et al. 1988), and the canyon mouse has a relatively long
weaning time and a very high relative birth mass compared
to other species of deer mice (Peromyscus;Modi 1984).
There is also some evidence to suggest that the extreme en-
vironments inhabited by some rock-specialist rodents influ-
ences their survival and tendency to remain philopatric. For
example, bushy-tailed woodrat yearling females experience
increased over-winter survivorship when they spatially ag-
gregate near their mothers on an outcrop (Moses and Mil-
lar 1992; Moses and Millar 1994). Similarly, although it is
thought that male gundis normally disperse out of their na-
tal group, many remained philopatric during the course of
a drought (Gouat 1988a; Nutt 2003; Nutt 2005). These ob-
servations suggest that the degree of sociality observed in at
least some rock-dwelling rodents may be influenced by ex-
treme environmental conditions and the need for extended
parental care.
High predation levels
Group living is thought to have evolved in some mammals
to alleviate high predation pressure on individuals (Jarman
1974; Waterman 1997; Hill and Lee 1998). It is possible
that rocky habitat tends to have a high abundance of pred-
ators, and that this is why some rock-dwelling mammals
are social. I have stated previously that topographically
complex rocky habitat is beneficial to prey because it helps
them to hide from predators. Such complex habitat may
also work against prey by concealing predators until it is
too late. It is perhaps because of high predation pressures
that so many rock-dwelling mammals have alarm calls and
utilize lookout posts (Mares and Lacher 1987). The cteno-
dactylid rodents are notable in this respect, as all five spe-
cies have characteristic alarm calls (George 1981).
The degree of predation pressure on each rock-dwelling
rodent will vary considerably depending on the predators in
the environment and the particular habitat. Not only is it
difficult to discern the exact level of predation pressure ex-
perienced by each rock-dwelling mammal, predators are
also likely to affect the sociality of each species in differ-
ent ways. For example, the American pika dens solitarily,
yet exhibits high levels of philopatry (Peacock 1997). By liv-
ing in close proximity to relatives, American pikas can in-
crease the nepotistic effect of alarm calling without having
to form large social groups (Sherman 1977; Smith 2001).
The cliff chipmunk also seems to den solitarily, yet this spe-
cies presumably minimizes predation pressure through the
use of communal foraging (Hart 1992). Finally, rock and
bush hyrax form heterospecific aggregations while basking,
apparently to increase the survivorship of young offspring
(Barry and Mundy 2002). Clearly, predation pressure has a
large influence on the type of aggregate behavior observed
in at least some rock-dwelling mammals.
The clumped distribution of a limiting resource
Some evidence suggests that many mammals live in groups
because of the clumped distribution of a limiting resource
such as food or shelter (Emlen and Oring 1977). For rock-
dwelling species, the distribution and abundance of rocks,
crevices, and outcrops is likely to be one of the most im-
portant resources influencing social behavior in this regard.
The size and distribution of rocky substrate is likely to af-
fect the sociobiology of rock-dwelling mammals in three dis-
tinct ways: (1) by influencing whether social groups form,
(2) by influencing the size of social groups and density of
animals that live on an outcrop, and (3) by influencing the
larger geographic spatial distribution of animals.How distribution of rocky substrate may influence social
group formation
Living in crevices may lead to high levels of sociality and
group living because crevices are a nonconsumable resource
that may easily be shared. Waser (1988) asserted that spe-
cies requiring nonconsumable resources, such as burrows
or elaborate refuges, tend to be philopatric. Philopatry is
expected to evolve in such a situation because the cost to
parents of sharing a refuge will be relatively low, whereas
the cost to offspring of dispersing and finding a new refuge
that is limited in number will be fairly high (Waser 1988).
Once natal philopatry occurs, sociality may evolve.
There is much evidence within Rodentia to support the
hypothesis that the use of elaborate refuges may lead to
group living. In New World hystricognath rodents, living in
groups appears to be more associated with the need to re-
duce burrowing costs than the need to reduce predation
pressure or the need for extended parental care (Ebensper-
ger and Cofré 2001). Similarly, group living in the bathy-
ergid mole-rats is thought to have evolved in environments
where burrowing costs are extremely high (Faulkes et al.
1997; Faulkes and Bennett, chap. 36 this volume). These
examples illustrate that sociality tends to evolve when bur-
rowing is costly. If burrowing is not an option at all (as
is the case for petrophilic rodents), then individuals will
be completely reliant on the availability of suitable crevices
and will be forced to clump together when these resources
are limited.How distribution of rocky substrate may influence population
density and group size
The distribution and abundance of crevices within an out-
crop is likely to influence the number of animals that can
live in a region and the size of social groups that may form.
The population densities of several rock-specialist rodents,
including the mountain viscacha, the cutch rock rat (Crem-
nomys cutchicus), the namtap, the bushy-tailed woodrat,
and the dassie rat, are all thought to be limited by theSocioecology of Rock-Dwelling Rodents 425