Why do yellow-bellied marmots call?

Correspondence and present address: D. T. Blumstein, Schools of Behavioural and Biological Sciences, Macquarie University, Sydney, NSW 2109, Australia (email: [email protected]). K. B. Armitage is at the Department of Systematics and Ecology, University of Kansas, Lawrence, KS 66045, U.S.A. When we see animals do things that are potentially risky, such as alarm calling around predators, we ask why. Why should an animal possibly identify itself to a predator when it could simply make a silent escape? Hamilton’s (1964) inclusive fitness (kin selection) hypothesis is often used to explain the evolution and maintenance of alarm-calling behaviour (Maynard Smith 1965; and references in Hauber & Sherman 1998). Mathematically, the logic is elegant: individuals surrounded by relatives acquire a genetic benefit both from warning or otherwise protecting relatives that are descendants as well as nondescendants. If the benefit from both ‘direct’ and ‘indirect’ pathways to fitness (Brown 1987) is sufficiently large, the benefit may outweigh any costs associated with performing the seemingly risky behaviour and the behaviour will evolve and/or be maintained. A reasonable empirical question that arises when looking at a system that may have evolved through kin selection is: do the animals have a way of assessing relatedness and modifying their behaviour appropriately, as would be predicted if they were attempting to maximize their inclusive fitness? We addressed this question indirectly by seeing if and how yellow-bellied marmot, Marmota flaviventris, alarm-calling behaviour was influenced by the presence or absence of both descendent and nondescendent kin (Blumstein et al. 1997). Hauber & Sherman (1998); questioned, on logical and methodological grounds, our assertion that yellow-bellied marmots respond in a way suggesting that they maximize their direct, and not indirect, component of inclusive fitness. Below we address their major criticisms and acknowledge on-going disagreement over the definition of kin selection (Sherman 1980; Shields 1980). Is total r the ‘simple weighted sum’ that Grafen (1982, 1984) showed could not be used to calculate inclusive fitness? In a word, no. We used total r as a measure of the opportunity for kin selection (strictly it is the sum of the weights that could be used to calculate inclusive fitness benefits). As Hauber & Sherman (1998) noted, Creel (1990, page 230) wrote that to calculate an individual’s inclusive fitness, the effect of one individual on others’