burrow was. This knowledge was also needed earlier, dur-
ing breeding, because estrous female California ground
squirrels travel to the burrows of males to mate, running a
gauntlet of other males eager to mate with them (unpub-
lished data, as well as Boellstorff et al. 1994).
Knowledge of the spatial layout and nature of the re-
sources in one’s surroundings has utility in many domains,
including both social and antipredator contexts. The bene-
fits of such spatial ability have been a source of natural se-
lection in some rodent species, shaping processes for learn-
ing, remembering, and using the home range (Gaulin and
Fitzgerald 1989). Nevertheless, different contexts, such as
predation, social interaction, and foraging may well differ in
the demands they place on spatial knowledge. For example,
the typical contexts for the study of spatial knowledge are
low-urgency ones, such as finding, storing, and retrieving
food. In these situations, some ground-dwelling sciurids
have exhibited remarkably precise cognitive-mapping abil-
ities (Devenport and Devenport 1994; Devenport et al.
2000). When a food source was moved with its conspicu-
ous visual landmark a mere 1.5 meters from the learned
location, animals initially ignored the nearby moved land-
mark and searched for the food in the previous location.
Similar spatial abilities may be used in low-urgency social
situations, such as keeping track of the home burrows of
other members of a kin cluster (Boellstorff and Owings
1995). Animals also exhibit evidence of cognitive mapping
under high urgency; alarmed sciurids use cues about their
distance and direction from refugia, often turning up to
180 degrees before initiating a run to a refuge after an anti-
predator call (Leger et al. 1979), or varying their decision
to run or call with their distance from a burrow (Noyes
and Holmes 1979; Sherman 1985; Bonenfant and Kramer
1996). However, when social contexts are as urgent as
most predatory ones — for example, during agonistic epi-
sodes — such high social and predatory urgency may not
provide the time or attention needed for detailed use of cog-
nitive mapping, and so may force the application of addi-
tional methods. These may include relying more heavily
on learned landmarks close to the destination (“beacons”;
Collett et al. 1986; Dinero, personal communication) or on
well-practiced movements along often-used routes (Stamps
1995).
Spatial learning can be useful in ways that go beyond
cognitive representations of the layout of resources. For ex-
ample, laboratory rats learn not only the spatial location of
food rewards, but also the appropriate motor activity, a di-
rection of turning that will lead them to food (Restle 1957).
Stamps (1995) has argued that many species may depend
heavily on such motor learning aspects of spatial learn-
ing, practicing movements over regularly used trails much
as humans refine their foot-racing performance for hurdles
events by running the track repeatedly. This learning may
prove especially useful when rapid locomotion through
cluttered surroundings is needed to deal with a conspecific
or predatory adversary that has not had the same route-
specific practice. Repeated use of the same route can also
increase an individual’s sensitivity to change along a trail,
thereby enhancing the chances of detecting a conspecific or
heterospecific behind a bush, for example, that wasn’t there
before.
Remembered cognitive, perceptual and motor knowl-
edge of surroundings are not the only sources of effective
spatial patterning of behavior. Animals may also exploit
spatially varying environmental features on the basis of
immediately recognized affordances rather than familiar-
ity. Unobscured vision, for example, indicates openness and
therefore affords vulnerability, motivating rats to make con-
tact with visibility-obscuring vertical surfaces (positive thig-
motaxis, e.g., Martínez et al. 2002), a response that is
intensified under fear motivation (Kelley 1985). Similarly,
burrowing rodents are strongly disposed to seek refuge in
dark burrow-like places, a disposition that we regularly ex-
ploit in laboratory experiments by making a specific place
dark where we want our squirrels to go (Rowe and Owings
1978; Rowe et al. 1986). California ground squirrels are
cognizant of nearby burrows and vertical surfaces such as
rocks, and vary their levels of foraging and vigilance with
the proximity of these environmental features (Leger et al.
1983). Similarly, dispersing yearling Columbian ground
squirrels (Spermophilus columbianus) follow cutbanks,
roads, and game trails into new terrain, a choice that ap-
pears to guide them to new ground squirrel habitat at
higher-than-chance rates (Wiggett and Boag 1989; Wiggett
et al. 1989).Means-end links: The social domain provides
means for dealing with predation
Perhaps the most familiar link between social and anti-
predator behavior comes from the fact that rodent social
systems provide a means for dealing with the threat of pre-
dation (Waterman and Fenton 2000). Our discussion of
antipredator calls and exploitation of social affordances has
already touched on the social domain as a means for deal-
ing with predators. The availability of such social affor-
dances has generated selection for increased sociality. For
example, black-tailed and white-tailed prairie dogs (Cyno-
mys ludovicianusand C. leucurus,respectively) are both
social species, but black-tails live in larger, denser groups
(Hoogland 1981; 1995). Hoogland has used variation in
group size and density within and between these two spe-
cies to explore the adaptive significance of such group liv-
ing. Three lines of evidence indicate that social living con-310 Chapter Twenty-Six