In accord with the hypothesis that these nocturnal prey
are communicating directly with the predator, larger (and
presumably less vulnerable) kangaroo rat species engage
in more frequent behaviors of foot drumming, sand kick-
ing, and approach to predators such as snakes than do
smaller kangaroo rat species (Randall et al. 1995; Ran-
dall and King 2001; Randall, chap. 31 this volume). Al-
though diurnal and living a solitary life, the eastern chip-
munk exhibits highly audible alarm calls that may serve
to alert kin (Burke da Silva et al. 2002) or to alert oth-
ers that, if they escape, reduces the patch quality to the
predator.
A series of trained-owl experiments at a site in the Negev
Desert, Israel, revealed the tendency to seek safety in num-
bers in a nonsocial gerbil species (Allenby’s gerbil) and also
revealed that the negative effect of a dominant competitor
(Egyptian sand gerbil) becomes nullified in the face of pre-
dation risk (overflights by barn owls). Whereas ordinarily
the gerbils would distribute their activity evenly between
two halves of an enclosure (in accord with an ideal free dis-
tribution), Allenby’s gerbil would actually bunch up into
one half or the other in response to increased numbers of
owl flights even when the owl flights were evenly distributed
between halves of the enclosure (Rosenzweig et al. 1997).
The larger Egyptian sand gerbil, by virtue of interference
competition — and disproportionate to its own numbers —
will drive the smaller Allenby’s gerbil into the half of the
enclosure free from the larger species. However, in the pres-
ence of increased predation risk from owl flights, the inter-
ference effect of G. pyramidumon G. a. allenbyievaporates
as responding to fear seems to trump interspecific competi-
tion (Abramsky et al. 1998).
Predation Risk and Seasonality
A striking difference between boreal and arid environments
emerges from the stronger temperature seasonality of bo-
real habitats and the greater precipitation seasonality of
arid habitats. For boreal and subarctic voles the winter pro-
vides long-lasting permanent snow cover that alters pre-
dation pressures and risk. Overwinter survival of the bank
vole suggests a late autumn period of high mortality before
the onset of permanent snow, and another spike in mor-
tality associated with the spring thaw (e.g., Ylönen and Vi-
itala 1985). This pattern does not necessarily demonstrate
changing predation rates and may simply represent the
mortality consequences of harsh environmental conditions.
However, a persistent blanket of snow provides rodents
with subnivean spaces that protect them from severe tem-
peratures and from the attacks of most predators. Large
owls, like the great grey owl (Strix uralensis), can attack
vole prey through thick snow cover, and least weasels and
possibly small stoats (Mustela erminea) are able to enter the
subnivean cavities of voles, but in general the snow hinders
most predators’ hunting. Winter predation risk is further
reduced by the absence of migratory raptors. While it of-
fers increased protection from most predators, the snow
cover may also protect and facilitate the hunting success of
the vole specialists, like the least weasel. T. Oksanen et al.
(2001) suggested that this winter tightening of the one-on-
one predation between voles and the weasel shifts the pat-
tern of vole population dynamics from one of annual fluc-
tuation to the regular and multi-annual cycles observed
between northern and southern Fennoscandia (Hansson
and Henttonen 1985, 1988).
Many rodent species of the northern hemisphere may
overwinter socially in aggregations (West 1977; Wolff and
Lidicker 1981; Ylönen and Viitala 1985), although good
empirical evidence for most species is lacking. The main
benefit of social aggregation in the winter is probably the
saving of energy through communal huddling (Madison
1984; Wolff 1980b). Furthermore, if your nest mate /mates
remain in the nest when you go foraging, then you return
to a warmer nest compared to the cold nest of a solitary
individual (Vickery and Millar 1984; Wolff and Lidicker
1981). Staying aggregated, however, may attract predators
and increase predation risk, especially because the snake-
like least weasel can search the subnivean space and is likely
to detect more easily the odor or heat signatures of aggre-
gated voles. However, during the winter the cost-benefit ra-
tio between huddling and predation risk may favor winter
aggregations as the benefits of huddling go up and the pre-
dation risks go down relative to huddling during other sea-
sons. Thus, the common pattern is that most voles are ter-
ritorial or solitary during the breeding season and social
during the winter (Madison 1984).
Deserts lie primarily at 30 degrees latitude and close to
the west sides of continents, creating diverse seasonal pat-
terns that influence predation risk on desert rodents. Most
if not all deserts lie along migratory bird routes, including
those of raptors. The GUDs of gerbils increase dramatically
at a Negev Desert site during the spring raptor migration
(Brown et al. 1994a). Deserts see sharp seasonal shifts in
temperature, although these are less dramatic than the shifts
that take place in boreal landscapes. At a Sonoran desert
site, rattlesnakes drove kangaroo rats (D. merriami) to pre-
fer the open microhabitat during the summer, whereas
ground squirrels (Spermophilus tereticaudus) enjoyed rela-
tive safety from diurnal raptors rendered scarce by high
temperatures. These same squirrels have increased GUDs338 Chapter Twenty-Eight