Behavioral Characters for the Higher Classification of Orb-Weaving Spiders Author(s): William G. Eberhard Source: Evolution, Vol. 36, No. 5 (Sep., 1982), pp. 1067-1095 Published by: Society for the Study of Evolution Stable URL: http://www.jstor.org/stable/2408084 . Accessed: 11/02/2011 13:36 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=ssevol. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Society for the Study of Evolution is collaborating with JSTOR to digitize, preserve and extend access to Evolution. http://www.jstor.org Evolution, 36(5), 1982, pp. 1067-1095 BEHAVIORAL CHARACTERS FOR THE HIGHER CLASSIFICATION OF ORB-WEAVING SPIDERS WILLIAM G. EBERHARD Smithsonian Tropical Research Institute and Escuela de Biologia, Universidad de Costa Rica, Ciudad Universitaria, Costa Rica Received August 25, 1980. Revised June 22, 1981 A number of studies have shown that substantiate this has yet to be performed, behavior patterns can be useful taxonomic although Risch (1977) has made a prom- characters (see review by Mayr, 1958; also ising start by finding that two congeneric Evans, 1966; Crane, 1975; Michener et species' webs were more similar to each al., 1978; Eickwort and Sakagami, 1979; other in a number of details than to those Greene, 1979). The webs and web-build- of two species from other genera. ing behavior of orb-weaving spiders are It is also possible that some details of complex, apparently stereotyped charac- construction behavior are employed by ters, and as such offer promise of being spiders in unchanged form to construct a useful in indicating taxonomic affinities. variety of different web forms, and that As Levi (1978a, 1978b) has noted, how- these behaviors are thus slow to evolve ever, this promise has not been fulfilled. even while the webs themselves change The gross, relatively superficial web char- rapidly. This paper describes several such acters such as the presence or absence of behaviors and shows that they are conser- stabilimenta, or open versus closed hubs vative enough to characterize the classical which have been studied to date have not subfamily and family groupings which proved to be useful indicators of subfa- have been based on adult morphology and milial relations. Webs are directly in- can thus be used to indicate relationships volved in orb weavers' interactions with between them, a topic on which there is a number of aspects of their environments currently substantial disagreement (e.g., (particularly prey), and relatively minor Lehtinen, 1967, 1975; Levi, 1978a, 1980; changes in environmental factors could Robinson and Robinson, 1978, 1980; result in selection for changes in web form. Opell, 1979). Levi argues (echoing Darwin, 1859) that Obviously the more characters used in at least some aspects of web design might constructing a system of relationships the thus be expected to be evolutionarily non- greater the likelihood of the results being conservative and of little use in indicating correct. A great deal is known about the higher taxonomic relationships. morphology of orb weavers, but no com- This does not, however, eliminate the prehensive surveys of characters are yet possibility that webs and web-building available (but see Levi, 1980), so they can- behavior may be useful as taxonomic not yet be included. This paper is meant characters. It is possible that some details to provide useful data for later syntheses, of web design with apparently low func- not to give the final word on the classifi- tional significance may be more conser- cation of orb weavers. vative than other more obviously func- tional characters. I have the impression Choice of Characters that many web characters are not scat- When one has collected data on a pre- tered randomly among the webs of ara- viously untried character or set of char- neoid species, and that their patterns of acters, the first step in their analysis is to occurrence will be of at least some use in compare their distribution with previous systematics, particularly at generic and taxonomic schemes based on other char- tribal levels. The analysis necessary to acters. If the new characters are function- 1067 1068 WILLIAM G. EBERHARD ally independent of the older ones-as dence, and the most probable directions would appear to be the case here since the of transformation are given in Appen- morphology of male and female genitalia, dix 1. eye positions and structures, cheliceral morphology, cephalothorax form and oth- MATERIALSAND METHODS er morphological characters used by other Specimens of spiders are deposited in workers would seem to have little func- the Museum of Comparative Zoology, tional relation to the details of leg move- Cambridge, Mass. 02138. The specimen ments and thread manipulation during numbers mentioned in the text and ap- web construction and attack and court- pendices refer to numbered labels includ- ship behavior-then there are several pos- ed in individual spiders' vials. At present sible results. If the previous schemes are it is unfortunately impossible to identify all incorrect, the distribution of the new many (most?) orb-weaving spiders at the characters will not be in accord with any species level, and most specimens are giv- of them, and the new characters can be en only generic names. This situation does used to construct a new set of relation- not seem likely to change radically in the ships. If, on the other hand, the previous near future, and it thus seemed wise to schemes are at least approximately cor- proceed with the presentation of the data rect, then the degree to which the new in this paper rather than wait for more characters "fit" will depend on the rates complete identifications. Unless otherwise of evolution of these characters. Those noted all references to subfamilial and which have evolved very slowly will be tribal groupings are based on the monu- uniform over many different groups, those mental work of Simon (1892). The ulob- which have evolved relatively rapidly will orid names follow Opell (1979). vary even within given taxa, and still oth- The techniques of observation and their ers will more or less match the taxonomic limitations are described in the Appendi- scheme. Since some of the behavioral ces. As might be anticipated, many details characters examined in this study do have of web-building behavior are extremely distributions similar to the classical group- stereotyped within a given species (see ings of Simon (1892) based on adult mor- Appendix 1). This uniformity makes web- phology, this scheme is probably at least building behavior an attractive set of approximately correct. characters to study because relatively brief There were other characters which were observations suffice to characterize a constant in all the groups studied (e.g., species. starting sticky spiral from the edge rather than the center or any other part of the RESULTSAND DISCUSSION web), and others which varied within giv- Observations of at least 148 species in en subfamilies (e.g., pulling motions of at least 55 genera are presented in Appen- legs IV on sticky spiral as it was pro- dices 2 and 3 and summarized in Table 1. duced). These behaviors are not included The data are arranged according to the here since the basic objective of the study groupings of Simon (1892). Both substan- was, after testing the classical scheme, to tial concordance within subfamily and attempt to use the behavioral characters family groups with respect to a number of to indicate relationships between subfam- characters and clear differences between ilies and families. Thus only those char- subfamilies and families are evident. These acters which appear to be constant or patterns constitute confirmation of the nearly constant within subfamilies or fam- classical groupings. The one distinction ilies and which also differ between them which is not confirmed is that between are discussed. Detailed descriptions of Tetragnathinae and Metinae, as there were these characters, the different states which no consistent differences between species they assume, their functional indepen- of these two groups. SPIDER BEHAVIOR AND TAXONOMY 1069 TABLE 1. Summary of data in Appendices 2 and 3 and Robinson and Robinson, 1980. The data are relatively scarce for some charactersfor Theridiosomatidae and Anapidae and their characterization here is tentative. Parentheses indicate character states thought to be secondarily derived within the group; in all cases fewer than 10% of the species observedfor that group have the presumed secondary state, and the species