species; Blumstein and Armitage 1997b) and between so-
cial complexity and ground squirrel life-history tactics, in-
cluding: the proportion of females breeding within a group,
age of first reproduction, length of gestation, litter size,
length of the lactation period, offspring survival, and female
body mass (twenty-five species: Blumstein and Armitage
1998). Alarm-call repertoire size was positively correlated
with social complexity (Blumstein and Armitage 1997b),
supporting Marler’s (1977, p. 46) suggestion that the great-
est elaboration of communication would be expected where
“trends towards increasing interindividual cooperation con-
verge with the emergence of social groupings consisting of
close kin.” This overall relationship remained in phylo-
genetically controlled contrasts, though in considering the
genera independently, the relationship between social com-
plexity and alarm-call repertoire size was upheld only within
Marmota,not within Cynomysor Spermophilus. Blumstein
and Armitage (1997b) conclude that social complexity has
contributed to the evolution of alarm-call repertoire size
in the ground squirrels, but that other factors, including
morphology, the acoustic properties of the environment,
and the specific functional needs dictated by the organism’s
ecology are also significant in determining communicative
complexity. Increased social complexity was also accompa-
nied by a decrease in the proportion of females within a
group that bred, an increase in the age of first reproduction,
a decrease in litter size, and enhanced survival of offspring
(Blumstein and Armitage 1998). Neither the duration of
gestation or lactation varied consistently with changing so-
cial complexity, though it is apparent that sociality imposes
a trade-off between per capita reproductive output and off-
spring survival (Blumstein and Armitage 1998).
The elucidation of factors underlying sociality in ground
squirrels is complicated by variation in the criteria investi-
gators have employed to quantify sociality. While this prob-
lem is not unique to comparative work on ground squirrels
(e.g., Lacey 2000), it is necessary to consider the limitations
of previous attempts to make progress in the future. Barash
(1973) and Rayor and Armitage (1991) contrasted species
in terms of social interactions, identifying life-history and
ecological correlates of more frequent and amicable inter-
actions among members of each species under considera-
tion. The schemes put forth by Armitage (1981) and Mich-
ener (1983a), while broader taxonomically, are based upon
qualitative criteria that in effect prejudge the importance
of delayed dispersal and the spatial overlap of kin, respec-
tively. While providing a continuous rather than categori-
cal metric of social classification, Blumstein and Armitage’s
(1997b, 1998) social complexity index also presumes that
advancing sociality can be equated with enhanced overlap
of sex /age cohorts within the species considered. As such,
their identification of life-history traits associated with in-
creased social complexity (Blumstein and Armitage 1998)
is circular. Further, it is unclear whether the factors emerg-
ing from contrasts of social systems across the Marmotini
as a whole can explain the narrower range of variation in
sociality expressed within certain genera, such as Spermo-
philus. Given these problems with broader comparative
classifications, it is useful to revisit the ecological and be-
havioral underpinnings of sociality and to critically evalu-
ate evidence as to how such factors contribute to the evolu-
tionary appearance and maintenance of sociality.Ecological Factors Promoting SocialitySpatial and temporal overlap among individuals is required
for benefits to accrue via sociality (Holmes and Sherman
1983). Thus a comprehensive examination of sociality in
the ground squirrels must first identify the factors promot-
ing the aggregation of individuals into groups (Ebensper-
ger 2001a). Ecological factors figure prominently in setting
the stage for sociality (Alexander 1974; Slobodchikoff and
Shields 1988).
Contagious distributions of individuals can be the prod-
uct of habitat heterogeneity, such that individuals are con-
strained to co-occupy suitable habitat patches (e.g., Malizia
1998; Stallman and Holmes 2002). While comparative data
are not available to examine the correlation between over-
all habitat patchiness and sociality, there is little doubt that
the distribution of burrow systems, underlain by suitably
drained and arable subsoils, promotes the aggregation of in-
dividuals. Subterranean burrows represent critical resources
for ground squirrels in that burrows facilitate escape from
predators (Blumstein 1998), and provide a stable and rela-
tively secure microenvironment during periods of reduced
activity and /or increased vulnerability such as resting, rear-
ing young, and hibernation (King 1984). Indeed, burrows
serve the same functions as expansible and defensible nests,
which Alexander et al. (1991) argued to be a primary fac-
tor promoting eusociality in naked mole-rats and other
social animals. The spatial distribution of burrows is inti-
mately linked to soil geomorphology (Moss 1940; Svendsen
1976) and land use (Henderson and Gilbert 1978). Thus in-
dividuals may be clustered in areas that contain or at least
provide suitable soil architectures for the development of
burrow systems.
Several lines of empirical evidence suggest that burrow
systems may ultimately influence the expression of social-
ity in ground-dwelling squirrels. Holmes (1984a) invoked
the wide dispersion of suitable hibernacula as a factor pro-
moting monogamy in hoary marmots (Marmota caligata).
Armitage (1988) turned to burrow systems to explain vari-
ation in the expression of yellow-bellied marmot sociality.348 Chapter Twenty-Nine