are able to excavate burrows to locate the smaller, more
evenly dispersed underground vegetation without the coop-
eration of conspecifics (Jarvis et al. 1994).
In contrast to the fossorial mole-rats, the costs of obtain-
ing food seem to impose significant restrictions on group
formation in many species of semifossorial desert rodents
(Daly and Daly 1975a, 1975b; Randall 1993, 1994; Shen-
brot et al 1999). For seed-eating rodents, heterogeneous
topography and scattered vegetation cause patchiness in
characteristics of the soil surface and the subsequent distri-
bution of food (Reichman and Price 1993). This patchy dis-
persion of food is especially true for seeds that are distrib-
uted by wind. Green vegetation in desert environments is
relatively sparse, of low quality, and patchily distributed,
which is probably why folivory is limited in desert rodents
(Shenbrot et al. 1999). Furthermore, temporal availability
of food can have major effects on efficiency of food har-
vesting (Price and Joyner 1997). In most desert conditions,
therefore, food may be too dispersed, low in quality and
temporally unpredictable to support more than a single an-
imal, resulting in a solitary social structure (Daly and Daly
1975a, 1975b; Eisenberg 1975; Shenbrot et al. 1999).
Comparison of arid-adapted species in different habitats
illustrates how constraints of food abundance and distribu-
tion may affect group formation. The great gerbil (R. opi-
mus) and the fat sand rat (Psammomys obesus) are ecolog-
ically similar. The two species have similar body mass, feed
primarily on green vegetation, and are diurnal, but P. obe-
susis solitary and R. opimusis social (table 31.1). The main
difference between the two species is that P. obesusinhabits
very arid areas of the Middle East and North Africa where
vegetation can be quite limited and competition for food is
high while R. opimusinhabits more productive habitats in
Central Asia that can support social groups (Daly and Daly
1974; Rogovin 1996; Tchabovsky et al. 2001).
Group sizes and structure in social desert rodents can be
constrained by food availability (Ågren et al. 1989a, 1989b;
Randall 1994; Rogovin, Moshkin et al. 2003; table 31.1).
The Mongolian gerbil (Meriones unguiculutus) forms larger
groups and attains higher population densities where vege-
tation is more abundant and predictable compared with
smaller groups in more arid regions with little herbaceous
growth (Xia et al. 1982; Orlenev 1987; Ågren et al. 1989a,
1989b). Social groups of the Indian gerbil (Tatera indica)
are much larger in urban areas near human habitation
where resources are plentiful and consistent compared with
solitary or male-female pairs in arid grasslands (Idris and
Prakash 1985). InR. opimus,large social groups form when
herbaceous vegetation becomes plentiful and population
densities are high. When food becomes limited and densi-
ties are low, social groups are small and many females are
solitary (Tchabovsky et al. 2001; Randall et al. 2005). So-
cial structure is also based on food type in the southern
African striped mouse (Rhabdomys pumilio;Schradin and
Pillay 2004; Schradin and Pillay 2005b). Social groups
form in the desert habitat because of a stable, year-round
food source of succulent plants. In grasslands, the mice feed
on a more ephemeral food source of seeds and are solitary
(Schradin and Pillay 2005b).PredationIndividuals in social groups often benefit from an increased
ability to detect and escape from predators (Ebensperger
2001a). Individual risk of predation decreases with an in-
crease in group size because of increased alternative targets
and early detection of predators (Betram 1978; Inman and
Krebs 1987; Lima and Dill 1990). Ground squirrels under
predatory risk in open, riskier habitats exhibit larger group
sizes and more cooperative defenses compared with ground
squirrels in more vegetated, safer habitats (Dunbar 1989;
Blumstein and Armitage 1997b). However, this relation-
ship does not hold for desert rodents. Few species have
group predator defenses, possibly because a majority are
nocturnal (70 – 90% in a local fauna; Shenbrot et al. 1999).
Nocturnal activity may eliminate the group benefit in
predator detection because long-distance visual identifica-
tion of predators is ineffective in the dark.
Both diurnal great gerbils, R. opimus,and subterranean
naked mole-rats, H. glaber,however, have evolved group
defenses against predators (Lacey and Sherman 1991; Ran-
dall et al. 2000). Alexander (1991) proposed that group
living in the naked mole-rats evolved as a response to pre-
dation, and the well-developed defenses against predators
suggest that predation was a factor in the evolution of so-
ciality in R. opimus. The gerbils benefit from mutual vigi-
lance and the use of alarm signals to alert group members of
approaching danger (Randall and Rogovin 2002). Rogovin
et al. (2004) found that a common predator of R. opimus,
the monitor lizard (Varanus greseus caspius), visited dis-
persed groups of gerbils more frequently than areas where
family groups were close together, but there was no differ-
ence by group size and composition. The alarm-call system
of gerbils may be more effective when family groups live in
close proximity and neighbors can hear each other. Rhom-
bomys opimushas an elaborate system of alarm communi-
cation, consisting of three different vocalizations and seis-
mic footdrumming signals (Randall et al. 2000; Randall
and Rogovin 2002). The calls communicate response ur-
gency to gerbils outside the burrow, and the footdrumming
probably communicates to family members inside. Merio-
nes unguiculatusfootdrums, but does not call, when dis-
turbed, as a possible alarm signal (Ågren et al. 1989a).Environmental Constraints and the Evolution of Sociality in Semifossorial Desert Rodents 371