Mono- culture techniques and plant breeding progi-ams have changed the chemical and physical defenses of plants, and soil, air, and water pollutants have made plants more vulnerable to attack by herbivores and disease agents. Pesti- cides applied to supplement natural plant defenses ai'e yet another environ- mental element affecting target and nontarget organisms alike. The growth Dr. Catherine E. Eastman is an Associate Entomologist and Dr. Marcos Kogan is an Ento- mologist, Section of Economic Entomology, Illi- nois Natural History Survey. of cities and industry has affected the environment as well by displacing the original plant cover and by releasing enormous amounts of industrial and urban wastes. Thus, we as biological researchers must concern ourselves with problems associated with in- herently unstable agroecosystems, with the reduction of suitable habitats and the consequent decline of wildlife pop- ulations, and with the difficulties in determining basic biological relation- ships among organisms in an environ- ment increasingly subject to human manipulation. The presentations in this sympo- sium are examples of the challenges presented to researchers investigating complex biological systems. In a broad sense most of the talks this morning have been concerned with stresses af- fecting plants and plant defense biol- ogy. Such stresses include herbivore injury, infections by disease agents, competition with other plant species, and a range of abiotic factors, among them temperature and moisture ex- tremes. Stress from one source, such as frost damage or insect feeding, may make the plant more vulnerable to other stresses, such as invasion by pathogenic organisms. It is in response to the net selective influences that are exerted by these stresses that plant defenses are evolved. In the last 25 years the defensive prop- erties of plant production of secondary metabolites have received increasing attention. The roles that these com- pounds may play in the producer plant's environment may be quite com- plex (Rosenthal & Janzen 1979). For ex- ample, flavonoids may serve both to screen plants from ultraviolet radia- tion and to attract pollinators. Phyto- alexins presumed to be produced by plants in response to invasion by bac- 239 240 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 teria or fungi have also been shown in recent work by Survey researchers to deter feeding by the Mexican bean beetle. And some metabolites, such as the cucurbitacins, which are repellent or outright toxic to many herbivores, may also serve as feeding excitants for other phytophagous species. Yet despite the importance of this line of research, a plant's particular chemical profile alone cannot explain plant-herbivore interactions. The microclimate and the types of carnivores present in the plant's environment also contribute to the degree of suitability of a plant or plant part as a host for a particular herbivore. Pesticides are superimposed on plant defenses when such defenses are deemed inadequate for crop production goals. Their effect in the environment ripples out from the immediate target pest on the host plant to include non- target organisms and for varying dis- tances in the host plant's abiotic en- vironment. Yet the stability of these pesticides often is dependent on that same environment, which determines whether the pesticide will be subject to microbial or ultraviolet degradation or whether it will persist for months or years in a biologically active form. That astute observer of scientific endeavor and human nature, Mark Twain, is reported to have said that "Researchers have already cast much darkness on the subject, and if they continue their investigations, we shall soon know nothing at all about it." The pursuit of the principles underlying biological relationships often seems to uncover more questions than answers. But those who are both charged and blessed with the task of conducting such investigations can seek comfort in the knowledge that each attempt sheds a little more light on the nature of these i-elationships and of life itself. LITERATURE CITED Hays, R. G. 1980. State science in Illinois: the scientific surveys. 1850-1978. Southern Illi- nois Universitv Pi'ess, Carbondale. Mills, H. B., G. C. Decker, H. H. Ross. J. C. Carter, G. W. Bennett, T. G. Scott. J. a Avars. R. R. Warrick, and B. B. East. 1958. A century of biological research. Illinois Natural History Survey Bulletin 27:85-234. Rosenthal. G. A., and D. H. Janzen. eds. 1979. Herbivores: Their interaction with secondary plant metabolites. Academic Pi-ess, New York. Long-term Biological Research In a Rapidly Changing Environment Lorin I. Nevling, Jr. This is a great weekend in Illinois for those interested in plants and ani- mals. This morning the Field Museum opened Plants ofthe World, the largest permanent exhibition of plant models ever presented. At this time, just a few miles from here, the Board of Trustees of the Pacific Tropical Botanical Gar- den sits in session. Tomorrow, the Chi- cago Botanic Garden in Glencoe will dedicate a magnificent bronze statue of Linnaeus, perhaps the single most in- fluential person in the history of sys- tematic biology. The most significant event, however, is that in which we have the honor to participate - recog- nition of 125 years of service by the Illi- nois Natural History Survey to the peo- ple of Illinois. As with many great institutions, the Survey had no precise beginning but evolved through time. Clearly, its genesis began with the proposal that a Natural History Society of Illinois be organized at the Illinois State Normal University. In June of 1858, the society was formed, its constitution providing that specimens be collected and be de- posited in the museum of the univer- sity. In due course, a state charter was issued, and in 1871, as a quid pro quo for financial assistance, the society relinquished ownership of the museum to the state. Subsequently, the Illinois Museum of Natural History at Normal was rechristened the Illinois State Lab- oratory of Natural History. The Direc- Dr. Lorin I. Nevling, Jr., is Director of the Field Museum of Natural History, Chicago, Illi- nois, and serves as a member of the Board of Natural Resoui-ces and Conservation, the govern- ing board of the Illinois Natural History Survey, Dr Nevling presented this address at the 12.5th anniversary celebration luncheon. tor of the State Laboratory, Stephen A. Forbes, who also served as State Ento- mologist, received a professional ap- pointment to the Illinois Industrial University (subsequently the Univer- sity of Illinois), where we are today. The Laboratory moved with him, as did the Office of the State Entomologist. The two offices remained separate until 1917, when the General Assembly wisely combined them into the Illinois Natui'al History Survey. The word "sur- vey" clearly implies census, but Forbes, as first Chief of the Survey, interpreted it as the relationship between living organisms and the environment, a for- ward-looking ecological concept that has directed the Survey to the present day and continues to set the course for the future. Clearly, during its long history in- numerable persons have made out- standing efforts on behalf of the citi- zens of Illinois through the medium of the Illinois Natural History Survey. Some have devoted entire professional lives to the Survey. Much of their effort has been obscured by time although a legacy remains in the form of publica- tions, collections, and accumulated data. As a unit of state government subject to the whims of political change, the Survey has drifted through an array of reporting structures, most of which were not especially helpful in furthering its activities. Some of these arrangements were, in fact, hostile. A major organizational break- through was made when the Illinois Department of Energy and Natural Resources was formed. Several govern- mental units, including the Illinois Natural History Survey, with inter- related interests and goals were assem- bled under this department. 241 242 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Perhaps for the first time in its long history, the Survey received positive support from the Director to whom it reported. Former Director Beal and current Director Witte must know that we appreciate their support, financial to the extent possible, but equally im- portant, their active interest in the pro- grams of the Survey. Neither has been passive, and both have insisted on pro- grams based on efficiency and effec- tiveness. This insistence has caused internal examination and, in some in- stances, the development of new and exciting directions. The responsibilities of the Survey are founded in law and include: • to maintain a biological survey; • to study the biological processes of ecological systems; • to acquire, organize, and utilize in- formation about the entire fields of entomology, zoology, and botany; • to recognize needs for, and provide recommendations about, the protec- tion, development, and use of biolog- ical resources, including agriculture; • to publish and furnish information about biological resources and their enjoyment, use, and management ac- cording to scientific principles. Clearly, to respond to such a com- plex set of activities a diverse, dedi- cated, thoroughly professional staff is required. Today, the staff numbers about 200. The Survey is organized administratively into five scientific research sections so as to respond effi- ciently to the various demands. These sections are Aquatic Biology, Botany and Plant Pathology, Economic Ento- mology, Faunistic Surveys and Insect Identification, and Wildlife Research. Laboratory and field facilities are scat- tered appropriately throughout the state. Each section relies heavily on the collections and accumulated data of the Survey. The collections, in toto, number about 6,000.000 items and are the larg- est and most comprehensive for the state. The larger segments of the collec- tions are: 1. Insects 4,875,000 2. Fish 500,000 3. Plants 210,000 4. A specialized library of 34,000 bound volumes. Specimens do not of themselves constitute a collection — they must be organized in some fashion, preserved, and maintained. The Survey's respon- sibility extends to the use of collections through research. The term, research, can mean a multiplicity of kinds of effort which, for the sake of simplicity, can be thought of in the case of the Survey as being of two nonmutually exclusive types: basic research and applied research. The Survey is actively engaged in both. When the term, research, is used in the scientific community, basic I'esearch is almost always implied. Basic reseai'ch stresses original scholarship, i.e., it adds to the base of human knowledge. Typically, the questions that are to be answered are posed by the individual researcher, and it is this ability which makes the Survey scientist most akin to a university research professor. The Survey makes a heavy investment in basic research because our future prob- lem-solving capability will depend sig- nificantly on the bedrock of knowledge provided by basic research. For example, is it important to maintain a large insect collection and research progi-am? Up to 3,000 animal species, mostly insects, are occasionally or potentially harmful. About once every 3 years, a new insect pest of major importance enters Illinois. With- out comprehensive collections and trained personnel, prompt identifica- tion would not be possible, and prompt identification /s one of the keys to effec- tive control. Some of the i-ecent pests which have found Illinois to be a land of opportunity are the western corn rootworm, gypsy moth, crucifer weevil, and a Mexican bean beetle that feeds on soybeans. Some other basic research pro- gi'ams of the Survey are the fioristic and faunistic surveys which have made Illinois one of the biologically best September 1985 125 Years of Biological Research 243 known areas in the world. The Survey's data base is unquestionably the finest of any state. One of the benefits of the Survey's long existence is the oppor- tunity to repeat surveys at intervals and, thus, to monitor changes that have occurred in populations and nat- ural habitats within the state. Eventually and irreversibly, we will find it necessai-y to undertake a biolog- ical sui'vey of the United States. When that day comes, Illinois will provide the keystone leadership for the nation. Two examples of the importance of resurveys, both from ornithology, show the value of such activities. Beginning with classical work in 1885, The Orni- thology ofIllinois, published in two vol- umes in 1889 and 1895, the Survey has produced a series of bird studies une- qualed in any other state. In 1906, the Survey began a series of statewide cen- suses of birds in Illinois. These surveys were repeated 1956-1958 and found about the same number of breeding birds in Illinois. But in 1909, 18 species constituted 70 percent of the breeding birds, whereas in 1957 only 9 species constituted 70 percent. These changes were brought about by habitat changes, and these changes are still occurring rapidly. Canada geese were first studied by the Survey in 1941, and these studies continue today. As a result, we now understand the condition factors of the birds and their relation to seasonal stress, management techniques, and various diseases and parasites. In 1962, a Survey scientist rediscovered the then-presumed-extinct giant Canada goose and provided the rationale neces- sary to reestablish this subspecies over much of its former range. Research on the soil insecticides aldrin and heptachlor conclusively demonstrated that the epoxides of these insecticides are translocated from the soil and are stored in the oil of soybeans and in the oils and waxes of pumpkins. Aldrin and heptachlor were widely used as soil insecticides in the 1950's and 1960's, and the Survey's research has had a tremendous impact nationally and internationally in con- trolling the unwarranted use of these chemicals. Other ecological research includes the study of ponds, streams, rivers, and more specialized bodies of water, such as reservoirs and cooling lakes. In spec- ialty areas, there are, for example, studies offish larvae, genetics, and be- havior; insect behavior, including over- wintering, migration, and dispersal; as well as insect-host interactions. Applied research is the application of basic research findings to a problem to develop a resulting product or an- swer The questions are usually thrust on, or assigned to, a researcher, and it is the kind of research most often car- ried on in the corporate world under the title, research and development. As the needs of the people of Illinois drive many research questions, much of the Survey's work is of this nature and en- compasses many diverse elements, in- cluding: Studies carried out under a grant or contract for governmental agen- cies or business firms when their needs fit Survey interests and ex- pertise. Usually, these studies are designed to answer specific man- agement questions. Major plant-related activities, in- cluding the identification of plants and plant diseases for the public and industry. Two examples are the discovery of the causal agent in brittleroot disease of horseradish and its transmission by the beet leafhopper, and the concept — now well accepted in corn-producing states - that black cutworm larvae that damage young corn plants in the spring come from eggs oviposited on the weeds in a field before the corn is planted. Effects of soil and air pollutants on plant growth and development and crop yield and quality. Studies on the prevention and conti-ol of diseases of greenhouse plants, woody ornamentals, and 244 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 shade trees. Specific problems, such as oak wilt and walnut and syca- more anthracnose, are being ad- dressed. In zoology some projects are: Identification services for the public and industry, management plans for fish, aquaculture, or the best use of wildlife resources for the preservation of the prairie chicken. Development of insect pest man- agement programs by integrating cultural, biological, and chemical controls into crop production sys- tems. A marvelous achievement was the publication of the manual. Soybean Insect Identification and Management in Illinois. Studies of the transmission of dis- eases and parasites between wild and domestic animals and between wild animals and man. For years, Survey scientists have carried on interdisciplinary research. This is the end of the rainbow for most research institutions. At the Survey, the rainbow is just down the hall or up the stairs. A particularly impressive piece of interdisciplinary long-term research is being continued on the Illinois River system. From the first years of its for- mation, the Survey has worked on this river, relating changes in water quality and river use to changes in the plank- ton, benthos, and fishes. Diversion of Lake Michigan waters through the sys- tem, channelizations with locks and dams for navigation, enrichment by sanitary wastes from Chicago, and draining of the floodplain lakes for agriculture have caused dramatic changes in the aquatic communities. The Survey's investigations of these changes have attracted an impressive array of scientists, have made the Illi- nois one of the worlds best-studied rivers, and recently have gained na- tional support to intensify the investi- gations for another two decades under the National Science Foundations Long- term Ecological Research program. No matter what the nature of the research carried out by the Survey, the results must be made available to the appropriate publics. This publication of results is accomplished in a variety of scholarly journals, books, monogi-aphs, brochures, and special reports. In the last year alone, staff activi- ties have resulted in five books, 17 book chapters, 98 scientific journal articles, 85 project reports, and 94 papers pre- sented at professional meetings. Newsletters and press releases are sent weekly to newspapers and radio and television stations throughout the crop season to keep the public abreast of developing problems. There are pro- gi-ams and workshops for school chil- dren and specialized groups, such as the professional pest consulting indus- try. Every person in the state is touched directly or indirectly by Survey re- search results, such as recommenda- tions concerning the protection of crops from crop pests, on reducing disease in ornamental plants, and for the man- agement of fish and wildlife popula- tions. Survey scientists also attend the meetings of and are active in the af- fairs of national and international scientific organizations. In short, the Survey is everywhere and has developed substantial local, national, and inter- national reputations. Previously, I mentioned the posi- tive step of placing the Survey under the Department of Energy and Natm-al Resources. We were placed there along with our sister organizations, the Illi- nois State Water Survey and the Illi- nois State Geological Survey. A history of cooperative efforts among the three Surveys existed, but cooperative efforts usually were on a scientist-to-scientist rather than Survey-to-Survey basis. Within recent years, a series of changes in the leadership of the Depart- ment and the three Surveys has oc- curred. The principals were Michael B. Witte, Director: Robert E. Bergstrom, Acting Chief of the Geological Survey; Stanley A. Changnon, Chief of the Water Survey: and Paul G. Risser, Chief of the Natural History Survey. September 1985 125 Years of Biological Research 245 The chorus was comprised of the mem- bers of the Illinois Board of Natural Resources and Conservation. In my opinion, a synergistic reaction has taken place among the parties in such a way that the Surveys are changing more than at any time in their com- bined histories. Change is a frighten- ing thing to all of us, but we are convinced that the opportunities are unlimited. Stimulated by Director Witte and the Board, the Surveys presented pro- posals for interdisciplinary research that were inter-Survey in nature. These proposals were innovative and had far- reaching consequences. Only a few could be funded, but the effort has pro- duced a new spirit of cooperative re- search among the Surveys. What are the demands on the Sur- vey likely to be in the future? A few can be predicted. We will: Monitor the health of species pop- ulations and ecosystems. We will recognize significant changes, sep- arating long-term trends from short-term fluctuations, especially for species population numbers and distribution as well as for ecosys- tem characteristics, such as pri- mary production and nutrient cy- cling and retention. Respond to specific problems as these issues arise, but also main- tain a steady basic research pro- gram driven by scientific questions. Draw together diverse disciplines to address, in a coherent fashion, complex topical problems, such as hazardous waste management and wildlife populations in landscapes continually managed by man, and organize information for decision- making that combines ecological with economic considerations. Develop the ability not only to describe the natural resources of the state, but to predict the conse- quences of alternate management schemes and possible demographic and energy scenarios. Provide a central location where one can obtain information about natural resources or receive guid- ance as to the location of natural- resource data and information. Maintain a competitive edge in the quality of natural-resource research, not only to solve our prob- lems effectively and manage our resources wisely, but to continue to attract and retain the very best scientific staff. Finally, we will need to move ful- ly into the computer age so that the incredible data locked into collec- tions can be fully utilized. It seems to me that this will be the only way to maintain open-end excellence with closed-end funding. Our im- mense data base, if united with those of our sister Surveys of Water and Geology, will form the most po- tent environmental data base avail- able anywhere. I am certain that Stephen Forbes would be proud of all those, past and present, who have contributed their talents to bring the Illinois Natural History Survey closer to the idea of understanding the relationships among living organisms and the environment. No one can tell what the most impor- tant problems of society will eventually turn out to be, but we can be sure that many of them will be tied to a deeper understanding of ourselves and our environment. Lake Tahoe: A Microcosm for the Study of Change Charles R. Goldman The concept of "The Lake as a Microcosm" was first proposed by Dr Stephen Alfred Forbes, first Chief of the Illinois Natural History Survey. This work (Forbes 1925), presented be- fore the Peoria, Illinois Scientific Asso- ciation in 1887, has become a classic among ecologists and is often cited as an example of the interdependence of organisms within a lake. In developing this theme to its maximum, Forbes suggested that the organisms of Illinois lakes were largely independent of their terrestrial surroundings. It can be con- vincingly argued that in 1887 the plants and animals inhabiting Illinois lakes were in fact a lot more indepen- dent of their surroundings than they now are. It is clear, however, from reading his monumental work, The Fishes ofIllinois (Forbes & Richardson 1908), that he was aware of the sewage pollution from Chicago entering the Illinois River system. Today, man is having a dramatic effect on most of the world's watersheds, which in turn play an increasingly dominant role in estab- lishing the physical characteristics of lakes. These characteristics include color and transparency as well as such biologically mediated factors as fertil- ity and trophic status. Eighteen years before Forbes' arti- cle appeared in the United States, S. A. Forel (1869) was founding the science of limnology at Lausanne, Switzerland, on the shores of beautiful Lake Geneva. His brilliant ti'eatise, Monographie Limnologique, provided a bench mark for the study of the drastic changes that were to alter this lake as well as the Lake of Zurich during the century Dr. Charles R. Goldman is Director of the Lake Tkhoe Research Group, Division of Environ- mental Studies, University of California, Davis. to follow (Forel 1892, 1895). Immediate- ly following the 1983 International Congress of Limnology in Lyon, France, an entire week was devoted to assess- ing the status of Lake Geneva and at- tempting to develop a strategy to bring nutrient loading from both Swiss and French sources under control (Fig. 1). Perhaps by coincidence or as a result of scientific information exchange, John Le Conte (1883a, b, 1884) mea- sured the transparency of Lake Tahoe near the crest of the Sierra Nevada in 1873. Le Conte's observations of Lake Tahoe appeared in a regional stage- coach journal, The Overland Monthly. Lake Tahoe, bordered by two states rather than two countries, has been successful in achieving total sewage diversion from its basin, yet continues to decline in water quality from an array of nonpoint nutrient sources (Fig. 2). Disturbance of the Tkhoe basin has been very recent in the geologic sense of time. The first major perturbation occurred about 1846 when most of the mountainsides surrounding the lake were cleared of virgin timber to be used for shoring up the mines of the Com- stock in Nevada. The second disturb- ance came during the years immediate- ly followng World War II after new growth had reestablished the forest. This period was marked by rapid human population growth (Fig. 3) and a great deal of road and housing con- struction. Despite the efforts of conser- vation groups and state and federal agencies to regulate development, growth continues to this day. As the tourist and resident population of the basin increased, from about 10,000 in 1955 to nearly 80,000 in 1979, real estate prices rose rapidly, and the year- round use of the basin has continued to increase. With this gi'owth, the 247 248 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Fig. 1. - Lake Geneva, nestled in the Alps, is bordered by both Switzerland and France It has suffered from heavy nutrient loading from both countries for more than a century. Photo by the author. Fig. 2. - Looking southeast across Lake Thhoe from 19,000 feet (5.800 ml. The enormous volume of the lake has served to buffer the nutrient input from it.-* disturbed watershed. Still, the primary productivity has increased at the rate of 6.3 percent per year Photo by Robert C. Richards. September 1985 125 Years of Biological Research 249 LAKE TAHOE BASIN POPULATION GROWTH Fig. 3. - Population gi-owth in the Lake Tahoe basin from about 1950 through 1977. The population has been divided into winter, summer, and permanent residents. The sum of both sea- sonal and year-round residents is presented as total residents. deterioration of Lake Tahoe, like that of many of the world's lakes, has accel- erated. This paper continues the docu- mentation of the changing lake and attempts to isolate some of the major factors involved. Through a better understanding of whole-system re- sponse, it should be possible to make more intelligent management deci- sions and eventually stabilize the fer- tility of lakes at some acceptable level. EUTROPHICATION OF LAKES Naumann (1919) used the concept of trophic status to distinguish between the oligotrophic mountain lakes and the more productive eutrophic lakes of the lowlands. Many of the world's lakes, like Tahoe, have experienced a steady increase in fertility as their watersheds have been developed for agriculture, silviculture, or urban communities. Documentation of the change has often included a variety of chemical mea- sures of fertility, such as nitrogen and phosphorus concentrations, changes in the species composition and abundance of algae, and decreases in hypolimnetic oxygen concentration during stratifica- tion. Observations of oxygen depletion under winter ice cover or following algal blooms were first made possible by the well-known Winkler technique for measuring dissolved oxygen. The measurement of primary productivity has provided a highly sensitive integra- tion of chemical, physical, and biolog- ical conditions in lakes (Goldman 1963; Goldman & Wetzel 1963). Comparison of the contrasting vertical profiles of in situ measures of primary productivity from lakes of different trophic status has previously been presented (Gold- man 1968). These profiles show how the euphotic zone is compressed as algal growth shades out the deeper water. Perhaps the most publicized exam- ple of eutrophication in North America has been Lake Erie, which, during the last several decades, has undergone a dramatic increase in fertility and gen- eral deterioration in water quality. This eutrophication resulted from a combination of urban and agricultural discharge and caused serious seasonal oxygen depletion in much of the lake. Efforts to reduce nutrient loading ap- pear to be meeting with success, and improved conditions have been re- ported in recent yeai-s. Because water levels have also been higher than average in the Great Lakes in recent years, simple flushing of the system may also be a factor in their improve- ment. It will be important to follow their trophic status through different hydrologic conditions in the years ahead. In addition to Erie, others of the Great Lakes have also increased in fer- tility, with a concomitant change in the fisheries (Beeton 1969). The invasion of the sea lamprey and alewife had dra- matic effects on Great Lakes fisheries, but industrial and domestic pollution also have caused great concern. Despite continuing public and governmental agency concern about heavy metals and industrial organics accumulating in fisheries products, there is a general impression that control measures are currently returning high dividends for the investments in pollution control 250 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 and fisheries management. There re- mains, however, the need for cautious concern. Although increased fertility doubtlessly provides the potential for an increase in the standing crop of fishes, large populations of exotic pred- atory salmon may begin to exert a neg- ative impact on their food resources. The alewife population, in particular, has undergone dramatic reductions, and other fishes may be subject to simi- lar predation. Further, the large reduc- tion in alewives may be expected to have influenced zooplankton populations. Perhaps the first example of eutro- phication (Ruttner 1963) was the Zur- ichsee in Switzerland, which became eutrophic in 1896 and showed a dif- ference in fertility between its upper and the lower basins. In what was prob- ably the first attempt at lake restora- tion, sewage treatment was initiated and has been credited with greatly im- proving the condition of this lake. Another large Swiss lake, the Bielersee, became eutrophic following the diversion of the river Aare into the lake and the concomitant increase in nutrient loadings after 1930 (Tschumi et al. 1982). Despite phosphorus reduc- tions of more than 50 percent, Bieler- see has shown little improvement. Nitrogen loading remains high from a drainage area that has been increased by 2.6 times. It is apparent that the phosphorus levels are still sufficient to sustain eutrophic conditions at the cur- rent level of nitrogen input. Even Siberia's Lake Baikal, the world's oldest and deepest freshwater body, has not been immune to the im- pact of man's activities (Fig. 4). The harvest of the slowly gi'owing, mixed deciduous and coniferous forests along its shores and the discharge from a large cellulose plant and other indus- tries along the Selenga River were re- sponsible for catalyzing what was cer- tainly the most publicized, if not the first, environmental movement in the Soviet Union. Fortunately, the lake's enormous volume (23,000 km^) provides ^«l Fig, 4. - Lake Baikal, USSR, is the oldest and deepest lake in the world. Despite its enormous volume, it has been subject to pollution from forest cutting and industrial development. The 500-ton limnological research vessel Vereschagin lies at anchor. Photo by the author. September 1985 125 Years of Biological Research 251 tremendous dilution of pollutants (Goldman 1973). The largest lake in Europe, Ladoga, near Leningrad, is also subject to pulp mill pollution as well as other industrial contamination, and progi-essive eutrophication has been reported. Although many ofthe world's major lakes have experienced varying degrees of eutrophication, detailed studies of change at various levels of the food chain have usually been lacking. One long-term data set developed by Lund (1964), which featured the annual sili- con depletion cycle of diatoms in Lake Windermei-e, England, is noteworthy although the lake does not appear to have changed much during the 16-year record. The public is usually first alerted to pollution by the appearance of dead fish along the shore or by the develop- ment of large blooms of blue-gi-een algae, which can give a once-blue lake the green appearance of a lawn. An unusually warm summer, drought, or the application of commercial fertili- zers to the watershed may trigger a severe algal bloom and sudden public awareness that a problem exists. It is likely, however, that the problem of excessive nutrients has been develop- ing over a considerable time. When the change occurs over many years, only the keenest observers notice it, and without long-term collection of good data, change is difficult or impossible to prove. In the western United States the most detailed studies of the progi-ess of eutrophication have been done for Lake Washington by Professor W T. Ed- mondson and for the earliest stages of change in Lake Tahoe, by the author. The Lake Washington studies used measures of chlorophyll a concentra- tion as evidence for the increasing fer- tility of the system, while the Tahoe studies measure the change in the rate of photosynthesis over time. Concentrations of chlorophyll a in Lake Washington increased steadily with the algal concentration as sewage loading increased between 1950 and 1963. The blue-green alga, Oscillatoria rubescens de Candolle, became the dominant form in Lake Washington, as it had in Lake Zurich a century earlier. With the collection of sewage from around the lake perimeter and its diver- sion to Puget Sound, beginning in 1963, transparency returned as chlorophyll values gi'adually declined during the next decade (Edmondson 1972a, b). With the exception of these last two studies and a long series of fisheries records and chemical measurements from the Great Lakes, there is, in gen- eral, a lack of long-term data sets on the limnological changes that have oc- curred in North American lakes exper- iencing eutrophication. The value of long-term studies in better understand- ing how aquatic ecosystems function is now becoming generally recognized (Goldman & Home 1983; Likens 1983; Goldman & de Amezaga in press). The Lake Tahoe data set is the basis for the following analysis of the first stages of eutrophication in an ultraoligotrophic lake. LAKE TAHOE Lake Tahoe was known only to the Paiute Indians until its discovery by General Fremont in 1844. Mark Twain in Roughing It described Lake Tkhoe as "surely the fairest sight the whole world affords" and marveled at its clari- ty. In addition to Mark Twain's obser- vations and the historical records on the Tahoe fishery, probably the most important single data point was that of transparency obtained by John Le Conte in September 1873. Without this simple measure of water clarity, taken with an ordinary white dinner plate serving as a Secchi disc, there would have been endless dispute as to the transparency of Lake Tahoe before de- velopment began in earnest. The lake is the tenth deepest in the world (505 m) and is uniquely situated for study of the early stages of eutro- phication. This ultraoligotrophic lake 252 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 has a small (800 km^) watershed for a lake with 500 km^ of surface area. Its enormous volume of 156 km^, with a retention time of about 700 years, pro- vides a large buffering effect for nutri- ents entering the lake from tributary streams draining the forests, roads, parking lots, golf courses, and building sites of its highly disturbed watershed. Due to the long residence time of the water most of the nutrients entering the lake remain there with little flushing action and only slow sedimen- tation to reduce them. In addition to stream-borne nutrients and sediment, air pollution is now visible, and storms bring acid rain to the basin as well as an important, but as yet inadequately measured, dry fallout (Fig. 5). Located at the crest of the Sierra Nevada, Lake Tahoe was first recog- nized as a graben fault basin by Le Conte (1875). It was probably formed during the upthrust of the mountain range between 3 and 9 million years ago and has a morphometry similar to that of a giant bathtub. The 450-m con- tour line of depth is very close to shore around most of its perimeter. Impor- tant to the lake's trophic status is its relatively infertile watershed consist- ing largely of decomposed granite de- rived from the great Sierra Nevada batholith as well as some Pi'ecretaceous metamorphic rocks, Pliocene volcanics, and a few cinder cones from the Holo- cene (Hyne et al. 1972). Ice damming of the Truckee River outlet of the lake occurred during the last period of Pleis- tocene alpine glaciation. At that time the water in the lake reached levels as much as 150 m higher than it is today. Because of Lake Tahoe's extremely low productivity, which, since 1959, has ranged from 0.12 to 0.26 g C m"^ day\ the more conventional measures of fertility, such as oxygen depletion, chlorophyll a, and change in chemical composition, have had little meaning. Oxygen levels have remained at or near saturation all the way to the bot- tom of the deepest part of the lake Fig. 5. - Air pollution is now frequently visible in the Tahoe basin during inversions. The smog layer is visible from Rubicon Point on the west side to Tahoe Keys at the south end. Photo by Robert C. Richards, February 1976. September 1985 125 Years of Biological Research 253 (505 m), and there has been no detect- able trend in the nutrient content of the surface waters from year to year. The increased nutrient loading has been so diluted and so quickly incor- porated into algal growth that annual increases in such important inorganic nutrients as nitrogen and phosphorus have not been detectable. However, a depletion of nitrate has been found in the surface waters during the main gi'owing season, resulting in a distinct "nitra cline" (Paerl et al. 1975). Follow- ing the pattern of vertical stratification of nitrate through the following fall and winter gave the first proof that waters in Tkhoe actually mixed all the way to the bottom. However, for prac- tical purposes this mixing occurs only during particularly stormy, cold win- ters. Deep mixing can provide a gi'eatly elevated nitrogen concentration for the spring growing season and is an impor- tant factor in determining year-to-year variations in productivity. When mix- ing occurs, internal nutrient loading from this deep nutrient storage is like- ly to be reinforced by runoff from the very storms that cause the mixing. The deep waters of the lake provide a large reserve of nitrate during years when winter mixing is incomplete (Goldman 1981). My data collection for Lake Tahoe began in 1958 and 1959 with the first measurements of primary productivity, using the sensitive carbon- 14 method (Steemann-Nielsen 1952) with the mod- ification of Goldman (1963). Ti-anspar- ency and an array of other limnological measures were also included as well as extensive bioassays of nutrient-limit- ing factors (Goldman 1964, 1981; Gold- man & Armstrong 1969; Arneson 1979). Intensive sampling began in 1967 and has continued to this time. The mea- surement of primary productivity pro- vides a sensitive integration of the physical, chemical, and biological fac- tors controlling photosynthesis by the free-floating phytoplankton component of this low-fertility environment. In ad- dition, the data have enabled us to show significant change long before conventional methods would have de- tected it. The primary productivity of the lake has more than doubled in a 20-year period at an average rate of in- crease of over 6 percent per year (Fig. 6). Peak years of productivity in this progression took place in 1975 and 1980. After those years of particularly high primary productivity, there fol- lowed 2 consecutive years of lowered productivity. If only a single lake had been studied, one might easily and er- roneously have concluded that weather or changing meteorological conditions, such as were caused by El Nino (Strub et al. in press) in the western United States, were responsible. But data col- lection using exactly the same technol- 1958 1962 1966 1970 1974 1978 1982 Fig. 6. - Annual primary productivity in Lake Tahoe, California-Nevada, between 1958 and 1982. Each year represents the integi-ation of weekly or biweekly measurements made using the Carbon-14 method at a series of depths e.xtending from the lake surface to 105 m. Approx- imately 35 sampling dates for each year are rep- resented in this figure. The curve represents a least-squares fit over this time span. ogy had been started at another subal- pine lake. Castle Lake, in 1959 and has continued without interruption for 25 years. This small (50 ha) lake, like Tahoe, has a very small watershed rel- ative to its surface area and shows nitrogen limitation. No upward trend in productivity is evident in this lake (Goldman & de Amezaga in press). In attempting to correlate the precipita- 254 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 tion at Lake Tkhoe with the annual pri- mary productivity, we discovered that there was no correlation if a direct regression was run. However, ifwe plot- ted the percentage of change from the previous year's productivity, a positive relationship emerged (Fig. 7). This find- ing suggests that the lake's "memory" of the previous year's productivity is more important in establishing the positive relationship with rainfall than is the irregular but upward trend in fertility evidenced by the productivity curve (Fig. 6). The relationship between rainfall and primary productivity in Lake Tkhoe, however, is somewhat more com- plicated than this explanation implies. Not only does rainfall influence the an- nual nutrient input to the system, but it is also related to storms which cause internal nutrient loading of the eu- photic zone and deep mixing of the nitrogen-enriched hypolimnion of the lake. Further, a rainy year, because of increased cloud cover, causes a reduc- tion in solar radiation and in temper- ature. In 1982, for example, high rain- fall as a result of El Nino was accom- panied by one of the darkest years on record at Lake Tahoe (Fig. 8). This darkness may have decreased the ex- pected productivity from nutrient run- off and contributed to the reduction of the average depth of the euphotic zone from the previous year by almost 2 m (Fig. 9). Therefore, 1982 did not fit the regression of more typical years. The high nutrient levels from the heavy runoff in 1982 coupled with deep mix- ing enabled the phytoplankton to use the decreased available light more effi- ciently in both Castle Lake and Lake Tahoe than they had in the previous year (Goldman & de Amezaga in press). Another evidence of the value of long-term data collection was appai'ent when it was discovered that the two cladoceran members of the endemic zooplankton population, Daphnia and Bosmina, had disappeared from the lake. The opossum shrimp, Mysis relicta loven, introduced as a food for the resi- dent lake and rainbow trout popula- tions as well as for the Kokanee sal- mon, emerged as the major suspect in the ensuing investigation (Richards et al. 1975; Goldman et al. 1979). Follow- ing a crash in the Mysis population, first Bosmina and then Daphnia have made weak but promising returns to the plankton of the lake (Fig. 10). Ob- servations of other western lake sys- tems indicate that Mysis will coexist -10% 40 80 120 160 PRECIPITATION ( cm woter year"') Fig. 7. - Percentage change in annual (1 Januai">'-31 December) primai-j' productivity in Lake Tkhoe is regi-essed on total precipitation for the corresponding "water year" (1 Octo- ber-30 September) for the period 1969-1982. Two unusual yeai-s. 1975 and 1982, have been ex- cluded from the regi-es- sion. (Pi-om Goldman & de Amezaga in press). September 1985 125 Years of Biological Research 255 with cladocerans ifthe fertility level is appreciably higher than Tahoe's and if warm water provides thermal refuges for the cladocerans. One way to extend the information time series for any lake study is to ex- amine lake sediments. Lakes are, after all, reservoirs of history in the sense that they usually store in their sedi- mentary records evidence of the events that have occurred on their watersheds. Erosional deposition from the building Annual Solar Radiation a: < 256 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 I%7|I968|I969|I970| 1971 |l972|l973|l974|l975|l976!l977|l978il979|l98C| 1981 |e82|l983 DAPHNIA Fig. 10 - Variations in populations oiDaph- nia, Mysis, Bosmina, Epischura, Diaptomus, and Kellicottia in Lake Tahoe between 1967 and 1983. Mysis values from the years 1968 to 1977 are based on the per- centages of lake trout containing Mysis dur- ing creel censuses by the California Depart- ment of Fish and Game. Beginning in 1975, My- sis values were deter- mined on the basis of a lakewide average (data through 1980 from Gold- man [1981]). of the Roman road, Via Cassia, about 2,000 years ago is evident from a sedi- ment core (Cowgill & Hutchinson 1970). Sediment from erosion, ash from vol- canic eruptions, as well as that portion of the biota which is resistant to decay may lie undisturbed in sediments for thousands of years. Further, pigments (Vallentyne 1960) and evidence of water chemistry may also be retained in the paleolimnological record. One of the arguments presented for the continued development of the Tkhoe basin was that the lumbering ac- tivities of the 1860's did not appear to have done serious damage to the water quality of Lake Tahoe. By examining sediments, we were able to demonstrate that damage done during the initial re- moval of timber from the basin was small in comparison to the disturbance that has accompanied modern develop- ments. Samples taken in midlake with a large box-coring device clearly show that much more nitrogen is now enter- ing the lake relative to carbon (Fig. 11 A). Fig. IIB shows the carbon con- tent of nearshore sediments in the vicinity of extensive lumbering activity during the 1860's. Although there are insufficient lead dates to confirm the 1860 level, it is indicated by estimated sedimentation rates. Similarly, the input of another important algal nutri- ent, iron, has gi-eatly increased with erosion from the steep slopes of the basin (Fig. 12). In addition, the diatom composition has shifted from centric forms to increasing numbers of pen- nates (Fig. 13). Finally, Pb-210 dating of the sediments indicates a more rapid rate of accumulation during the last several decades than occurred during the period of forest cutting in the 1860's (Byron & Goldman 1984). As we look back in time, so must we look forward. The future of Lake Tkhoe can be predicted with increasing cer- tainty as the years of data are added, one to aiiother What seemed to some to be speculation or doomsmanship in the 1960's has been transformed into highly significant regressions. Fmther, visible periphyton growth now covers the once clean littoral zone so that, in spring, any observer can see a green margin around 'Rihoe (Fig. 14). Prima- ry productivity correlates significantly September 1985 125 Years of Biological Research 257 Dry Weight of Carbon (g m~^ yr"') 20 40 60 80 Total Nitrogen/ Total Carbon 06 .07 08 .09 .10 .11 12 LAKE TAHOE ORGANIC CARBON IN GLENBROOK BAY SEDIMENT CORE 20- Q. a LAKE TAHOE SEDIMENT CORE NITROGEN TO CARBON RATIO Fig. 11. - (A) The dramatic increase in nitrogen deposited as sediment in Lake Tahoe is evident from a midlake core sample. The higher nitrogen-to-carbon ratio in recent years is obvious (Byron & Goldman unpublished). (B) High organic carbon input is evident from this neai-shore sediment core collected in Glenbrook Bay. This area was subjected to extensive timber clearing during the 1860's. Both charcoal and sawdust are to be found in the sediments here. at the 1 -percent level with the decline years (Fig. 15). With each year that the in transparency and, if projected into the future, predicts a lake of very ordi- nary transparency within about 40 productivity continues its rise, the Araphidinate Pennales/Centrales Biologically Available Iron ( mg/g DW Sediment| 10 20 30 40 50 60 '70 80 90 IRON COMPOSITION Fig. 12. - Distribution of biologically avail- able iron in a midlake core sample from Lake Tahoe. The rise at 20 cm probably represents the 1860 lumber removal, and the increase above 5 cm is probably due to recent development in the basin (Byron & Goldman unpublished). 258 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Fig. 14. - Periphyton covers the littoral zone of Lake Tkhoe diu-ing the spring months. This under- water photograph was taken near Sunnyside (northwest Tahoe) on 8 May 1982 at a depth of 4 m. When this growth breaks loose, it coats the beaches with a brown slime. Photo by Stanford Loeb. regression line steepens and the num- ber of future years with clear waters in "the lake in the sky" is reduced. The damaged watersheds must be repaired now through the application of innova- tive biological engineering and further development must be more strictly regulated. Otherwise, future genera- tions will look back and wonder why, with so much knowledge and technol- ogy available, we failed to preserve the quality of this remarkable lake micro- cosm whose exceptional transparency became a legend. ACKNOWLEDGMENTS This work, done over a period of 25 years, would not have been possible 1970 LAKE TAHOE ANNUAL SECCHI DEPTHS 1990 2010 2030 2050 10 JAN - DEC P < .001 Fig. 15. - The regi-es- sion constructed from the average annual Sec- chi depth in Lake TUioe from 1968 through 1982. The regi-ession line as extended is admittedly futuristic as it extends to a theoretical zero point, which occm-s some- time after the middle of the next century. With each additional data point, the regiession line steepens and the degi-ee of steepness pi\5- vides some insight into alternative futures. September 1985 125 Years of Biological Research 259 without the th-eless efforts of the Tkhoe Research Group. Robert C. Richards has been particularly invaluable in the collection of data during the last 17 years of this effort and has provided valuable photographic documentation. Data reduction and figure preparation were largely accomplished by Patricia Arneson and Evelyne de Amezaga. Word processing of the manuscript was kindly provided by Carol Barnes, Anne C. Forcella, and Meryllene Smith. George Malyj provided important assistance with various aspects of coor- dination. This paper is dedicated to the mem- oiy of my father, Marcus Selden Gold- man (1894-1984), who first interested me in the streams and lakes of Illinois and, for more than 50 years, enthusias- tically endorsed the varied activities of the Illinois Natural History Survey. He was a close friend of many of its staff, an enthusiastic amateur ichthyologist (Goldman 1977), a life member of the Izaak Walton League of America, and a staunch supporter of conservation ac- tivities everywhere. LITERATURE CITED Arneson, P. A. 1979. Effects of nutrient enrich- ment on the natural phytoplankton of Lake Tahoe, California-Nevada. M. S. Thesis. Uni- versity of California, Davis. Beeton, a. M. 1969. Changes in the environment and biota of the Great Lakes. Pages 150-187 in Eutrophication: causes, consequences, cor- rectives. National Academy of Sciences, Wash- ington, D. C. Byron, E. R., and P. Eloranta. In press. Recent historical changes in the diatom community of Lake Tahoe, California-Nevada, U. S. A. Internationale Vereinigung fur Theoretische und Angewandte Limnologie Verhandlungen. , and C. R. Goldman. 1984. Recent sedi- mentation and the fertility of Lake Tahoe. Tahoe Research Group, Institute of Ecology, University of California, Davis. Cowgill, U. M., and G. E. Hutchinson. 1970. Chemistry and mineralogy of the sediments and their source materials. In lanula: an account of the history and development of the Lago di Monterosi, Latium, Italy. American Philosophical Society Transactions 60:.37-101. Edmondson, W. T. 1972a. The present condition of Lake Washington. Internationale Vereini- gung fur Theoretische und Angewandte Limnologie Verhandlungen 18:284-291. 1972b. Nutrients and phytoplankton in Lake Washington. Pages 172-193 m G. E. Likens, ed.. Nutrients and Eutrophication. American Society of Limnology and Ocean- ography Special Symposia Vol. I. Forbes, S. A. 1925. The lake as a microcosm. Il- linois Natural History Survey Bulletin 15:537-550. (Originally read on 25 February 1887 before the Peoria Scientific Association and published in its Bulletin.] , and R. E. Richardson. 1908. The FLshes of Illinois. Illinois State Laboratory of Natural History. FoREL, S. A. 1869. Introduction a letude de la faune profonde du lac Leman. Societe Vaudoise des Sciences Naturelles Bulletin (Lausanne) 10:217. . 1892. Le Leman: monogiaphie limnolo- gique. Tome I. Geogi-aphie, hydrogi'aphie, geo- logie, climatologie, hydrologie. F. Rouge, Lausanne. (Reprinted 1969. Slatkine Reprints, Geneva.) . 1985. Le Leman: monographie limnolo- gique. Tome II. Mechanique, hydraulique, thermique, optique, acoustique, chimique. F. Rouge, Lausanne. (Reprinted 1969. Slatkine Reprints, Geneva.) Goldman, C. R. 1963. The measurement of pri- mary productivity and limiting factors in freshwater with Carbon- 14. Pages 103-113 in M. S. Doty, ed., Proceedings of the Conference on Primary Productivity Measurement, Marine and Freshwater. U. S. Atomic Energy Commission, Division of Technical Informa- tion Report TID-7633. . 1964. Primary productivity and micro- nutrient limiting factors in some North Amer- ican and New Zealand lakes. Internationale Vereinigung fur Theoretische und Angewandte Limnologies Verhandlungen 15:365-374. 1968. Aquatic primary production. Amer- ican Zoologist 8:31-42. 1973. Will Baikal and Tahoe be saved? Cry California. Journal of California Tomor- row, Winter 1973-74, 9(l):19-25. 1981. Lake Tahoe: two decades of change in a nitrogen deficient oligotrophic lake. Inter- nationale Vereinigung fur Theoretische und Angewandte Limnologie Verhandlungen 21:45-70. , and R. Armstrong. 1969. Primary pro- ductivity studies in Lake Tahoe, California. Internationale Vereinigung fur Theoretische und Angewandte Limnologie Verhandlungen 17:49-71. 260 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 , and E. de Amezaga. In press. Primary productivity and precipitation at Castle Lake and Lake Tahoe during twenty-four years, 1959-1982. Internationale Vereinigung fur Theoretische und Angewandte Limnologie Verhandlungen. , and A. J. Horne. 1983. Limnology. McGraw-Hill, New York. , M . D. Morgan, S. T. Threlkeld, and N. Angeli. 1979. A population dynamics analy- sis of the cladoceran disappearance from Lake Tahoe, California-Nevada. Limnology and Oceanogi-aphy 24:289-297. , and R. G. Wetzel. 1963. A study of the primary productivity of Clear Lake, Lake County, California. Ecology 44:283-294. Goldman, M. S. 1977. In praise of little fishes. David R. Godine, Boston. Hyne, N. J., R Chelminski, J. E. Court, D. S. GORSLINE, and C. R. Goldman 1972. Quater- nary history of Lake Tahoe, California- Nevada. Geological Society of America Bulle- tin 83:1435-1448. Le Conte, J. 1875. On some of the ancient glaci- ers of the Sierra, Nevada. American Journal of Science and Arts, 3rd Ser, 10:126-139. . 1883a. Physical studies of Lake Tkhoe - I. Overland Monthly, I., 506-516. 1883b. Physical studies of Lake Tahoe - II. Overland Monthly II., 595-612. . 1884. Physical Studies of Lake Tahoe - III. Overland Monthly 41-46. Likens, G. E. 1983. A priority for ecological re- search. Ecological Society of America Bulletin 64:234-243. Lund, J. W. G. 1964. Primary production and per- iodicity of phytoplankton. Internationale Ver- inigung fur Theoretische und Angewandte Limnologie Verhandlungen 15:37-56. Naumann, E. 1919. Nagra synpunkter angaende planktons okologi. Med. sarskild hansyn till fytoplankton. Svensk Botanisk Tidskrift 13:129-158. Paerl, H. W., R. C. Richards, R. L. Leonard, and C. R. Goldman. 1975. Seasonal nitrate cj'cling as evidence for complete vertical mixing in Lake Tahoe, California-Nevada. Limnology and Oceanography 20:1-8. Richards, R. C, C. R. Goldman, T. C. Frantz, and R. WiCKWiRE. 1975. Where have all the Daph- nia gone? The decline of a major cladoceran in Lake Tahoe, California-Nevada. Interna- tionale Vereinigung fur Theoretische und Angewandte Limnologie Verhandlungen 19:835-842. RuTTNER, F. 1963. Fundamentals of limnologj'. English translation of 3rd ed. by D. G. Frey and F. E. J. Fry. University of Toronto Press, Toronto, Canada. Steemann-Nielsen, E. 1952. The use of radioac- tive carbon (C-14) for measuring organic pro- duction in the sea. Conseil International pour FE.xploration de la Mer Journal 18:117-140. Strub, p. T, T Powell, and C. R. Goldman. In Press. Climatic forcing: effects of El Nino on a small lake. Science. TscHUMi, P.-A., B. Bangerter, and D. Zbaren. 1982. Zehn Jahre Limnologische Forschung am Bielersee (1972-1981). Vierteljahrsschrift der Naturforschenden Gesellschaft in Zurich 127/4:337-355. Vallentyne, J. R. 1960. Fossil pigments. Pages 83-105 in M. B. Allen, ed.. Comparative bio- chemistry of photoreactive systems. Academic Press, New York. A Comparison of the Embryonic Development of Northern, Florida, and Reciprocal F^ Hybrid Largemouth Bass in Different Thermal Environments David P Philipp, Christine Kaminski, and G. S, Whitt The two subspecies of largemouth bass, Micropterus salmoides salmoides andM s. floridanus, naturally occur in different geogi'aphic and climatic re- gions of the United States, with hy- brids occurring in a zone of intergi'ada- tion between these regions. Genetic dif- ferences between these stocks are re- flected by differing physiological re- sponses to the thermal environment. To determine the extent to which these differences contribute to natural geo- gi'aphic separation of the subspecies, we produced the embryos of the four genetic stocks of largemouth bass (M. s. salmoides, M. s. floridanus, M. s. sal- moides 9 X M. s. floridanus d, and M. s. floridanus 9 x M. s. salmoides d) in vitro, using artificial fertilization tech- niques. The developmental success of the embi-yos and the schedule of embry- ogenesis for each stock were compared at each of a series of temperatures. The developmental success of each stock at each incubation temperature was deter- mined by total hatching percentage. In addition, the thermal requirements for embryonic development of each of these four stocks were compared by deter- mining the a —threshold temperatures of development as well as the number of thermal developmental units re- Dr. David P. Philipp is an As.sociate Aquatic Biologist, Section of Aquatic Biologj', Illinois Natural History Survey, and an Assistant Pro- fessor of Animal Science. University of Illinois, Urbana-Champaign; Ms. Christine Kaminski is a Junior Technical Assistant. Section of Aquatic Biology, Illinois Natural History Survey; and Dr G. S. Whitt is a Professor of Genetics and Develop- ment, University of Illinois, Urbana-Champaign, and a Research Affiliate, Section of Aquatic Biol- ogy, Illinois Natural History Survey. quired to reach each of 22 key embry- onic stages. Significant differences in the thermal requirements for embryo- genesis exist among these different stocks of largemouth bass. The implica- tions of these findings on current and future largemouth bass management programs are discussed. The morphological differences be- tween the northern largemouth bass, Micropterus salmoides salmoides (La- cepede), and the Florida largemouth bass, M. s. floridanus (Le Sueur), were discussed by Bailey & Hubbs in 1949. Since that time a number of studies designed to further assess the differ- ences between the subspecies have been conducted (Hart 1952; Clugston 1964; Addison & Spencer 1972; Zol- czynski & Davies 1976; Inman et al. 1977; Cichra et al. 1981; Smith & Wil- son 1981; and Wright & Wigtil 1981). These studies have shown that a varie- ty of biological differences exist be- tween these two subspecies. Our recent electrophoretic survey of the genetic structure of largemouth bass populations in the United States (Philipp et al. 1981, 1982, 1983) has quantified the degree of genetic diver- gence between these two subspecies. We have described a fast and reliable means of identifying pure northern, pure Florida, or intergrade populations of largemouth bass, that of the electro- phoretic determination of the allele fre- quencies at the Idh-B and Aat-B loci. Meristic and morphometric counts rou- tinely used in the past by fisheries biologists to distinguish these types of populations (Bryan 1969; Addison & Spencer 1972; Buchanan 1973; Inman 261 262 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 et al. 1977; Moyle & Holzhauser 1978; and Bottroff & Lembeck 1978) are am- biguous and, hence, unreliable. In at least one study (Pelzman 1980) elec- trophoretic techniques also proved un- reliable, since analyses were inappro- priate and data were misinterpreted. Our biochemical genetic analyses of largemouth bass populations (Philipp et al. 1981, 1982, 1983) also demon- strate that the intergrade zone between the ranges of the two pure subspecies, as it exists today (northern Florida, Mississippi, Alabama, Georgia, South Carolina, North Carolina, Virginia, and Maryland), is much more extensive than that described by Bailey & Hubbs (1949). As a result, the previously assigned genetic status of the stocks of largemouth bass used as representa- tives of the two pure subspecies in many previous studies appears ques- tionable. In the absence of genetic con- firmation of the stocks used, the data generated by these earlier studies and their resulting conclusions must be considered very cautiously. In 1959, Florida largemouth bass, or at least largemouth bass containing some portion of the genome of M. s. floridanus, were introduced into cer- tain waters in California (Sasaki 1961). The subsequent establishment of large- mouth bass populations with a sub- stantial proportion of the gene pool con- tributed by the Florida subspecies has been well documented (Smith 1971; von Geldern & Mitchell 1975; Bottroff & Lembeck 1978; Moyle & Holzhauser 1978). The populations of largemouth bass which existed in California prior to 1959 were the result of introductions of northern largemouth bass imported from Illinois in 1891 (Shebley 1917) and were not the result of immigi-ation and natural selection, since the state of California is well outside the native range of largemouth bass (MacCrimmon & Robbins 1975). It is not surprising that the introductions of largemouth bass from Florida were successful in southern California, since the climatic conditions more closely resemble those of Florida than of Illinois. The apparent success of the recent introductions of M. s. floridanus in southern California waters and the de- mand from fishermen for more and larger largemouth bass have apparent- ly provided the impetus for a number of states to initiate Florida largemouth bass programs. These progi-ams range from controlled research with limited introductions to large-scale propaga- tion and widespread stockings. Unlike California, many of these states al- ready contained populations of natui-al- ly established largemouth bass. Due to the lack of reliable, quantitative data concerning the genetic differences of the native and introduced stocks and the relative fitnesses of these stocks in different environments, the long-term effects of these introductions upon the existing largemouth bass fisheries can- not be accurately predicted at this time. However, the impact of these stocking programs on the genetic in- tegi-ity of the native largemouth bass populations in these states could be catastrophic. We have postulated I Philipp et al. 1981, 1982, 1983) that when alleles present in the Florida sub- species are introduced into a popula- tion of the northern subspecies, the unique genie combinations initially present may become irreversibly al- tered. In addition, some of the Florida alleles are likely to be less fit for these new environments. Although the result- ing populations may be genetically suf- ficient in terms of short-term survival, the long-term effect of this genetic mix- ture would be a lowering of fitness in the recipient population. For these rea- sons, we have recommended that pro- grams designed to introduce the Flor- ida subspecies into states outside of peninsular Florida but within the na- tive range of the largemouth bass be halted until appropriate research has determined the effects on the recipient populations. Factors which affect year-class strength among largemouth bass popu- lations are complex. It has been sug- gested that spawns hatched early in a given year may suffer substantially September 1985 125 Years of Biological Research 263 less mortality than those hatched later in the season (Aggus & Elliot 1975). Therefore, in mixed populations, differ- ential thermal effects upon the repro- ductive behavior and the rate and success of development between nor- thern and Florida largemouth bass em- bryos may play a crucial role in deter- mining the relative contribution of each stock to the total year-class pro- duction. The current study was de- signed to assess thermal effects upon the rate and success of the development of embryos of geneticaly defined stocks of both pure subspecies, M. s. salmoides and M. s. floridanus, and of both recip- rocal Fj hybrids. MATERIALS AND METHODS Parental Stocks Northern largemouth bass were collected from Clinton Lake, Illinois. Electrophoretic analyses of individuals from this population showed the fre- quency of the northern Idh-B allele (Idh-BM and the sum of the northern Aat-B alleles (Aat-B' and Aat-B^) both to be 1.000, confirming that this popu- lation consisted of pure M. s. salmoides (Philipp et al. 1981, 1982, 1983). Flor- ida largemouth bass were collected from Lake Dora, Florida. Electrophor- etic analyses of individuals from this population showed the frequency of the Florida Idh-B allele (Idh-B^) and the sum of the Florida Aat-B alleles (Aat-B^ and Aat-B-*) both to be 1.000, confirm- ing that this population consisted of pure M. s. floridanus (Philipp et al. 1981, 1982, 1983). Adult males and females from only these two genetical- ly confirmed populations were used as brood stock throughout this study. Production of Embryos Florida largemouth bass (FLMB) and Florida 9 x northern d F, hybrid largemouth bass (F x N) embryos were produced at the Florida Game and Freshwater Fish Commission, Eustis Fisheries Research Laboratory. North- ern largemouth bass (NLMB) males were collected in December 1980 from Clinton Lake, Illinois, and were held indoors at 10 °C at the Illinois Natural History Survey (INHS). On 2 February 1981, these fish were marked with a right pectoral clip, transported to Flor- ida, and allowed to reach reproductive readiness in outdoor earthern ponds at the Richloam State Fish Hatchery. During February 1981 ripe male NLMB, retrieved from the hatchery, and ripe male FLMB, collected from Lake Dora by electrofishing, were brought to the Eustis laboratory, where they were held in indoor flow-through raceways. On 21 and 26 February 1981, mature, ripe female FLMB were col- lected from Lake Dora, Florida, by elec- trofishing and were brought to the Eustis laboratory. The eggs from indi- vidual FLMB females were manually stripped into a bowl, mixed, and split into two aliquots in separate petri dishes. One aliquot was fertilized with sperm from a single FLMB male and the other with sperm from a single NLMB male, using methods described in Childers (1967) and Philipp et al. (1979). Northern largemouth bass (NLMB) and northern $ x Florida d F, hybrid largemouth bass (N x F) emoryos were produced in essentially the same man- ner, with the following exceptions. FLMB males, collected from Lake Dora in February 1981, were marked with a left pectoral clip and were air shipped to the Illinois Natural History Survey to be held indoors at 10 °C until they were stocked outdoors in INHS earthen ponds on 10 March 1982 and were allowed to reach reproductive readiness. During May 1981, FLMB males retrieved from INHS ponds and NLMB males collected from Clinton Lake by electrofishing were brought to the INHS laboratory and held indoors. On 13, 22, and 28 May, mature, ripe NLMB females were collected from Clinton Lake by electrofishing and were brought to the INHS laboratory. NLMB and N x F embryos were pro- duced in the laboratory, using the pro- 264 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 cedures described previously for pro- ducing FLMB and F x N embryos. Rearing of Embryos Procedures and equipment used for rearing the FLMB and F x N embryos were identical to those used for rearing the NLMB and N x F embryos. The very dense egg-sperm mixtures, pro- duced as has been described, were allowed to stand in minimal water for 5 minutes to allow for completion of the fertilization process. These newly fer- tilized eggs were transferred to plastic containers holding about 2-3 cm of water. The eggs were thinly spread within these containers and incubated for 1 hour at 24 °C until initial cleav- age. For each cross, samples of 100 nor- mally cleaving eggs (2-4 cell stage) were then removed, transferred to each of a number of glass finger bowls con- taining 24 °C water, and allowed to adhere to the glass. The embryos in each finger bowl were acclimated for 1 hour to a specified test temperature, the bowls were covered with nylon net- ting to allow water circulation but pre- vent loss of embryos, and the bowls were immersed in an aerated, filtei-ed, 75-liter constant temperature bath at the test temperature. The temperature of each incubation bath was monitored continuously, using dual water temper- ature probes and a calibrated 12-channel recorder (Chino Works, Ltd., Tokyo, Japan, Model EW 1200). Morphological development of each set of embryos reared at each temper- ature was visually monitored, using a dissecting microscope. Prior to retinal pigmentation, each set of embryos was monitored every 2-4 hours. After retinal pigmentatin had progi-essed, the frequency of visual observation was reduced to every 6-8 hours. The times required to reach each of 22 key morph- ological stages were recoi'ded for the embryos at each test temperature. Dead eggs or embryos were counted, removed, and recorded at each visual inspection. A photogi'aphic record of the morphological development of these embryos was made, using an Olympus JM dissecting microscope with an Olympus photogi-aphic attach- ment and an Olympus OM-2 camera. Data Analysis The success of development was determined by calculating for each set of embryos the percentage of eggs which hatched and of these the propor- tion which appeared normal. The a -threshold temperature of development, a theoretical thermal value below which embryonic develop- ment ceases (Childers 1967), was deter- mined. For each set of embryos raised at each temperatm-e, the time required to reach each of 11 readily identifiable morphological stages was determined, and the cumulative average tempera- ture of development was calculated from the thermal record for each of these stages. For each of the four gen- etic stocks (NLMB, N x F, F x N, and FLMB) at each developmental stage, average cumulative temperature was plotted versus the inverse of develop- mental time in hours. The regi-ession equation describing the lineai- relation- ship for each of the four stocks was determined from these values (10-12 data points per stage for the FLMB and F X N embryos and 16-18 data points per stage for the NLMB and N x F embryos). The x -intercept determined from each equation is the a -threshold temperature determined for that stock, using the values for that development- al stage. The final a -threshold temper- ature for each genetic stock was deter- mined as the average of the values for these 11 stages. The number of thermal develop- mental units (TDU) required for an em- bryo to reach a given stage of develop- ment is defined as the number of degree-hours above the a -threshold temperature which must be accumu- lated. TDU values were calculated, using the final value for the a -thresh- old temperature together with the times of development and average cum- ulative temperatures for each set of September 1985 125 Years of Biological Research 265 embryos of each genetic stock at each of 22 morphological stages of develop- ment. RESULTS Patterns of Development No qualitative differences in mor- phogenetic events could be detected among the four genetic stocks of large- mouth bass studied (NLMB, N x F, F X N, and FLMB). Therefore, the follow- ing morphological features and their developmental sequence hold for the embryogenesis of all four stocks. The cortical reaction occurs immed- iately upon fertilization, with the cor- tical layer becoming raised by 1 minute after fertilization. There follows a con- fluence of cytoplasm around the zygote nucleus and the resultant formation of the germinal disc. Cleavage is telo- lecithal, the initial formation of the two-cell stage occurring within 1 hour at 24 °C. Rates of development during the rest of this study depended upon temperature and will be discussed in detail later. Fig. lA through IP illus- trate the periods of morphogenesis of one set of embryos, FLMB embryos raised at 24.2° ± 0.4 °C. These pictures illustrate the patterns of development for each of the four stocks of large- mouth bass embryos studied (NLMB, N X F, F X N, and FLMB). The develop- mental progression paralleled that described for other centrarchid species (Morgan 1951; Baton 1959; Champion & Whitt 1976; Taubert 1977). Early cleavage (Fig. lA) continues, eventually resulting in blastula forma- tion (Fig. IB). Epiboly commences (Fig. IC) and progresses through the yolk plug stage (Fig. ID) prior to the forma- tion of a body axis (Fig. IE). Embryonic development continues with the forma- tion of increasing numbers of somite pairs (Fig. IF). The optic cup and peri- cardial cavity continue to develop. The embryonic heart begins beating, and true circulation with colorless blood starts shortly afterward (Fig. IG). The tail increases in length, and body con- tractions increase in number and sever- ity. Red blood cells containing hemo- globin appear just prior to hatching (Fig. IH). Hatching occurs fairly syn- chronously for most eggs at normal temperatures (90 percent of the eggs hatched in a 3-4 hour period at 24 °C). The newly hatched embryo is still quite underdeveloped (Fig. II), having only limited and unguided movement. After hatching, the heartbeat becomes more vigorous, and the first pigment gran- ules become visible in the retina (Fig. IJ). Pectoral and pelvic fins form as the retina becomes quite darkly pigmented (Fig. IK). The tapetum lucidum devel- ops, to give the eye at first a silvery and later a golden appearance (Fig. ID. During this period the rudiments of many of the internal organs, such as the liver, urinary bladder, intestine, and swim bladder, are developing. Fol- lowing this stage the jaw begins to form (Fig. IM) and eventually stai'ts to open and close regularly (Fig. IN), as body pigmentation commences. At this point, the embryos begin to swim off the substrate with increasing regulari- ty (Fig. 10), eventually reaching a free- swimming stage and active feeding as yolk sac absorption is completed (Fig. IP). Our monitoring of the develop- ment of largemouth bass ceased at this point and did not include an investiga- tion of the subsequent fry and finger- ling stages. Effect of Temperature on Developmental Success The percentage of eggs from which embryos (normal and abnormal) hatched was calculated for each set of embryos reared. These results are given in Tkble 1. The results for the NLMB and FLMB embryos are also compared gi'aphically in Fig. 2. Comparing the two pure subspecies, the FLMB em- bryos exhibited optimal hatching rates at temperatures (20-28°C) substan- tially higher than those exhibited by the NLMB embryos (17 °-24°Cl (Table 1, Fig. 2). These temperature ranges are somewhat lower than the values reported by McCormick & Wegner (1981). In addition, the lower thermal limits for successful hatching were higher for FLMB embryos than for 266 September 1985 125 Years of Biological Research 267 NLMB embryos. Conversely, the upper thermal limits for successful hatching were lower for NLMB embryos than for FLMB embryos. Less than 15 percent of hatched embryos appeared morpho- logically abnormal at each test temper- ature except for the NLMB and N x F at 30.5 °C (100 percent abnormal in each case), the NLMB and N x F at 30.3 °C (68.6 percent and 75.0 percent abnormal, respectively), and the FLMB and F X N at 18.5 °C (16.3 percent and 18.5 percent abnormal, respectively). The tendency for both hybrids to exhibit developmental patterns more similar to that of the maternal than that of the paternal subspecies was not- able (Table 1). Although the relation- ship between temperature and hatch- ing percentage for N x F embryos par- alleled that for NLMB embryos, the ab- solute hatching percentage for the N x F embryos was lower than that for the NLMB embryos at most temperatures. The relationship between temperature and hatching percentage for the F x N embryos paralleled that for the FLMB embryos as well. However, the absolute hatching percentage for the F x N em- bryos was greater than that for the FLMB embryos at most temperature. Thermal Limits of Embryogenesis The a -threshold tempei'ature of development were calculated for each of the four stocks of largemouth bass (NLMB, N x F F X N, and FLMB) using data from each of 11 distinct developmental stages (body axis; 10, 15, 20, and 25 somites; heartbeat; 50- percent hatching; onset of, light, and even retinal pigmentation; and jaw movement). We chose to monitor these 11 stages because they were easily identifiable visually. Because the rates of development for the NLMB and N x F embryos were similar at all of the temperatures studied, the data for these two stocks were combined for a -threshold temperature calculations. Data for the FLMB and F x N embryos were similarly combined (Table 2). The overall a -threshold value for each stock was calculated as the aver- age of the values calculated individual- ly from the data for each of the 11 indi- vidual developmental stages used (Table 2). The overall a- threshold tem- perature determined for the NLMB and N X F embryos, 12.62 + 0.27, was significantly higher (P<0.01, Wilcox- on's signed-ranks test, Sokal & Rohlf 1973) than that determined for the FLMB and F x N embryos, 11.52 + 0.77. In addition, the stage-specific a -threshold temperatures calculated for the NLMB and N x F embryos were higher than those calculated for the FLMB and F x N embryos at each of the developmental stages used (Table 2). The upper thermal limits of develop- ment can be estimated by extrapolat- ing from the decrease in hatching per- centage for a few of the trials at ele- vated temperatures, >30°C (Table 1). NLMB and N x F embryos did not hatch at temperatures above 30.5 °C, and even at temperatures between 30 ° and 30.5 °C there was an extremely high percentage of morphological de- formities. These results agree closely with those reported by McCormick & Wegner (1981) for northern lai'gemouth bass embryos. FLMB and F x N em- bryos, however, successfully hatched at 30.8 °C, although at 31.7 °C no success- 26 28 30 32 INCUBATION TEMPERATURE (°C) Fig. 2. - The effect of temperature on the hatchint,' success of northern largemouth bass ( • I and Florida largemouth bass (O ). 268 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Table 1. — Hatching percentages of largemoutti bass eggs incubated at various temperatures. September 1985 125 Years of Biological Research 269 ful hatching was observed. It appears, therefore, that the upper thermal limit for successful embryonic development of FLMB and F x N embryos is 0.5°-1.0°C higher than that for the NLMB and N x F embryos. Thermal Requirements of Embryogenesis Using the overall a -threshold tem- peratures calculated for the NLMB and N X F embryos and for the FLMB and F X N embryos, 12.62 + 0.27 and 11.52 ± 0.77, respectively, the numbers of thermal developmental units (TDU) re- quired to reach each of 22 development stages were calculated and are shown in Table 3. For the first 12 stages (mid- blastula through end of hatching), em- bryos from all stocks required compar- able thermal input (TDU). However, for the remainder of development, the FLMB and F x N embryos required substantially greater thermal input (TDU) than did the NLMB and N x F embryos, ranging from 11.3 percent more for the onset of retinal pigmenta- tion to 29.0 percent more for the final yolk absorption stage. DISCUSSION The environmental thermal regimes which exist among the aquatic com- munities throughout the United States are extremely varied. Temperature plays a potentially major role in the processes of natural selection, geneti- cally tailoring stocks of largemouth bass to specific environments (Childers 1975). Genetic differences between the two subspecies of largemouth bass, M. s. salmoides and M. s. floridanus, have been assessed and documented (Philipp et al. 1981, 1982, 1983). Indeed, a varie- ty of physiological and behavioral dif- ferences most likely result from these genetic differences, each stock reacting somewhat differently to thermal condi- tions. Because these physiological and genetic differences among the two sub- species and their hybrids determine their relative fitness in a given environ- Table 3. — Developmental unit requirements of the four stocks. 270 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 ment, a quantitative assessment of thermal response differences among these stocks is critically needed for effectively formulating present and future management programs for large- mouth bass. Although Swingle (1956) reported that most largemouth bass spawning occurs between 68° and 75 °F (20°-24°C), Chew (1974) documented spawning of FLMB in Lake Weir, Florida, at 59 °F (15 °C). Fluctuations in water temper- ature during the spring spawning sea- son are usually not as severe in lakes in peninsular Florida as they are in lakes in more northerly regions. Large- mouth bass spawns in Florida lakes during the earliest portion of the sea- sonal spawning period may be less like- ly to result in the exposure of eggs or embryos to very cold temperatures (<12°C) than would correspondingly early spawns in northern regions. It has been suggested that, in mixed pop- ulations, the Florida subspecies tends to spawn earlier in the season, at lower temperatures, than does the northern subspecies of largemouth bass (Hun- sacker & Crawford 1964; Bottroff & Lembeck 1978; Moyle & Holzhauser 1978). Observations of spawning in ponds at the Illinois Natural History Survey during 1981 agreed with this suggestion (unpublished results); fur- ther experimentation is currently under way to verify this finding. Interestingly, the a -threshold tem- perature of development of the FLMB (11.52 ± 0.77) is significantly lower than that of the NLMB (12.62 ± 0.27). This relationship suggests a survival strategy of delayed spawning in the northern subspecies which presumably reflects a more variable, lower mean water temperature during spawning periods in the north. In addition, the northern subspecies apparently can use environmental incubation temper- atures more efficiently. This use is evidenced by the significantly lower number of thermal developmental units required by NLMB embryos to reach the free swimming and actively feeding stage (2,031 + 229) than that required by the FLMB (2,620 ± 203). Only at temperatures below about 16.2 °C would FLMB eggs require less incubation time than NLMB eggs to reach the free swimming and activelj' feeding stage. At "normal" incubation temperatures NLMB embryos reach this stage sooner than do FLMB em- bryos. For example, at 20 °C NLMB em- bryos would reach the free swimming and actively feeding stage at 272 hours, whereas the FLMB would require 309 hours to I'each this stage. This more rapid development gives NLMB em- bryos a clear advantage over FLMB embryos at temperatures normally en- countered during incubation periods in the wild(17°-25°C). Interestingly, in comparing the em- bryological development of the two subspecies, major diffei-ences in num- bers of thermal devlopmental units re- quired occur at retinal pigmentation and later These periods of embryonic development are associated with organ- ogenesis and are periods during embiy- ogenesis in which many of the genes encoding metabolic enzymes become activated (Philipp et al. 1979). This observation is consistent with the hypothesis that many of the metabolic processes up to the retinal pigmenta- tion period may be under the control of maternal enzymes or m-RNA mole- cules synthesized during oogenesis. This observation also suggests that the two subspecies may have diverged in certain of their gene regulatory proc- esses and now respond differently to temperature. The differences in the in- itial timing and levels of enzyme ex- pression in developing embryos of these four stocks (NLMB, N x F, F x N, and FLMB) strongly support this sugges- tion (Philipp et al. 1983; Parker, Philipp, & Whitt unpublished results). The two subspecies also differ in hatching success at various incubation temperatures. The peak of successful hatching for the NLMB occurs at lower temperatures (17 -22 ''O than that for the FLMB (21 °-24 °C). In addition, the FLMB embryos apparently survive higher incubation temperatures than September 1985 125 Years of Biological Research 271 do the NLMB (Fig. 2). These relation- ships probably provide the NLMB and FLMB embryos an advantage in their respective environments. It is interesting that the N x F embryos have consistently lower hatch- ing percentages than their thermal counterpart NLMB embryos but that the opposite holds true for the F x N and FLMB embryos (Table 1). This set of relationships is an example of the inherent differences between reciprocal hybrids. We have postulated that these nonadditive asymmetrical responses by reciprocal hybrids result from differ- ences in the interactions of the paternal genes with the maternal gene effector molecules (Whitt et al. 1977; Philipp et al. 1983). Our analyses reveal that the de- veloping embryos ofNLMB and FLMB react differently to different tempera- tures. We feel that these differences have I'esulted from these two subspe- cific genomes having evolved indepen- dently in response to different thermal selective pressures. We postulate that to better survive colder climates, NLMB have evolved a reproductive strategy of delayed spawnng until they encounter higher temperatures. This strategy helps to prevent premature spawns, which could be destroyed by severe cold. In addition, NLMB have evolved schedules of embryonic gene expression, and thus mechanisms of controlling embryonic metabolism and morphogenesis, that operate efficiently and rapidly at lower temperatures. These patterns of gene expression also allow NLMB embryos to reach maxi- mal hatching rates at lower tempera- tures than those required by FLMB embryos. However, during the pro- tracted spawning seasons in penin- sular Florida, early spawns produce individuals as much as 3 months ear- lier than late spawns. FLMB appear to have evolved a reproductive strategy which allows them to spawn at lower temperatures than those at which NLMB spawn. These early fish have distinct competitive advantages over fish spawned later FLMB embryos have also evolved an increased tolerance to incubation temperatures higher than those of NLMB embryos. Thus, FLMB embryos are better able to tolerate the warm water temperatures found in peninsular Florida during the large- mouth bass spawning season. The divergence of developmental response to temperature in the two sub- species of largemouth bass has signifi- cant implications for management pro- gi'ams. Fitness probably varies gi'eatly, depending upon the geographic loca- tion of the population, the physical characteristics of the body of water being considered, and the weather con- ditions during a given spawning sea- son. However, it is not unreasonable to assume that, in the long run, stocks of largemouth bass introduced into inap- propriate thermal environments will not perform as well as stocks which are introduced into thermal environments for which they have been genetically tailored. Our findings support the recommendation that management programs which result in mixing the two subspecies be discontinued. Pre- cautions must be taken to protect the genetic integi'ity of the two subspecies. Specifically, we recommend that no largemouth bass containing any por- tion of the genome of the Florida subspecies be propagated for intro- duction into waters north of the inter- grade zone as described by Philipp et al. (1981, 1982, 1983). We also recom- mend, of course, that the introduction of largemouth bass containing any por- tion of the genome of the northern subspecies into waters of pensinsular Florida be prohibited. For maximum effectiveness, it is imperative that sound genetic prin- ciples be incorporated into current and future fisheries management progi-ams (Smith & Chesser 1981; Philipp et al. 1981, 1982, 1983). Individual genetic stocks which constitute a species must be identified, characterized, and recog- nized as distinct units requiring indi- vidual consideration in management efforts. The genetic integi'ity of discrete stocks must be preserved, because as 272 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Ryman (1981) has so aptly stated, "Genetic variation in natural popula- tions constitutes a biological resource that must be properly managed so as not to reduce future opportunities for use of the resource." ACKNOWLEDGEMENTS The authors wish to thank the Flor- ida Game and Freshwater Fish Com- mission for their support and assist- ance in this study from its inception. We especially wish to thank F Gerry Banks, Dennis "Smokie" Holcomb, For- rest J. Ware, and Edward Zagar for per- mitting field collection and providing laboratory space at the Eustis Fisher- ies Research Laboratory, and Charles Starling and Harrell Revels for assist- ance at the Richloam hatchery. We also wish to thank a number of staff mem- bers of the FGFWFC who helped col- lect fish, assisted with experiments, lent equipment, and offered valuable advice: Joseph Crumpton, William Cole- man, William Johnson, A. Michael Wicker, Charles Mesing, Richard Krause, Levi J. Jenkins III. James Bitter, Robert Wattendorf, and Paul Shafland. Finally, we wish to thank Dr. Bruce Taubert. Henry R. Parker. Nancy Frye, Lynn Dettman, Charles Stone, Sheila Magee, and Todd Powless for field and laboratory assistance and Suzanne Peratt for technically prepar- ing the manuscript. This project was supported by funds from the Illinois Natural Histoiy Survey and by Federal Aid for Fish Restoration funds from the Illinois Department of Conservation, Project F-35-R. LITERATURE CITED Addison, J. H., and S. L. Spencer. 1972. Pi-elimin- ary evaluation of three strains of largemouth bass, Microplerus salmoides (Lacepede), stocked in ponds in south Alabama. Annual Confer- ence of the Southeastern Association ofGame and Fish Commissioners Proceedings 25:366-374. Aggus, L. R., and G. V. Elliot. 1975. Effect of cover and food on year-class strength of large- mouth bass. Pages 317-322 ;;; R. H. Stroud and H. Clepper, eds.. Black bass biology and management. Sport Fishing Institute, Wash- ington, D.C. Bailey, R. M., and C. L. Hubbs. 1949. The black basses [Micropterus) of Florida, with descrip- tion of a new species. University of Michigan Museum of Zoology Occasional Papers 516. Balon, E. 1959. Spawning of Lepomis gibbosus (Linne 1758) acclimatized in the backwaters of the Danube and its development during the embryonic period. Zeitschrifft fur Fisheries Deren Hilfswissen 8:1-27. Bottroff, L. J., and M. E. Lembeck. 1978. Fish- ery trends in reservoirs of San Diego County, California, following the introduction of Flor- ida largemouth bass, Micnipterun aalmoides floridaniis. California Fish and Game 64:4-23. Bryan, C. F. 1969. Variation in selected nieristic characters of some basses, Micropterus. Copeia 1969: 370-373. Buchanan, J. P. 1973. Separation of the sub- species of largemouth bass, Micropterus sal- moides salmoides and M. s. floridanus and intergrades by use of meristic characters. Annual Conference of the Southeastern Association of Game and Fish Commissioners Proceedings 27:608-619. Champion, M. J., and G. S. Whitt. 1976. Dif- ferential gene expression in multilocus isozyme systems of the developing gi-een sun- fish. The Journal of Experimental Zoology 196:263-282. Chew, R. L. 1974. Early life history of the Florida largemouth bass. Florida Game and Fresh Water Fish Commission Fishery Bulletin 7. Childers, W. F 1967. Hybridization of four species of sunfishes (Centrarchidae'. Illinois Natural History Survey Bulletin 29:159-214. 1975. Bass genetics as applied to culture and management. Pages 362-372 in R. H. Stroud and H. Clepper, eds.. Black bass biol- ogy and management. Sport Fishing Insti- tute, Washington, D.C. Cichra, C. E., W. E. Neill. and R. L. Noble. 1981. Differential resistance of northern and Florida largemouth bass to cold shock. Annual Conference of the Southeastern Asso- ciation of Fish and Wildlife Agencies Proceed- ings 34:19-24. Clugston, J. P. 1964. Growth of the Florida largemouth bass, Micropterus salmoides flor- idanus (Le Sueur), and the northern large- mouth bass, M. s. salmoides (Lacepede). in subtropical Florida. American Fisheries Soci- etv Ti'ansactions 93:146-154. September 1985 125 Years of Biological Research 273 Hart, J. S. 1952. Geographic variations of some physiological and morphological characters in certain freshwater fish. University of Toronto Biological Series 60, Ontario Fisheries Re- search Laboratory Publication 72. HuNSACKER, D., and R. W. Crawford. 1964. Pref- erential spawning behavior of the largemouth bass. Copeia 64:240-241. Inman, C. R., R. C. Dewey, and P. P. Durocher. 1977. Growth comparisons and catchability of three largemouth bass strains. Annual Con- ference of the Southeastern Association of Game and Fish Commissioners Proceedings 30:1-17. MacCrimmon, H. R., and W. H. Robbins. 1975. Distribution of the black basses in North America. Pages 56-66 in R. H. Stroud and H. Clapper, eds.. Black bass biology and manage- ment. Sport Fishing Institute, Washington, D. C. McCormick, J. H., and J. A. Wegner. 1981. Responses of largemouth bass from different latitudes to elevated water temperatures. American Fisheries Society Ti-ansactions 110:417-429. Morgan, G. D. 1951. Life history of the bluegill sunfish, Lepomis macrochirus, of Buckeye Lake (Ohio). Denison University Journal of the Scientific Laboratories 42:21-59. MoYLE, R B., and N. J. Holzhauser. 1978. Effects of the introduction of Mississippi silverside iMenidia aiidens) and Florida largemouth bass [Micropterus salmoides floridaniis) on the feeding habits of young-of-the-year large- mouth bass in Clear Lake, California. Amer- ican Fisheries Society Transactions 107:574-582. Pelzman, R. J. 1980. Impact of Florida large- mouth bass, Micropteruti salmoides floridaniis, introductions at selected northern California waters with a discussion of the use of meristics for detecting introgression and for classifying individual fish of intergi'aded populations. California Fish and Game 66:133-162. Philipp, D. R, W. F. Childers, and G. S. Whitt. 1979. Evolution of differential patterns of gene expression: A comparison of the tempor- al and spatial patterns of isozyme locus ex- pression in two closely related fish species (northern largemouth bass, Micropterus sal- moides salmoides, and smallmouth bass, Micropterus dolomieui). Journal of Experi- mental Zoology 210:473-488. , , and 1981. Management implications for different genetic stocks of largemouth bass (Micropterus salmoides) in the United States. Canadian Journal of Fish- eries and Aquatic Sciencies 38:1715-1723. , , and . 1982. Biochemical genetics of largemouth ba.ss. Electric Power Research Institute, Palo Alto, CA. , , and . 1983. A biochemical genetic evaluation of the northern and Florida subspecies of largemouth bass. American Fisheries Society Ti-ansactions 112:1-20. Ryman, N. 1981. Conservation of genetic re- sources: experiences from the brown trout {Salmo trutta). Pages 61-74 in N. Ryman, ed.. Fish gene pools. Preservation of genetic re- sources in relation to wild fish stocks. Ecologi- cal Bulletin (Stockholm) 34. Sasaki, S. 1961. The introduction of Florida large- mouth bass into San Diego County. Califor- nia Fish and Game, Inland Fisheries Administration, Report 61-11. Sacramento, CA. Shebley, W. H. 1917. History of the introduction of food and game fishes into the waters of Cali- fornia. California Fish and Game 3(1):3-12. Smith, G. 1971. Florida largemouth bass in southern California. Florida Wildlife 25(41:30-33. Smith, M. H. and R. K. Chesser. 1981. Rationale for conserving genetic variation of fish gene pools. Pages 13-20 in N. Ryman, ed., Fish gene pools. Preservation of genetic resources in relation to wild fish stocks. Ecological Bul- letin (Stockholm) 34. Smith, R. P, and J. L. Wilson. [1981.1 Growth comparison of two subspecies of largemouth bass in Tennessee ponds. Annual Conference of the Southeastern Association of Fish and Wildlife Agencies Proceedings 34:25-30. SoKAL, R. R., and F. J. Rohlf. 1973. Inti'oduction to Biostatistics. W. H. Freeman and Co., San Francisco, CA. Swingle, H. S. 1956. Appraisal of methods offish population study - Part IV determination of balance in farm fish ponds. 21st North American Wildlife Conference Ti-ansactions 21:289-322. Taubert, B. D. 1977. Early morphological develop- ment of the green sunfish, Lepomis cyanellus, and its separation from other larval Lepomis species. American Fisheries Society Ti-ansac- tions 106:445-448. von Geldern, C. E., Jr., and D. F. Mitchell. 1975. Largemouth bass and threadfin shad in Cali- fornia. Pages 426-449 in R. H. Stroud and H. Clepper, eds.. Black bass biology and mange- ment. Sport Fishing Insitute, Washington, D. C. Whitt, G. S., D. R Philipp, and W. K Childers. 1977. Aberrant gene expression during the development of hybrid sunfishes (Pcrci/fjrmes, Teleoslei). Differentiation 9:97-109. Wright, G. L., and G. W. Wigtil. 1981. Compar- ison of growth, survival, and catchability of Florida, northern, and hybrid largemouth bass in a new Oklahoma reservoir. Annual Conference of the Southeastern Association of Fish and Wildlife Agencies Proceedings 34:31-38. ZoLCZYNSKi, S. J., and W. D. Davies. 1976. Growth characteristics of the northern and Florida subspecies of largemouth bass and their hybrid, and a comparison of catchability be- tween the subspecies. American Fisheries Society Transactions 105:240-243. Evolution of Reproductive Behaviors in Percid Fishes Lawrence M. Page Percid fishes are freshwater deriva- tives of a marine perciform, perhaps an anadromous serranid (McCully 1962; Collette & Banarescu 1977). They ap- pear to be relatively recent (i.e., most evolved during the Pliocene or later) but have undergone a tremendous radi- ation in North America, where they now comprise about 151 species in five genera. Among North American fresh- water fishes, only minnows (Cyprini- dae), with about 225 species, are more diverse. Together, percids and minnows constitute almost half of the temperate North American freshwater fish fauna. In addition to being diverse gi'oups, these fishes often occur- in large popula- tions and are important in determin- ing the ecological characteristics of North American streams and lakes. Elsewhere, percids are present only in temperate Eurasia, where there are 14 species in six genera. Two genera, Stizostedion and Perca, occur in both North America and Eurasia (Table 1). Several reasons exist for studying the reproductive habits of organisms. Among these are that (1) reproductive habits are inherently interesting; (2) reproduction is that activity toward which all life processes ultimately are directed and, therefore, understanding its variations is important if we wish to protect or manage species; and (3) re- productive characteristics provide in- formation useful in estimating evolu- tionary (phylogenetic) history. Percids are especially interesting in this last regard because, as a recent and diverse gi'oup, transitions among their various types of reproductive behaviors remain evident today. Reconstructing the evolu- tionary history of reproductive behav- iors of percids results in a better under- standing of the relationships among species. ACKNOWLEDGMENTS I am grateful to an anonymous re- viewer and especially to James D. Wil- liams for constructive comments on the manuscript. I wish to thank B. M. Burr, D. A. Etnier, and W. C. Starnes for un- published information on the spawning habits of E. parvipinne, E. (Nanostoma) spp., E. trisella, and E. (Catonotus) sp. TYPES OF REPRODUCTIVE BEHAVIORS Accompanying the evolution of a large diversity of percids has been the evolution of a variety of reproductive Table 1. — Percid taxa and their distributions. Dr. Lawrence M. Page is an Ichthyologist, Sec- tion of Faunistic Surveys and Insect Identifica- tion, Illinois Natural History Survey. Taxa 276 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 strategies, and in the most advanced states, parental care of the eggs. Among the 76 species of percids for which spawning behaviors are known, six types of behavior are recognized, some of which can be subdivided fur- ther, as discussed below. These six types, named for the mode of egg- deposition, are termed (1) broadcast- ing, (2) stranding, (3) burying, (4) at- taching, (5) clumping, and (6) cluster- ing. Broadcasting Broadcasting is the primitive mode of reproductive behavior among fishes and is retained in primitive percids. Eggs and sperm are discharged in large numbers, often in a frenzy of activity involving several individuals. A coarse substrate, usually composed of rocks or plants, is chosen to provide crevices and other hiding places for the fertilized eggs. In its most primitive state, broad- casting makes no other behavioral pro- visions for the eggs, either before or after spawning. Stizostedion vitreum (Mitchill) and iS. canadense (Smith) broadcast their eggs between March and June in streams and in areas of lakes with suf- ficient water movement to ventilate the eggs (Eschmeyer 1950; Nelson 1968). Spawning usually occurs in the even- ing (Ellis & Giles 1965) or at night (Eschmeyer 1950), and territories are not established (Ellis & Giles 1965). Courtship consists of an individual of either sex approaching another indi- vidual from behind or fi'om the side and pushing against it. The approached fish either withdraws or, if sexually re- sponsive, makes quick darts forward and upward. This increased activity often elicits the attention of other indi- viduals, and soon a compact group forms, which periodically rushes for- ward and upward. Eggs and sperm are released during a forward rush of ac- tivity (Ellis & Giles 1965). Spawning groups vary from two individuals to several individuals of both sexes (Fig. 1). Eggs are adhesive for a few hours, and they may adhere to rocks or plants for a short time; ultimately they fall to the bottom and into interspaces among the components of the substrate (Fig. 2), where they presumably are less ex- posed to predation (Colby et al. 1979). S. marinum (Cuvier) and Percarina demidofjft Nordmann appear to exhibit similar, perhaps identical, behavior (Berg 1949) although descriptions of spawning are vague. Stizostedion lucioperca (Linnaeus) exhibits a modification of this behavior in that eggs and sperm are broadcast over a circular pit previously con- structed by the male. The pit is lined with gi-avel, shells, or plant material (often roots) and is guarded before and after spawning by the male, who also fans the fertilized eggs (Kryzhanovshy et al. 1953). Mating in this more special- Fig. 1. - Stizostedion vitreum congregated on spawning gj-ounds at hake Gogebic Micliigan. 4 May 1948. From Esch- meyer (19501. (Photo pro- vided by the Institute for Fisheries Research. Fisheries Division. Mich- igan Department of Nat- ural Resources.) September 1985 125 Years of Biological Research 277 Fig. 2. - EggsofSriz- ostedion vitreuin on spawning ground at Lake Gogebic, Michigan, 12 May 1948. From Esch- meyer (1950). (Photo pro- vided by the Institute for Fisheries Research, Fisheries Division, Mich- igan Department of Nat- ural Resources.) ized, nesting broadcaster may be mon- ogamous (Deedler & Willemsen 1964). Broadcasters Group A. Spawn over open substrate Stizostedion vitreum (Eschmeyer 1950) S. canadense (Nelson 1968) S marinum (Berg 1949) Percarina demidoffi (Berg 1949) Gi'oup B. Spawn over male-constructed pit S. lucioperca (Kryzhanovshy et al. 1953) Stranciing Stranders have the unique habit of encasing their eggs in long gelatinous strands. A ripe female is followed by several males through beds of vegeta- tion (sometimes tree roots or debris) in slowly flowing or standing water. The males release sperm as she extrudes a convoluted egg strand (Worth 1892; Ti-easurer 1981). The strand is gelatin- ous, transparent, hollow, and arranged in bellowslike transverse folds. When first laid, the shape of the egg mass is similar to that of the ovarian cavity (Fig. 3), but eventually it stretches (Fig. 4) and may reach a length of 2.4 m and a width of 10 cm (Hardy 1978). Egg stranding is known to occur in Perca /Zai;esce«s(Mitchill) (Worth 1892; Har- rington 1947; Hergenrader 1969) and P fluviatilus Linnaeus (Seeley 1886:27; Wheeler 1969:322; Treasurer 1981). It also may occur in Gymnocephalus cer- nua (Linnaeus), as described by Seeley (1886:36) and Wheeler (1969:321), al- though some authors (e.g., Nikol'skii 1961:369; Muss 1978:161) discuss spawning in Gymnocephalus without mentioning egg strands. Stranders Gymnocephalus cernua (Seeley 1886) Perca flavescens (Worth 1892) P fluviatilus (Seeley 1886) Stranding is probably a direct deriv- ative of broadcasting. The eggs become encased in gelatin in the ovaries and when emitted remain part of a cohesive mass rather than becoming independ- ently scattered. The derivation of stranding from percid spawning behav- iors other than broadcasting is less likely, because each is specialized in a way that seems to preclude the evolu- tion of stranding. For a female to put all of her eggs in one highly visible basket, which stranding certainly does, and then abandon them would seem to be a stra- tegy quickly selected against unless the eggs somehow are protected intrin- 278 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 -P '-'W- fS^^-^<- 1 T 1 Fig. 3. - Eggs of Perca flauescens encased in ovaries (left) and in gelatinous strands after being spawned (right). From Worth (1892). sically against predation and disease. Having toxic, or at least distasteful, eggs would be an obvious strategy, but observations by Seeley (1886:27) of predation by "birds and various fishes" on perch eggs put this method in doubt. Treasurer (1983) reported low mortality of perch eggs and attributed it to the improved ventilation and midwater position (thereby avoiding siltation) of the strand. Burying Egg-burying behavior is similar to broadcasting except for the important distinction that the release of eggs oc- curs just below, rather than above, the surface of the substrate. In egg bury- ing the female works her body partially below the surface of the substrate and, with her genital papilla buried and a male mounted on her back, expels eggs (Fig. 5). The substrates usually used are loose gi'avel, sand, or mixed gi-avel and sand. Egg burying is characteristic of many darters, including all species of Perciria for which spawning observa- tions have been published and many species ofEtheostoina. Some egg buriers are thought to be territorial (Winn 1958), but other than the indirect pro- tection associated with territoriality, eggs receive no parental care. Among darters, burying represents the primi- September 1985 125 Years of Biological Research 279 Fig. 4. - Egg strand of Perca flavescens. From Pearse & Achtenberg (1921). Fig. 5. - Egg-burying behavior. With a male mounted on her back, a partially buried female deposits eggs below the surface of the substrate. Eggs are fertilized as they are laid and then are abandoned by both parents. A B Fig. 6. - Egg-attaching behavior (A) The female selects the site of egg deposition. (B) The male mounts the female and releases sperm as eggs are attached to plants. The eggs subsequently are abandoned. 280 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 tive form of reproductive behavior; among percids, it represents an egg- hiding modification of broadcasting behavior. Buriers Percina (Alvordius) maculata (Petravicz 1938) P. (A.) peltata (New 1966) P (A.) notogramma (Loos & Woolcott 1969) 7? (Ericosrna) evides (Page et al. 1982) P (Hypohomus) aurantiaca (Howell 1971) P (Cottogaster) copelandi (Winn 1953) P (Percina) caprodes (Winn 1958) Etheostoma (Litocara) nianguae (Pflieger 1978) E. (Psychromaster) tuscumbia (Koch 1978) E. (Etheostoma) tetrazonum (Pflieger 1978) E. (E.) variatum (May 1969) E. (Doration) stigmaeum (Winn 1958) E. (Nothonotus) juliae {James 1983) E. (N.) rufilineatum (Stiles 1972) E. (N.) camurum (Mount 1959) E. (N.) bellum (W D. Voiers personal communication) E. (N.) tippecanoe (Trautman 1981) E. (Fuscatelum) parvipinne (B. M. Bun- personal communication) E. (Ozarka) cragini (Distler 1972) E. (Oligocephalus) spectabile {Vfinn 1958) E. (O.) caeruleum (Winn 1958) E. (O.) radiosum (Scalet 1973) E. (O.) swaini (Ruple et al. 1984) Attaching Attaching behavior is a derived be- havior known among percids only in the genus Etheostoma. The female se- lects the site of egg deposition (presum- ably in some species within a male's territory), typically a plant or large rock, and with the male following her, elevates to the site. As she does so, the male follows and mounts, the two vi- brate, and eggs and sperm are released (Fig. 6). Usually one to three adhesive eggs are released during each spawn- ing act and are pushed by the female onto a plant, rock, or other object. At- taching eggs to plants seems to involve behavior identical to that of attaching eggs to rocks; the substrate used is that which is I'eadily available in the habi- tat. Eggs are abandoned and receive no direct parental care, although as in some burying species, territorial be- havior may provide some protection. Attachers Etheostoma (Etheostoma) blennioides (Fahy 1954) E. (Nanostoma) zonale (Winn 1958) E. (N.) coosae (O'Neil 1981) E. (N.) baileyi (Page personal observa- tion) E. (N.) simoterum (Page & Mayden 1981 1 E. (N.) duryi (Page et al. 1982) E. (N.) barrenense (Winn 1958) E. (N.) rafinesquei (Winn 1958) E. (N.) sp. (Red Snubnose) (B. M. Burr personal communication) E. (N.) sp. (Lowland Snubnose* (B. M. Burr personal communication) E. (loa) vitreum (Winn & Picciolo I960) E. (VaiUantia) chlorosomum (Page et al. 1982) E. (Belophlox) okaloosae (Collette & Yerger 1962) E. (Villora) edwini (Williams 1976) E. (Ozarka) boschungi (Qoschnng 1979) E. (O.) trisella (W. C. Starnes personal communication) E. (Oligocephalus) lepidum (Strawn 1956) E. (O.) asprigene (Page et al. 1982) E. (O.) grahamj (Strawn 1956) E. (O.) ditrema (Seesock et al. 1978) E. (Boleichthys) exile (Winn 1958) E. (B.I fusiforme (Fletcher 1957) E. (BJgracile {Braasch & Smith 1967) E. (B.) pi-oeliare (Burr & Page 1978) E. (B.) fonticola (Schenck & Whiteside 1977) E. (B.) microperca (Burr & Page 1979) The presence of both egg-burying and egg-attaching species within three subgenera o{ Etheostoma (Etheostoma, Ozarka, and Oligocephalus) suggests that egg attaching is derived directly from egg burying and has arisen inde- pendently in several unrelated gi'oups September 1985 125 Years of Biological Research 281 of darters. In fact, at least two species which usually attach their eggs have been observed to bury eggs under cer- tain conditions (Petravicz 1936; Page & Mayden 1981). Egg attaching appar- ently also has arisen independently among species within subgenera; its distribution among species of Ozarka is inconsistent with the distribution of morphological synapomorphies (Fig. 7). 282 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 the species from extinction. Suitable spawning sites may be a limiting factor for the only extant population of E. sellare Clumping Egg clumping was first reported in 1939 for E. maculatum Kirtland (Eaney & Lachner 1939) but was unreported for other darters until observed in 1981 in E. aquali Williams and Etnier and E. microlepidum Raney and Zorach (Page et al. 1982). In preparation for spawning, the male selects a cavity under a large rock as a territory and future nesting site. Ultimately, a fe- male swims into the cavity being guarded by the male, wedges herself into the interface between the stone and the gravel substrate beneath the stone, and deposits eggs. As she does so, they are fertilized by the male. After spawning, the female leaves, and the male remains to guard the clump of eggs (Fig. 8). The adhesive eggs adhere to both the nest stone and the under- lying substrate material; if a stone is lifted from the water, a clump of adhe- sive eggs remains attached to it (Fig. 9). Egg clumping is known only among species of the subgenus Nothonotus of Etheostoma. Clumpers Etheostoma (Nothonotus) maculatum (Raney & Lachner 1939) E. (N.) aquali (Page et al. 1982) E. (N.) microlepidum (Page et al. 1982) Egg clumping is almost certainly a direct derivative of the egg-burying be- havior (Page et al. 1982) characteristic of other species of Nothonotus [known in E. juliae Meek, E. rufilineatum (Cope), E. camurum (Cope), and E. tip- pecanoe Jordan and Evermann]. The transition from burying in some spe- cies of Nothonotus to clumping in others requires only that the male establish a territory beneath a stone and that a female deposit her eggs in the interface area between the stone and the under- lying substrate. ^2^ Fig. 8. - Egg-clumping behavior. (A) The male establishes a territory centered under a large stone. (B) The female enters the territory and wedges herself between the stone and the underlying substrate. (C)The male positions himself next to the female and fertilizes eggs as they are released by the female. (Dl The male remains to guard the clump of eggs. With 14 described and 1 undescribed species, the subgenus Nothonotus is one of the two largest subgenera of darters, and intrasubgeneric relationships are unclear. The egg-clumping species are members of the £. maculatum lineage within the subgenus (Fig. 101, but it is unknown at present at what point on the lineage egg clumping arose. E. rufilineatum has been observed to bury its eggs (Stiles 1972), and although spawning was not observed, Bryant (1979) felt confident that E. acuticeps Bailey also buries its eggs. Spawning behavior data on E. rubrum Raney and Suttkus and E. moorei Raney and Suttkus should pinpoint the origin of egg clumping in Nothonotus. Although Etheostoma sanguifluum (Cope) has been considered a subspecies of E. maculatum Kirtland by some September 1985 125 Years of Biological Research 283 f > <^- • si- Fig. 9. - A male Etheostoma microlepidum and his clump of eggs on the underside of a stone removed from East Fork Stones River, Rutherford County, Tennessee, on 6 May 1981. The male was guarding the eggs prior to their removal from the stream. authors (Zorach & Raney 1967; Wil- liams & Etnitr 1978), the distinctive pigmentation of the first dorsal fin (especially the bright red spots at the front and rear of the fin) of the male of E. sanguifluum is shared with E. aquali Williams and Etnier but is ab- sent in all other species of the sub- genus Nothonotus. Recognition of a closer relationship between E. sangui- fluum and E. aquali, than of either taxon to E. maculatum (Fig. 10), re- quires the elevation of £. sanguifluum to specific status. The form heretofore known as E. maculatum vulneratum (Cope) also possesses the distinctive dorsal fin pigmentation and should be named E. sanguifluum vulneratum (Cope). Clustering The first report of egg-clustering behavior in darters was by Seal in 1892 in a report on E. olmstedi Storer. Since then it has been documented in an additional 15 species in the subgenera Boleosoma and Catonotus ofEtheostoma listed below. The male of an egg-clustering spe- cies establishes a territory centered about the cavity under a large (usually fiat) stone. The cavity and the under- side of the stone are cleared of silt and debris by fin-wagging activities of the male. A ripe female enters the cavity and, following courting by the male, rolls to one side and, once inverted, rises and lays eggs on the underside of the stone. The male inverts and ferti- lizes the eggs, which are arranged by the female in a single-layer cluster (i.e., eggs are rarely laid on top of one another) on the stone. The female leaves; other females sequentially may add eggs to the nest. The male remains and guards the eggs to hatching (Fig. 11). In species o{ Boleosoma, a log may be substituted for a stone. Egg clustering is similar to egg clumping in that eggs are amassed under a stone and guarded by the male. 284 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 E September 1985 125 Years of Biological Research 285 Clusterers ,J D Fig. 11. - Egg-clustering behavior. (A) The male establishes a territory centered under a large stone. (B) The female enters the territory and selects a site for egg deposition on the under- side of the stone. (CiThe female inverts, rises, and deposits eggs on the stone; as she does so, the male follows and fertilizes the eggs. (D) The male remains to guard the cluster of eggs. and fertilize each release of eggs (usually 1-5 eggs); (B) those in which the female, once inverted, remains in- verted for a prolonged period of egg lay- ing (several minutes to a few hours) but the male only briefly (a few seconds) and periodically inverts; and (C) those in which both male and female remain inverted for an extended period. Ob- viously, gi'oup B represents an advance- ment over group A, and group C rep- resents an advancement over group B. Group A includes the primitive species of the subgenus Catonotus; B, the ad- vanced species of Catonotus (Page 1975b); and group C, species of the subgenus Boleosoma. Winn (1958) distinguished between behavior types B and C. Group A. Female and male briefly invert E. (Catonotus) squamiceps (Page 1974) E. (C.) olivaceum (Page 1980) E. (C) neopterum (Page & Mayden 1979)^ Group B. Female only has prolonged inversion E. (C.) flabeUare (Winn 1958) E. (C.) sp. = Duskytail darter (Etnier personal communication)'' E. (C.) kennicotti (Page 1975a) E. (C) obeyense (Page et al. 1981)^ E. (C.) virgatum (Kornman 1980)=> E. (C.) smithi (Page & Burr 1976) E. (C.) striatulum (Page 1980)^ E. (C.) barbouri (Page et al. 1982)* Group C. Female and male have pro- longed inversions E. (Boleosoma) olmstedi (Atz 1940) E. (B.) nigrum (Winn 1958)^ E. (B.) podostemone (Jenkins 1980)^* E. (B.) perlongum (Lindquist et al. 1981) E. (B.) longimanum (Page et al 1981)=» ^ Although known to be a clusterer, group assignment is predicted from the behavior of closest relatives and is not based on actual observations. The phylogenetic sequence of breed- ing behaviors among Catonotus is con- cordant with a phylogeny of the subgenus based on morphology in that three synapomorphies unite the E. flabeUare Rafinesque lineage as ad- vanced and apart from the E. squami- ceps Jordan species group (Fig. 13). Both Catonotus and Boleosoma must have evolved from egg-attaching pre- cursors. Although egg clustering itself could be considered a synapomorphy uniting these two subgenera, they dif- fer markedly in their morphology and do not appear to be sister groups (Page 1981). Boleosoma shares derived char- acteristics with other groups of Ethe- ostoma (Fig. 14) and almost certainly had an origin independent from that of Catonotus. It appears to be most closely related to E. (Ion) vitreum (Cope), an 286 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 egg-attaching species. The sister-group of Catonotus should first be sought among egg-attaching species; however, if the attaching precursor now is ex- tinct, an egg-burying group may be the closest extant relative. EVOLUTION OF PERCID SPAWNING BEHAVIORS Combining the evolutionarj' sequen- ces discussed above produces the phyl- / Fig. 12. - Male Etheosloma neoplenim and his cluster of eggs on the undei-side of a stone i-emoved from Birdsong Creek, Benton County, Tfennessee, on 15 April 1978. The male was guarding the eggs prior to their removal from the stream. September 1985 125 Years of Biological Research 287 00 288 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 hybrids. A phylogenetic classification necessitates the recognition of the sub- families Etheostomatinae (containing the tribes Luciopercini, Etheostomatini, and Romanichthyini) and Percinae (Table 1). Each step in the evolution of spawn- ing behaviors represents, for the spe- cies in which it evolved, a survival advantage over its precursor Pit broad- casting concentrates the eggs in an area subsequently guarded by the male Doration Valllantia loa Boleosoma O) September 1985 125 Years of Biological Research 289 Fig. 15. - Hypothesized evolutionary relationships among percid breeding be- CluiTIDine haviors. y - Clustering t p - Clustering t a - Clustering Attaching Burying pit-Broadcasting open-Broadcasting Stranding and thus increases the probability of their survival by reducing predation. Also, S lucioperca males fan the eggs, presumably to pi'event them from be- ing covered by silt, and thereby reduce mortality due to anoxia. Burying eggs reduces exposure to predators and probably to parasites, enhancing survivability. However, eggs can be buried only in flowing water habitats, usually in gravel but some- times in sand, where burying does not subject the eggs to especially low oxy- gen levels such as would occur if eggs were buried in slow-flowing or standing water. Buriers are constrained behav- iorally to flowing water and coarse substrates. Consequently, stream mod- ifications, such as impoundments and channelization, prevent successful spawning by these species. In the primitive state, species of Nothonotus bury their eggs in the sand- gravel mixture on the downstream side of a large stone in fast water. Clump- ing, as discussed above, evolved in a species of Nothonotus which buried its eggs under a partially elevated stone, thereby enhancing crypticity. The ad- vantage of male egg guarding followed. The derivation of attaching from burying (or from any other behavior) is difficult to envision. However, the presence of both behaviors within several (at least three) subgenera of Etheostoma leaves no doubt that one is derived directly from the other The strong positive correlation between primitive morphology and burying (e.g., all species oi Percina are buriers) and between advanced morphology and attaching (e.g., in the subgenus Boleichthys) clearly indicates that the direction is from burying to attaching. For darters living in slow-flowing or standing water (e.g., Boleichthys species), the advantage of attaching is obvious; with demersal and adhesive eggs, the only alternatives to attaching are bury- ing eggs in an oxygen-poor substrate (often rotting vegetation and mud) or migi'ating to suitable habitats for bury- ing. Similarly, species living in bedrock pools [e.g., E. simoterum (Cope)] where 290 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 little suitable egg-burying substrate ex- ists, were probably better fit after they changed to egg attaching (Page & Mayden 1981). That so many species have converted from burying to at- taching (at least 9 of the 13 subgenera of Etheostoma for which the spawning behavior of at least one species is known, are known to contain egg- attaching species) seems to reflect the fact that few alternatives exist. Clum- ping is an alternative, but it was available evolutionarily only to species which bury their eggs near (and even- tually under) large stones. Among dar- ters, clumping evolved only in Nothon- otus; other fishes with behavior similar to egg clumping in darters are species of Cottus (Smith 1923) and Noturus (Mayden & Burr 1981). The evolution of egg attachment permitted groups of darters to invade new habitats (e.g., bedrock pools, sand and mud bottomed pools, swamps) that otherwise could have been invaded only with periodic (presumably annual) emigi'ations into other habitats suitable for egg burying. This emigration would be difficult for advanced darters, which lack gas blad- ders. A few darters have followed the latter strategy [e.g., Percina cymato- taenia (Gilbert and Meek), a primitive darter which has a gas bladder and lives in vegetated backwaters, moves into riffles to spawn (Pflieger 1975: 299)]; however, the gi'oups which live and spawn in the same habitat have been most successful in diversifying (e.g., Boleichthys, Nanostoma). Clustering darters are sophisticated attachers which concentrate their eggs in a hidden space and then guard them through hatching. The single-layer ar- rangement of eggs permits the guard- ing male to tend each egg. Although the close proximity of the eggs pro- motes the spread of infection, and of predation if a predator succeeds in gaining access to the nest, the advan- tages of clustering apparently out- weigh the disadvantages. Further- more, males of both egg-clustering darters and egg-clustering minnows (Pimephales) develop swollen flesh on the head and nape during the breeding season. The swollen flesh is thought to have a secretory, presumably fungi- cidal or bactericidal, function related to protecting the eggs during nest guarding (Cross 1967; Smith & Mur- phy 1974). Clustering darters and min- nows are some of the most widespread and abundant fishes [e.g., Etheostoma nigrum Rafinesque, E. olmstedi, E. flabellare, Pimephales notatus (Rafmes- que), and P promelas Rafinesque] in North America. Two earlier discussions on the di- versity and evolution of percid repro- ductive behaviors were those of Winn (1958) and Balon et al. (1977). The lat- ter assigned percids to seven reproduc- tive "guilds" that conceptually encom- passed both spawning and ontogenetic characteristics. Species that herein are classified as open substrate broadcast- ers, stranders, and attachers are vari- ously classified by Balon et al. (1977) as open substrate lithophils (rock spawn- ers), phytophils (plant spawnersi, and phyto-lithophils. Buriers are termed brood hiding lithophils, and dumpers and clusterers are nest spawning spe- leophils (cave spawners). McElman (1983) argued subsequently that S. vitreum and S. canadense should be classified as litho-pelagophils because of the buoyancy, positive phototaxis, swimming ability, and surface suspen- sion properties of their newly hatched young, and that S. luciopei-ca should be termed a phyto-pelagophil because of the swimming ability of its newly hatched young. Page & Swofford (1984) argued that Balon et al. (1977) assigned some darters [i.e., P caprodes (Rafin- esque), E. blcnnioides. E. vitreum, and possibly P shumardi (Girard) and Am- mocrypta spp.] to guilds incorrectly. Winn (1958) dealt exclusively with darters and concentrated primarily on the 14 species for which he presented original data. He recognized the least complex behaviors as primitive and the most complex (especially clustering) as derived, and discussed these evolution- September 1985 125 Years of Biological Research 291 ary trends; development of territoriali- ty; use of substrates other than gravel for spawning; occupation of smaller streams; maturation of fewer, larger eggs; decrease in promiscuity; develop- ment of parental care; an increase in sexual dimorphism (including a rela- tive increase in the size of the male); a decrease in body size; and a decrease in the age of sexual maturity from 2-3 years to 1 year. With data now avail- able on many more species, these trends still seem accurately to characterize darter evolution, except that the only darters thought to be nonterritorial (and the only species in which females are larger than males) are species of ad- vanced (i.e., Ozarka and Boleichthys), not primitive, subgenera, and sexual dimorphism in color is most extreme among egg buriers (Page 1983; Page and Swofford 1984), behaviorally the most primitive darters. LIFE HISTORY CORRELATES OF REPRODUCTIVE SPECIALIZATIONS Some of the morphological corre- lates of the various reproductive behav- iors of darters are discussed by Page & Swofford (1984). Life history character- istics also show trends correlated with reproductive specialization. Some trends are in the direction of r-selected characteristics, and some are in the direction of k-selected characteristics (MacArthur & Wilson 1967; Pianka 1978:122). One dramatic trend accompanying the evolution of percids is to lay fewer, larger eggs (a k-selected trait). Species of Stizostedion lay up to 2,500,000 eggs (Filuk 1962) averaging 0.55-1.10 mm in diameter (Kryzhanovshy et al. 1953). Species of Perca lay up to 210,000 eggs averaging 1.0-2.1 mm (Thorpe 1977). Darters lay only up to a few thousand eggs averaging 0.7-2.7 mm (Page 1983:162). The largest eggs among per- cids are those of the highly evolved egg- clustering species in which parental care of the eggs warrants a large ex- penditure in the production ofeach egg. Incubation periods seem to have changed little. Stizostedion and darters take about 25-30 days to hatch at 10 °C and 8-10 days at 20 °C (Marshall 1977; Page 1983:168). In a more r-selected mode, advanced percids (darters) I'eproduce early (most at 1 year, a few at 2 years), whereas species of Stizostedion do not reach maturity until 2-6 years in the male and 3-6 years in the female (Marshall 1977). Longevity (r-selected) ranges from 20 years in S. vitreum (Scott & Grossman 1973:772) to only 1.5 years in several species of Etheostoma (Page 1983:169). Some of the most advanced species of darters have become semel- parous; they mature at 1 year, repro- duce, and die before a second spawning season (Page 1983:169). Semelparous species include egg attachers [E. micro- perca Jordan and Gilbert and E. pro- eliare (Hay)] and egg clusterers [E. striatulum Page and Braasch). Some darters exist in large dense populations - to 33 individuals per square meter (Page 1983:170-171); the larger piscivorous percids exist in much smaller populations. Decreasing body size (r-selected) has been a pervasive element of percid evolution and was discussed at length for darters by Page & Swofford 1984). In brief, darters feed on small benthic organisms and have become more effi- cient consumers by becoming smaller. Gonstraints on decreasing body size in- clude living in gravel runs (correlated with egg burying); the smallest darters are pool inhabitants, including many egg-attaching and egg-clustering spe- cies. S. vitreum reaches a maximum length of about 107 cm and a maxi- mum weight near 11.5 kg(Golby et al. 1979). Those values are 30 times the maximum length and 19,000 times the maximum weight of E. fonticola (Jor- dan and Gilbert), the smallest species of darter at a maximum length of 3.5 cm and weight of 0.6 g (Page & Burr 1979). 292 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 SUMMARY About one-fifth of all North Amer- ica freshwater fishes ai'e percids (the walleye, sauger, yellow perch, and 148 species of darters). In addition to being diverse, percids often occur in large populations and are important in deter- mining the ecological characteristics of North American streams and lakes. Elsewhere percids occur only in tem- perate Eurasia (14 species). Knowing the reproductive requirements of these fishes is fundamental to understanding their positions in aquatic environ- ments, protecting them, managing them, and predicting the impacts of proposed environmental alterations, such as impoundments and dredging. Accompanying the evolution of a large diversity of percids has been the evolution of a variety of reproductive strategies. Among the 71 species of per- cids for which spawning behaviors are known, six types of behavior are recog- nized: broadcasting (including pit broad- casting), stranding, burying, attaching, clumping, and clustering (including alpha beta, and gamma clustering). Broadcasting is the discharging of large numbers of eggs and sperm, usually over rocks or plants. It is the most primitive type of reproductive be- havior among fishes and is character- istic of especially primitive percids (Stizostedion vitreum, S. canadense, and probably S. inarinum and Percarina demidofji). In pit broadcasting, found in S. lucioperca, eggs are broadcast over a pit constructed and guarded by the male. Stranding is the unique habit of en- casing eggs in long gelatinous strands. Stranding appears to be a derivative of broadcasting and is known in Perca and Gymnocephalus. Burying is a derivative of, and is similar to, broadcasting, except that eggs are released just below rather than above the surface of the substrate. Among percids burying is restricted to darters. Among darters it is the primi- tive reproductive behavior and is known in 12 subgenera of Percina and Etheostoma. Attaching behavior involves depos- iting eggs individually on plants or rocks over a wide area. It is derived from burying and, among percids, is known only in the genus Etheostoma. Attaching is found in at least nine subgenera and obviously has evolved independently several times. Clumping is amassing eggs under a large stone, which subsequently is guarded by the male. It also is derived from burying but has evolved only within the subgenus Nothonotus. At present it is known in E. maculatum, E. aquali, and E. microlepidum. Clustering is found in two unre- lated subgenera, Boleosoma and Cat- onotus, of Etheostoma. It is a highly evolved derivation of attaching behav- ior in which eggs are concentrated in a single-layer cluster on the underside of a stone and are guarded by the male. Clustering evolved independently in Boleosoma and Catonotus. Estimating the evolutionary his- tory (phylogeny) of reproductive behav- iors results in a better understanding of the relationships among percids. The phylogeny of breeding behaviors sup- ports the contention that darters, the most primitive ofwhich are bm-iei-s, ai-e derivatives of, and among the species discussed here, probably the sister gi'oup of Stizostedion. A phylogenetic classification necessitates recognition of the subfamilies Etheostomatinae (containing the tribes Luciopercini, Etheostomatini, and Romanichthyini') and Peixinae. The unique pigmenta- tion of the first dorsal fin (bright red spots at the front and rear of the fin) of males of Etheostoma sangitifluum (Cope) and E. aquali Williams and Etnier is a synapomorphy indicating that the two are sister taxa. Thus, E. sa/igiiifluiim cannot remain a subspe- cies of E. maculatum Kirtland and is recognized here as a species. The sub- species known as E. maculatum vulner- atum (Cope) becomes E. sanguifluum vulneratum (Cope). September 1985 125 Years of Biological Research 293 Life-history correlates of reproduc- tive specialization among percids in- clude the production of fewer, larger eggs; earlier reproduction; semelpari- ty; reduced longevity; denser popula- tions; and reduced body size. LITERATURE CITED Atz, J. W. 1940. Reproductive behavior in the eastern johnny darter, Boleosoma nigrum olmstedi (Storer). Copeia 1940:100-106. Balon, E. K., W. T. Momot, and H. R. Regier. 1977. Reproductive guilds of percids: results of the paleogeographical history and ecolog- ical succession. Fisheries Research Board of Canada Journal 34:1910-1921. Berg, L. S. 1949. Freshwater fishes of the U.S.S.R. and adjacent countries. Vol. 3. (English ti-ans- lation published by Isreal Progi-am for Scien- tific Ti-anslations, Jerusalem, 1965). Boschung, H. 1979. Report on the breeding hab- its of the slackwater darter (Percidae: Etheo- stoma boschungi) in the Cypress Creek water- shed. Submitted to U.S. Department of Agi"iculture Soil Conservation Service Au- burn, Alabama. Braasch, M. E., and P W. Smith 1967. The life history of the slough darter, Etheostoma gracile (Pisces, Percidae (. Illinois Natual History Survey Biological Notes 58. Brigham, A. R., W. U. Brigham, and A. Gnilka. 1982. Aquatic insects and oligochaetes of North and South Carolina. Midwest Aquatic Enterprises, Mahomet, IL. Bryant, R. T. 1979. The life history and compara- tive ecology of the sharphead darter, Elhe- ostoma acuticeps. Tennessee Wildlife Re- sources Agency Technical Report 79-50. Burr, B. M., and L. M. Page. 1978. The life history of the cypress darter, Etheostoma pro- eliare, in Max Creek, Illinois. Illinois Natural History Survey Biological Notes 106. , and . 1979. The life history of the least darter, Etheostoma microperca, in the Iroquois River, Illinois. Illinois Natural His- tory Survey Biological Notes 112. Colby, P J., R. E. McNicol, and R. A. Ryder. 1979. Synopsis of biological data on the wal- leye Stizostedion v. vitreum (Mitchell 1818). Food and Agricultural Organization of the United Nations Fisheries Synopsis 119. CoLLETTE, B. B. 1963. The subfamilies, tribes, and genera of the Percidae (Teleostei). Copeia 1963:615-623. , and P. P. Banarescu. 1977. Systematics and zoogeography of the fishes of the family Percidae. Fisheries Research Board of Canada Journal .34:1450-1463. and R. W. Yerger. 1962. The American percid fishes of the subgenus Villora. Tulane Studies in Zoology 9:213-230. Cross, F. B. 1967. Handbook of fishes of Kansas. University of Kansas Museum of Natural His- tory Miscellaneous Publications 45:1-357. Deedler, C. L., and J. Willemsen. 1964. Synop- sis of biological data on pike-perch Lucioper- ca lucioperca (Linnaeus) 1758. Food and Agi'icultural Organization of the United Na- tions Fisheries Synopsis 28. DiSTLER, D. A. 1972. Observations on the repro- ductive habits of captive Etheostoma cragini Gilbert. Southwestern Naturalist 16:439-441. Ellis, D. V., and M. A. Giles. 1965. The spawn- ing behavior of the walleye, Stizosledion vitreum (Mitchell). American Fisheries Socie- ty Ti-ansactions 94:358-362. Eschmeyer, R H. 1950. The life history of the walleye, Stizostedion vitreum vitreum (Mitch- ill), in Michigan. Michigan Department of Conservation Institute for Fisheries Research Bulletin 3. Fahy, W. E. 1954. The life history of the northern gi-eenside darter, Etheostoma blennioides blen- nioides Rafinesque. Elisha Mitchell Scientific Society Journal 70:139-205. FiLUK, J. 1962. Nachkriegsstudium uber Biologie und Fang des Zanders des Frischen Haffs. Zeitschrift fuer Fischerei und deren Hilfswis- senschaften 10:705-709. Fletcher, A. M. 1957. A rare darter-spawning. Aquarium 26:202-203. Hardy, J. D., Jr. 1978. Development of fishes of the Mid-Atlantic Bight. Vol. III. Aphredoderi- dae through Rachycentridae. U. S. Fish and Wildlife Service Biological Services Progi-am FWS/OBS-78/12. Harrington, R. W, Jr. 1947. Observations on the breeding habits of the yellow perch, Perca flaveseens (Mitchill). Copeia 1947:199-200. Hergenrader, G. L. 1969. Spawning behav- ior of Perca flaveseens in aquaria. Copeia 1969:839-841. Howell, J. F, Jr. 1971. The life history, behavior, and ecology of Percina aurunliaca (Cope). Ph.D. Dissertation. University of Tennessee, Knoxville. HuBBS, C. 1971. Survival of intergroup hybrids. Japanese Journal of Ichthyology 18:65-75. Jaivies, P W. 1983. The life history of the yoke darter, Etheostoma juliae. Master's Thesis. Southwest Missouri State University, Springfield. Jenkins, R. E. 1980. Ethei>stt>ma podoslenume Jor- dan and Jenkins, Riverweed darter Page 682 in D. S. Lee, C. R. Gilbert. C. H. Hocutt, R. E. Jenkin.s, D. E. McAllister, and J. R. Stauf- fer, Jr., eds., Atlas of North American freshwater fishes. North Carolina State Museum of Natural History, Raleigh. 294 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Knapp, L. W. 1976. Redescription, relationships and status of the Maryland darter, Etheos- toma seltare (Radcliffe and Welsh), an endangered species. Biological Society of Washington Proceedings 89:99-117. Koch, L. M. 1978. Aspects of life history of Etheostoma (Psychromaster) tuscumbia in Buf- fler Spring, Laudei'dale County, Alabama. M. S. Thesis. University of North Alabama. Florence. KoRNMAN, L. E. 1980. Life history aspects of the striped darter, Etheostoma virgatum (Jordan) 1880, from Clear Creek, Rockcastle County, Kentucky. M. S. Thesis. Eastern Kentucky University, Richmond. Kryzhanovshy, S. G., N. N. Disler, and E. N. Smirnova. 1953. Eco-morphological principles of development in percoid fishes (Percoidei). Trudy Instituta Morfologii Zhivotnykh Aka- demii Nauk SSSR 10:3-138. LiNDQUiST, D. G., J. R. Shute, and P. W. Shute. 1981. Spawning and nesting behavior of the Waccamaw darter, Etheostoma perlongum. Environmental Biology of Fishes 6:177-191. Loos, J. J., and W. S. Woolcott. 1969. Hybrid- ization and behavior in two species o{ Percina (Percidae). Copeia 1969:374-385. MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography. Princeton Uni- versity Press, Princeton, NJ. Marshall, T. R. 1977. Morphological, physio- logical, and ethological differences between walleye (Stizostedion vitreum vitreum) and pikeperch (S. lucioperca). Fisheries Research Board of Canada Journal 34:1515-1523. May, B. 1969. Observation on the biology of the variegated darter, Etheostoma variatum (Kirt- land). Ohio Journal of Science 69:85-92. Mayden, R. L., and B. M. Burr. 1981. Life history of the slender madtom, Noturus exilis, in southern Illinois (Pisces: Ictaluridae). Univer- sity of Kansas, Museum of Natural History Occasional Papers 93. McCully, H. H. 1962. The relationship of the ftr- cidae and the Centrai'chidae to the Serranidae as shown by the anatomy of their scales. American Zoologist 2:430. McElman, J. F. 1983. Comparative embryonic ecomorphology and the reproductive guild classification of walleye, Stizostedion ritreum, white sucker, Catostomiis commersoni. Copeia 1983:246-250. Mount, D. I. 1959. Spawning behavior of the blue- breast darter, Etheostoma camurum (Cope). Copeia 1959:240-243. Muss, B. J. 1978. Freshwater fish of Britain and Europe. Collins Clear-Type Press, London and Glasgow. Nelson, W. R. 1968. Reproduction and early life history of sauger, Stizostedion canadense, in Lewis and Clark Lake. American Fisheries Society Ti'ansactions 97:159-166. New, J. G. 1966. Reproductive behavior of the shield darter, Percina pettata peltata, in New York. Copeia 1966:20-28. Nikol'skii, G. V. 1961. Special ichthyology. Israel Program for Scientific Translations, Jerusalem. O'Neil, R E. 1981. Life history oi Etheostoma coosae (Pisces: Percidaei in Barbaree Creek, Alablama. Tulane Studies in Zoologj' and Bot- any 23:75-83. Page, L. M. 1974. The life history of the spottail darter, Etheostoma squamiceps, in Big Creek, Illinois, and Ferguson Creek, Kentucky. Illi- nois Natural History Survey Biological Notes 89. 1975a. The life history of the stripetail darter, Etheostoma kennicotti, in Big Creek, Illinois. Illinois Natural History Siu^ey Bio- logical Notes 93. 1975b. Relations among the darters of the subgenus Catonotus of Etheostoma. Copeia 1975:782-784. 1980. The life histories of Etheostoma olivaceum and Etheostoma striatulum, two species of dailers in central Tennessee. Illinois Natural History Survey Biological Notes 113. 1981. The genera and subgenera of dar- ters (Percidae, Etheostomatini). University of Kansas, Museum of Natural Histor\' Occa- sional Papers 90. 1983. Handbook of dartere. T. F H. Publi- cations, Inc.. Neptune City. NJ. and B. M. Burr. 1976. The life history of the slabrock darter. Etheostoma smithi, in Ferguson Creek. Kentucky. Illinois Natural History Survey Biological Notes 99. , and . 1979. The smallest species of dai-ter (Pisces: Percidae). Amei'ican Midland Naturalist 101:452-453. , W. L. Keller, and L. E. Cordes. 1981. Etheostoma iBoleosomal longimanum and£. (Catonotus) obeyense, two moi'e darters con- firmed as egg-clusterers. Kentucky Academy of Science Ti'ansactions 41:35-36. , and R. L. Mayden. 1979. Nesting site of the lollypop darter. Etheostoma neopterum. Kentuckv Academv of Science Ti-ansactions 40:56-57. and . 1981. The life history of the Tfennessee snubnose darter. Etheostoma simo- terum. in Brush Creek. Tennessee. Illinois Natural History Survey Biological Notes 117. M. E. Retzer, and R. A. Stiles. 1982. Spawning behavior in seven species of dartere (Pisces: PercidaeK Brinileyana 8:135-142. , and D. L. Swofford. 1984. Morpholog- ical correlates of ecological specialization in darters. Environmental Biology of Fishes 11:139-159. Pearse, a. S.. and H. Achtenberg. 1921. Habits of yellow perch in Wisconsin lakes. U. S. Bureau of Fisheries Bulletin 36(1917-1918):293-366. Petravicz. J. J. 1936. The breeding habits of the least darter. Microperea punctulata Putnam. Copeia 1936:77-82. Petravicz. W. R 1938. The bi-eeding habits of the black-sided darter. Hadroptcrus maciilaliis Girard. Copeia 1938:40-44. September 1985 125 Years of Biological Research 295 Pflieger, W. L. 1975. The fishes of Missouri. Missouri Department of Conservation. 1978. Distribution, status, and life his- tory of the Niangua darter, Etheostoma riian- gtiae. Missouri Department of Conservation Aquatic Series 16. Pianka, E. R. 1978. Evolutional^ ecology. 2nd ed. Harper & Row, Publishers, New York. Raney, E. C, and E. A.Lachner. 1939. Observa- tions on the life history of the spotted darter, PoecUichthys maculatus (Kirtland). Copeia 1939:157-165. RuPLE, D. L., R. H. McMichael, Jr., and J. A. Baker. 1984. Life history of the gulf darter, Etheostoma swaini (Pisces: Pfercidae). Environ- mental Biology of Fishes 11:121-130. ScALET, C. G. 1973. Reproduction of the orange- belly darter, Etheostoma radiosum cyanorum (Osteichthys: Percidae). American Midland Naturalist 89:156-165. ScHENCK, J. R., and B. G. Whiteside. 1977. Repro- duction, fecundity, sexual dimorphism, and sex ratio of Etheostoma fonticola (Osteich- thyes: Percidae). American Midland Natural- ist 98:365-375. Scott, W. B., and E. J. Grossman. 1973. Fresh- water fishes of Canada. Fisheries Research Board of Canada Bulletin 184. Seal, W. P. 1892. Observations on the aquaria of the U.S. Fish Commission at Central Sta- tion, Washington, D. C. U. S. Fish Commis- sion Bulletin 10(1890):1-12. Seeley, H. G. 1886. The freshwater fishes of Europe. Cassell & Co., Ltd., London. Seesock, W. E., J. S. Ramsey, and F L. Seesock. 1978. Life and limitation of the coldwater darter {Etheostoma ditrema) in Glencoe Spring, Alabama. Association of Southeastern Biologists Bulletin 25:56. Smith, B. G. 1923. Notes on the nesting habits of Cottus. Michigan Academy of Science, Arts, and Letters Papers 2:221-222. Smith, R. J. F, and B. D. Murphy. 1974. Func- tional morphology of the dorsal pad in fathead minnows {Pimephales promelas Rafinesque). American Fisheries Society Ti-ansactions 103:65-72. Stiles, R. A. 1972. The comparative ecology of three species o{ Nothonotus (Percidae Etheos- toma) in Tbnnessee's Little River Ph.D. Disser- tation. University of Tennessee, Knoxville. Strawn, K. 1956. A method ofbreeding and rais- ing three Texas darters. Part IL Aquarium Journal 27:12-14, 17, 31-32. Thorpe, J. E. 1977. Morphology, physiology, be- havior, and ecology o{ Perca fluviatilis L. and P. flavescens Mitchill. Fisheries Research Board of Canada Journal 34:1504-1514. Trautman, M. B. 1981. The fishes of Ohio. Rev. ed. The Ohio State University Press, Columbus. Treasurer, J. W. 1981. Some aspects of the repro- ductive biology of perch Perca fluviatilis L. Fecundity, maturation and spawning behav- ior. Journal of Fish Biology 18:729-740. . 1983. Estimates of egg and viable embryo production in a lacustrine perch, Perca fluviatilis. Environmental Biology of Fishes 8:3-16. Wheeler, A. 1969. The fishes of the British Isles and north-west Europe. Michigan State Uni- versity Press, East Lansing. Williams, J. D., and D. A. Etnier. 1978. Etheos- toma aquali, a new percid fish (subgenus Nothonotus) from the Duck and Buffalo rivers, Tennessee. Biological Society of Washington Proceedings 91(21:463-471. , and H. W. Robison. 1980. Ozarka, a new subgenus of Etheostoma (Pisces: Percidae). Brimleyana 4:149-156. Williams, J. S. 1976. Spawning behavior oi Ethe- ostoma edwini (Pisces: Percidae). Association of Southeastern Biologists Bulletin 23:107. Winn, H. E. 1953. Breeding habits of the percid fish Hadropterus copelandi in Michigan. Copeia 1953:26-30. . 1958. Comparative reproductive behavior and ecology of fourteen species of darters (Pisces-Percidae). Ecological Monographs 28:155-191. , and A. R. Picciolo. 1960. Communal spawning of the glassy darter Etheosto/na vitreum (Cope). Copeia 1960:186-192. Worth, S. G. 1892. Observations on the hatching of the yellow perch. U. S. Fish Commission Bulletin 10(1890):331-334. Zorach, T, and E. C. Raney. 1967. Systematics of the percid fish, Etheostoma maculatum Kirtland, and related species of the subgenus Nothonotus. American Midland Naturalist 77:296-322. Wind Drift and Migration of Thrushes: a Telemetry Study William W. Cochran and Charles G. Kjos ABSTRACT Nocturnal migi'atory flights of four species of thrushes (Catharus spp.) that migrate through Illinois were investigated by radiotelemetry. Winds aloft were measured by theodolite and weather balloon during the birds' migi-atory flights. Analyses of the migi'atory flights of individual birds as they encountered different wind conditions indicated that individuals maintained a constant heading and airspeed regardless of the speed and direction of the wind, cloud cover, or land- marks. Lateral wind drift was mitigated and ground speed was maintained above about 6 m/s by the selection of flight altitudes where winds were more favorable. When winds were extremely unfavorable, birds landed. Path deviations caused by variable winds mask the remarkable ability of thrushes to maintain particular headings. Thus, unless winds are scrupulously accounted for, radiotelemetry, radar, and other observations of natural migration are of limited value in the study of orientation. The spectacular nocturnal migi-a- tions of songbirds between breeding and wintering areas separated by many thousands of kilometers have in- spired several decades of reseach that seem to have produced more questions than answers. A persistent question concerns how migi'ants respond to winds, particularly winds capable of blowing them off course, and another question concerns the effects of over- cast skies. This paper is about how migi'ating thrushes (Catharus spp.) re- spond to wind and overcast skies. Thorpe (1949) described a migra- toi-y journey as consisting of a start, a body, and a termination with each part representing different orientation prob- lems for the migi-ant. The spring and fall destinations of thrushes passing through the area of this study were Mr. William W. Cochran is an Associate Wild- life Specialist, Section of Wildlife Research, Illi- nois Natural History Survey, and Mr Charles G. Kjos is a Fish and Wildlife Biologist, United States Fish and Wildlife Service, Department of the Interior, Federal Building, ¥m-\. Snelling, Twin Cities, Minnesota. many hundreds to 1 or 2 thousand kilometers distant, and we consider the 200- to 700-km flights reported here representative of the "bod}'" of thrush migrations. These flights were noctur- nal, typically about 300 km in length, and usually separated by one or more nights without flight. Allen (1939) pointed out that the motion of a bird over the ground is the sum of the bird's motion relative to the air and the motion of the air relative to the ground (wind). Although re- searchers have differed greatly about how birds react to winds, they have accepted the physical reality stated by Allen. The question has never been whether or not birds are affected by the wind but rather what, if anything, they do about it? Data are presented for Swainson's thrushes [C ustulatus (Nuttall)], veeries [C. fuscescens (Stephens)], gi'ay-cheeked thrushes (C. ryiinirnus (Lafresnaye)], and hermit thrushes [C. guttatus (Pal- las)]. Graber et al. (1971 ) delineate the wintering and breeding ranges of these species as well as their occurrence and 297 298 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 progress through IlUnois during migra- tion. METHODS Radio-tagged thrushes were followed by means of direction-finding equip- ment carried in an automobile. Approx- imations of the thrushes' flight paths were obtained by plotting on U.S. Geo- logical Survey maps (scale 1:250,000) the places and times where thrushes crossed roads. We refer to a line be- tween two plotted road crossings as a track segment, the azimuth of a track segment as a track direction, the distance represented by a track seg- ment divided by the time difference between two plotted road crossings as a track segment, and azimuth of a track segment as a track direction, the distance represented by a track seg- ment divided by the time difference between its end points as a track speed, and the track direction and speed col- lectively as the track vector Two or more connected track segments are estimated flight vector. Where we refer to the actual path of a bird we use course, groiuid speed, and path vector as counterparts to track direction, track speed, and track vector, respectively. Where we refer to actual bird perform- ance, the terms heading airspeed, and flight vector are used without qualify- ing them as estimated. Some authors have used track to describe the direc- tion of a path, but the correct word is course Therefore, to avoid confusion, we have substituted pa^/i approximation in those cases where track could properly have been used as a noun. Directions are given in degrees, clockwise from true north. Altitudes are referenced to the earth's surface. Times are central standard. When referring to the behavior of individual thrushes, we often use the term preferred heading as a short and convenient substitute for the mean of a tightly clustered set ofmeasured head- ings. In addition, preferred implies some- thing special about the direction. Al- though we do not speculate on what is special, the preferred headings of thrushes always pointed toward some part of the breeding or wintering area of the species (as seasonally appro- priate), but not always away from the wintering area from which an indi- vidual could have been coming. Others have used the term similarly to des- cribe the means of sets of headings (usually of different individuals) with- out implying that particular goals (such as nesting sites) were in the mean direction. The methods used to determine the geographic (road crossing) points, which define track segments, and to measure or estimate bird altitudes are discussed in the Appendix. Fovu- different meth- ods were used to estimate the headings corresponding to track segments. Occa- sionally two of the methods were used to estimate a heading for a single track segment. For an extensive discussion of these methods, described briefly below, refer to the Appendix. Of the 71 heading (and airspeed) estimates in Tkble 1, 40 were obtained by constructing vector triangles in which an estimated flight vector was the vector difference between the track vector and wind vector Estimated head- ings obtained with this method are re- ferred to as calculated heading esti- mates. The method required knowledge of the birds' altitudes and the winds at the birds' altitudes. Winds were mea- sm-ed over increments of altitude. There- fore, the altitude of a bird was required. This method also provides an airspeed estimate. Six heading estimates were made by the departure method. During the first 100 or so seconds of flight, thrush- es climbed through winds from tree level to no more than 150 m altitude. Also, during the period after sunset when migi-atory flights usually began, the measured winds from surface to 150 m were sometimes light. The de- parture method is actually a special case of the more general vector triangle method but with a small crosswind component. Therefore, measurement of the initial track direction and confir- September 1985 125 Years of Biological Research 299 mation of light crosswinds are required. The initial track direction with a small allowance for crosswind provided an estimated heading. One way by which the track direction was measured was to determine the place where a bird crossed over a road within about 2 km of the starting point. The potentially higher track direction error associated with the short track segment was off- set by a higher accuracy of the plotted points, i.e., the departure point was known to within 10-20 m and the cross- over point could be determined to with- in 30-40 m due to the low bird altitude and the reduction of map errors through adequate preparation. Nineteen heading estimates were determined by the head-null method. This method required measurement of the azimuth of the null plane of the transmitter antenna radiation pattern. In the transmitter attachment method used in this study (Cochran 1972) and in a better method (Raim 1978), the transmitter is mounted on a bird's back, centered between the points where the wings attach to its body (Fig. 1). The 30-cm x 0.013-mm wire anten- na exits the transmitter flush with a bird's back and extends backwards over and beyond the tail. When a bird is in flight, the antenna and the bird's head- ing are both in the same vertical plane and the antenna is approximately hori- zontal. No horizontally polarized ener- gy (radio waves) is radiated in this plane. In practice, this arrangement caused the signal to fade completely and then return as the tracking vehicle was driven on a road which passed through this plane. The azimuth of the plane (the bird's heading) was bounded by the azimuth of the last bearing taken as the signal faded and the first taken when the signal returned. The six remaining heading esti- mates were made using the cross-polar- ization method, which required the observer to be positioned approximate- ly under a bird. A receiving antenna pointed up at the bird gave the maxi- mum signal when the antenna elements were aligned parallel with the trans- mitting antenna and zero signal (a null) when at right angles (cross polar- ized) to the transmitting antenna. Ranges to overhead birds were equal to their altitudes and thus, never greater than a few kilometers. At such short distances signal maxima were strong and the nulls sharp. When a signal was nulled, the azimuth of the receiving antenna elements was at right angles to the transmitter antenna with the latter in the same plane as the bird's heading, as described above for the head-null method. Both the head-null and cross-polar- ization methods resulted in measure- ments with a 180° ambiguity, because a bird may be headed in either of two azimuths which define the same plane. To resolve the ambiguity, we made flight vector estimates by calculation from the track vector and wind vectors at all conceivable bird altitudes. Transmitter Bias It is axiomatic that the weight of the transmitter (approximately 6 per- cent of a thrush's body weight) affected the flight of the birds we followed, but we do not know whether the effect was faster-than-normal airspeeds (at normal angles of attack with higher-than-normal energy use), normal airspeeds (at higher- than-normal angles of attack and ener- gy use), or slower-than-normal airspeeds (at higher-than-normal angles of attack and normal energy use). The fore-aft centering of the transmitter may have determined which of these effects per- tained. We tried to attach each trans- mitter with its center as near as pos- sible to the places where a bird's wings attach to its body, but in practice the looseness of a thrush's skin and the way in which the bird was held during at- tachment sometimes resulted in an at- tachment that was slightly forward of the desired spot. Some of the variabili- ty in the airspeeds reported in this paper may have been due to variations in transmitter placement. The mean of the 17-airspeed sample of Cochran et al. (1967) was 13.1 m/s. Calculations for most of those samples 300 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 were made from plots on highway maps, which later were found to have a 7-per- cent scale error. Adjustment for this er- ror brings their mean airspeed down to about 12.2 m/s, but this is still 20 per- cent higher than the 10.1 m/s mean for the airspeed estimates we now report. A species-by-species comparison be- tween the two studies shows the mean airspeeds of Cochran et al. '1967) to be higher by 16, 39, and 25 percent for Swainson's, veery, and gray-cheeked Fig. 1. - (Above) Swainson's thrush carrying a 1.3-g radio transmitter of the type currently in use (Cochran 19801 (Below) a 2.2-g transmitter of the type used in the study reported in this paper September 1985 125 Years of Biological Research 301 thrushes, respectively. Therefore, the airspeed difference cannot be attrib- uted to a difference in species mix in the two studies. Most of the birds in this study were fitted with 2.2- to 2.4-g transmitters by C. Kjos, whereas in the study by Cochran et al. (1967), 2.6- to 2.8-g transmitter attachments were about equally divided among R. Graber, W. Cochran, and G. Montgomery. A consistent difference in fore-aft center- ing of the transmitter is one explana- tion for the difference between the mean airspeed from the two studies. However, it seems more likely that the heavier transmitters in the earlier study required birds to fly slightly faster to provide additional lift. Although we believe that thrushes' airspeeds are affected by the attach- ment of transmitters and that radio- telemetry is not ideal for studying ab- solute airspeed, the speed analyses in this paper are directed toward finding whether thrushes change their air- speed, altitude, and course direction in response to changes in wind and not toward finding how fast they fly. Anal- yses of responses to changes in wind are unaffected by the bias discussed above. It is possible that transmitter at- tachment left or right of a bird's center of gi-avity affects the direction of flight. We were more successful in left-right centering of transmitters than we were in fore-aft centering. Furthermore, our analyses are concerned with changes or lack of changes in direction of flight in relation to wind. Therefore, we do not believe that variations in transmitter attachment could have affected our analyses. The question of the effect of a transmitter, no matter how well cen- tered, on the orientation mechanisms involved in migi-ation is one which we cannot address directly. The migi-atory flights we observed were within the broad limits of what is generally ac- cepted as normal with regard to prog- ress and orientation toward the breed- ing or wintering areas. As we will show, the thrush migrations we ob- served were, with few exceptions, char- acterized by headings that were nearly constant regardless of the direction or speed of the wind or the presence or absence of cloud cover. We cannot con- ceive of how a transmitter could impart to a bird the urge or ability to main- tain a particular heading under a vari- ety of wind conditions while depriving it of an urge or ability to behave in some other way. SUPPLEMENT TO RESULTS OF THE 1965-1966 STUDY (Cochran et al. 1967) We will first review and discuss the flights of two thrushes reported in Cochran et al. (1967). Although the authors of that paper provided maps showing path approximations, some of which were crooked enough to preclude significant compensation for wind, they did not materially discuss wind drift. The flight numbering they used is retained here, and the path approxi- mations for the two thrushes are shown in their Fig. 7. For Swainson's #5, Cochran et al. (1967:'Ikble 1) gave only average values for heading estimates and track direc- tions. Detailed data are given here. This Swainson's passed the airport at Springfield, Illinois, when and where winds aloft were being measured by the U.S. Weather Bureau. Several measure- ments of the bird's altitude and loca- tion were made, and from these and the winds-aloft data we calculated esti- mated headings (±10°) and estimated airspeeds (m/s shown in parentheses) of 295°(10.5), 298°(12.9), and 305 "(12.4) for track directions and speeds of 220°(9.5), 252 "(7.8), and 286°(7.4), re- spectively. The sky was clear for this entire flight. The next evening was completely overcast, and the bird was tracked for 83 minutes after it took off The head- ing was estimated at 309° (±10°) for a track direction of 302 ° The spread of 302 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 estimated headings for this thrush (295°-309°) was 14° less than one- fourth the 82° spread in track direc- tions. That a single heading, 300°, was within the error range of all estimated headings, that all estimated headings pointed to the breeding range of the species, and that at least one track direction, 220°, pointed well outside the breeding range are consistent with constant-heading behavior. If the differ- ences in estimated headings were not due to measurement errors, it is note- worthy that the greater the deviation of a track direction from a 300 ° pre- ferred heading, the greater also the deviation of the corresponding esti- mated heading from 300°. Instead of altering heading to compensate par- tially for wind drift, this Swainson's may have flown headings which actual- ly increased drift. However, looked at another way, the indicated clockwise heading shifts (295° to 298° to 305°) may have resulted from a slowly al- tered heading in response to, and to compensate for, the counterclockwise drift in track directions (up to 80°) earlier in the flight. Cochran et al. (1967:222) mention- ed a curvature in the flight path of gi'ay cheek #19. They stated that the change occurred under clear skies but did not elaborate. Detailed data are given here. A Ill-minute track segment was in a direction of 314° at a gi'ound speed of 15.8 m/s and was within 35 km of the Peoria, Illinois, airport when winds were being measured there. The alti- tude of the bird was between 300 and 1,300 m during this segment. Depend- ing on the altitude assumed, estimates varied from 332° to 342° for heading and from 11.3 to 10.1 m/s for airspeed. Later, during the last 53 minutes of the flight, the altitude was between 100 and 300 m with a track direction of 358 ° and a gi'ound speed of 18 m/s. This portion of the flight was not well situated relative to the schedules and locations of U.S. Weather Bureau winds- aloft measuring stations. Fortunately, there were no frontal systems in the area, and the temporal and spatial variations in wind measured at the various reporting stations followed a consistent pattern, allowing the use of the 2315 CST winds measured at Madison, Wisconsin, about 65 km from the flight. The heading and airspeed estimates were 338° at 10.7 m/s and 348° at 11.5 m/s for altitudes of 300 and 150 m, respectively. Whereas be- tween the earlier and last parts of the flight the track direction shifted 44°, the heading estimates changed less than 17 °, and the heading appeared to have been consistent in a direction of 338°-342° The data indicate that as the wind direction shifted clockwise, the heading may have shifted clockwise also. A clockwise heading change, opposite to that which would produce compensa- tion for wind drift, could be called negative compensation. One special case of negative compensation is orien- tation at a fixed angle to the wind, as suggested by Vleugel (1962). Another special case is downwind orientation as reported by Gauthreaux & Able (1970). Gray-cheeked thrush #19 did not fly downwind nor did it fly a constant angle to the wind, a 100° clockwise shift in winds being associated with only 5 ° heading change, if any. Paradoxically, the indicated 5 "" clock- wise heading shift (337° to 342°) may have resulted from a slowly altered heading in response to the early counter- clockwise di'ift (track direction = 314°). This case illustrates, as did the data for Swainson's #5, how the same data ma^' be interpreted as partial negative (instantaneous) compensation behavior or as positive compensation behavior achieved by a slow response. RESULTS, 1967-1972, ON EFFECTS OF WIND The flights of 36 thrushes were re- viewed in a process of selecting the 14 for which data ai-e presented in Thble 1. Of the 22 that were excluded, either winds did not vary significantly during September 1985 125 Years of Biological Research 303 H S T3 S m oo oo OO OOOO O O OOOO OQ CCCOC OCOOO t--00 IO'tJ'COtJ'CDCDCDCO lOCDCDCD COO-^ ''tlOlO-^tO COlCCOCO"^ 0505 ^a>a:)odh2cgK^.-H oiocooo Q^^^ ^cdoaiao Qoaio;^ o5o ocow^oiooco cgoooio ooio oaoocoo o£c5aiiH oo COCOCOCOCOCOCOCO COCOCOCO COCOCO ,-i,-i^^^ . _„^ So ^pLC^ in CO aioqtoo '""o S^oiooc ""c^S^o loco TtaJooJ lo CO oa)0[> Qco T^'^Ttaiai QoicNi>o 1—I^Hi—(i—I 1—li-Hi—If—I.—I C'*c^ oicMTj-r-iojcoo COCOCOCO CO CO OOOCO coco OOO^i-f ^(MNC^^ o September 1985 125 Years of Biological Research 305 H S & Q5 pa 00 0) 306 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 the period of observation, or wind mea- surements were not available sufficient- ly near the flight path to permit analy- ses (e.g., see Fig. 2). For the 14 selected thrushes, track direction data were significantly ( 1 ) duinng a single night's flight, (2) between different nights, or (3) during ascent or descent. From the numerous track segments measured for each of the 14 individuals, we se- available for flight in winds differing lected (for analysis and presentation) 15( 'Departed 1955CST VECTOR SCALE ( Fig. 2. - Tiack of veery 6705, area winds aloft, and the position of a weak front. Windsaloft data are by the U.S. Weather Bureau (USWB) for 2315 CST, 25 May 1967, except for P, where the lower set of winds is for 1730 CST. Letters P, R, M, F, and I are the geographic locations of the places where the USWB measured winds aloft (drawn nearby). Winds are for increasing altitudes, going clockwise from surface wind (marked S) to 150. 300, 800, 1,300, 1,800, 2,300, and 2,800 m. The posi- tion of the front was taken from the USWB surface chart for midnight, 25 May. Ti-ack speeds (one- or two-digit numbers between points marking the bird's pathl were computed from straight-line distances between points. Bird altitudes shown are ±20 percent. X is the position of the bird at 2315 CST and also the approximate position of the warm front at that time. September 1985 125 Years of Biological Research 307 segments for which headings could be estimated with minimal error by vector subtraction of winds aloft measui-ed close to the bird's paths and near their altitudes or by the head-null or cross- polarization methods. Track directions and heading esti- mates for each individual bird were normalized about their respective means so that the data for all birds (Tkble 1) could be pooled to examine the relationship between courses and head- ings (Fig. 3). When the use of two me- thods resulted in two heading estimates for a single track direction, as indi- cated by a double entry in the heading estimate column in Tkble 1, the mean of the two estimates was used in con- structing Fig. 3. Course-heading plots are useful, because constant-course (complete com- pensation for wind) and constant-head- ing (no compensation for wind) behav- iors, in variable crosswinds, would pro- duce distributions along the vertical and horizontal axes, respectively. Dis- tributions representing headings which partially compensate for lateral wind drift, or which are at a constant non- zero angle to the wind, are difficult to interpret in course-heading plots. How- ever, the point is moot for the thrush data, because heading estimates are distributed along the horizontal axis. The actual data do not consist of courses and headings but of approxima- tions to these, namely, track directions and heading estimates. For simplicity, track direction versus heading esti- mate plots (e.g.. Fig. 3) will be called TD-HE plots. Track directions were bet- ter estimates of courses (+1° or 2°) than heading estimates were of head- ings (±3-10°). This fact biases against track variations the impression given by a TD-HE distribution unless the variations are significantly greater than the error range of heading esti- mates. It is for this reason that flights were excluded from the analysis, as mentioned above, if wind variations were too small to alter courses by an amount greater than the probable error range in estimated headings. The other consequence of the 6 -20 ° error range in heading estimates is that small heading changes, whether in i-esponse to wind or for other rea- sons, are masked. Some of the vertical scatter of points in the TD-HE plot (Fig. 3) may thus represent real heading changes which are small compared with corresponding variations in track directions. Swainson's Thrush 6905 — An Example The flight of Swainson's 6905 was chosen as an example for discussion because it was observed flying in winds which varied from night to night, dur- ing the same night, and significantly and rapidly with altitude on one night. Track directions and heading estimates for this bird are distinguished by a sep- arate symbol in Fig. 3. Swainson's 6905 was netted at 1130 on 15 May in a 0.5-ha woods located on the west edge of Champaign, Illinois. It weighed 36.7 g. A 2.4-g transmitter was attached, and the bird was released within 25 minutes of its striking the net. It remained in the woods until the evening of 23 May, when it departed on a flight which ended the next morning in a park near the center of Galesburg, Illinois. It departed that evening, flew all night, and landed at twilight in a 40-ha woodlot about 18 km southeast of Waterloo, Iowa, where it was renet- ted on the same morning for the pur- pose of replacing the transmitter with one having a fresh battery. It weighed 34.7 g, not including the transmitter weight. The transmitter was not re- placed, because the adhesive held it so firmly that the bird might have been wounded had the transmitter been pulled off. The bird was released in the woodlot and was monitored in the wood- lot until the battery failed 2 days later. The heading estimates for track segments Tl, T2, and T3 (Fig. 4) show a progressive clockwise shift totaling 9° into winds coming from the bird's right side. Similarly, the heading esti- mates for T4 and T5 indicate an 11° 308 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Fig. 3. - Plot of devi- ations of track directions and estimated headings, for individual thrushes, from the respective mean value of track direction and etimated heading. Datp are from Table 1. • = SWAINSON'S THRUSH 690 O = VEERY 6901 + = CLOCKWISE - = COUNTERCLOCKWISE -30 ' -10 -40 -20 +10 ' +30 +20 mo +50 TRACK DIRECTION - DEVIATION FROM MEAN (degrees) counterclockwise shift into winds with a component coming from the bird's left side. Also, the airspeed estimate for T5 was greater than that for T4. These changes are consistent with a compensa- tory response to lateral wind drift by a gradually altered heading and airspeed or by partial negative compensation. During the first 40 minutes of the second night's flight (Fig. 5) the track direction changed by more than 90°. The bird departed while we measured winds aloft 5 km west of the bird. The first two bearings to the bii'd were taken from this place 1 and 5 minutes after the bird took off Subsequently, seven additional bearings were taken as the tracking vehicle was driven on a circuitous route under and around the path of the bird. These bearings, and the times and places from which they were taken, are shown on the maps of Fig. 6. Although bearings were taken continually, only those above, taken carefully while the vehicle was stopped at known map points, could be used for analysis. Between 21 and 26 minutes after the bird's takeoff, while stopped, we made repeated attempts to measure the elevation angle to the bird but failed because the angle was below 30°, where ground reflections cause erratic readings. The series of bearings and the times, coupled with the winds mea- sured in the same airspace as the bird, provided data for evaluation of com- puter simulated paths of a bird's hypo- thetical constant-heading, constant-air- speed ascent through the winds aloft that we measured. The process was iterative, a matter of computingpaths for different headings, airspeeds, and ascent rates and noting which com- bination best fit the observed bear- ing-time data. Sixty iterations revealed that a 290° heading, 8.1-m/s airspeed, and 0.66-m/s ascent rate fit all bearing- time data within 5 percent. Fig. 6A, 6B, and 6C illustrate that headings of 280° and 300°. airspeeds of 7.2 and 9.0 m/s, and ascent rates of 0.56 and 0.81 m/s, respectively, cause poor fits to the bear- ing-time data. lb understand this analysis, it is important to note that there is no guai-- antee. no matter how many parameter combinations are tried, that a partic- ular model will produce a path that will match (fit) the bearing-time data. If we had tried all combinations of headings, airspeeds, and ascent rates in increments of 1°. 0.1 m/s. and 0.01 m/s. respectively, and had found none which, with the constant parameter model, produced a path which nearly September 1985 125 Years of Biological Research 309 10m/Sec VECTOR SCALE 50 Km MAP SCALE Fig. 4. - The approximate 23-24 May path of Swainson's 6905, winds aloft vectors measured along the path (shown near where measured), and vector representation of estimated headings (Ught hues) and track vectors (heavy lines with "T" designators). Ti-ack speeds (numbers along lines connect- ing solid points) are in meters per second for the straight-line distance between points. Bird altitude measurements (underlined) are in meters above the surface and are accurate to ±20 percent, or better Vector plots are for track segments drawn as solid lines and are labelled chronologically Tl, T2, etc. Winds-aloft vectors are labelled with altitude in lOO's of meters (S = surface). Each estimated heading in the vector plots (light lines) is labelled with the altitude (in lOO's of meters) of the wind used for its computation, and corresponding wind vectors are shown to the same scale separately with the wind vector sets. Tl and winds taken at departure go together, estimated heading vectors shown for T2 and T3 are from the second set of winds, taken along the path, and the vectors for T4 and T5 use the last set of winds, taken just after the bird landed. fit the observed directions and times, we would have been required to reject the model, our bearing-time data, our wind data, or some or all of these. Similarly, had the model provided a fit only for unbelievable parameters, e.g., an airspeed of 40 m/s, a rejection would have been required. We could then have constructed more complicated behavi- oral models ad nauseum in hopes of finding one which fit the bearing-time- wind data. That the simplest model (constant behavior, i.e., constant head- ing, airspeed, ascent) produced a path consistent with our bearing-time-wind data and that the airspeed, heading, and ascent rate required to produce the fit were close to those observed for this 310 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 bird in other portions of its migi-ation, but we reject these on gi-ounds of sim- constitute powerful support for both plicity in hypothesis (Barker 1957: model and parameters. Of course, there Chapter 5). are numerous other models, involving Later in the flight, as the bird de- variable parameters, which can pro- scended in the vicinity of the Missis- duce fits to our bearing-time-wind data, sippi River, the track direction again n ''^''^'^ Rapids 41^° lOm/Sec VECTOR SCALE 50 Km MAP SCALE Fig. 5. - The approximate 24-25 May path of Swainson's 6905. Other information is shown as in Fig. 4, except that two sets of estimated headings are shown for TV, the upper set for winds measured at departure and the lower set for winds measured immediately to the north. I September 1985 125 Years of Biological Research 311 290 9.0 312 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Table 2. — Track directions and speeds calculated for real wind data (Table 4) and a bird with tieading 290° and airspeed 8.1 m/s. ..... , Track Direction ^ , o . Altitude Direction Relative to Track Speed (meters) ^i^g,.^^^^ 290° ""'s' Surface September 1985 125 Years of Biological Research 313 : MAGNITUDE OF DIFFERENCE BETWEEN MEAN ESTIMATED HEADING S TRACK DIRECTION THAT A BIRD DID MAINTAIN(-«- ) OR COULD HAVE MAINTAINED AT A DIFFERENT ALTITUDE ( •- ) GROUNDSPEED COMPONENT IN DIRECTION OF MEAN ESTIMATED HEADING 1 1 1 1 1 1 10 12 14 16 18 20 Fig. 7. - Twelve thrush choices between low progress speed with small deviation of course from a preferred heading and high progress speed with a larger deviation of course from a preferred heading. The dot is the option rejected, and the arrowhead is the option chosen. In each case the bird passed through the altitude where the wind provided the rejected op- tion and levelled off at the altitude providing the chosen option. PROGRESS SPEED m/sec W 12 10 H 8 6 t- 2- .'/. ~1 T" 10 20 30 40 1^ 50 ^ I I 50 70 80 DIFFERENCE BETWEEN TRACK DIRECTION & MEAN ESTIMATED HEADING - DEGREES Fig. 8. - Estimated airspeed versus differ- ence between track direction and mean estimated (preferred! heading. Data are from Table 1. alysis. Birds were usually presented with different wind profiles near the beginnings and ends of flights. These differing wind profiles and the lack of accurate, continuous heading and speed data make it impossible to apportion the observed drift reduction among such possibilities as the winds becom- ing more favorable, the birds selecting altitudes with more favorable wind, or perhaps the birds slowly changing headings. The effect of changing wind is demonstrated by the first night's flight of Swainson's 6905 (Fig. 4). Un- fortunately, we do not know exactly where the wind change occurred, but Table 3, — Track direction, relative to mean heading estimate, for an early and a late track segment in the same night's flight. Species, Year Interval Number (minutes) Early Segment Later Segment Deg'' (TSr Deg*" (TSf He6910 Sw6902 Sw6905 Gc7003 Sw7005 VeTOOS Sw7012 He7015 He7018 Ve7103 Sw7118 Gc7207 Mean 136 98 370 334 183 353 113 86 257 161 128 63 51 360 265 288 203 -89 -1-23 -28 -41 -21 -1-19 -H25 -21 -53 -1-38 -28 -H28 -1-51 -23 -21'' -40 (22) (14.5) (9.5) (8.0) (15) (14.4) (19.8) (23) (14) (9.2) (9.0) (9.7) (16.3) (11.2) (19.3)'' (11.3) -47 -1-6 + 7 -H26 + 7 -11 -1-18 -8-' -25 -1-1 -3 -H42 -1-15 -1-14 -1-16 + 6 (24) (13.7) (8.0) (10.7) (16) (18.3) (10.6) (28)-' (9.8) (10) (13.2) (13.8) (17.9) (10.6) (15) (9.4) 34.3' 14.1 15.8' 14.3 ' Between the midpoints of the two compared track segments. '' Difference in degrees between track direction and mean estimated heading; plu.s indicates track direction clockwise from the mean esti- mated heading, minus counterclockwise. ' Track speed in meters per second. '' Data from track segment not given in Table 1. ' Mean of the absolute value of differences, sign ignored. at an altitude of about 800 m the bird had achieved a track direction (seg- 314 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 16 li* - 12 10 H ^ 8 6 - I— "^ 2 ...«!•.. o o o* ° o p* • < o.* • o • = DATA FROM TABLE 1 0= DATA FROM OTHER BIRDS "lO S W 16 18 20 22" 24 2^ 28 TRACK SPEED, m/sec. Fig. 9. - Estimated airspeed versus track speed. ment T3) closed to its preferred head- ing. Then, presumably upon encounter- ing changing wind, it reduced its alti- tude, and the track direction and speed changed markedly. A little later, at a still lower altitude, the difference be- tween track direction and mean esti- mated heading was again reduced. It is tempting to assume that the process of finding an optimal altitude takes time and that, when different, less favorable, winds are encountered, a bird must repeat the process. If a bird repeats the process when less favorable winds are encountered, changing winds would tend to hinder a bird's efforts to find an altitude with winds optimal for its preferred heading and would tend to negate the compensatory effects of slowly altering its heading. Thus, the analysis derived from Tkble 2 probably understates thrushes' ability to reduce lateral drift. However, the data of Tkble 2 suggest that thrushes reduce drift over a period of time, that such reduc- tion is accomplished without sacrifice of ground speed, and that the main method employed is altitude selection rather than change in heading or air- speed. ATYPICAL BEHAVIOR Of the 36 thrushes tracked, veei-y 6901 was the only one to make a large heading change not attributable to storms. Data for it are included in Table 1 even though the wind varied little during the 3.5 hours the bird was tracked. The data plotted for this bird in Fig. 3 are given a special symbol. The veery did not maintain a constant course, a constant heading, or a con- stant heading relative to the wind direction. Instead, the bird turned clockwise (a heading change of at least 60°). as evidenced by both track direc- tions and heading estimates. The turn could not have been an attempt to fly downwind, because the heading con- tinued to change by 30° or more past the downwind direction. A delayed response to lateral wind drift is ruled out, because the indicated heading change was in the wrong direction. The sky was cloudless throughout the flight. We offer no explanation for the behav- ior of this veery. Veeries 7008 and 7103 encountered thunderstorms and landed after chang- ing heading away from head winds. September 1985 125 Years of Biological Research 315 Table 4. — Theodolite measurements and computed winds alott for the start of the 24 IVIay flight of Swainson's thrush 6905. 316 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 overcast skies provided tests for the ef- fect of cloud cover. Veeries 7008 and 7103, Swainson's 7118, and hermit 7015 (Table 1) collec- tively made eight transitions between clear and overcast skies. Similarly, Swainson's 6905 passed through a region with complete overcast (Fig. 5). Swainson's #5 and gi-ay cheek #19 (Cochran et al. 1967) were observed under clear and overcast. In addition to these, three thrushes not included in Table 1 were observed under clear and overcast skies. These represent 15 tran- sitions between clear and overcast sky, none of which was associated with a heading change. None of the 34 thrush- es tracked during this study or the 21 thrushes studied by Cochran et al. (1967) was observed to alter heading during changes in cloud cover. Migi-ation was initiated under over- cast skies by Swainson's 7012 and her- mits 7015 and 7018 and by Swainson's #5, gray cheek #19 (Cochran et al. 1967), and one of the three thrushes observed under oveixast in this study but not included in Table 1. We believe that the data show that thrushes are not affected by overcast skies present at the initiation of, or encountered dur- ing, migratory flight. Griffin (1973) reported oriented flight of small birds in, above, and in between cloud layers. We were able to measure the altitude of flight and cloud ceiling for 12 thrushes tracked when there was a solid cloud layer. Of these 12, including the 10 discussed above, all flew below the clouds. A veery (Cochran 1972:Fig. 6) may be an excep- tion. It was about 30 km east of the west shore of Lake Michigan at day- bi'eak. The western lakeshore region was shrouded in heavy fog, but we could not tell how far the fog extended over the lake. The veery may have been flying over the fog, but certainly was unable to see the western shore even had it been high enough for line-of- sight viewing. The veery 's track direc- tion over Lake Michigan was within 10° of that observed during the pre- ceding 560 km of flight, and it thus seems probable that this veerj' was well oriented in the absence of cues in the same sense that Griffin suggests for birds over clouds at more typical cloud altitudes. DISCUSSION Wind di-ift is an important factor in the study of the orientation and navi- gation of migi'atoiy birds. In a practical sense, the responses of birds to wind determine the kinds of observations needed and how these should be inter- pi'eted. If individual birds maintained constant courses, a knowledge of their orientation behavior could be obtained directly, because track directions are measurable by most techniques. How- ever, few of the many radar and visual studies give precise wind-altitude data, and none provides rigorous proof that individual birds maintain a constant course in variable winds. Analyses in all but a few of these studies have re- lied on assumptions about mean goals (intentions) of mixtui-es of unknown species (the radar swarm) arranged into sets flying in different winds. Even if species could be identified, assump- tions about the goals of observed indi- viduals could seldom be more precise than a quadrant. No matter how rigor- ous the mathematical and statistical treatment of models, radar observa- tions, or visual observations, the validi- ty of any conclusions can never exceed that of the underlying assumptions. The constant-heading hypothesis finds partial support in several radar studies (Richardson 1976; Williams et al. 1977; Emlen & DeMong 1978). How- ever, a majority of radar studies, at least 26 according to Alerstam (1976), indicate that the drift of the swarm is less than passive, i.e., there is partial compensation or no drift at all. If par- tial compensation is a reality for "aver- age" behavior, we point out that it may result from mixes of individuals and species behaving differently, for exam- September 1985 125 Years of Biological Research 317 pie, some compensating completely; some partially; some flying downwind; and some, like the thrushes, flying a preferred heading. Bloch et al. (1981), in a study of European migi-ants, em- phasize the mixture aspect but state that "Songbirds compensate drift due to a side wind at least partly by alter- ing their heading; large individuals ap- pear to be more successful in this than small individuals." Their separation of radar targets into five classes by echo signature and size analyses, although useful for some purposes, required assumptions about goal directions as in less sophisticated radar studies. How- ever, the goal directions of migrants in Europe may be far more restricted than those of the western hemisphere thrushes, which as species breed and winter over a wide range of longitude. In reference to radar studies in Illi- nois, Bellrose (1967:305) states that "all our evidence indicates that they correct within a few degi'ees for wind drift." It is difficult to reconcile our thrush data with this conclusion about the seasonal average behavior of the average trans- Illinois migi-ant. That thrushes select altitude for favorable winds would have shown up in radar data as pseudo drift (Nisbet & Drury 1967; Alerstam 1976), which increases apparent drift in the analysis of radar data and would add to rather than account for the discrep- ancy between our findings and Bell- rose's findings. Graber et al. (1971) consider thrushes to be common mi- grants in Illinois, and both tower-kill data (Graber 1968; Seets & Bohlen 1977) and flight-call data (Graber & Cochran 1960) indicate that thrushes are among the most common migi'ants through Illinois in May and September Therefore, thrushes should have been well represented in Bellrose's radar samples. Thrush species might be among the few that do not compensate totally, and it may be that their inclu- sion in Bellrose's averages was respon- sible for the "few degrees" of drift allowed for the average bird. We have shown that thrushes re- duce lateral drift as a night's flight progresses and that the reduction is ac- complished primarily by the selection of an altitude with the least unfavor- able wind and secondarily, if at all, by small, slowly implemented changes in heading and/or airspeed. Our data also suggest that when cross and opposing winds are the only options, keeping progress speed above about 6 m/s takes precedence over reduction of lateral drift, the latter having first priority only when progress speed exceeds 10 m/s. Thrushes landed when winds pro- vided no options other than a progress speed of less than 2 m/s. The ascent rates we measured for thrushes, between 0.3 and 0.7 m/s, are much less than the climbing capability of up to 4.4 m/s reported for passerines by Able (1977). Although high rates of ascent may not be as aerodynamically efficient as lower rates, the overall ef- ficiency during a night's flight - in terms of distance covered per unit ener- gy — would be improved for birds that quickly find the best winds available. If energy efficiency is improved by find- ing the best winds available as quickly as possible, and if thrushes are capable of faster ascents than those we ob- served, then their relatively slow ascent may indicate that the "measurement" process by which thrushes assess the effects of winds is an integi-ative one re- quiring a significant amount of time. This idea finds additional support in the considerable time (20-50 minutes) taken by thrushes 6809, 6810, and 7103 to respond (by landing) to ex- tremely unfavorable winds and possibly in thrushes slowly altering their head- ings, as discussed below. Some of our data suggest that if the optimal wind has a significant cross- wind component, as it often does, thrushes respond by slowly altering their heading and perhaps airspeed as well. We reject the alternate interpreta- tion, negative compensation, because there is no apparent benefit in a bird's increasing its lateral drift. Our data 318 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 are not accurate or continuous enough for precise analysis, but they do rule out rates of change gi-eater than a few degi-ees per hour Corrections at such small rates could produce significant compensation for drift on long over- water flights in consistent wind but would be of limited benefit in the vari- able wind of the midwestern United States, particularly in spring, and for typical 6-hour overland thrush flights. We speculate that the ultimate degree of correction sought by slowly altering heading or airspeed may be set at lim- its imposed by progress speed, as sug- gested for corrections achieved by the selection of altitude for optimal wind. Some of the thrushes we observed ascended, during the first hour or so of flight, through a variety of unfavorable winds and then descended to an alti- tude with winds more favorable than those encountered at higher altitudes. Some remained at a higher altitude with a wind no better than one en- countered at a lower altitude. Still others ascended for a shorter period and levelled off at an altitude with an acceptable wind without testing higher altitudes for more acceptable winds. We note also that the ascent at the in- itiation of migi-atory flight is not repeated later in flight even if winds change significantly. Instead, after the initial ascent, altitude selection is restricted to descent or to small upward adjustments. These observations are consistent with a strategy of seeking a tolerable compromise through mini- mum effort rather than continuously seeking the most favorable wind avail- able at all times. We believe that the energy expenditure required for major ascents, coupled with the low probabili- ty that major wind changes will occur during a night's flight, represent an energy-cost-to-wind-benefit ratio which favors the acceptance of a satisfactory compromise at hand as opposed to the continual seeking of a better com- promise. That lateral drift is reduced, by whatever means, and that low progi'ess (or ground) speed results in the termin- ation of flight, suggest that thrushes are aware of some aspects of their path vector, can assess the speed and direc- tion of the wind relative to their head- ing, or some combination of both. Bruderer (1982:11) states that the ground is the "most simple and prob- able" reference by which the course is estimated. Bellrose (1967:306) specu- lated that birds may be capable of ob- taining information about both speed and direction of wind, relative to their heading, from the "gust-form of air." If small scale accelerations of air (gusts) were used by thrushes to evaluate wind, it would seem that the evalua- tion could be done more quickly than our data indicate. For this reason and also because linear landscape featm-es appear to be responsible for faster- than-usual course evaluation, as dis- cussed below, we favor the explanation that thrushes use the gi'ound as a ref- erence for evaluating their situation much as a man might use objects on shore when paddling a canoe in a strong current. Chicago (population about 7 mil- lion) and nearby Lake Michigan did not influence the flight headings of thrushes. Veery 7008 flew for 20 km over south and central Chicago before changing course and landing after en- countering a thunderstorm. Tvvo thrushes (gi-ay cheek §1 , Cochran et al. 1967; a veery, Cochran 1972: Fig. 6) flew essentially straight paths over Chicago and then out over Lake Michi- gan. Four others, not reported here because their flights were in unchang- ing winds, had straight paths over the Chicago metropolitan area. Unfortun- ately, none of these thrushes could be shown to be on courses markedly differ- ent from their preferred headings; therefore, we cannot dismiss the possi- bility that a bird flying in a strong crosswind as it encountered Chicago or Lake Michigan would use these major land features as references for course evaluation or correction, as is shown below for thrushes crossing major rivers. September 1985 125 Years of Biological Research 319 Of nine thrushes that we have tracked across the Mississippi River, the course of only one, Swainson's 6905 (Fig. 5), changed in the vicinity of the river. The change in course is entirely attributable to the birds descent to an altitude having a markedly different wind. The river and nearby Burlington, Iowa (population about 33,000), were prominent landscape features which may have enabled the bird to deter- mine quickly that its precrossing course of about 254 ° was well off its approximately 295 ° preferred heading. Later in the same night, this thrush passed directly over Iowa City, Iowa (population about 35,000), and near Cedar Rapids, Iowa (population about 85,000), on a course that was 20°-30° to the right of its preferred heading without noticeably responding. How- ever, during this time the bird was already flying so low that it could not have safely responded by descending. The paths of the other eight thrushes did not vary in the vicinity of the river, but their courses were within 20° of their preferred headings when they crossed it. Of 17 thrushes that we have tracked crossing the Illinois River, Swainson's #5 (Cochran et al. 1967) was the only one whose course changed in the vicini- ty of the river. Its approximately 250° course prior to encountering the river was well off its 300 ° preferred heading. A descent was detected near the river. The course shift could have been entire- ly due to a change in winds at the lower altitude, but altitude data were too crude to preclude other possibilities. The observations of Swainson's 6905 and #5 suggest that features of large rivers may provide better ground refer- ence than the relatively featureless in- tervening landscape, despite a profu- sion of lighted small towns and cities. Perhaps cities are too bright or too new on an evolutionary time scale, or per- haps thrushes respond to linear land- scape features only (e.g., rivers), as sug- gested by Bruderer (1982), but not to point features (e.g., cities), as proposed by Rabol (1974). Bingman et al. (1982:49) hypothesized "that migrants could utilize the Hudson River as a topogi'aphical reference by which to perceive wind drift from a preferred track and correct at least partially for these effects." However, they concluded that the birds altered their headings to achieve courses closely paralleling the river. That rivers are not followed by thrushes, even temporarily, is contrary to their conclusions and in agreement with James (1955). Cochran et al. (1967:224) state, "Flights of less than an hour's duration sometimes occur during or before thun- derstorm activity. These flights are downwind at low altitudes." Data gath- ered since 1966 do not support the downwind aspect of their conclusion. Instead, upon encountering a storm or squall line, thrushes adopted a heading that precluded upwind flight (and therefore was sometimes downwind) and always landed after changing heading, sometimes within minutes and sometimes after as long as an hour. We have not yet analyzed all the data bearing upon the question of thrush behavior in inclement weather and wish here only to remove any sugges- tion that downwind headings are spe- cifically chosen. However, regardless of wind direc- tion, flights of radio-tagged wood thrushes [Catharus mustelina (Gmelin)] observed in Illinois during the last phase of spring migration to their nest sites were flying approximately down- wind and at low altitude (Cochran un- published data). At times and geo- graphic localities where a low-altitude mix of radar or visual "targets" con- sisted of a considei'able proportion of birds in the site-search phase, down- wind distributions of track directions similar to those reported by Gauth- reaux & Able ( 1970) might be observed. Cloud cover had no effect on the headings of individual thrushes, whether encountered before or during flight. In this regard our thrush obser- vations are consistent with the radar 320 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 observations reported by Bellrose & Graber (1963). Able (1982) discussed the grovi'ing body of evidence for well- oriented flight under (and in) overcast and briefly reflected on what this may mean regarding orientation mechan- isms. He concluded that "we appear no closer to understanding the mechan- ism by which this is accomplished than when Griffin (1973) discussed the prob- lem nearly a decade ago." The data pre- sented here add to the evidence for oriented migratory flight under over- cast skies but do not reveal the mech- anisms involved. How can thrushes find their way between specific nesting and wintering areas separated by thousands of kilom- eters if they are largely at the mercy of the wind? We do not know, but winds do not make the feat as difficult as one might conclude at first glance. For ex- ample, we followed three thrushes for more than one night's flight and more than 400 km total distance each. The path of Swainson's 6905 was as serpen- tine as any we have observed, with track directions which varied more than 100°. Yet, after two nights of flight its azimuth from where we began observations at Champaign, Illinois, to near LaPorte City, Iowa, where we left it, was 308 °, only"l3 ° different from its preferred 295° heading. The overall path of Swainson's 7118 was 335°, only 5 ° different from its preferred heading of 340°; that of veery 7008 was 350°, not measurably different from its 352 ° preferred heading. Long migrations made up of many shorter single night's flights are sub- ject only to net drift according to the prevailing wind components of the var- ious geographic areas traversed. The ef- fects of weather systems will tend to cancel one another. Prevailing winds would necessarily shape any evolved system of navigation. The surprise in our data was that the net drift was remarkably small for such short path distances as 400-700 km in spite of the fact that the birds selected for presen- tation here drifted more than most of those that we observed. Thrushes' constant-heading behav- ior, with wind drift mitigated by their seeking an altitude with, at worst, a not-too-unfavorable wind, represents a long-distance migi-atory stragety that, in view of the probability that net drift is negligible, is moi'e conservative of energy than strategies involving com- plete or partial compensation for wind. We point out that nothing short of nearly perfect compensation for wind drift, especially for long migi'ations, can substantially reduce a distance- time penalty incurred by a bird in a homing process near the end of a migi-a- tory journey. Factors in addition to energy conservation are probably im- portant in successful migi-atory strat- egies. For example, partial-compensa- tion behavior may have an advantage over both complete-compensation and constant-heading behaviors on long overwater flights. Perhaps the advan- tages of partial-compensation behavior versus complete-compensation and con- stant-heading behaviors on long over- water flights could be examined on theoretical gi'ounds, but we have not done so. We only speculate that the slow altering of heading, subject to restrictions on minimum acceptable progi-ess speed, may have survival value for long overwater flights. Such behavior had such a small effect dur- ing short overland flights that we can- not be certain we observed it: therefore, if thrushes do alter headings and air- speeds slowly, the significance of such behavior would seem to lie elsewhere in their migi-atory journeys. CONCLUSIONS 1. Individual thrushes do not main- tain a constant track direction (path over the gi-ound) unless the wind they are flying in is also constant. 2. Individual thrushes maintain a constant heading during a night's flight and from night to night to an ac- curacy equal to or better than the ±3 ° precision of our best measurements, liable to the exceptions in 3 and 4 below. September 1985 125 Years of Biological Research 321 3. Thrushes may change their head- ing in response to lateral wind, but if they do so, it is at a rate of change of less than about 3° per hour. 4. If thrushes alter their airspeeds to reduce lateral di'ift, they do so by less than about 2 m/s. 5. Thrushes' airspeeds (mean = 10.1 m/s, SD = 1.6) are independent of and less variable than their gi'ound speeds (mean = 13.7 m/s, SD = 4.6). 6. Thrushes turn to avoid a head- wind component (prior to landing) when they encounter thunderstorms. 7. Thrushes mitigate the effects of lateral winds by flying at altitudes where winds are not too unfavorable (see 8 and 9). After takeoff, thrushes ascend until they find suitable winds, but if these are not found below 2 or 3 km, they descend and either land or accept a compromise at a lower alti- tude, never below about 75 m. As winds change during a night's flight, thrushes adjust their altitude accordingly, usual- ly by descending and rarely by ascend- ing - but never to the altitude reached during the initial ascent. 8. If winds at all altitudes above about 75 m have unfavorable head and side components so that progi-ess speed cannot be maintained above 2 or 3 m/s, thrushes land. 9. Thrushes choose winds for mini- mum lateral drift provided progi'ess speed does not fall below about 6 m/s. When the choice is between lateral drift (up to about 60°) and progress speed (below about 6 m/s), thrushes accept the lateral drift. 10. The ascent rate of thrushes at the beginning of flight is usually in the range of 0.5-1.0 m/s. This rate is less than their climbing capability and may reflect the time needed for assessment of winds aloft. 11. Primarily because of (7) above and secondarily because of (3), (4), and the fact that wind speed typically de- clines during the night, thrushes' track directions become closer to their pre- ferred headings toward the end of a night's flight. 12. Thrushes initiate and maintain constant (preferred) headings during migi-atory flights under, and during transitions between, clear, partly cloudy, and completely overcast skies, but they seldom, if ever, fly above the lowest unbroken cloud layer. 13. By implication from (7) above, thrushes either directly sense informa- tion about wind speed and direction (e.g., from anisotropies of small-scale turbulence), or they evaluate some aspect(s) of their progress over the ground, or both. 14. Prominent linear landscape features, such as large rivers, may enable thrushes to more effectively sense information about their progress over the ground. ACKNOWLEDGMENTS This project received vital support from individuals who gave long hours at unconventional times to assist with bird netting, equipment construction, nightly transmitter signal monitoring, and/or the driving of the tracking vehi- cle during migi'ations: W. L. Anderson; R. Applegate; J. A. Bailey; W. W. Coch- ran, III; T. C. Dunstan; W. R. Edwards; J. K. Hughes; G. B. Joselyn; Barbara Kermeen; Mary Ann Kjos; L. D. Mech G. G. Montgomery; Judy Montgomery D. Morpurgo; R. Phillips; A. J. Raim G. W. Swenson, Jr.; G. W. Swenson, III and R. O. Watson. Richard R. and Jean W Graber provided valuable consulta- tion, R. P. Larkin and G. C. Sanderson critically reviewed the manuscript, and D. L. Swofford advised us on statistical analysis. This research was supported by National Science Foundation gi'ants NSF-GB 3155 and GB 6680. APPENDIX DETERMINING LOCATIONS OF FLYING THRUSHES Cochran et al. (1967:214) describe a method of determining the locations 322 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 of flying thrushes by driving a track- ing vehicle under them. This and other location methods require that the ob- servers get close to a flying bird. From a knowledge of the approximate track speed and direction of a flying bird, a forward extrapolation of its flight path was plotted on a highway map. The places and times where the bird was likely to cross roads were noted. One of these road crossover places was se- lected by considering the average speed the vehicle could be expected to make in reaching the chosen place and the need to get there before the bird. Each new location of the bird was used to up- date the projected path and select the next crossover place. A general discus- sion of tactics and strategy are given in Cochran (1972:53-58). We point out here that tactics for keeping up with a bird are considerably diffei'ent and more easily accomplished than those for making the repeated close passes necessary for gathering the best data. Cochran et al. (1967: Fig. 4) employ- ed a horizontally aimed eight-element yagi receiving antenna rotatable through 360 ° of azimuth. The antenna was equipped for reception of both hori- zontal and vertical polarization. This antenna provided horizontal directivity to a bird if the elevation angle to the bird was less than about 45 °. For eleva- tion angles less than 45° the signal was strongest (a peak) when the anten- na was directed at the bird's azimuth, and it gi-adually weakened for azimuths on either side of the peak, reaching the minimum for azimuths in the vicinity of 90° fi'om the peak. A second, much weaker peak occurred when the anten- na was pointed directly away from the bird. As a bird entered a "cone" above the vehicle (elevation angles gi'eater than about 45°), the relation between signal strength and antenna azimuth could not be interpreted, peaks and nulls appearing at numerous and vary- ing azimuths. As a bird left the over- head "cone of uncertainty," the familiar relationship between antenna azimuth and signal strength returned. Position- ing the vehicle for such an overhead passage (crossover) required selecting a place along the bird's projected path, as discussed above. At this place I ahead of the bird), the bearing to the bird was approximately 180 ' different from the bird's previously plotted track direc- tion. In practice, the process of position- ing the vehicle was iterative. The first location where the vehicle was stopped was only approximately con-ect. As the bird moved closer, bearing changes were noted, and the vehicle was moved forward or backward in an effort to place the vehicle directly in the oncom- ing bird's path. The process was some- times frantic during the last moments, requiring rapid turning about or driv- ing at high speeds in reverse. Some- times the procedui'e failed, and the bird crossed the road well ahead of or behind the vehicle, a situation which, we will show, evolved into a method for determining both altitude and loca- tion. Here we are interested in how the accuracy of the information gathered in this simple crossover procedure was estimated. In our study we made the receiving antenna moveable in eleva- tion as well as azimuth (Cochran 1972: Fig. 4). With this alt-azimuth antenna mount, the cone of uncertainty was theoretically eliminated. In practice, because of the awkwardness of an alt- azimuth mount when directed in the near-overhead region, a cone of uncer- tainty remained for elevation angles above about 75°. By simple geometry it can be shown that (Da bird must pass directly overhead for entry and exit bearings (into and out of the cone of uncertainty) to differ by 180 "', and (2) for a bird's passage anywhere through the cone, the assumption that it passed directly overhead can err by no more than the altitude of the bird times the cosine of the elevation angle defining the edge of the cone. For example, for cones defined by 45 ° and 75 ^ elevation angles and a bird at an altitude of 100 m, the bird must pass over a point on the road within +70 and -1-26 m. re- September 1985 125 Years of Biological Research 323 spectively, of the position of the vehicle. Except on the rare occasions when crossovers were detected while we were within sight of mapped landmarks, our errors in positioning the vehicle on a map were typically ±50 m due to map plotting and vehicle odometer pi-ecision and accuracy limitations. Therefore, for birds flying at 100 m or below, there was little point in worrying about cone entry and exit bearings. This situation was fortunate, because the rapidly changing angles to birds that flew past at low altitudes did not allow us to get more than crude bearings. At higher bird altitudes cone entry and exit bear- ings are more important. The errors discussed above for a bird at 100 m altitude become 700 and 260 m for a bird at 1,000 m altitude. However, when thrushes were flying high, there was time to take bearings carefully and to position the vehicle so that the bird passed near the axis of the cone. We conservatively estimate that, with the alt-azimuth mounted antenna used in our study, the crossover positions we plotted were never more than 200 m, and usually less than 100 m, from the actual crossing point. Estimating the time of a bird's passage overhead required taking the mean of the times of its entry into and exit from the cone of uncertainty. A bird flying at a ground speed of 500 m/min (about 8.3 m/s) at an altitude of 1,000 m would pass through a cone de- fined by a 75° elevation angle in 62 seconds. In practice, these timings were never precise, especially for birds fly- ing high, mainly because the transi- tion into and out of the cone was never as distinct as presented here for pur- poses of explanation. It usually took five to ten transmitter pulses (about 5-10 seconds) to take a bearing and, failing in this, 10-20 additional sec- onds to determine that readings made no sense (the bird was in the cone). A similar delay occurred in determining that a bird had left the cone of uncer- tainty. Therefore, timings were usually late by 20 or 30 seconds. However, be- cause track speeds were determined be- tween successive crossovers, the sys- tematic errors due to the delays tended to cancel one another. We believe our crossover timing erroi", as it affected track speed calculations, was rarely greater than 30 seconds, and from limi- tations imposed by the geometry of a pass, never as great as 1 minute. The accuracy of the track directions and track speeds, as measured between successive crossovers, is estimated as follows. Assuming a worst-case 300-m error at right angles to a 20-km track segment, the error in computed track direction would be 0.85°. For a 20-km track segment flown at 8.3 m/s gi'ound speed, a 30-second timing error would cause a 0.2-m/s eri'or in computed track speed. It was possible to measure altitude from two or more elevation and azi- muth angles taken before and/or after a bird passed nearly overhead. Geomet- rically, one side and two angles define a triangle. In this case the length of the side was provided by track speed times the time interval between angle mea- surements. However, it was difficult to measure accurately the elevation angle to birds approaching head on or reced- ing (tail view), because these aspects gave mixed polarization, which magni- fied errors caused by signal energy reflected from the ground. Errors in the elevation angle of only 10-15° can cause errors of altitude measurement as great as 50 percent. Therefore, we used this triangle method only when circumstances precluded the use of the method described below. We have mentioned that birds cross- ing roads ahead of or behind the vehi- cle provided a way of measuring alti- tude at the crossover point. The method is described and illustrated in Cochran (1972), but will be briefly recounted here. It requires that the track direc- tion be known, preferably to within 10°, from prior data. The method is best described by example, starting with the tracking vehicle northwest of a bird holding a northerly course. The 324 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 tracking vehicle is driven east as if a crossover place were to be determined, but is held back so that the bird will pass in front of the vehicle. While the bird passes in front of the vehicle, the transmitting antenna (and bird) pi'es- ent a side aspect, which gives pure horizontal polarization. Polarization is important, because the best elevation angles are taken when the polarization of the received energy is purely hori- zontal. When the azimuth to the bird is straight ahead (as the bird crosses the road), the time, odometer reading, and elevation angle to the bird are recorded. The vehicle continues moving east while azimuths are taken, and soon the azimuth is north (a second odometer reading is recorded at this time). Because the track direction is north, the vehicle is now at the place where the bird recently crossed the road (the crossover place is deter- mined). The time of the crossover has ah'eady been recorded. The altitude of the bird as it crossed the road is the distance between the two odometer readings times the sine of the previous- ly recorded elevation angle. The method is valid even if the path direction is not at right angles to the road, but the desired side aspect of the bird is, in practice, lost if path-road angles are less than about 45 °. The trick is to be at a place on the road (when the bird crosses it) which gives an elevation angle in the useful range of 30-70°. Potential timing and positioning errors are not significantly different for this cross-ahead method from those for the crossover (overhead) method. The primary error in timing a crossing equals the product of the sine of the er- ror in the first (presumed dead ahead) bearing, the distance from the place this bearing was taken to the road- crossing point, and the reciprocal of a bird's track (ground) speed. For a 5° bearing error, a 1-km distance, and an 8-m/s track speed, the timing error would be 10.9 seconds. The error in computing the place of crossing equals the error in distance measurement plus the product of the sine of the sum of the errors of the assumed track direc- tion and second bearing taken, the track speed, and the time interval be- tween first and second bearings. In the example above, an 80-m distance eiTor, a combined track direction and second bearing error of 10° (these sum alge- braically and may cancel I, and a time interval of 41 seconds (vehicle speed of 55 mph), would cause a maximum crossover positioning error of 137 m (80 + 57). Altitude error equals the prod- uct of the difference between the tangents of the true and measured ele- vation angles and the sum of the dis- tance and the distance error. For the above example, if the bird's actual cross- over altitude was 1,000 m, if a 45° elevation angle was erroneously mea- sured as 35 °, and if there was an 80-m error in the measured distance the vehicle traveled between measurement points, then the calculated altitude would be 644 or 756 m for vehicle dis- tance short or long, respectively. The sensitivity of altitude compu- tation to elevation angle error results in highly variable accuracy of altitude etimates (Table 1). We tested the accu- racy of radio measurement of elevation angle against optical measurement to a transmitter attached to a kite and found that between 30° and 60° eleva- tion, given time for numerous vertical sweeps of the antenna, errors were less than +3°. In practice we seldom noted a consistency of elevation angle mea- surements which would suggest that their accuracy was within the few de- gi'ees indicated by the kite tests. One exception was Swainson's 7005. which had a continuous (instead of pulsed) transmitter that gi'eatly facilitated measurement of elevation angles (and azimuth bearings as well). The futility of using U.S. Weather Bureau (USWB) winds-aloft data is il- lustrated by the impossibility of anal- yzing the night of Veery 6705 (Fig. 2), which was tracked before we obtained winds-aloft measuring equipment. This veerv flew at different altitudes in September 1985 125 Years of Biological Research 325 winds which varied greatly with alti- tude, time, and geogi'aphic location. The veery descended as it passed through a slowly moving warm front with large shifts of the lower-strata winds north of the front. A case for a variety of behaviors could be made by choosing particular winds-aloft data from those available from the USWB for different times and places. Only constant-course behavior (complete compensation) can be excluded. We invite the reader to apply compass and rule to data of Fig. 2, which is scaled accurately enough for this purpose. Sometimes birds did fly close to USWB stations when winds aloft were being measured, e.g., thrush- es #5 and #19 (Cochran et al. 1967) dis- cussed in the text. The probability of this occurring was small in the 1960's and is lower now, because times when and places where the USWB measures winds aloft have greatly decreased. For example, we did not measure winds aloft when Swainson's 6905 passed south of Peoria, Illinois, around mid- night (Fig. 4), because we counted on obtaining the 2315 USWB measure- ments. When W8 went to the USWB for their data, we were told that they no longer measured winds at 2315 hours, a severe disappointment, as this wind information would have allowed us to analyze the middle portion of the flight. Even when USWB winds-aloft measure- ments were available for the place and time of a bird's passage, they were of limited use when wind shear was high, because they are computed for incre- ments of about 500 m (above about 800 m). Some USWB wind data we exam- ined showed a 180° shift between two adjacent 500-m levels! We do not imply that USWB personnel were ever less than helpful and cooperative; in fact we found that they would, for cost, given an hour or two of advance notice by telephone, measure winds aloft at times we specified. On several occasions, when we saw that a bird was headed toward one of their stations, we took advantage of this service. The widely distributed network of USWB stations with its capability for winds-aloft mea- surement represents a useful and acces- sible resource. We calculated winds aloft from the- odolite measurements of elevation and azimuth angles to ascending helium- filled balloons. We took readings at 15, 30, 45, and 60 seconds, and at 30-second intervals thereafter. These intervals provided wind data for altitudes of 30, 100, 150, and 195 m and at about 100-m intervals at higher altitudes. The theodolite was calibrated by sight- ing the north star or by a corrected magnetic compass sighting when the sky was cloudy. The horizon was estab- lished by a bubble level. A light was attached to the balloon for night read- ings, but this was unnecessary for mea- surements started 30 minutes after sunset or before sunrise, near the usual beginning and ending times of thrush flights. We tested how accurately bal- loons conformed to the assumed ascent rate by measurements from two theodolites. For six test ascents actual altitude did not deviate from the stan- dard rates (for a 30-g balloon) by more than 2 percent. Measurements taken at the beginning of the second night of flight of Swainson's 6905 (Fig. 5) are listed in Table 4. Note that we missed the 45-second reading in this set. Miss- ing one of the first few readings was common because of the difficulty in keeping the theodolite pointed at the nearby balloon as the balloon rapidly changed azimuth and elevation. We measured winds aloft before and after flights and as often as circumstances permitted during a flight. Calculated heading estimates (those with associated track speeds in Table 1) were obtained by using standard trigo- nometric equations (on a computer) in- stead of the equally valid but more time consuming gi-aphical method il- lustrated in Fig. 4 and 5. The accuracy of track and wind vectors we measured was usually good enough for the calcu- lation of approximate path vectors (near where winds were measured) to an accuracy of ±1° or 2° and 0.5-1.0 326 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 m/s. Unfortunately, altitude measure- ments were usually too inaccurate to permit an unambiguous choice of winds to use for the calculations. Therefore, lack of accurate altitude data and/or lack of wind data at particular places precluded the use of most of the track- ing data gathered during the study. Calculated heading estimates (± 3 ° :Table 1 ) resulted under ideal con- ditions when wind did not change sig- nificantly over the range of altitude estimates. We applied a 3° uncertainty to all heading estimates as a conserva- tive lower limit to the error range for calculations from measurements taken under these ideal circumstances. Ap- propriate (higher) error ranges were ap- plied when added uncertainty resulted from less than ideal track or wind vec- tor data or when wind varied signifi- cantly over the range of altitudes esti- mated for a bird. Headings were ex- cluded from the data set when the estimated errors exceeded about +10°. TAKING HEADINGS BY METHODS NOT REOUIRING WIND AND ALTITUDE DATA To understand these methods, it is useful to visualize the aspect, i.e., the projection, of a trailing wire antenna (on a bird) on an observer's plane of vi- sion, as viewed from various places. For purposes of discussion, we consider the antenna wire to be straight and to pass directly over the bird's tail parallel to the earth. Viewed from directly ahead of or behind a bird (at the bird's alti- tude) the wire is invisible (has a point aspect). Viewed from ahead or behind, but from the gi'ound, the wire appears as a vertical (to earth) line which sub- tends a smaller viewing angle than if the wire were turned broadside to the viewer (wholly in his plane of vision). This apparent shortening of the wire is called "foreshortening." The maxi- mum amount of energy which can be received from a wire antenna is propor- tional to the ratio of the foreshortened aspect to the broadside aspect (the wire in the observer's plane of vision), or simply, to the sine of the viewing angle (Va), which is 0° from head on and 90° for a broadside view. We say "maximum amount of energy which can be re- ceived" because, for an antenna pointed at a bird, the signal actuallj' received depends on the relative orientation be- tween the receiving antenna elements and the aspect presented by the trans- mitter antenna wire. In the example above (a view from the ground ahead of the bird) the wire antenna on the bird appears foreshortened and ver- tical. If a receiving antenna with ele- ments vertical to the gi-ound is pointed at the vertically appearing transmitter wire antenna, it will receive all the energy available. Conversely, if the receiving antenna elements are hori- zontal, none of the available energy* will be received. In contrast, viewed from the side there is no foreshortening (sine 90° = 1); so the maximum amount of energy will be received, but only if the receiving antenna is held with elements horizontal to match the horizontal aspect of the wire. Matching polarization maximizes signal and crossing polarization nulls the signal. More concisely, the received signal strength is proportional to the cosine of the angle between the relative as- pects of the receiving antenna ele- ments and the transmitting antenna wire, which we call the crossing angle (Ca). Thus, from any point the received signal is proportional to the product of sine(Va) and cosine(Ca). Therefore, re- ceived signal strength is maximum when Va=90 and Ca=0, and zero when Va = or Ca = 90. We stress the logical use of and and or in the preceding statement. It is unfortunate for discus- sion's sake that Va and Ca cannot in all cases be conveniently expressed in an orthogonal system relative to the earth's surface. However, the following state- ments are true for head-on, side, and underneath positions of an observer, the special cases where these angles bear an orthogonal relationship to the earth's surface. September 1985 125 Years of Biological Research 327 1. The strongest signals are received when an observer is to the side (Va=90) using horizontal (to earth) polarization (forcing Ca=0). 2. A weaker signal is available from directly ahead of (or behind) a bird (Va = elevation angle to bird) and is maximized by using vertical polariza- tion (forcing Ca= 0). 3. A zero signal is received for hor- izontal (to earth) polarization (forcing Ca=0) only when directly ahead of (or behind) a bird. 4. A zero signal is received for a bird overhead (or below) by forcing Ca=90. THE HEAD-NULL METHOD This method required measurement of the azimuth to the bird while condi- tions (2) and (3) above apply. Although it is theoretically possible to verify geo- metrically a directly ahead (or behind) position by noting a finite signal with vertical polarization (2 above) and a zero signal with horizontal polarization (3 above), it is impossible to confirm a zero signal, because no matter how sen- sitive a receiver is, a signal always dis- appears into the receiver's background noise before the signal becomes zero. In practice, the head-null method required driving the tracking vehicle across the dead-ahead-of-the-bird position and noting the disappearance and reap- pearance of the horizontally polarized signal and the corresponding azimuths to the bird (obtained by using the ver- tically polarized antenna). Dead ahead (of the bird) refers to the bird's heading and not to the projection of its course (probable path). When the tracking vehicle is a small distance (relative to the distance to the bird) on either side of dead ahead of the bird, the strength of signals from both horizontally and vertically polarized receiving antennas is very sensitive to the elevation angle and the angle subtended at the bird by the tracking vehicle and the dead- ahead azimuth. Both of these angles are small and change rapidly as the distance from the oncomng bird de- creases. The net result of the angular asymmetry thereby introduced between points equidistant on each side of the dead-ahead-of-the-bird position, com- pounded by the difference in distance to the oncoming bird at these two points, is that the azimuth for Ca=0 (corresponding to the dead-ahead vehi- cle position and from which the bird's heading is determined) is not the mean of the azimuths of the disappearance and reappearance of the horizontally polarized signal. The asymmetry and distance problems are compounded on a road which angles toward the bird and are mitigated on a road which angles away from the bird, the degree of compounding or mitigation depend- ing on the angles and distances in- volved. For these reasons and because it is practically impossible to correct quantitatively for them, this method is best executed at a high vehicle speed which approximately "freezes" the over- all geometry while the azimuths of disappearance and reappearance are being measured. This "freezing" is more closely approximated for a bird several kilometers distant, because the angles and distance to the bird are limited to smaller percentage changes during the time required for vehicle movement. Unfortunately, the signals from a more distant bird are weaker due to distance and small Va and eleva- tion angle, which move the places for signal disappearance and reappear- ance farther apart and require the vehicle to cover a greater distance dur- ing the measurements. Additionally, it is harder to measure bearings accu- rately when travelling at high speed, and roads which were not at right angles to a bird's heading, as was the usual case, further complicated execu- tion and interpretation. Therefore, the high precision typical of null measure- ments could not be realized in practice with the head-null method. The error range for the head-null estimates used in this paper (Table 1 ) was determined from the disappearance and reappear- 328 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 ance azimuths, which were definite bounds to the actual heading. We never purposely executed this method by crossing behind a bird, because falling behind was too often associated with permanently losing contact with a bird. An optimal tactic was to cross in front for a head-null estimate of head- ing and to turn around and obtain a crossover location and an altitude esti- mate by the means previously discussed. THE CROSS-POLARIZATION METHOD The cross-polarization method of obtaining a heading estimate is concep- tually simple, a matter of getting under the bird and as in (4) above, rotating the receiving antenna about a vertical axis until a null indicated Ca was equal to 90°. This method was a natural companion to the crossover positioning method described previous- ly. When the bird was in the cone of uncertainty, it was also in position for rotating the vertically pointed yagi antenna to the null position (Ca=90°) and reading this position directly on the azimuth compass rose. In practice this maneuver was difficult, especially for birds at lower altitudes, because birds were not in the cone of uncertain- ty long enough to allow us to take good measurements. Actually, the cone of uncertainty represented a zone of con- fusion for the antenna operator, and it was not until late in the study that we became skillful in executing the cross- polarization method. We consider +4° to be a sufficiently conservative base error range for these measurements, which were precise to 1 ° or 2 ° relative to the azimuth of the vehicle. The major source of error was in the vehicle's azimuth. The three cross-polarization heading estimates for Swainson's 7118 were obtained from an aircraft circling high above the bird. In general, an air- craft would be an excellent platform from which to use the cross-polarization method if a belly-mounted dipole an- tenna could be rotated from inside the aircraft. In the case of Swainson's 7118 we used a hastily rented aircraft and pilot and were forced to use an antenna clamped to the footstep. The aircraft had to be circled without banking it (using rudder only) alternately clock- wise and counterclockwise, because the antenna was not mounted symmet- rically on the airframe. The mean of 10-12 readings, (half taken during clockwise circling), which ranged over 10° in azimuth, was used for each of the estimated headings. Frequent gs'ro- scope calibration was required. Only three heading estimates were obtained because of time taken in navigating, fueling, locating the bird, and deter- mining its altitude by the crude prox- imity method described in Cochran (1972). The last heading estimate was obtained as the bird descended through a large wind shear, and the associated track segment is without a track speed (Table 1) because the end point of the segment is where we located the bird on the gi'ound without having detected the exact landing time. IN SUMMARY The procedures and methods are described above as set-piece operation, the making of a specific measurement at a specific time or when a certain situation pertained. In reality, bird tracking was more chaotic and required the gathering and recording of a large amount of potentially redundant data without knowing beforehand which would be good enough and which would be acquired at the right places and times to be useful for analysis. For example, we mentioned taking an ele- vation angle when the azimuth to a bird indicated that it was crossing a road ahead of the vehicle. In practice, azimuths and elevations were contin- ually taken and recorded as the vehi- cle sped toward the bird. As likely as not, no good set of measurements was September 1985 125 Years of Biological Research 329 available for the dead-ahead position, but was available when the bird was at some small (less than 20°) angle to the right or left of dead ahead. The geom- etry of all of the methods described is somewhat tolerant of deviations from the ideal situation as long as the devia- tions are taken into account. More often than not, attempts to execute the various procedures failed. We were for- tunate that the data from the relative- ly small percentage of successful execu- tions were sufficient to address the question of wind drift in thrushes. The equipment used in this decade- old study was primitive by the stand- ards of 1983, but the geometries of the methods remain fixed. Today, gi'eater quantities of more accurate data could be obtained by the use of microcomputer control of several vehicle-mounted an- tenna systems, each specialized for gathering data appropriate to the geometry of a particular method. LITERATURE CITED Able, K. P. 1977. The flight behavior of individ- ual passerine nocturnal migrants: a tracking radar study. Animal Behaviour 25:924-935. 1982. The effects of overcast skies on the orientation of free-flying nocturnal migi-ants. Pages 38-49 m F. Papi and H. G. Wallraff, eds., Avian navigation. Springer-Verlag, New York. Allen, F. H. 1939. Effect of wind on flight speeds. Auk 56:291. Alerstam, T. 1976. Bird migi-ation in relation to wind and topogi-aphy. Ph.D. Dissertation. Uni- versity of Lund, Sweden. Barker, S. F. 1957. Induction and hypothesis. Cornell University Press, Ithaca and London. Bellrose, F. C. 1967. Radar in orientation re- search. Pages 281-309 in D. W. Snow, ed.. Pro- ceedings of the XIV International Ornithological Congi"ess. Blackwell Scientific Publications, Oxford and Edinburgh. , and R. R. Graber. 1963. A radar study of the flight directions of nocturnal migi'ants. Pages 362-389 in Proceedings XIII Interna- tional Ornithological Congress. BiNGMAN, V. P., K. P. Able, and P. Kerlincer. 1982. Wind drift, compensation, and the use of landmarks by nocturnal migi'ants. Animal Behaviour 30:49-53. Bloch, V. R., B. Bruderer, and P. Steiner. 1981. Flugverhalten nachtlich ziehender Vogel - Radardaten uber den Zug verschiedener Vogel- typen auf einem Alpenpas. Vogelwarte 31:119-149. Bruderer, B. 1982. Do migi-ating birds fly along straight lines? Pages 3-14 in F. Papi and H. G. Wallraff eds.. Avian navigation. Springer- Verlag, New York. Cochran, W. W. 1972. Long-distance tracking of birds. Pages 39-59 in Sidney R. Galler, Klaus Schmidt-Koenig, George J. Jacobs, and Richard E. Belleville, eds., Animal orientation and navigation. National Aeronautics and Space Administration Special Publication 262. 1980. Wildlife telemetry Pages 507-520 in Sanford D. Schemnitz, ed.. Wildlife man- agement techniques manual. Wildlife Society, Washington, D.C. , G. G. Montgomery, and R. R. Graber. 1967. Migi-atory flights of hylocichla thrushes. Living Bird 6:213-225. Emlen, S. T, and N. J. DeMong. 1978. Orienta- tion strategies used by free-flying bird mi- gi'ants: a radar tracking study. Pages 283-293 in K. Schmidt-Koenig and W. T Keeton, eds., Animal migration, navigation, and homing. Springer-Verlag, Berlin, Heidelberg. Gauthreaux, S. A., and K. P. Able. 1970. Wind and the direction of nocturnal songbird migra- tion. Nature 228:476-477. Graber, R. R. 1968. Nocturnal migration in Illi- nois. Different points of view. Wilson Bulletin 80:36-71. , and W. W Cochran. 1960. Evaluation of an aural record of nocturnal migi-ation. Wilson Bulletin 72:253-273. , J. W Graber, and E. L. Kirk. 1971. Illi- nois birds: Turdidae. Illinois Natural History Survey Biological Notes 75. Griffin, D. R. 1973. Oriented bird migration in or between opaque cloud layers. American Phil- osophical Society Proceedings 117:117-141. James, F. C. 1955. The influence of the Mississippi River on the nocturnal migj-ation of birds. M. S. Thesis Louisiana State University, Baton Rouge. Larkin, R. R, and D. Thompson. 1980. Flight speeds of birds observed with radar: evidence for two phases of migratory flight. Behavioral Ecology and Sociobiology 7:301-317. Nisbet. I. C. T, and W. H. Drury. 1967. Orienta- tion of spring migrants studies by radar Bird- Banding .38:173-186. 330 Illinois Natural History Survey Bulletin Vol. 33, Art. 3 Rabol, J. 1974. Correlation between coastal and inland migratory movements. Dansk Ornith- ologisk Forenings Tidsskrift 68:5-14. Raim, a. 1978. A radio transmitter attachment for small passerine birds. Bird-Banding 49:326-332. Richardson, W. J. 1976. Autumn migration over Puerto Rico and the western Atlantic: a radar study. Ibis 118:309-332. Seets, J. W., and H. D. Bohlen. 1977. Compara- tive mortality of birds at television towers in central Illinois. Wilson Bulletin 89:422-433. Thorpe, W. H. 1949. Recent biological evidence for the methods of bird orientation. Linnean Society of London Proceedings 160:85-94. Vleugel, D. A. 1962. Wind and orientation of mi- grating butterflies in comparison with birds. Pages 15-19 in XI International Kongress fur Entomologie Wien 1960. Williams, T.C, P. Berkeley, and V. Harris. 1977. Autumnal bird migi'ation over Miami studied by radar: a possible test of the wind drift hypothesis. Bird-Banding 48:1-10. Afternoon Session Sumnnary James R. Karr A 50th anniversary symposium or even the centennial meeting only 25 years ago would no doubt have been very different from that held in 1983 to commemorate the 125th anniver- sary of the Illinois Natural History Survey. Noteworthy differences might include the conceptual context of the scientific questions addressed and, especially, the increased technological dependency of late 20th-century biolog- ical science. But commonalities would also be obvious, the most important of which would be the firm foundation in natural history and the use of that knowledge in the interpretation of pat- tern and process in nature. Each paper in the afternoon session draws on its own unique combination of these and other factors. Goldman takes his title from the classic paper of Stephen A. Forbes and forges an im- pressive array of observations and in- sightful interpretations to account for changes in Lake Tbhoe. He demonstrates the importance of long-term research, of the integi'ation of observations of ap- parently unrelated phenomena, indeed of the perseverance required to accum- ulate sufficient information to demon- strate pattern in nature and the process that generates and, in this case, de- gi-ades natural resources of consider- able value. Goldman's work at Lake Tkhoe, like research that led to improvement of the Thames River (Gameson & Wheeler 1977) and Lake Washington (Edmond- son 1977), demonstrates that the causes of environmental degi'adation can be identified. The work at Tahoe Dr. James R. Karr is a Professor of Ecology, Ethology, and Evolution, University of Illinois, UrbanaChampaign, and a Research Affiliate, Section of Aquatic Biology, Illinois Natural History Survey. parallels studies of the Illinois River conducted by staff of the Illinois Natu- ral History Survey (Mills et al. 1966; Starrett 1972; Bellrose et al. 1983). The challenge for the Natural History Sur- vey in the future is to do for Illinois streams what Goldman has done for Lake Tahoe. But the lesson of Lake Tahoe is more than "the lake is dam- aged and may even face ecological de- struction" - the land itself may be destroyed as well through erosion and other degradation. In Illinois we must take action to protect rivers and the land they drain. Without them even the most innovative technological soci- ety faces overwhelming environmental crises. The second paper of the afternoon session is concerned with very different geogi'aphic and biological scales. The concern of Philipp, Kaminski, and Whitt about the genetic integrity of largemouth bass populations develops from a careful integration of natural history, population genetics, and evolu- tion. The application of new technol- ogies to identify genetic traits allows a more sophisticated view of the respon- sibility of fishery managers and, in- deed, of resource managers in a broader context. The folly of mixing gene pools of northern and southern largemouth bass is demonstrated thi'ough docu- mentation of the negative effects of careless stocking programs on region- ally adapted populations. We see modern technology, allowing sophisti- cated analysis of pattern, combined with good old-fashioned natui'al history to yield the potential for more informed management of natural resources. More conventional forms of management techniques, such as habitat manipula- tion, harvest regulation, propagation, and stocking, must be combined with genetic management. 331 332 Illinois Natural History Survey Bulletin Vol. 33. Art. 3 Page reviews reproductive behavior and its evolution in about 150 species of percid fishes. Spawning behavior provides the focus for an analysis of an array of life-history attributes that vary in concert with reproductive be- havior. The accumulation and integra- tion of detailed information on the biology of many species provide a foun- dation for the synthesis of theoretical insights and help to define manage- ment options and strategies. With Cochran and Kjos we emerge from the water to explore the biology of bird migration. Specifically, they deal with the problems posed by wind and overcast skies for thrushes on the long voyage between breeding and win- tering grounds. I can attest to the pre- cision of these navigational skills be- cause, like others, I have had migi'ant warblers and thrushes return to the same wintering territoiy in Panama after annual trips to breed in North America. Cochran and Kjos document the ability of thrushes to compensate for wind and cloud cover during migi-ation. With an insightful combination of technical sophistication and natural- history wisdom, avian migi-ation is not made simple, but it at least becomes more comprehensible. Cochran and Kjos show that thrushes have preferred flight directions and speeds for accom- plishing their navigational feats. They seek winds at a variety of heights that pi'ovide for these preferences. When winds are such that flight speeds are reduced to less than 2 m/s, the birds simply land and wait for better con- ditions. Throughout this set of papers, nat- ural-history information accumulated laboriously from tedious observations over extended periods combines with sophisticated applications of modern technology and biological theory to yield knowledge of factors that will, when taken collectively, provide the op- portunity for better management of natural resources as well as greater understanding of our natural world. As I listened to these presentations, several general thoughts came to mind. These thoughts involve inherent diffi- culties in the integi'ation of knowledge from various biological disciplines and the incorporation of that knowledge in a broader societal context. 1. The study of organisms and the interpretation of natural-history data are key responsibilities of the Natural History Survey. The job has been well done in the past but will become more difficult in the future because biolo- gists face the same problem that herbi- vores face in dealing with plant chem- isti'y while feeding and in avoiding predators. In the evolutionary battle between food and feeder, each is con- stantly trying to get one step ahead of the other. In a very real sense human society is on the same treadmill in the use of pesticides (evolution of pest re- sistance) and technolog>' (the inability of biological systems to change rapidly enough to survive the gi'owth of that technology). I am confident that, unless we destroy the biological systems of earth, including ourselves, no long- term panacea will be forthcoming be- cause of the evolutionary process. In the end. Homo sapiens as we know the species may simply become extinct through evolutionary processes much as H. erectus was replaced. This continuing evolutionary proc- ess is both good and • ad. Good because we can continue to enjoy the vitality, excitement, and challenge of biological research; bad because we will always be racing to keep up with insults to the integi'ity of our planet "s thin biological mantle while we try to extract goods and services from that mantle. 2. The biologist "s perspective must always be both microscopic and macro- scopic. Short-term or local solutions should not be used without careful evaluation of their probable long-term and regional (even national and inter- national) implications. Perhaps the most difficult task facing biologists will be the integi'ation of information ob- tained over a hierarchy of spatial and September 1985 125 Years of Biological Research 333 temporal scales. We must simultan- eously take a close look at detail while standing back to view systems as inte- gi-ated wholes. 3. Interactions between basic and applied biologists have often been less than satisfactory. Many theoreticians, often with limited field experience, have looked down upon managers, per- haps because of a disdain for mission- oriented research. Conversely, resource managers have been reluctant to eval- uate the merit of recent theoretical developments, perhaps because they are too "esoteric." That the two can ef- fectively merge, indeed depend on each other, is demonstrated by the papers presented here and by the research go- ing on at INHS. The complexity of im- portant resource issues requires the effective merger of the two camps and cooperation to their mutual benefit. 4. As I have noted several times, a substructure of systematic and natural- history information (of inherent value j on their own, as ai-e music and art) is essential as a backdrop for virtually all natural-resource decisions. Theoretical considerations alone are inadequate for informed resource management. 5. Finally, biologists must recog- nize their insights and integrate them into the social, political, and economic contexts in which societal decisions are made. Biologists can no longer advo- cate policies that ignore any of these realities. Similarly, economists, politi- cians, and others cannot ignore long- term trends in the degi'adation of life support systems as if they were not relevant to current and future condi- tions for human society. If biologists do not participate as equal partners, biol- ogy will be ignored, or perhaps worse, we will continue to base decisions on the input of those not familiar with the facts of biology. In summary, challenges to biolo- gists for the future are immense. Look- ing back on the last 25 years, I see growth and intellectual vitality. I ex- pect the future to be even more exciting and look forward to learning more about the gi-owth at the 150th anniver- sary celebration of the Illinois Natural History Survey. LITERATURE CITED Bellrose, F. C, S. p. Havera, F. L. Paveglio, Jr., and D. W. Steffeck. 1983. The fate of lakes in the Illinois River valley. Illinois Natural History Survey Biological Notes 119. Edmondson, W. T. 1977. Recovery of Lake Wash- ington fromeutrophication. Pages 102-109 in J. Cairns, Jr, K. L. Dickson, and E. E. Her- ricks, eds., Recovery and restoration of dam- aged ecosystems. University Press of Virginia, Charlottesville. Gameson, a. L. H., and A. Wheeler. 1977. Resto- ration and recovery of the Thames estuary. Pages 72-101 m J. Cairns, Jr., K. L. Dickson, and E. E. Herricks, eds.. Recovery and restora- tion of damaged ecosystems. University Press of Virginia, Charlottesville. Mills, H. B., W. C. Starrett, and F. C. Bellrose. 1966. Man's effect on the fish and wildlife of the Illinois River. Illinois Natural History Survey Biological Notes 57. Starrett, W. C. 1972. Man and the Illinois River Pages 131-169 m R. T. Oglesby, C. A. Carlson, and J. A. McCann, eds., River ecology and man. Academic Press, New York. Gardner. Gene, M.S., Assistant Supportive Scientist Hofmann, Joyce. Ph.D., Assistant Supportive Scientist McGiffen. Kathryn, M.S.. Assistant Supportive Scientist Peratt. Suzanne J., Assistant Supportive Scientist Solecki, Mary Kay. M.S., Assistant Supportive Scientist Suloway, Liane B.. M.S., Assistant Supportive Scientist Sweeney. Bernice P., Assistant Supportive Scientist Taft, John B.. M.S., Assistant Supportive Scientist Timmons. Randall G,, B.S., Assistant Supportive Scientist Ulazek, Eric F., B.S,, Assistant Supportive Scientist Wetzel, Mark J.. M.S., Assistant Supportive Scientist Butler. Randal L.. B.S.. Junior Professional Scientist Brooks. Thomas M., B.S., Junior Supportive Scientist Holt. Aieta A,. B.S., Junior Supportive Scientist Kasprowicz, Jeanine M., B.S., Junior Supportive Scientist Keene, Dennis. B.S-. Junior Supportive Scientist Malmborg. Patti L.. M.S., Senior Technical Assistant Brower, Joan L., B.S., Technical Assistant Heydon, Steven, MS.. Technical Assistant Keevin. Thomas M.. M.S.. Technical Assistant Kirts, Ann. B.S., Technical Assistant Miles, Charlene. Technical Assistant Possoa. Steve. Technical Assistant Swanson, Charles S., B.S.. Technical Assistant Wetzel, Faith D., B.S., Technical Assistant Crossett, Lorrie, B.S., Junior Technical Assistant Cummings, Kevin S.. M.A., Junior Technical Assistant Carney, Douglas A.. M.A., Junior Technical Assistant Irish, Jeffrey A., B.S., Junior Technical Assistant Pescitelli, Pamela A.. B.S.. Junior Technical Assistant Sandberg, Sherri L., B.S., Junior Technical Assistant Sherman, Renee, B.S-, Junior Technical Assistant Tuman. Marycarol, B.S., Junior Technical Assistant Vanderah. Glendy C, Junior Technical Assistant SECTION OF WILDLIFE RESEARCH Sanderson, Glen C, Ph.D., Wildlife Specialist and Head Bellrose, Frank C. Sc.D., Wildlife Specialist and Principal Scientist Graber, Richard R,. Ph.D., Ornithologist and Principal Scientist Emeritus Anderson, William L,. M.S., Wildlife Ecologist Edwards. William R,. Ph.D., Wildlife Ecologist Graber. Jean W., Ph.D., Ornithologtst Hanson. Harold C . Ph.D., Wildlife Specialist Nixon, Charles M., M.S.. Wildlife Ecologist Cochran, W.W.. Jr., B.S.. Associate Wildlife Specialist Havera, Stephen P., Ph.D., Associate Wildlife Ecologist Larkin, Ronald P.. Ph.D., Associate Wildlife Ecologist Quine. Douglas B,. Ph.D., Associate Biophysicist Warner, Richard E . PhD., Associate Wildlife Ecologist Westemeier. Ronald L , M.S., Associate Wildlife Ecologist Wood, Susanne G., Ph.D., Associate Chemist Burnett, Christopher D., Ph.D.. Assistant Wildlife Ecologist Hansen, Lonnie P., Ph.D., Assistant Wildlife Ecologist Robinson. Scott K., Ph.D., Assistant Wildlife Ecologist Witham. James H.. Ph.d., Assistant Wildlife Specialist Buhnerkempe. John E.. M.S., Associate Supportive Scientist Brown. Bruce W., Ph.D., Assistant Supportive Scientist Duzan, Ronald E., Assistant Supportive Scientist Gilluly. David, M.S.. Assistant Supportive Scientist Herkert, James, R.. M.S., Assistant Supportive Scientist Jackson. Cynthia G , B.A.. Assistant Supportive Scientist Jones. Jon Martin, M.A-, Assistant Supportive Scientist Pool. Robert R.. B.S,, Assistant Supportive Scientist Szafoni, Robert E., M.S., Assistant Supportive Scientist Crompton, Robert D.. Junior Professional Scientist Sects, James W.. Junior Professional Scientist Anderson, Elizabeth A,, Senior Technical Assistant Baum. Sharon E., B S., Senior Technical Assistant Belcher, H. Kathleen. B.S., Senior Technical Assistant Hunter. Katherine J.. Senior Technical Assistant Iko, William M,, B.S., Technical Assistant Martin. Brian H., B.S., Technical Assistant Raim, Arlo J., M.S., Technical Assistant Georgi, Michelle, Junior Technical Assistant Roat. Katie E., Junior Technical Assistant ADMINISTRATIVE UNIT Schwartz, Melvin. Administrative Assistant to the Chief and Unit Head Supporting Services Clark. Phyllis. Mail and Supply Room Supervisor Dillman. WUma G,, Fiscal Assistant Gross, Larry D., Assistant for Operations Hale. Sue M.. Grants and Contracts Officer McNamara. James, Operations Assistant Reed. Joyce. Junior Technical Assistant Rohl, Chris, Operations Assistant Sanders, Jacque, Personnel Officer Office of Communications Korb, Robert, Supervisor Duzan. Patty L., Word Processing Coordinator Hodgins, Audrey S,. M.A.. Technical Editor LeMere, Lloyd, Technical Illustrator McClellan, Shirley, B.S., Associate Technical Editor Steger. Eva, B.A., Assistant Technical Editor Technical Library Heister, Carla. M.A., Technical Librarian Lusk, Monica, Library Technical Assistant CONSULTANTS AND RESEARCH AFFILIATES: Aquatic Biology. Robert Costanza, Ph.D.. Professor of Wetland Ecology. Louisiana State University. Baton Rouge; Bruce Hannon. Ph.D., Professor of Geography, University of Illinois; James R. Karr, Ph.D., Professor ofEcology. Ethology, and Evolution, University of Illinois; Lewis L. Osborne. Ph.D., Assistant Professor ofUrban and Regional Planning. University of Illinois; John A. Tranquilli. Ph.D.. Regional Manager. Ducks Unlimited; and Gregory S. Whitt, PhD,, Pro- fessor of Genetics and Development, University of Illinois; Botany and Plant Pathology, Jean D, Schoknecht. Ph.D.. Associate Professor ofLife Sciences, Indiana Slate University, Terre Haute; Faunistic Sureys and Insect Identification, Roderick R. Irwin, B.A.. Chicago. Illinois; Wildlife Research, Nancy T. Burley, Ph.D.. Associate Professor ofEcology. Ethology, and Evolution, University of Illinois; Robert L, Jones, Ph.D., Professor of Soil Mineralogy and Ecology. University of Illinois; Willard D, Klimstra. Ph.D., Pro- fessor of Zoology and Director of Cooperative Wildlife Research, Southern Illinois University; and Norman D. Levine. Ph.D., Pro- fessor of Veterinary Parasitology. Veterinary Research, and Zolitgy, and Director of the Center for Human Ecology. University (jf Illinois; Entomology, Nelda Alger. Ph.D., Associate Professor of Genetics and Development, University of Illinois; Robert L Metcalf, Ph.D., Professor of Biology and Research Professor ofEntomology, University of Illinois; and Gilbert P. Waldbauer, Ph.D.. Professor of Entomology. University itf Illinois. Some Publications of the ILLINOIS NATURAL HISTORY SURVEY BULLETIN Volume 33. Article l.-Diptera, or Ti-ue Files, of Illinois I. Tabanidae, L. L. Pechuman, Donald W. Webb, and H. J. Teskey. April 1983. 122 p. Volume 33. Article 2. -An Annotated Bibliography of the Illinois Herpetological Literature 19601980, and An Updated Checklist of Species of the State. Michael A. Morris, Richard S. Funk, and Philip W. Smith. April 1983. 15 p. Volume 33. Article 4. -The Crayfishes and Shrimps (Decapoda) of Illinois. Lawrence M. 1985. 113 p. BIOLOGICAL NOTES I20.-Recreational Fishing in the Kankakee River, Illinois. Robert J. Graham, R. Weldon Larimore, and William F Dimond. June 1984. 13 p. 121.-Bibliogi'aphy of Illinois Vegetation. Paul G. Risser. August 1984. 51 p. 122.-The Life History of the Mud Darter, Etheostoma asprigene, in Lake Creek, Illinois. Kevin S. Cummings, James M. Grady, and Brooks M. Burr. December 1984. 16 p. 123.-The Life History of the Shorthead Redhorse. Moxosloma macrolepidotum, in the Kankakee River Drainage, Illinois. Michael J. Sule, and Thomas M. Skelly. 1985. 16 p. 124.-Illinois Birds: Vireos. Jean W. Graber, Richard R. Graber, and Ethelyn Kirk. 1985. 52 p. SPECIAL PUBLICATION Number 2.-Landscape Ecology: Directions and Approaches. Paul G. Risser, James R. Karr, and Richard T. T. Forman. November 1984. 18 p. CIRCULAR 49.-The Dunesland Heritage of Illinois. Herbert H. Ross. August 1963 (Reprinted May 1974). 28 p. 51. -Illinois Ti-ees: Selection, Planting, and Care. J. Cediic Carter. March 1977 (Third printing). 123 p. 52. -Fertilizing and Watering Trees. Dan Neely, and E. B. Himelick. December 1971 (Third printing). 20 p. 54. -Corn Rootworm Management in Canning Sweet Corn. W. H. Luckmann. J. T. Shaw, D. E. Kuhlman, R. Randell. and CD. LeSar. March 1975. 10 p. 55.-Observing, Photographing, and Collecting Plants. Kenneth R. Robertson. August 1980. 62 p. MANUSCRIPTS High quality manuscripts dealing with any aspect of natural history will be considered for publica- tion in one of the Illinois Natural History Survey publication series: Bulletin, Biological Notes, Circular, and Special Publication. The author need not be an employee of the Survey, but if not, will be required to pay printing costs. Manuscripts must follow the recommendations of the third edition of the Council ofBiology Editors Style Manual except that journal names in the Literature Cited section are to be spelled out in full. The Survey expects to publish only one or two manuscripts by non-Survey authors yearly. Please send three copies of manuscripts to be considered for publication to Office of the Chief, Illinois Natural History Survey, 607 East Peabody Drive, Champaign, Illinois 61820. List of available publications mailed on request No charge is made for publications of the Illinois Natural History Survey. A single copy of most publications will be sent free to anyone requesting it until the supply becomes low. Costly publications, more than one copy of a publication, and publications in short supply are subjects for special correspondence. Such correspondence should identify the writer and explain the use to be made of the publication or publications. Address orders and correspondence to the Chief Illinois Natural History Survey Natural Resources Building 607 East Peabody Drive Champaign, Illinois 61820