values. FISshould approach zero as the area sampled is ex-
panded, resulting in a balance between natal dispersal from
breeding groups and a spatial scale at which the genetic
influence of dispersal distances becomes limited geographi-
cally. For example, dispersal movements were much greater
within the colony of black-tailed prairie dogs that Hoog-
land (1995) studied than the low rate of immigration to
the colony from other colonies (Garrett and Franklin 1988;
Dobson et al. 1997). At larger spatial scales,FITmay become
strongly positive, as geographic isolation results in genetic
differentiation of regional subpopulations, as Wright (1969)
envisioned. Relationships between inbreeding, heterozygos-
ity, and spatial scale raise interesting questions that deserve
further research.
Estimates of effective population size for local popula-
tions made up of breeding groups are higher than estimates
from traditional models, which suggest effective sizes at
about half the census size or less (Nunney 1993; Nunney
and Elam 1994; Frankham 1996). Nunney (1999) criticized
Sugg et al.’s (1996) estimate for black-tailed prairie dogs
because the breeding-group model assumes nonoverlapping
generations, and generations overlap in prairie dogs. Thus,
he argued that the estimate may have been biased upward.
However, in a study of Tibetan plateau pikas (Ochotona
curzoniae), Dobson and Smith et al. (2000) applied the
breeding-group model to behavioral and demographic re-
sults for a species that has extremely limited overlap of
generations. Families of these highly social pikas exhibit
a variety of mating systems, but on average are slightly po-
lygynous (Smith and Wang 1991). They also exhibit male-
biased dispersal, though either sex may be philopatric (Dob-
son and Smith et al. 1998). In spite of the potential for high
rates of inbreeding, FILwas strongly negative and FLSindi-
cated strong genetic differentiation among family groups
(table 14.1). Effective population size was slightly differ-
ent depending on whether multiple paternity occurred (un-
known, though multiple copulations occurred; F. S. Dob-
son and A. T. Smith, personal observations). Estimates of
effective size were slightly less than the size of the census
population (fig. 14.2), and appeared much higher than ex-
pected from classical models of gene dynamics that do not
take social structure into account (Dobson and Smith et al.
2000).
Gene dynamics of nonrodent species with social breed-
ing have been reviewed by Storz (1999). The purpose of
his review was to evaluate whether social breeding groups
could lead to speciation, which seemed unlikely for most
mammals. However, significant genetic differences among
social breeding groups were found in several species. In par-
ticular, red howling monkeys (Alouatta seniculus) exhibited
significant breeding-group structure, based on F-statistics
estimated from allozyme data from a mixture of predis-
persal and postdispersal individuals of both sexes and sev-
eral ages. FLSaveraged about 0.14, FILabout 0.22, and
FISabout0.04 (Pope 1992). Red howlers exhibit polygyny
and predominant male dispersal, often with related males
dispersing together (reviewed by Pope 1992, 1998). Interest-
ingly, troops of six other primate species, such as macaques
(genusMacaca), exhibited generally low but substantial ge-
netic differences among troops (FLS0.04 0.08; Storz
1999). Two other primate species exhibited insignificant
genetic differentiation among troops. Finally, spatial groups
of two species of artiodactyls exhibited significant genetic
differentiation, perhaps due to social breeding groups. With
few exceptions and for estimates based on postdispersal es-
timates of F-statistics, Storz (1999) found substantial influ-
ences of social breeding groups on gene dynamics in species
exhibiting polygynous or promiscuous mating systems and
predominant male dispersal patterns.SummaryIn this review, I have examined whether social groups com-
posed of philopatric female kin create genetic structure
within local populations. An understanding of the hierar-
chical nature of gene dynamics and correct interpretations
of F-statistics are necessary to address this question. Thus
I’ve discussed the hierarchical nature of gene dynamics, and
of the F-statistics that Wright (1965, 1969, 1978) used to
describe them. These F-statistics, particularly FLS, can be
used to indicate whether social breeding groups create a
level of genetic structure within local populations. For a few
species of rodents and other mammals, there does appear to
be significant genetic structure at the level of social breed-
ing groups. It could hardly be otherwise, since philopatry
among females will create clusters of female kin, and kin
exhibit genetic correlations that are spatially concentrated.
As Chesser (1991a, 1991b) pointed out, these genetic cor-
relations are a form of genetic structure that can be mea-
sured with F-statistics and linked to higher levels of spatial
genetic structure. Further studies of mammals will reveal the
generality of sociogenetic structure in populations of highly
social mammals. Studies of less social species, such as the
groundbreaking work of vanStaaden et al. (1994) on Rich-
ardson’s ground squirrels, are needed to discern whether
sociogenetic substructuring extends beyond highly social
species, like black-tailed prairie dogs. In addition, this re-
view provides a working hypothesis for future studies of
relationships between the inbreeding coefficient, F,and its
expectation for randomly mating populations at different
spatial scales, as measured by hierarchical F-statistics.Gene Dynamics and Social Behavior 171