A Host Plant Is More Than Its Chemistry zen de t t w f :4 Field naturalists frequently ob­ 1983; Coley 1982, 1983a, h^Coñnor et serve that the intensity of insect herbi- al. 1983; Haukioja 1980; Heinrich & vory in a natural defoliation event is Collins 1983; Kareiva 1982; Lawton not spread uniformly over individual 1983; McClure 1983; Niemela et al. conspecific plants and may even vary 1982; Price et al. 1980; Rausher & within one plant crown. Until the early Papaj 1983; Schultz 1983; Schultz et al. 1960’s, the general explanation would 1982; Stanton 1982; Thompson 1983; have been that heterogeneity in physi- Washburn & Cornell 1981; Wint 1983). cal conditions, carnivory, and/or the ar­ I feel that the philosophy that gener- rival of the herbivores generate such ates them deserves máximum encour- lack of uniformity. During the past 20 agement. years, much research on the Chemical I have two practical reasons for at- defenses of plants has paved the way for tempting this leavening, though I am the now commonplace concept that sure that the reader will think of such heterogeneity of herbivory in a others. First, one has only so much defoliation event may also be caused by time and resource to expend on a given heterogeneity in the nutrient or de- study of the intensity of herbivory, and fense properties of plants or plant parts there is a very real question of whether (e.g, Kogan 1977). Indeed, a number of efforts should be focused on determin- studies have found this to be the case. ing the (potential) internal plant prop­ For example, squirrels browse much erties that drive the system or on de- more heavily on terpene-poor signing observations and experiments ponderosa pines than on their more to reveal the external factors crashing terpene-rich conspecifics a few meters down on the lowly Caterpillar. Second, away (Farantinos et al. 1981). Pana- in attempting to understand the eco- manian forest insects browse young logical and evolutionary distribution of leaves more intensely than they do con­ herbivores among their host plants, it specific oíd ones, presumably because is easy to forget that what might be of the greater nutrient valué and lesser termed the carnivory regime and cli- toughness of the former (Coley 1982, mate regime of a host plant individual 1983a, b). Chrysomelid beetles and or population are as much traits of a Finnish caterpillars vary their feeding plant as is its Chemical profile. Both of patterns in response to seasonal these regimes should count for much in changes in host leaf chemistry (Haré whether and to what degree a plant or 1983; Haukioja & Niemela 1979). plant part is a suitable host. When we What I would like to do in this ask why a Caterpillar feeds on only one essay is to give the pendulum a push particular host species, it may be as back toward a middle ground, remind- much that it is highly adapted to the ing all of us (including me) that when predator risks and desiccation regimes a herbivore moves onto a host plant it of that plant as that it is adapted to the gets the outside of the plant as well as plant’s internal chemistry. By the same what is inside. Such integrative studies token, when we ask what does a her- are now beginning to appear in the bivorous generalist have to do to be a literature (e.g., Abrahamson et al. generalist, it may be as much that it has to be able to withstand the preda- Dr. Daniel H. Janzen is a Professor of Biology, tion risks of living on various kinds of University of Pennsylvania, Philadelphia. backgrounds (e.g., Heinrich & Collins ñl/lcUúL foyiSfe./}/) ¡ <, C cfirc cfA tj ccjDpiU-eÁ hM -ij \(r i i ■íMi R ./1 4 /L tL í d 142 Illinois N atural History Survey Bulletin Yol. 33, Art. 3 1983) as to have the gut chemistry to I have chosen briefly to describe tolérate various kinds of food (e.g., four systems as illustrative case histor­ Ahmad 1983). Which ability carne first ies rather than to dwell on hypothetical may be lost in the decomposed pages structure. I do this because of my opin­ of time. ión that theory in evolutionary ecology I also have an apologetic reason for is intrinsically prone to outrun descrip- attempting to meld two sequential tion of what is actually happening out fashions. I think we erred in not there. recognizing two blended questions in the seminal essay that argued that since the herbivores did not eat the green world to the ground, or even THE ARENA down very much, they must be regu- lated by the carnivores (Hairston et al. The attitudes and examples in this 1960). Question one is why don’t all the essay derive from my experiences with herbivores eat up all the plants? The the herbivore array of a lowland tropi­ answer that lay undiscussed by Hairs­ cal forest, that of Santa Rosa National ton et al. (1960) is that most of the green Park, in northwestern Guanacaste world is inedible to any given species Province, Costa Rica (this site is des- of herbivore. Also lying dormant was cribed in detail in Janzen 1983a and the derivative evolutionary question of in Boza & Mendoza 1981). This mosaic why doesn’t any given herbivore species of deciduous forest, evergreen forest, evolve the ability to eat many kinds of semi-evergreen forest, and pastures plants? Question two is the real ques­ regenerating to forests occupies about tion in Hairston et al. (1960); why don’t 11.000 ha from 0 to 350 m elevation the herbivores that can readily and between the Pan-American Highway with impunity consume a species of and the Pacific Ocean. The area has a host plant eat their host to oblivion? 5-6-month dry season (approximately Every plant species has at least one December through April), and 1,000- herbivore that can eat it. To some de- 2.000 mm of rain falls during the re- gree a plant’s herbivores do consume it, mainder of the year. Portions of the thereby leaving resources for other park were an operating cattle ranch plant species, but to a large degree they from no later than 1710 to 1978. The do not, with the consequence that com- vegetation contains at least 680 species petition and the physical environment broadleafed plants (at least 400 species determine much of the structure of of perennial woody plants) and sup- vegetation arrays. The very same car- ports at least 3,000 species of cater- nivory and climatic regimes that pre- pillars plus several hundred species of vent herbivores from eating their hosts other animáis that eat living plant to oblivion are also the traits of the parts. There are checklists of plants potential new host that must be over­ (Janzen & Liesner 1980), birds (Stiles eóme when a herbivore evolutionarily 1983), reptiles and amphibians (Scott moves to, or incorporates, a new host. et al. 1983), mammals (Wilson 1983), Nothing I have said is new, but I and butterflies (DeVries 1983) for the feel that the emphasis is different from park. The plant distributions within that of contemporary ecology and evo­ this vegetation range from nearly lutionary biology; this emphasis may monospecific stands of very large trees be witnessed in two recent books on (e.g., 10-20-m-tall stands of Quercus coevolution (Futuyma & Slatkin 1983; oleoides Cham. and Schlecht., Hymen- Nitecki 1983). Virtually no attention aea courbaril L., Ateleia herbert-smithii is given to this subject, while the (Pittier) to highly mixed vegetation coevolution of herbivores and plant where as many as 200 species of woody chemistry plays a prominent role in plants may occur in 100 ha and adult examples and in generation of theory. conspecific crowns are usually sepa- September 1985 125 Years of Biological Research 143 rated by one to many allospecific through the forest for 5 minutes or less crowns. At Santa Rosa, herbivory by at any time during daylight hours. The caterpillars, the focus of this essay, is moths darted among the shrubs and characteristically highly heterogene- treelets at a height of about 1-3 m. ous among years and among individ­ They touched branchlets, twigs, and uáis, species, and age classes of plants buds with legs and the tip of the abdo­ (e.g., Janzen 1981). 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 palé green egg on the bud, newly expanding leaf, thorn, or ANATOMY twig it contacted (Fig. 1). It then flew OF A DEFOLIATION EVENT to other branches of the same plant or, about equally frequently, off to neigh- The event boring plants. Both species of Randia During the 1983 rainy season, a were beginning budbreak; a few indi­ representative defoliation event oc- viduáis were covered with a thin layer curred at Santa Rosa. The impact of the of newly expanding leaves, while others herbivores was highly heterogeneous. had only swelling buds. The outcome 1 briefly describe the ecology of this im- of this oviposition, by what must have pact as an example of a pattern that been several thousand moths in the could have been generated either by study area, was the deposition of tens the heterogeneity of internal plant of eggs to a thousand or more eggs on chemistry or by mortality factors exter- each Randia in the forest (of 214 plants nal to the plant (or both). In fact, the briefly examined, all had some eggs).