version of the breeding-group model revealed that the com-
bination of philopatric females and male polygyny resulted
in coteries of related adult females, and consequently signifi-
cant genetic differentiation occurred among coterie breed-
ing groups (table 14.1). Surprisingly, Sugg et al. (1996) also
found that effective population size was slightly greater
than the size of the adult census population (figure 14.2).
Dobson et al. (1997) compared estimates from the breed-
ing-group model, allozyme data, and pedigrees in a 10-year
study of black-tailed prairie dogs. All three methods of
estimating F-statistics yielded very similar estimates, thus
confirming results of the model. Effective population size
was also estimated from more detailed data (Dobson et al.
2004), using all three methods, and again estimates of ef-
fective population size were slightly greater than the census
population size (fig. 14.2).
Further work with black-tailed prairie dogs revealed dif-
ferences that might be expected from predispersal and
postdispersal estimates of F-statistics (table 14.1; Dobson
and Chesser et al. 1998). FLSwas higher in predispersal
(19%) than postdispersal (8%) estimates for both pedi-
grees, and for the allozyme data set (25% and 16%, re-
spectively). Greater genetic differentiation is to be expected
from predispersal estimates, because they are for the breed-
ing-group young, most of whom are full or half siblings.
Genetic differentiation among breeding groups was also
significant using separate estimates of FLSfor adult females
(producing post-dispersal estimates), using pedigree data
(12%). Adult males exhibited a much lower FLS(3%). The
difference in values for adult males and females reflects the
ubiquitous natal dispersal of males and the close kinship of
females. As might be expected, FILestimates were closer to
zero in postdispersal estimates, though they were still sig-
nificantly negative, and all FISvalues were close to zero.
These three F-statistics are linked to each other by the for-
mula: (1 – FIS) (1 – FLS) (1 – FIL). Thus, F-statistics tend
to change together. In spite of differences in FLSbetween
pre- and postdispersal estimates, it is clear that the social
structure of breeding groups creates genetic structure that
persists over generations.
Deviations of the inbreeding coefficient (F) from the ex-
pectation of random mating must be evaluated with cau-
tion. Patton and Feder (1981) estimated FILat a substantial
positive value at the level local groups of pocket gophers
(table 14.1), but felt that their estimate was biased by a sub-
stantial Wahlund effect (caused by pooling social breeding
groups; see also Sugg et al. 1996). Nonetheless, an expected
increase in the value of FIXat increasing spatial levels seems
reasonable (Dobson et al. 1997). Within breeding groups,
dispersal from the natal area may often result in negative FIL170 Chapter Fourteen
Figure 14.2 Estimates of effective population sizes and census population sizes (numbers of potentially breeding adults).
(A) From Sugg et al.’s (1996) estimates from results presented in Hoogland’s (1995) book on black-tailed prairie dogs.
(B) From Dobson et al.’s (2004) empirical data on black-tailed prairie dogs. (C) From Dobson et al.’s (2000) study of Tibetan
plateau pikas under two assumptions: complete single paternity (left hand estimate) and complete multiple paternity (esti-
mate to the right) within pika family breeding-groups.