Behav Ecol Sociobiol (1983) 12:121-128 Behavioral Ecology and Sociobiology 9 Springer-Verlag 1983 Dispersion of a Small-Island Population of the Spider Mctepeira datonn (Araneae: Araneidae) in Relation to Web-Site Availability Thomas W. Schoener and Catherine A. Toft Department of Zoology, University of California, Davis, California 956] 6, USA Received July 8, 1982 / Accepted November 29, 1982 Summary. Metepeira datona is the commonest orb- weaving spider on very small islands of the central Bahamas. Where studied, large spiders built high- er, wider and deeper webs than did small spiders, and they oriented webs modally so as to expose the least surface to the prevailing winds. About 17% of webs contained more than one spider, but large females almost never shared webs with one another. All individuals combined, and sex-age classes individually (including large females), were significantly more clumped in 2-dimensional space than random. Moreover, potential web sites for large females were about 10 x as common as occu- pied sites, were dispersed randomly, and were sig- nificantly less clumped than occupied sites. We conclude that a rudimentary coloniality exists and speculate on predator defense, prey capture, sharing of silk and structural support as adaptive functions. small islands surveyed during April-May, 1982 (Schoener and Toft 1983). Wherever observed, individuals of M. datona appear from casual observation to be clumped. Such clumping could reflect habitat patchiness and/or a rudimentary social tendency on the part of the spiders. The major purpose of this paper is to present a precise method to test whether clumping of potential web sites appears responsible for clumping of occupied web sites. This is the first paper of which we are aware that attempts to determine quantitatively the dispersion of avail- able sites for orb-weaving spiders (for related ap- proaches, see Colebourn 1974; Riechert etal. 1973; Riechert 1976). We also present data on structural habitat and web characteristics of var- ious size-sex classes in the population. We shall show that while potential sites are distributed vir- tually randomly, actual sites are highly clumped. The possible adaptive significance of such clump- ing is briefly considered. Introduction The Bahamas Archipelago includes thousands of tiny islands, many of which contain well-estab- lished populations of the orb spider Metepeira datona Chamberlin and Ivie. Such populations can reach very high densities on islands where lizards, the most dominant kind of vertebrate in the Bahamas (Schoener and Schoener 1983a, b), are absent (Schoener and Tort 1983). Here we report on a population of M. datona inhabiting a tiny (65.5 x 19.8 m) unnamed island free of the other common Bahamian diurnal orb spiders (Gastera- cantha cancriformis (L.), Argiope argentata (Fab.)) and of breeding vertebrates. This population was the second densest of any orb spider on the 93 Materials and Methods Members of the genus Metepeira build a web comprising an orb and an abutting amorphous barrier web (Levi 1977). Mete- peira also typically builds a retreat, where the spider is generally found. The retreat is a cocoon-like structure not much larger than the spider and fashioned of bits of vegetation or shell woven together. We measured the following web characteris- tics: (1) height: distance from the hub of the orb to the ground; (2) span: maximum horizontal distance, including support threads; (3) depth: perpendicular distance from the orb to the distal part of the barrier web; (4) orientation; compass angle of the plane of the orb (to the nearest 5~ (5) substrate: nature of the vegetation, rock or other web to which the given web is attached. Body length (to the nearest 0.5 mm) and, where possible, sex of each individual within the study plot were recorded, as well as the individual's location. Individuals smaller than e 2 mm could seldom be distinguished as to sex. Size ranges for adults were : ~ s ~3 mm; ~ s 3 5 ram. We erected the follow-