Rodent Societies: An Ecological & Evolutionary Perspective

Armitage (1988) noted that large patches of suitable habitat
support yellow-bellied marmot colonies composed of mul-
tiple matrilines, while small habitat patches (satellite sites)
support only a single matriline with few individuals. As
habitat area increases, more burrow systems are available
and the number of matrilines increases. The mean size of
matrilines, however, is inversely correlated with habitat
area, reflecting the opportunity of individual females to es-
tablish their own smaller matrilines where burrow systems
are available, and thereby avoid the potential cost of re-
productive suppression in a larger group (Armitage 1988).
Among yellow-bellied marmots certain burrows also con-
stitute preferred hibernacula that are used year after year,
and cooperating individuals may outcompete others for
these choice hibernacula (Armitage 1988). Selection may
thus favor social relationships in the face of competition for
burrows as resources (King 1984).
Basic habitat variables also interact with antipredator
benefits. Holmes (1991) has shown that pika (Ochotona
princeps) will extend their foraging into previously unoccu-
pied areas if rocky outcrops providing refuge from preda-
tion are extended into those areas. The ground squirrels are
also known to be risk-sensitive foragers, adjusting escape
responses relative to their distance from a burrow (Bonen-
fant and Kramer 1996) and the trade-off between foraging
and vigilance according to their proximity to the center ver-
sus edge of the group (Armitage 1962; Hoogland 1979b,
1981). Indeed, Hoogland’s (1981) study clearly implicates
the relative density of cover as a factor influencing the ex-
pression of coloniality in prairie dogs. Individuals aggregat-
ing together as a group enjoy a variety of antipredator bene-
fits (Sherman 1977; Hoogland 1981; Kildaw 1995). Thus at
least some component of antipredator benefit likely served
as a primary impetus for group living early in the evolution
of ground squirrel sociality.
The spatial distribution and general abundance of food
resources do not, in and of themselves, appear to have been
instrumental in promoting ground squirrel sociality. Food
is widely distributed within patches of suitable habitat, and
individuals do not appear to have been selected to cooper-
ate in the defense of food proper. The distribution of food
in time, however, is invariably clumped, with a discrete
growing season providing an abundance of resources at one
time, and a paucity of resources at another. If the climate
is such that the growing season is short, as it is for ground
squirrels occupying high latitude or high elevation sites,
then the attainment of adult body mass, and hence breed-
ing, may be delayed, and selection may favor delayed dis-
persal (1973, 1974a; Armitage 1981, 1999).
The harsh winter climate imposed by high elevation and
high latitude sites may also act directly to promote social
tolerance. Arnold (1990b) has demonstrated that social

thermoregulation is critical to successful hibernation in al-

pine marmots, and thus a harsh environment alone can se-

lect for social grouping (Armitage, chap. 30 this volume).

Group hibernation in alpine marmots reduces winter mor-

tality of all group members, though the presence of infants,

particularly in poor-quality hibernacula, imposes costs on

other group members (Arnold 1990b). These costs, how-

ever, are not borne equitably by all subordinate subadults:

full siblings of infants within the group lose more mass than

others, suggesting investment on their part in nonlittermate

siblings (Arnold 1990b). Thus in addition to direct fitness

benefits of reduced mortality and the increased likelihood

of territory inheritance, delayed dispersal may allow indi-

viduals to accrue indirect fitness benefits by assisting non-

descendant kin.

Kinship and Ground Squirrel Sociality

As summarized previously, clumped burrow resources, a

short growing season, a harsh winter environment, and in-

tense predation pressure set the stage for group living in the

ground squirrels. Michener (1983a), however, regarded kin

selection as the driving force behind the evolution of ground

squirrel sociality, in that the retention of daughters in the

natal area results in the formation of matrilines composed

of related females. Where relatives are available for social

interactions, as is commonly the case for ground squirrels

(Vestal and McCarley 1984), kin selection may promote

the evolution of nepotism. Favoritism toward kin among

ground squirrels appears in a variety of contexts. In both

Belding’s (Spermophilus beldingi; Sherman 1977) and

round-tailed ground squirrels (S. tereticaudus; Dunford

1977a), females with either descendent or nondescendent

kin in the local population are more likely to emit alarm

vocalizations in response to a terrestrial predator than fe-

males without kin present. Further, female Belding’s ground

squirrels are more likely to participate in chases of territo-

rial intruders with close female kin than with nonkin (Sher-

man 1981a). In Arctic ground squirrels, neighboring female

relatives have been reported to pool recently weaned young

into common burrows (McLean 1982). The most universal

expression of kin-biased behavior among ground squirrels,

however, appears in the context of social tolerance, mani-

fested both in terms of behavioral interactions and the use

of space. In Spermophilusalone, evidence of kin-differential

behavior has been obtained for thirteen-lined (S. tridecemli-

neatus;Holmes 1984b; Vestal and McCarley 1984), Frank-

lin’s (S. franklinii; Hare 2004), round-tailed (Dunford

1977a), Belding’s (Sherman 1980a, 1981a; Holmes and

Sherman 1982), Richardson’s (S. richardsonii;Yeaton 1972;

Michener 1973a), Arctic (S. parryii;McLean 1982), and

Ecology, Kinship, and Ground Squirrel Sociality: Insights from Comparative Analyses 349