creased density decreased the probability of successful com-
petition by young females for territories during the next
breeding season, prompting young females to emigrate and
settle in areas with lower density. The findings of Dobson
(1979) and Nunes et al. (1997) support the idea that com-
petition for environmental resources was an important fac-
tor in the evolution of female but not male dispersal in ro-
dents, and mammals in general.
Proximate causes of natal dispersal
A wide variety of factors have been proposed as triggers of
dispersal in rodents. Changes in the habitat, demography,
or population density of the home area or a nearby area may
influence the future prospects of animals in either of these
areas, and make it advantageous for them to transfer resi-
dence to a new area. Some of the potential triggers of dis-
persal are reviewed in the following.
Early developmental influences on dispersal behavior
For most rodent species, dispersal behavior is malleable and
responsive to environmental variables (e.g., Lambin 1994a).
However, some species exhibit rigid natal dispersal behav-
ior, in which the probability of dispersal is not influenced
by environmental conditions. These inflexible predisposi-
tions to disperse may be brought about by the action of hor-
mones during early development (Holekamp et al. 1984;
Ims 1989, 1990).
Androgens can have long-lasting influences on the brain
during early development, which in turn can influence be-
havior later in life. Holekamp et al. (1984) and Nunes et al.
(1999) suggested that androgens promote development of
a predisposition to leave the natal area in young male Bel-
ding’s ground squirrels. All male squirrels eventually leave
the natal area, regardless of ecological and social conditions,
whereas most females remain in their natal areas through-
out their lives. Holekamp et al. (1984) and Nunes et al.
(1999) observed that treatment of young female squirrels
with androgens at the time of birth caused them to disperse
at higher rates than control females, and suggested that an-
drogens promote development of an innate drive to leave
the natal area. This influence of androgens results in fairly
inflexible behavior among young male squirrels, with males
emigrating regardless of social conditions and resource
availability in the natal area. Environmental conditions in
the habitat of Belding’s ground squirrels fluctuate greatly
over the course of a year, but these seasonal changes are
highly predictable, and this predictability might obviate the
need for flexibility in the dispersal behavior of young males
(Smale et al. 1997).
Androgens have also been implicated in the development
of dispersal behavior in female grey-sided voles (Clethrion-
omys rufocanus;Ims 1989, 1990). In particular, female
voles from male-biased litters tend to leave their natal areas
at higher rates than do females from unbiased or female-
biased litters, and their likelihood of dispersal is not influ-
enced by social or ecological conditions. Ims (1989, 1990)
proposed that female voles in male-biased litters have an
increased probability of developing next to male fetuses
in utero, and thus may be exposed to androgens produced
by neighboring males (see vom Saal and Bronson 1978,
1980). Androgens might help organize dispersal tendencies
in these females. Lambin (1994a) did not observe an effect
of sex ratio on the dispersal behavior of Townsend’s voles
(Microtus townsendii), suggesting that the effects observed
by Ims (1989, 1990) may not be universal among vole spe-
cies. The adaptive benefit of androgenic effects on dispersal
for female grey-sided voles is not fully understood, but they
may help to maintain variability in the tendency to disperse
(Ims 1989, 1990; vom Saal and Bronson 1978, 1980).
Bondrup-Nielsen (1993) proposed that nutrition during
infancy influences dispersal in female meadow voles (Mi-
crotus pennsylvanicus). Female voles malnourished during
early development displayed greater behavioral tendencies
associated with dispersal as adults than did control females,
despite being well nourished at the time they expressed these
tendencies. Bondrup-Nielsen (1993) suggested that poor
nutrition during early development might be an indicator of
an unstable or unpredictable habitat, and might adaptively
promote development of behaviors causing female voles to
seek environments that are more stable.Availability of environmental resources
In some cases, dispersal appears to be dependent on avail-
ability of vacant habitat in which to settle, and dispersal
may be delayed or inhibited in the absence of obtainable,
high-quality habitat with sufficient resources to support
survival and reproduction (Solomon 2003). Rates of dis-
persal tend to decline when habitat is saturated, and com-
petition for available space, nesting areas, or food is high
(Wolff 1994a; Kokko and Lundberg 2001; Solomon 2003).
For example, inhibited dispersal under conditions of high
population density has been observed in field voles (Micro-
tus agrestis;Sandell et al. 1991) and prairie voles (M. och-
rogaster;McGuire et al. 1993). By contrast, dispersal rates
tend to be high when nearby habitat has nest sites available
for settlement (Wolff 1994a). For example, in removal stud-
ies of pocket gophers (Thomomys talpoides), areas from
which gophers were eradicated were rapidly recolonized
(Engeman and Campbell 1999). Moreover, continuous re-
moval of gophers did not result in long-term reduction of
population size (Sullivan et al. 2001). Similarly, removal
of yellow-bellied marmots (Marmota flaviventris) from a
colony resulted in increased recruitment of marmots intoDispersal and Philopatry 153