1985]). However, since the soil was easily dug and dens were
widespread in the study area, it seems unlikely that a limita-
tion on den sites forced maras into communal denning.
Some maras den solitarily rather than communally; it
seems likely that the interaction of resource dispersion, pro-
tection from predators, and the advantages of sociality, in-
cluding vigilance, communal suckling, and thermoregula-
tion, determine denning behavior. The relative importance
of each variable could be investigated by comparing mara
behavior in different parts of their range. Such comparisons
may shed light on the two possible evolutionary routes to
the mara’s social system: (1) from a monogamous ances-
tral system, modified to accommodate the advantages of a
creche, or (2) from a communal polygynous ancestor on
which monogamy has been imposed by, for example, habi-
tat becoming more patchy, sparse, and unpredictable. The
fact that coloniality and polygyny are widespread in the
mara’s caviomorph relatives (Kleiman 1977; Lacher 1981;
Ebensperger and Cofré 2001) suggests that the latter route
is the more probable.
Group Size and the Resource Dispersion Hypothesis
Arguments about the determinants of social group size of-
ten involve the richness of resource patches; for example,
the Resource Dispersion Hypothesis (see Macdonald and
Carr 1989; Johnson et al. 2002). In Herrera and Macdon-
ald’s (1987, 1989) capybara study area, each home range
included patches of each major habitat type, including a
body of water. However, in the low Llanos of Venezuela
there are vast expanses of bajio without permanent surface
water, and these are devoid of capybaras. Around water-
holes big enough to retain water well into the dry season,
capybara territories contiguously cover the land available.
As all other habitat types can exist in the absence of water-
holes, it appears that the waterhole is the key resource that
makes a territory viable and defendable for year-round sur-
vival. Components of a territory necessary to sustain cap-
ybaras throughout the changing seasons were (1) bushy
banco (higher-lying land) to provide food and harborage
during the height of the wet season, (2) ponds for drinking
water, thermoregulation, mating, and sanctuary from pred-
ators in the dry season, and (3) bajio for grazing during
the peak of the dry season. The size and shape of an eco-
nomically defendable territory with these minimal require-
ments will be determined by the dispersion of these three
resources. In such a minimum territory (sensu Carr and
Macdonald 1986), that is, one that supports the require-
ments of a minimum social unit, the resource patches may
be sufficiently rich to support a larger group. Insofar as
group size is determined by resource availability (as opposed
to, for example, predation), it will be limited by whichever
was the most limited of the three critical resources: banco,
bajio, or pond. In the groups, studied over 3 years, the close
correlation between adult group size and area of banco in-
dicates that this was the limiting resource.
In these circumstances, territory and group size might be
determined independently and respectively by the disper-
sion and abundance (richness) of available resources, a pos-
sibility described by the Resource Dispersion Hypothesis.
In minimum territories, group and territory size could be
explained solely by a correlation between patch dispersion
and richness; for example, if territories sufficiently large to
encompass widely spaced ponds and /or bajios in the dry
season automatically encompassed additional banco for the
wet season. In this case, the dispersion of one or two critical
resources in one season would be correlated with the rich-
ness of the third critical resource in the other season. An ob-
servation by Herrera and Macdonald (1989) that larger ter-
ritories were apparently configured to embrace two ponds,
and thereby acquire additional intervening banco (and ba-
jio), is compatible with this suggestion.
In the case of maras, the Resource Dispersion Hypothe-
sis may help explain why settlements develop in certain
locations. The vicinity of the lagoon and outstation repre-
sents a single rich patch: a carpet that can be utilised simul-
taneously by all the maras in a settlement during a bottle-
neck in food availability in the adjoining scrub, and possibly
a safer place for the pups. This rich patch is effectively in-
defensible from the other maras or sheep which flock to it.
The result is that year-round local carrying capacity for ma-
ras is elevated, by either or both of water or sheep dung, to
a level where one or more pairs can be supported within
daily travelling distance of a warren.
We have already seen that group size in capybaras is de-
termined by the size of particular habitat patches. According
to the Resource Dispersion Hypothesis, the development
of larger social groups is more likely where resources are
spatiotemporally heterogeneous in availability and patch
richness is high. However, the diverse habitats occupied
by capybaras make the species a good model for exploring
intraspecific variation in the consequences of resource dis-
persion. For example, Salas (1999) found larger groups in
a location with both greater habitat homogeneity and more
abundant (and less seasonal) resources, whereas rather small
groups are more likely where resources are homogeneously
dispersed and /or patch richness is low. Maras, in contrast,
appear to be confronted with conditions that lie at either
extreme of the resource dispersion continuum. In the dry
season maras depend on a single, rich patch, which can be
viewed as either shareable or undefendable but which isSocial Organization and Resource Use in Capybaras and Maras 401