Ecology, Behavior, and Conservation 17habitat and social structure is a strong but not a universal
pattern.
Social ground squirrels and prairie dogs provide a com-
pletely diff erent perspective regarding food type and social
behavior, but one that is still consistent with the view above.
These species generally feed on nondefensible food items
(other than secondary metabolites) that do not store well.
Instead, their strategy involves fat storage. However, some
species (e.g., Tamia s) use a combination of both food storage
and the accumulation of body fat to overcome periods of
food shortages and harsh environmental conditions.
Seed Dispersal
In the process of scatter-hoarding, squirrels—primarily Sci-
urus, but also members of several subtropical and tropical
genera (e.g., Sciurotamias)—will store hundreds if not thou-
sands of nuts or fruits per individual squirrel. Early observa-
tions led to the long-held notion that squirrels store more
than they need and then forget where many seeds and nuts
are cached, allowing some of them to germinate and be-
come established. More recent research, however, suggests
a far more complicated picture: one in which the squirrels
are considerably more effi cient at nut recovery, and one in
which germination and establishment occur under far more
limited circumstances than previously thought. This is not
to say that scatter-hoarding squirrels are not critical for seed
dispersal—they are. In fact, in many deciduous forests sys-
tems, as well as some subtropical and tropical forests, they
may serve as keystone agents of seed dispersal. Experimen-
tal evidence strongly suggests that individual squirrels re-
member precise locations of their stored nuts, most likely
based on spatial information, as is also shown for their
avian counterparts, the scatter-hoarding corvids (rooks,
jays, and crows). This is not to say that they don’t steal from
one another; pilfering is common and sometimes even ex-
treme. Yet these scatter-hoarders may have domain over
their own scatter-hoards in many situations.
The process by which scatter-hoarding squirrels dis-
perse seeds, nuts, and fruits is rather involved. It is perhaps
best understood in the eastern gray squirrel, which resides
predominantly in oak forests throughout much of the cen-
tral and eastern USA. Numerous experimental studies now
demonstrate that this species is highly selective with re-
spect to the nuts that are eaten and those that are stored.
Acorns of red oak species, for example, are signifi cantly pre-
ferred for scatter-hoarding over those of white oak. Typi-
cally, the acorns of red oak species are higher in fat content
and tannin levels (which reduce the palatability and digest-
ibility of plant foods), but they also exhibit dormancy prior
to germination. White oak acorns, in contrast, have lower
tannin and fat levels and germinate rapidly in the autumn,
sometimes while still attached to the tree. Behavioral ex-
periments by M. Steele, P. Smallwood, and others show that
in the autumn, gray squirrels selectively consume white oak
acorns, but at the same time they also selectively cache red
oak acorns, primarily because of their delayed germination
schedules and reduced perishability in the cache. Tannins
and fats secondarily infl uence the squirrels’ eating prefer-
ences, but overall the primary determinant of their fall
caching decisions is the germination pattern.
Indeed, early germination in white oak acorns appears
to have exerted a strong selective pressure on eastern gray
squirrels. Several decades ago J. Fox demonstrated that when
faced with heavy crops of white oak acorns, gray squirrels
cache these more perishable food stores, but, when doing
so, they carefully excise the small embryo at the apical end
of the acorn with a few quick scrapes of their incisors. In-
terestingly, this embryo-excision of white oak acorns has
been demonstrated recently in at least two other species
of Sciurus (fox squirrels [S. niger] and Mexican gray squirrels
[S. aureogaster]) from North America and at least one species
of Sciurotamias (S. davidianus [Père David’s rock squirrel])
from China. Moreover, in experiments with captive-raised
eastern gray squirrels having no previous experience with
acorns, they also attempt embryo-excision, but often per-
form it incorrectly or on the wrong part of an acorn. This
suggests that the behavior may be largely innate. Although
numerous other rodents—such as mice (Peromyscus), chip-
munks (Tamias striatus), and southern fl ying squirrels (Glau-
comys volans)—selectively cache red oak acorns, these spe-
cies do not appear to perform embryo removal. Hence this
excision behavior may be unique to one or a few lineages of
squirrels that regularly scatter-hoard acorns.
Numerous other nut characteristics (e.g., nut size, nut
mass, insect infestation, and food abundance) also infl uence
the food-hoarding decisions of scatter-hoarding squirrels in
many predictable ways. Ultimately, many of these behav-
ioral decisions in turn infl uence where nuts are cached and
the likelihood that the nuts will germinate and establish
if they are not recovered. For example, we currently know
that the sites frequently selected by eastern gray squirrels
for scatter-hoarding are often coincidentally optimal for
germination, because these are also the sites where acorns
store well. We also know that the probability of seedling
establishment increases during bumper-crop (or mast) years,
when animals store large quantities of seeds and nuts and
some are not recovered. What we don’t yet understand is
how other processes, such as interactions with competitors
(potential cache pilferers) and predators, indirectly infl uence
cache recovery and seedling establishment. Despite this
fascinating interrelationship between squirrels and trees, we
still have a great deal to learn before we fully understand