als may place themselves at greater risk of predation in
warning fellow group members of the presence of a pred-
ator (Sherman 1985). While kin selection may ultimately
explain such behavior —where potential benefits associ-
ated with self-sacrifice are directed toward kin (Holmes and
Mateo, chap. 19, this volume; Dunford 1977a; Sherman
1977)— several alternative mechanisms must be considered
to explain cooperation among nonkin (Connor 1995), and
may also contribute to the benefits accruing via social in-
teractions among relatives.
Individuals interacting cooperatively within a group may
simultaneously enhance each other’s fitness, as is the case
where nonrelatives participate in assemblages of hibernat-
ing individuals (Arnold 1990). Such elements of social be-
havior are best considered mutualisms. Individuals that
manifest apparent altruism but are repaid later in kind by
their beneficiary are offsetting any costs of their initial sac-
rifice through reciprocal altruism (Trivers 1971). It is con-
ceivable that an individual placing itself at greater risk
by alarm calling in one instant will be repaid by receiving
a warning from another individual at a later time. Recent
demonstrations that Richardson’s ground squirrels (Hare
1998a) and yellow-bellied marmots (Blumstein and Daniel
2004) discriminate among callers as individuals, and that
Richardson’s ground squirrels selectively attend to callers
that have proven reliable in the past (Hare and Atkins 2001),
are consistent with the recognition of individuals and the
memory of how those individuals behaved in past encoun-
ters, as required for the evolution of reciprocal altruism
(Trivers 1971; Axelrod and Hamilton 1981). Data are not
available for any ground squirrel species, however, to ad-
dress the extent to which individuals modify the expression
of apparent altruism according to that of their social con-
temporaries. Even outside the context of strict reciprocity,
costs associated with alarm signaling could be offset by
benefits associated with receiving signals from others where
alarm signaling becomes a social convention within a group
(i.e., byproduct mutualism; Connor 1995), though no data
are available to address this hypothesis. Similarly, no data
are available to address the extent to which an individual’s
apparent personal sacrifice may enhance its social status
within the group, thereby improving the apparent altruist’s
lot by virtue of direct fitness payoffs associated with that en-
hanced social standing (Zahavi 2003).
Future Directions
Comparative analysis has implicated a variety of ecological
factors, and has identified components of both direct and
indirect fitness contributing to the expression of ground
squirrel sociality. Sociality imposes tangible costs on indi-
viduals (Alexander 1974; Wilson 1975; Hoogland 1979a),
including reproductive suppression (Blumstein and Armi-
tage 1998). Such costs are offset, however, by benefits de-
rived from (1) increased survival of young, (2) potential
breeding opportunities as a subordinate, (3) potential for
territory inheritance and reproductive benefits associated
with becoming a territory holder, (4) participation in social
thermoregulation during hibernation (e.g., Arnold 1990),
and (5) alloparental care (Armitage 1999). Further research,
however, is necessary.
While both King (1984) and Armitage (1988) highlighted
burrow systems as resources critical to survival and repro-
duction, no systematic study has examined how the disper-
sion and availability of burrow systems influences social re-
lationships, either within or among ground squirrel species.
Similarly, further insight is likely to be gained by applying
the comparative analysis of vegetative cover, both to the
density and size of ground squirrel colonies, just as Hoog-
land (1981) did for prairie dogs. Such analyses could make
fruitful use of Geographical Information Systems to address
the patchiness of available habitat at multiple scales, to de-
termine whether habitat availability at the landscape level
correlates with the expression of sociality.
Considerable effort should also be devoted to consider
plasticity in the expression of sociality within certain ground
squirrel species, so that the factors underlying such vari-
ability might become apparent. Woodchucks, for instance,
have traditionally been regarded as solitary (Bronson 1964;
Ferron and Ouellet 1989a), though Meier (1992) reported
considerable overlap in space use and frequent and amic-
able interactions among female kin and their coincident
adult male. Chris Maher (personal communication) has ob-
served similar social interactions among woodchucks, and
thus both the behavior of woodchucks themselves, and
comparative schemes treating woodchucks as solitary, bear
further examination (fig. 29.3). For Columbian ground
squirrels, both litter size (Murie et al. 1980; Risch et al.
1995) and age of first reproduction (Zammuto and Millar
1985; Murie 1995) have been shown to vary with eleva-
tion. It remains to be seen, however, whether those differ-
ences in life-history traits translate into differences in so-
cial behavior.
Future comparative analyses require a well-supported
phylogeny of the ground-dwelling squirrels. Molecular phy-
logenies are available for marmots (Steppan et al. 1999),
and are now beginning to appear for a wide range of sciurid
species (Harrison et al. 2003; Herron et al. 2004). The phy-
logenies resolved by such analyses, however, are often at
odds with traditional phylogenetic schemes based largely
upon morphological traits (e.g., Hafner 1984). Once phy-
logenetic relationships among species are established with
confidence, the mapping of specific social traits (e.g., joint352 Chapter Twenty-Nine