directly with body size, seasonality, or reproductive effort,
but was highly correlated with age of first reproduction,
such that delayed reproduction could be equated with en-
hanced sociality. While body size emerged as a highly sig-
nificant predictor of reproductive effort, the lack of any cor-
relation between sociality and reproductive effort or body
size suggested to Armitage that reproductive effort was a
product of size-dependent energetics, while sociality was a
consequence of other factors. Among those factors, Armi-
tage (1981), like Hoogland (1981), suggested antipredator
benefits of group living as the necessary precursor to soci-
ality, and allowed that kin selection may have played a role
in the evolution of ground squirrel sociality, given evidence
of kin-biased patterns of alarm calling (Dunford 1977a;
Sherman 1977). In more direct accord with his own analy-
sis, however, Armitage (1981) indicated that a nutritionally
poor environment (as would be the case with short grow-
ing seasons) could result in delayed maturity, which in turn
would delay dispersal and promote sociality via interac-
tions among individuals of different sex /age cohorts. To Ar-
mitage, then, sociality is the product of life-history trade-
offs, particularly in the timing of an animal’s development
toward reproductive independence, which is shaped ulti-
mately by extrinsic environmental factors.
Michener (1983a) also recognized five grades of social-
ity, and presented a similar classification scheme to that of
Armitage (1981), but focused more directly on the role of
kinship in defining sociality. Based upon the spatial over-
lap of sex /age cohorts, and the relatedness of females within
groups, Michener delineated grades as: 1 asocial, 2
single-family female kin cluster, 3 female kin cluster with
territoriality, 4 female kin cluster with male dominance,
and 5 egalitarian, polygynous harems. In comparing spe-
cies, Michener (1983a) noted that while more social species
tend to possess large body size, the active seasons of those
species were not necessarily short. Delays in dispersal and
reproductive maturity were characteristic of the more social
species, though the primary correlate of advancing social
grade in Michener’s analysis (accounting for some 84% of
the variance) was the temporal overlap between adults and
subadults. Michener recognized that in addition to delayed
dispersal, factors postponing the fall immergence of adults,
or even year-round activity (as is apparent in black-tailed
prairie dogs), could enhance overlap and thereby promote
sociality. The driving force, though, was kin selection, as
Michener noted that the clusters of females forming the
core of social groups were related females comprising ma-
trilines. Thus from her comparative assessment Michener
recognized the following emergent trends characterizing the
evolution of ground squirrels from a presumptive ancestral
asocial state to advanced societies composed of multiple
polygynous harems: (1) the retention of daughters in the
female’s home range, (2) the relaxation of distinctions be-
tween litters of adjacent females, and (3) male territorial-
ity extending beyond the breeding season and ultimately
expanding in scope to incorporate the ranges of multiple
females.
Rayor and Armitage (1991) contrasted patterns of spa-
tial overlap and social behavior among highly social Gunni-
son’s prairie dogs (Cynomys gunnisoni), less social Colum-
bian ground squirrels, and relatively asocial thirteen-lined
ground squirrels. They found that amicable social inter-
actions between mothers and their offspring and among
littermates were more prevalent with complex sociality.
Further, spatial overlap was greater and was maintained for
a longer period of time in more social species. Such trends
were not restricted to littermates: increased spatial overlap
and cohesive behavioral interactions with young from
other litters were also apparent with advances in sociality
(Rayor and Armitage 1991). Thus while sociality in ground-
dwelling squirrels is associated with the retention of off-
spring within the mother’s range, it is the extension of co-
operative behavior among individuals that is the earmark of
increasing social complexity.
The terms socialityand social complexityare often used
in the literature without explicit quantification, though
comparative analyses by Blumstein and Armitage (1997b,
1998) present a quantitative and continuous index of social
complexity. The index itself is based on information theory
(Shannon and Weaver 1949), whereby a higher index ex-
presses the increased amount of information necessary to
describe a more complex social construct. Their overall
index is, in effect, based on a grand sum of the number of
bits of information required to describe the variation in the
frequencies with which certain demographic (age /sex) co-
horts appear within groups of individuals of each species
of interest, multiplied by a factor that increases with delays
in dispersal. Complexity, then, is equated with variability in
the probability of interaction of individuals in various sex /
age cohorts, indirectly capturing factors that have been used
historically to describe sociality (e.g., group size, Eisenberg
1981, spatiotemporal overlap of sex /age cohorts, Michener
1983a). The multiplicative factor that scales this informa-
tion index for each species reflects the period over which
young are retained in the natal area, and hence amplifies the
overall social complexity score where kin have the poten-
tial to overlap in space and time (Blumstein and Armitage
1997b).
While recognizing that their index of social complexity
does not reflect the actual nature of behavioral interactions,
Blumstein and Armitage used the indices calculated for
multiple Cynomys, Marmota,and Spermophilusspecies to
examine correlations between social complexity and com-
municative complexity in alarm vocalizations (twenty-twoEcology, Kinship, and Ground Squirrel Sociality: Insights from Comparative Analyses 347