hibernation, delayed dispersal) as attempted by Blumstein
and Armitage (1997b, 1998) onto those phylogenies may
ultimately provide the greatest insight into the evolution of
the traits in question (Dobson 1985).
Our efforts to understand ground squirrel sociality must
also extend beyond the application of comparative meth-
ods. Indeed, while comparative methods have proven in-
valuable in suggesting both proximate and ultimate hy-
potheses to test, comparative data are correlative and thus
do not necessarily reveal underlying causation. This prob-
lem is particularly evident in attempts to uncover the un-
derlying basis of ground squirrel sociality, as the retention
of young within the natal area may be the product of at-
traction to the resources of the natal area and direct fitness
benefits, but inescapably results in, and is thus correlated
with, the clustering of female kin (Michener 1983a; Armi-
tage 1988). To determine whether such clustering is the
result of an active attraction to kin proper, as opposed to
a mutual attraction of those individuals to a common area,
manipulative experimentation is required. A transplant of
individuals to new sites using methods similar to those
of Wiggett and Boag (1993a) or Michener (1996), so that
groups composed exclusively of kin or of nonkin are
formed, could begin to tease out the role that kinship itself
plays in determining aspects of spatial overlap and social
behavior. The removal of close kin within a colony of Rich-
ardson’s ground squirrels by Davis (1984c) increased inter-
ruptions of foraging with bouts of vigilance, increased chas-
ing and fleeing, reduced sharing of core areas by neighbors,
and decreased breeding success relative to what was doc-
umented on the other half of the same colony, where re-
movals left females with close kin as contiguous neighbors.
The absence of replication, particularly in light of extensive
badger predation within the nonkin area, and potential bias
introduced into spatial relationships via the attraction of
individuals to familiar burrow systems within their natal
areas, preclude attributing the behavioral asymmetries doc-
umented by Davis to relatedness per se, but suggest that a
similar approach holds promise.
Subsequent investigations must also explore the general-
ity and strength of certain patterns of kin-biased behavior
and the presumptive costs associated with those. Data sug-
gesting costs associated with ground squirrel alarm call-
ing have been presented only for Belding’s ground squir-
rels (Sherman 1977, 1985), and though cited extensively,
even these data must be interpreted cautiously. Statistically
signifiicant differences were evident only for pursuits, and
predators killed equal numbers of noncallers and callers
(Sherman 1985). Similarly, kin bias in cooperative chas-
ing (Sherman 1981a) merits further empirical study. The
greater likelihood of displacement in chases of nonkin rel-
ative to close kin, as indicated for Richardson’s ground
squirrels by Davis (1984b), might give the appearance of
cooperation between kin when both of them are simply re-
sponding to intrusions of nonkin on their territories.
Our efforts to explain sociality in ground squirrels are
also hampered by our limited understanding of the func-
tional aspects of social discrimination. Kin discrimination
clearly functions to promote nepotism in some species, but
may serve other purposes in more solitary species (Mateo
2002). While in the past it was commonplace to equate kin
discrimination and kin-differential behavior with kinship-
based sociality, such conclusions are not always warranted.
We must work toward a greater understanding of the roles
that individual (Hare 1998a), neighbor (Hare 1998b), and
even group-member discrimination (Hare 1992, 1994) play
in promoting the fitness of individuals within ground squir-
rel societies. Progress toward understanding the function of
social discrimination at any level is less likely to be made by
studying sociality overall than by observational and manip-
ulative studies focusing on how such discrimination con-
tributes to the expression of specific social traits within and
among species (e.g., joint hibernation, allogrooming, play,
territory defense, alarm calling).
It is particularly revealing that three of what are com-
monly regarded as the most social ground squirrel species
within their respective genera (Olympic marmots in Mar-
mota,black-tailed prairie dogs in Cynomys,and Colum-
bian ground squirrels in Spermophilus) fail to show kin-
biased behavior in certain contexts (Barash 1973; Hoogland
1986; Hare and Murie 1996, respectively). Clearly, advanc-
ing sociality cannot be predicated entirely upon kin selec-
tion, but must involve social benefits that accrue via other
mechanisms. Just as kin selection recognizes the fitness pay-
offs implicit in the broader inclusion of collateral kin overEcology, Kinship, and Ground Squirrel Sociality: Insights from Comparative Analyses 353Figure 29.3 Intraspecific variation in sociality requires further examination, as
is evident from field studies of woodchucks (Marmota monax). Photo by J. Hare.