Most species of gerbil footdrum when they are disturbed
(Randall 2001). Whether they drum in response to preda-
tors is unknown.
Solitary desert rodents have individual predator defenses
that depend on auditory and olfactory information for de-
tection of predators at night (Randall 1993, 1994; Randall
et al. 1995; DeAngelo 2002; Busch 2003). In compari-
son with semifossorial rodents in mesic environments, arid-
adapted rodents have a more specialized ear, which includes
expanded auditory bullae and specialized cochlea for hear-
ing the low-frequency sounds made by predator movements
(Webster and Webster 1984; Webster and Plassmann 1992).
Low-frequency hearing for predator detection seems to be
a highly valuable adaptation, because it has evolved inde-
pendently in at least four different lineages of semifossorial
desert rodents: kangaroo rats in North America, (Webster
and Webster 1984), gerbils in Africa and Asia (Plassmann
et al. 1987; Pavlinov 1988), jerboas (Dipodidae) in Asia
(Pavlinov and Rogovin 2000), and vizcacha rats (Tympa-
noctomys barrerae) in South America (Ojeda et al. 1999).
Desert rodents also exhibit fast and efficient locomotion
to escape predators (Shenbrot et al. 1999). Bipedal loco-
motion, which provides an advantage in speed to escape
predators in the initial stages of locomotion in open habi-
tats, has evolved independently in the deserts of Asia (Dipo-
didae), Africa (Pedetidae), Australia (Muridae), and North
America (Heteromyidae; Mares 1993; Randall 1994; Shen-
brot et al. 1999). Species of larger body mass usually show
the most specialization for a fast, bipedal gate, whereas
medium- or smaller-sized rodents usually hide in shelter
burrows or under vegetation (Rogovin 1999; Shenbrot et al.
1999).
Flexible strategies of reproduction
Mammalian behavior often reflects opportunistic responses
to changes in the characteristics of the environment (Austad
1984; Wrangham and Rubenstein 1986; Lott 1991). In par-
ticular, mixed strategies allow animals to deal with uncer-
tainty (Haccou and Iwasa 1995; Flaxman 2000). For in-
stance, although marmots are usually polygynous, sparse
resources and spacing of females can lead to strictly mo-
nogamous mating (Holmes 1984a). Social behavior is also
related to the distribution and predictability of resources
(Bradbury and Vehrencamp 1976; Waser 1981; Wrangham
and Rubenstein 1986); temporal and spatial patterns of re-
source availability can affect group sizes (Pulliam and Ca-
raco 1984; Johnson et al. 2002).
Behavior of desert rodents seems especially opportunis-
tic and flexible. Instead of a strictly seasonal pattern of re-
production, many species reproduce at any time of the year,
during variable periods of rainfall and peak food availabil-
ity (Schmidt-Nielsen 1975; Randall 1993, 1994). Most des-
ert rodents are sexually mature before their first year of life
and thus are able to take advantage of favorable conditions
for reproduction. These adaptations are common to differ-
ent lineages of desert rodents; fast maturation and irregular
reproductive schedules are reported in rodents from deserts
in North and South America, Africa, Asia, and Australia
(Happold 1975; Prakash 1975; Taylor and Green 1976;
Breed 1990; Jones 1993; Randall 1993, 1994).
Flexible behavior also appears during mating. In kanga-
roo rats, mating ranges from being exclusive between male
and female neighbors to competition between two or more
males for access to a female (Kenagy 1976; Randall 1987a,
1987b, 1989b, 1989c, 1991a; Randall et al. 2002). Exclu-
sive mating usually occurs at low densities as a result of spa-
tial isolation when a single male overlaps the home range
of a single female. Multiple males are much more likely to
converge at the home area of an estrous female and com-
pete for access to her when population densities are higher
and the operational sex ratio favors males (Kenagy 1976;
Randall 1987b, 1989b, 1991a; Randall et al. 2002). Ex-
clusive mating can also occur at the beginning of the breed-
ing season, when a female enters estrus outside of regular
breeding activity, or when females occupy territories that
are spaced far apart (Randall 1993; Randall et al. 2002).
Because the timing of female receptivity is variable and
short, males must continuously monitor females for signs of
a mating opportunity (Randall 1989b, 1991a). Although
vaginal swelling can last for several days, behavioral estrus
is brief, lasting only 4 – 6 hours per cycle (Eisenberg 1963a;
Daly et al. 1984; Randall 1991a; Yoerg and Shier 2000;
Randall et al. 2002). Estrus is not evident in the appearance
of the vaginal area (Wilson et al. 1985), but may be de-
tected by olfactory cues (Roberts, chap. 22 this volume).
During my studies of mating behavior, I found that females
mated from two to eight nights after the vagina became
swollen and open, and judging by visits of males to estrous
females, neither the males nor I could predict behavioral es-
trus until females accepted advances by the males (Randall
1989b; Randall 1991a; Randall et. al 2002). Moreover, fe-
males of some species, such as D. merriami,do not seem
to signal their reproductive readiness ahead of time by scent
from either their dorsal gland or their urine (Randall 1993);
D. merriamiand D. agilisfemales, but not males, prefer
conspecific odors when in reproductive condition (Daly
et al. 1980). These results suggest that males may obtain
limited information in advance about the reproductive sta-
tus of females in these species.
Male kangaroo rats may gain information about whether
an estrous female is ready to mate only at close proximity,372 Chapter Thirty-One