Anim. Behav., 1982, 30, 700-713 SIBLING ASSOCIATION AMONG SCHOOLING TOAD TADPOLES: FIELD EVIDENCE AND IMPLICATIONS BY BRUCE WALDMAN Section of Neurobiology and Behavior, Cornell University, Ithaca, New York 14853 Abstract. The results of recent laboratory studies suggest that the ability to recognize kin may be a widespread phenomenon among diverse animal groups, but the question of how such recognition abilities influence individuals' behaviour under natural conditions has not always been considered. As an assay for sibling recognition, I measured the sibship composition of experimental schools formed by tadpoles of the American toad (Bufo americanus) in their natural habitat. Sibling cohorts were obtained from amplectant pairs in the laboratory and reared in groups either (1) in separate tanks, apart from non-siblings; or (2) in a common tank, separated from non-siblings by a screen partition across which water could pass. Tadpoles were dye-marked to indicate sibship identity. Pairs of sibling groups were then mixed together in a bucket and released in natural outdoor ponds, where they formed schools. I found significant differences in sibship composition among schools in 79 % of all sampling periods; overall, 64 % of the schools sampled were significantly biased in favour of one or the other sibship. The formation of sibling schools appears to result from a behavioural recognition mechanism rather than from differential habitat selection. Tadpoles preferentially formed schools with familiar siblings over both familiar and unfamiliar non-siblings, suggesting that sibling discrimination is not based on familiarity alone. Although experience may affect the ontogeny of sibling recognition, sibling preferences are apparently formed early in development, perhaps prior to hatching. Grouped individuals may in some situations increase their inclusive fitness by associating with kin: aposematic advertisement, alarm signalling in response to predators' attacks, and kin-influenced growth regulation are suggested as three specific advantages conferred by the formation of sibling schools. Tadpoles of many anuran amphibians aggregate in social groups generally similar in form to fish schools (reviewed in Wassersug 1973). Some aggregations may form as larvae indivi- dually respond to microclimatic gradients (Brattstrom 1962; Beiswenger 1977), but social factors probably are important for the main- tenance of these aggregations (Beiswenger' 1972; Waldman, unpublished data). Moreover, larvae of at least two anuran species preferentially associate with siblings in laboratory tests (Bufo americanus (Waldman & Adler 1979; Waldman 1981), Rana cascadae (Blaustein & O'Hara 1981)), suggesting that tadpole schools might consist of kin groups. In this paper, I present evidence that tadpoles of the American toad (B. americanus) do form sibling schools in the field, discuss the mechanism by which tadpoles recognize siblings, and suggest some factors that may favour schooling with one's siblings. Tadpoles that school, whether with kin or non-kin, may gain many of the advantages conferred by group living (Alexander 1974; Bertram 1978). Increased feeding ability or decreased predation risk appear the most likely advantages for schooling tadpoles. Tadpoles can collectively stir food material from the substrate, creating local areas of food abundance and thus facilitating feeding (Richmond 1947; Bragg 1965; Beiswenger 1972). Anti-predator advan- tages have not been as well documented for tadpoles, but several mechanisms are possible: schooling individuals may benefit by concealment from predators, increased capabilities to detect and evade predators, or by confusing or satiating predators (reviewed in Radakov 1973; Shaw 1978; Hobson 1979). Some schooling fishes suffer less predation than their solitary con- specifics (e.g. Neill & Cullen 1974; Seghers 1974; Major 1978); presumably other school- ing orgamsms gain simiIar advantages. Although the propensity to school, even with unrelated conspecifics, may thus be selectively advantageous, individuals that preferentially associate with kin might accrue additional benefits. For instance, if schooling has evolved in part as a 'selfish herd' response to predators (Williams 1964; Hamilton 1971), individuals consistently located at the edge of a school (perhaps because larger conspecifics displace them) might be more vulnerable to predation than they would be alone outside the school. Unless they are somehow compensated for their added risk, these peripheral individuals should 700