The Striped Mouse (Rhabdomys pumilio) From the Succulent Karoo,
South Africa: A Territorial Group-Living Solitary Forager With Communal
Breeding and Helpers at the Nest
Carsten Schradin and Neville Pillay
University of the Witwatersrand
The authors studied the striped mouse (Rhabdomys pumilio) in the semiarid succulent karoo of South
Africa. Mice forage alone, but they live in groups that share a common nest. Groups consist of 1 to 4
breeding females, 1 to 2 breeding males, and their offspring of both sexes, which remain in their natal
group even after reaching adulthood, participating in territorial defense and nest building without
showing signs of reproductive activity. Interactions are typically amicable and take place inside or in
front of the nest. In contrast, encounters with mice from other groups are aggressive. Group living in the
succulent karoo is possibly due to ecological constraints imposed by habitat saturation because of a
year-round stable food supply as well as associated benefits of philopatry.
For researchers of animal behavior, the study of differences in
social organization has always been a fascinating topic (Lott,
1991). Closely related species were often studied to find explana-
tions for differences in social behavior (e.g., Crook, 1964; Jarman,
1974; Reburn & Wynne-Edwards, 1999; Schradin, Reeder, Men-
doza, & Anzenberger, 2003). Differences in social organization
occur, for example, in closely related species of the vole genus
Microtus (Parker, Phillips, & Lee, 2001). Some species are soli-
tary, such as the California vole (Microtus californicus; Salvioni &
Lidicker, 1995) and the meadow vole (Microtus pennsylvanicus;
Parker et al., 2001; Webster & Brooks, 1981), whereas in the
common vole (Microtus arvalis), females live in groups (Dobly &
Rozenfeld, 2000). Differences in social organization can even
occur within species, and the social flexibility of a species is often
associated with its ability to inhabit different habitats (Lott, 1991).
One example is the prairie vole (Microtus ochrogaster), which can
be monogamous or solitary promiscuous or polygynous depending
on the habitat it occupies (Roberts, Williams, Wang, & Carter,
1998). Another rodent species that inhabits various habitats is the
striped mouse (Rhabdomys pumilio; as this genus is monotypic, we
refer to this species as Rhabdomys hereafter; but see also Rambau,
Stanyon, & Robinsom, 2003), a diurnal murid rodent with a body
weight of 40–80 g. It is widely distributed in southern Africa and
can be found in different habitats, such as grassland, marsh,
forests, semideserts, and deserts (Kingdom, 1974).
Many field studies have been conducted on Rhabdomys
(Brooks, 1982; Choate, 1972; David & Jarvis, 1985; Perrin, Ercoli,
& Dempster, 2001; Willan & Meester, 1989; Wirminghaus &
Perrin, 1993), all of which used the indirect method of capture,
mark, and recapture. In the grasslands of Zimbabwe, a female and
her pups of the last litter stay in one nest, whereas the males
occupy separate areas (Choate, 1972). The same pattern apparently
exists in the grasslands of KwaZulu-Natal Province, South Africa,
where female Rhabdomys have exclusive territories, which are
aggressively defended against other females, and male territories
overlap several female territories (Perrin et al., 2001; Willan, 1982;
Willan & Meester, 1989; Schradin & Pillay, 2003b). Additionally,
field studies in grassland areas in several other South African
localities, such as Pretoria (Brooks, 1974) and KwaZulu-Natal
midlands (Wirminghaus & Perrin, 1993) as well as in Acacia
habitat near Cape Town (David & Jarvis, 1985), and in semisuc-
culent thorny scrub in the Eastern Cape Province (Perrin, 1980b)
indicate that Rhabdomys lives solitarily, except for mothers and
We showed in an earlier study that male Rhabdomys exhibit
high levels of paternal care in captivity (Schradin & Pillay, 2003d).
As a direct response to the presence of pups, fathers increase the
time spent in the nest nearly threefold. Males lick and huddle pups
in the nest, and they show this behavior to the same extent as
females. Also, males retrieve pups that have been positioned
outside the nest. These behaviors do not concur with observations
that Rhabdomys is a solitary species and males associate with
females only for mating and are not associated with juveniles in
the field (Willan, 1982). Males of many rodent species show
Carsten Schradin and Neville Pillay, Ecophysiological Studies Research
Group, School of Animal, Plant and Environmental Sciences, University of
the Witwatersrand, Johannesburg, South Africa.
This study was supported by the Swiss National Science Foundation, the
Schweizerische Gesellschaft fu ¨r Naturwissenschaften, the Fonds zur Fo ¨r-
derung des akademischen Nachwuchses (Zu ¨rcher Universita ¨tsverein), and
the University of the Witwatersrand. Animal Ethical clearance numbers are
AESC 2001-32-3 and 2002-14-3. We thank Northern Cape Department of
Agriculture, Land Reform, Environment and Conservation, Kimberley,
South Africa, for their assistance and Klaas van Zyl, Enrico Oosthuysen,
and their staff at Goegap Nature Reserve, Springbok, South Africa, for
their support during the study. M. Burmeister, B. Britz, R. Gutzat, J.
Matthee, F. Mattle, J. Schradin, and M. Schubert assisted during different
parts of the field study. Jens Schradin provided the photos for Figure 1.
Important comments by A. Mu ¨ller, A. Rusu, and R. Weinandy helped to
improve the manuscript.
Correspondence concerning this article should be addressed to Carsten
Schradin, Ecophysiological Studies Research Group, School of Animal, Plant
and Environmental Sciences, University of the Witwatersrand, Private Bag 3,
Wits 2050, Johannesburg, South Africa. E-mail: firstname.lastname@example.org
Journal of Comparative Psychology
2004, Vol. 118, No. 1, 37–47
Copyright 2004 by the American Psychological Association, Inc.
0735-7036/04/$12.00 DOI: 10.1037/0735-7036.118.1.37
paternal care under artificial conditions in the laboratory without
indications from the field, and these behaviors might be regarded
as laboratory artifacts (Dewsbury, 1985). Alternatively, paternal
care might be a true alternative strategy shown under special
ecological conditions, but not under others (Dewsbury, 1985).
There is some indication that the social structure of Rhabdomys
is flexible. In the Kalahari, Rhabdomys seems to exhibit a more
communal than solitary lifestyle (Nel, 1975). For small mammals,
group living could be a strategy that provides some benefits in arid
environments (Dean & Milton, 1999), but not in more mesic
environments. For example, southern Africa squirrels (Sciuridae)
living in arid environments are typically group living, probably
because of benefits accrued through predator avoidance and ther-
moregulatory benefits provided by nest sharing, but are solitary in
more mesic areas (Waterman, 1995). One reason for Rhabdomys
being solitary in grasslands but social in dry habitats might be
differences in food abundance: In grasslands, most vegetation
consists of grass that is apparently not consumed by Rhabdomys,
which instead feeds on grass seeds and other foods, such as berries
and herbs (Curtis & Perrin, 1979; Perrin, 1980b). Thus, Rhab-
domys in the mesic grasslands may actually be surrounded by
mostly unpalatable food, and palatable food is patchily distributed
and scarce, which then would demand large home ranges, low
population density, and a solitary lifestyle (Ostfeld, 1990). In
contrast, food is abundant in spring in the semiarid succulent
karoo, and mice gain weight during this period, probably ensuring
survival during the forthcoming dry season (summer) when the
food availability decreases (Schradin & Pillay, 2003a). Even dur-
ing the dry summer, mice have access to a stable year-round food
source that consists of succulents and Zygophyllum retrofractum
bushes (personal observation). Similarly, Acacia tree seeds are a
long-lasting food source of high nutritional value in the Kalahari,
and mice here are communal (Nel, 1975; personal observation). A
larger food supply might lead to smaller home ranges, higher
population density, and possibly sociality (Ostfeld, 1990).
Because of larger food abundance in dry areas, we predict that
Rhabdomys is more social there than in the mesic grasslands.
Furthermore, for paternal care to be an adaptive strategy, males
would need to be associated with their potential offspring. To test
these predictions, we studied a Rhabdomys population in a dry
environment, the succulent karoo of South Africa.
Study Area and Period
The study was conducted in Goegap Nature Reserve near Springbok in
the Northern Cape Province of South Africa. The area is semiarid, and
rainfall, which averages 160 mm per year, falls mainly in winter (Jackson,
1999). The vegetation type is succulent karoo (Acocks, 1988), consisting
mainly of Zygophyllum retrofractum bushes (see Figure 1A). There are
large, open sandy patches, containing different species of small succulents.
In spring (August–September), these sandy areas are covered by approxi-
mately 600 different species of wild flowers (information provided by
Goegap Nature Reserve). There are more than 4,000 species of plants in the
succulent karoo (leRoux, Schelpe, & Wahl, 1997), which has been iden-
tified as 1 of 25 global biodiversity hotspots (Myers, Mittermeier, Mitter-
meier, Fonseca, & Kent, 2000).
The study was performed from September 2001 to January 2002, during
2 weeks in April 2002, and from September 2002 to December 2002. The
study area was initially 80 m ? 60 m but was enlarged in December 2001
by an additional 60 m ? 40 m to include two more Rhabdomys groups. In
September 2002, the study area was further increased to 200 m ? 150 m.
Trapping and Marking of Animals
Rhabdomys was live trapped using metal traps (26 ? 9 ? 9 cm) baited
with a mixture of bran flakes, currants, sea salt, and salad oil. A total of 235
mice (114 males and 121 females) was trapped in 2001, 149 mice (84
males and 65 females) in April 2002, and 234 mice (124 males and 110
females) from September 2002 to November 2002. Traps were placed in
the shade under bushes where mice had been observed previously. Trap-
ping was done only in the morning and afternoon, but not during the hottest
times of the day. Traps were checked continuously. Trapped mice were
sexed, weighed, and individually marked with hair dye. Each mouse
received a number written on its side with black hair dye (Rapido, Pine-
town, South Africa; see Figure 1B). Mice were retrapped about every 5–6
weeks to refresh the markings and to mark juveniles and emigrants onto the
grid. We decided to mark mice permanently in 2002 by toe clipping, the
standard method for studies in small mammals (e.g., Jackson, 1999;
McGuire, Getz, & Oli, 2002). Toe wounds were disinfected with alcohol,
and no recaptures had infected wounds. The decision to use toe clipping
was not made lightly. However, we decided not to mark animals with
transponders because the large number of mice that needed to be marked
over subsequent years (more than 200 a year) would have led to contam-
ination of the study site with electronic waste, a concern raised by the
authorities of the nature reserve where we conducted our studies. Also,
there is no difference in survival probability between toe clipping and
marking with transponders (Braude & Ciszek, 1998). We did not consider
eartags as these are known to lead to increased parasite load (Ostfeld,
Miller, & Schnurr, 1993) and to be unreliable, as they are often lost (Harper
& Batzli, 1996; Wood & Slade, 1990). Instead, we first tried permanent
marking by ear punching, but this method turned out to be unreliable, as the
punch marks either split or closed again. Thus, we decided to use toe
clipping, which does not appear to have deleterious effects on survival or
body weight in other species (Braude & Ciszek, 1998; Korn, 1987; Wood
& Slade, 1990).
We recorded the location of all nest sites. Five nest sites were investi-
gated and the structure recorded. We presented white tissue in front of
nests, which resident mice used for nest construction, thus enabling us to
identify active nests.
Mice inhabiting one nest site were regarded as one group, and groups
were numbered from G1 to G9. Group composition was determined
through observations in front of nests during early morning and late
afternoon, when mice emerged from or withdrew into nests. At each nest,
we recorded the identity of individual subjects and how many mice were
present at the nest. During the short field trip in April 2002, we did not
mark mice with hair dye and determined only group size, but not sex ratio.
Groups were located one at a time. As mice were not toe clipped in 2001,
subjects observed in 2001 were not recognized in 2002. However, groups
inhabiting the same home range and/or nesting site in 2002 as in 2001 were
regarded as the same group.
Scrotal males that were present since the beginning of the breeding
season or that weighed over 60 g in October–December were regarded as
the breeding males of the groups. At the beginning of the breeding season,
females weighing over 60 g were regarded as being pregnant and thus
breeding. We also recorded the number of same-aged juveniles in each
group. The maximum litter size observed in 12 wild-caught pregnant
SCHRADIN AND PILLAY
in the morning sun in front of their nests. Note that the numbers written with black hair dye on their sides are
to enable individual recognition.
A: Field study site in the succulent karoo in December. B: Group of striped mice basking together
SOCIAL SYSTEM OF THE STRIPED MOUSE
Rhabdomys was 9, with a mean of 5.3 (Schradin & Pillay, 2003a). Using
these data and the number of same-aged juveniles in a nest, we calculated
the minimum number of breeding females per group as the number of
same-aged juveniles divided by 9 (maximum litter size) as well as the
approximate number of breeding females as the number of same-aged
juveniles divided by 5.3 (mean litter size).
During 2001 (but not 2002), mice were observed directly using 10 ? 42
binoculars during the day. Focal-animal sampling was performed during
the morning (6:00 a.m. to 11:00 a.m.) and afternoon (4:00 p.m. to 7:30
p.m.), the main activity periods of Rhabdomys (Schradin, 2003). Every
time a marked mouse was located, it was observed and carefully followed
for a distance of 3–10 m until it disappeared from view. Observation time
totaled 210.0 hr, during which focal observations were performed for 52.1
hr (21.0 hr for males and 31.1 hr for females). For the remaining time, no
focal animal was present, and we searched for the next focal animal. A total
of 477 focal observations were performed, ranging from 1.0 min to 45.0
min in duration, with a mean of 6.5 min. In total, 19 adult males resident
at the start of the breeding season (male breeders), 33 males born during the
breeding season, 23 adult females resident at the start of the breeding
season (female breeders), and 37 females born during the breeding season
were observed as focal animals. All social interactions between focal and
other mice and the identity of the actor and receiver were recorded. All
behaviors were recorded as events, as the complex and rapid movements of
mice made recording of behavioral states impractical. The following be-
havior patterns were recorded: aggressive interactions (chasing) and so-
ciopositive interactions (sniffing at each other, sitting in body contact, and
grooming one another).
In addition to focal-animal observations, observations of mice were
performed during mornings and afternoons in front of nests (see above).
These observations were done in front of all known group nests at the end
of the breeding season (48 observations from November 2001 to January
2002 and 30 observations in December 2002). Observations started when
the first mouse left the nest and continued for 30 min thereafter, when
normally most group members had already left the nest. All social inter-
actions among group members were recorded.
During focal-animal sampling, the locations of focal mice were re-
corded. For this, a map of the study area had been drawn and divided into
2 ? 2-m grids. Bushes (N ? 95), which were individually marked, were
used as landmarks. We recorded when a focal animal changed its position
into another square of the grid.
To determine home-range size, we applied the minimum polygon
method (Kenward, 1987). The number of grids (representing 2 ? 2 m) and
half grids within this convex polygon was summed to calculate home-range
size in square meters. Home ranges were determined only for mice that had
been used as focal animals on at least five occasions and that were
observed for at least 30 min in total. Obvious excursions by mice were not
taken into account for home-range analyses (three cases of 3 different
mice). Excursions were noted when the focal mouse had been observed
only once in a location that was more than 10 m from other locations where
the same mouse had been observed during another time. Data were ob-
tained from 18 mice (8 males and 10 females). Group home ranges were
determined by using all data available of all known group members,
independent of how often each individual had been observed. The center of
group home ranges was determined as the center of gravity of the polygon.
Data are described as means plus or minus standard errors of the means.
We applied nonparametric tests (Siegel & Castellan, 1988) throughout
using InStat (GraphPad Software, San Diego, CA); all tests were two-
tailed. The Wilcoxon matched-pairs signed-ranks test is abbreviated as
Wilcoxon test, and the Fisher’s exact test as Fisher test. All correlations
were performed using the Spearman rank correlation (rs).
Nesting Sites and Nests
Nests were typically situated inside dense Zygophyllum retro-
fractum bushes. Nests in bushes were above ground, had an oval
shape with a diameter of about 25 cm, and were lined with soft
hay, resembling a bird’s nest. Some nests also were underground
inside burrows that were originally dug by Littledale’s whistling
rats (Parotomys littledalei).
Observations of nest construction were done ad hoc. An adult
breeding male was observed carrying hay inside the nest. Juveniles
of approximately 40 days of age as well as breeding females were
observed to transport white tissue inside nests, which we provided
outside nests (see the Method section).
Group compositions are given in Table 1. There was usually
only 1 breeding male per group, although three groups comprised
2 big scrotal males (G3 and G6 in 2001; G3 in 2002). Normally,
there was more than 1 breeding female in the group, with a range
of 1 to 4. Table 2 shows the number of potentially reproducing
females in each group for both years.
Breeding Animals and Their Adult Offspring
From the middle of October onward, groups contained young
adult mice in addition to breeding animals (see Table 1). Offspring
of both sexes remained in their natal group after reaching adult-
hood. Of the young adult males, 71.4% ? 19.8% that were
observed at nests in December 2002 had been observed at the same
nest the previous month (n ? 9 groups, 66 mice). For females, the
value was 80.8% ? 20.6% (n ? 9 groups, 60 mice), and there was
no difference between the sexes (Wilcoxon test), T(N ? 8) ? 9,
p ? .25. Emigration of juveniles or young adults into other groups
was never observed, but 2 old adult males were observed emigrat-
ing into a group of breeding females in October 2001 (G1 and G2).
Groups were still large in April (see Table 1), consisting of both
males and females, which indicates that offspring remain within
their group for several months.
Rhabdomys reach sexual maturity at an age of 1–3 months
(Brooks, 1982). Because the breeding season lasts 3 months
(September–November), it would have been possible for offspring
born at the start of a breeding season to reproduce in the season of
their birth. However, significantly more breeding males (34 of 34)
were scrotal than their sons (2 of 17; Fisher test, p ? .0001) in
October (data from 2001 and 2002 combined), and the same
pattern occurred in December 2001 (scrotal males: 14 of 19; sons:
2 of 33; Fisher test, p ? .0001). In October, 28 of 48 breeding
females showed signs of sexual activity (open vagina or being
pregnant), but only 1 of 9 daughters did so (Fisher test, p ? .03).
In December, there was no significant difference between female
breeders with an open vagina (2 of 22) compared with daughters (1
of 26; Fisher test, p ? .59).
SCHRADIN AND PILLAY
The difference in breeding status between breeders and their
adult offspring cannot be explained by differences in body weight
alone because breeders weighed as much at the start of the breed-
ing season as their adult nonreproductive offspring did at the end
of the breeding season. In September, 90% of males with a body
weight between 40 and 50 g were scrotal, and 100% of females
within the same weight range had an open vagina. This pattern
changed dramatically during the following months: When non-
breeding adult offspring attained 40–50 g body weight, the breed-
ers had gained additional weight (all 11 breeders with a body
weight below 50 g in September weighed more than 50 g in
October; Wilcoxon test), T(N ? 11) ? 0, p ? .0001. The propor-
tion of reproductively active mice with a body weight of 40–50 g
was significantly higher in September than in October and Decem-
ber (Fisher test, p ? .0001).
Social Interactions Among Group Members
Mice spent the morning and afternoon in front of the nest
engaged in sociopositive interaction (see Figure 1B) but spent the
rest of the day foraging alone. No agonistic interactions were
observed between group members. Data collected in 2001 revealed
that mice were much more often observed to perform sociopositive
interactions in front of the nest (120 interactions in total, morning
and afternoon combined) than during focal-animal sampling dur-
ing the day (25 interactions; Fisher test, p ? .0001).
Data collected in front of nests in December 2002 revealed that
breeders, mainly breeding males, initiated more social interactions
than adult offspring (see Figure 2). However, the overall difference
only approached significance (Friedman test), Fr(N ? 7 groups) ?
7.172, p ? .066.
For both sexes, the variance in home-range size was high
(males: 990 ? 447 m; females: 960 ? 520 m). There was a
positive correlation between individual home-range size and group
home-range size, rs(N ? 18) ? .568, p ? .02. When we controlled
for differences in group territory size by comparing males and
females of the same groups in a paired design, there was no
Composition of Focal Groups, Arranged According to Different Age Classes
Group and age class
October NovemberDecemberJanuaryAprilSeptember OctoberNovember December
17155 1217 21
are separated by a slash. Mice inhabiting the same area in 2001 and 2002 were regarded as the same group.
Juveniles had a body mass below 30 g. Total numbers are shown, and if known, sex is indicated as male–female. Juveniles from different litters
SOCIAL SYSTEM OF THE STRIPED MOUSE
difference in home-range size between the sexes (Wilcoxon test),
T(N ? 6) ? 5, p ? .313. Home ranges of mice of one group
overlapped largely with each other (see example in Figure 3) but
not with the home ranges of mice from other groups. Overlap with
group members was 91.0% ? 11.6%, whereas overlap with indi-
viduals from other groups was 13.0% ? 11.8% (Wilcoxon test),
T(N ? 16) ? 0, p ? .0001. Figure 4 shows the group home ranges
Number of Potentially Reproducing Females
Group and year
Adult females at the start
of the breeding season
Minimum number of
Approximate number of
The minimum number of breeding females is the number of juveniles divided by 9 (i.e., the maximum litter size). The approximate number of breeding females is the number of juveniles divided
by 5 (i.e., the average litter size). Dashes indicate that no data were available.
aThis group was located only after the breeding season in 2001; so, no data are available prior to this time. Only a single female was observed in the former territory of this group in 2002.
females (F) nested alone but occupied part of the home range of G6.
breeding males, adult sons, breeding females, and adult daughters from
seven different groups in front of their nests in December 2002. Values are
shown for 30-min focal observation periods in front of nest.
Mean (? SEM) number of social interactions initiated by
ranges are indicated in gray, and 2 females’ ranges are indicated in black).
Compare with Figure 4, in which this group’s home range is represented
with a gray dotted line in the middle of the figure.
Overlapping home ranges of the mice of one group (3 males’
SCHRADIN AND PILLAY
of the seven focal groups in 2001. There was a significant positive
correlation between group home-range size and number of mice in
the group at the end of the observation period in 2002, rs(N ? 7) ?
.857, p ? .04.
A total of 48 aggressive interactions were observed. Figure 4
shows the locations of aggressive interactions in relation to home-
range boundaries. Two interactions at the lower end of the study
area were excluded from further analyses, as home-range bound-
aries of adjacent groups were not known. Aggressive encounters
were significantly more likely to occur near home-range bound-
aries than in home-range centers (binomial test, p ? .001).
In 12 encounters, both mice were known: In 2 cases, they were
from the same group, and in 10 cases, they were from different
groups (binomial test, p ? .05). It was possible to establish the sex
of the mouse initiating the encounter in 26 aggressive encounters:
Males initiated encounters on 19 occasions, and females on 7
occasions (binomial test, p ? .05). It was possible in some cases
to determine the sex of the attacked mouse: Males attacked other
males on 4 occasions and females on 10 occasions (binomial test,
p ? .10), whereas females attacked a male once and another
female on 3 occasions.
In a previous study, we reported high levels of paternal care in
captive Rhabdomys (Schradin & Pillay, 2003d), which did not
match the solitary life pattern described for this species in the field.
We predicted that Rhabdomys would exhibit a social lifestyle in
the succulent karoo because of higher food abundance compared
with the mesic grasslands. In particular, we expected paternal care
to be an alternative male reproductive strategy (Dewsbury, 1985).
relation to group territories are included. At one place (indicated by the gray dot in the upper part of the figure),
three aggressive interactions were observed during different times.
Group home ranges of the seven focal groups. Aggressive interactions (black dots) observed in
SOCIAL SYSTEM OF THE STRIPED MOUSE
Our data indicate that Rhabdomys is highly social in the semiarid
succulent karoo, where it lives in permanent groups that include
breeding males that interact highly amicably with other group
Communal Breeding, Polygyny, and Paternal Care
Most females were pregnant at the beginning of the breeding
season and also showed an open vagina as additional external signs
of reproductive activity. In addition, the number of same-aged
juveniles counted in front of nests exceeded the mean and maxi-
mum litter size of single females, indicating that parturition of
group females is synchronized. Communal breeding (or plural
breeding) in rodents occurs when several mature group members
breed together, whereas communal nesting occurs when reproduc-
ing females use the same nest (Hayes, 2000). Thus, both commu-
nal breeding and communal nesting occur regularly in Rhabdomys
in the succulent karoo, although single-breeding females were also
observed. To date, however, we do not know whether and to what
extent group females cooperate in rearing one another’s young. It
is notable that 1 heavily pregnant female was observed to nurse
juveniles that were approximately 10 days old in front of the group
nest (personal observation). These juveniles could not have been
her offspring because her own offspring from a previous litter
would have been significantly older (at least 20 days old). Thus,
there is anecdotal evidence for allonursing by communal breeding
Three factors could drive the selection of communal nesting in
the succulent karoo. The first is predator avoidance. Unpublished
data (Schradin, 2002) of time-lapse videotaping in two natural
nests revealed that always at least one of the group members was
awake during the night, leading to increased vigilance. Even light
disturbance caused the entire group to quickly leave the nest.
However, this advantage does not explain why communal nesting
occurs in the succulent karoo but not the grasslands. The second is
thermoregulatory benefits of nest sharing (Howard, 1950), which
is known to be associated with communal breeding in rodents
(Carter & Roberts, 1997). As the breeding season starts 1 month
earlier (spring) in the succulent karoo than in the mesic grasslands,
communal nesting may be more beneficial in the succulent karoo
because of lower night temperatures. The third factor is communal
breeding, which is known to lead to fitness benefits in house mice
(Ko ¨nig, 1993, 1994a, 1994b). Communal breeding usually occurs
when cooperatively breeding females are close kin and the social
system is egalitarian (Gerlach & Bartmann, 2002). For example,
communal breeding of sisters leads to increased fitness in the
wood mouse (Apodemus sylvaticus; Gerlach & Bartmann, 2002).
An egalitarian social system is also likely to occur in Rhabdomys
from the succulent karoo because daughters delay their reproduc-
tion until the next breeding season when females of similar body
weight form breeding groups. Also, we did not observe any direct
indication of a dominance hierarchy among cooperatively breeding
females, indicating an egalitarian society. In contrast, reproductive
skew exists in mother–daughter pairs of wood mice, in which the
dominant mother can exploit maternal care from her daughter,
making it advantageous for the daughter to leave instead of breed-
ing communally with her mother (Gerlach & Bartmann, 2002).
This pattern is likely to apply for Rhabdomys from the mesic
grasslands, where the breeding season is more than twice as long
as that in the succulent karoo, and daughters accordingly start
breeding at a very young age (i.e., with a body weight below 30 g).
As experienced breeding females (i.e., mothers) weigh normally
around 40 g (Brooks, 1974; Perrin, 1980a), daughters will be better
off leaving the natal nest and breeding on their own than breeding
communally with their mother as a subdominant female, thereby
experiencing high reproductive skew.
Contradictory to previous assumptions, paternal care not only is
associated with monogamy in mammals but also occurs in polyg-
ynous species (Komers & Brotherton, 1997). In fact, paternal care
in monogamous mammals might have evolved from a group-living
ancestor with 1 breeding male and several breeding females, in
which the male participated in infant care (Brotherton & Komers,
in press). Rhabdomys in the succulent karoo is socially polygynous
(paternity has not been determined thus far) and displays paternal
care. Males living in stable social groups have the potential to
increase their fitness by showing paternal care, simply because
they are permanently associated with pups and sleep in the same
nest. It might not be very costly for males to groom and warm pups
(Schradin & Pillay, 2003d), and males may benefit by being
associated with receptive females. Nevertheless, paternal care
might have a great impact on male fitness. We have data showing
that pup development is better under biparental than exclusive
maternal care (Schradin & Pillay, 2003c). In contrast, females do
not form social groups but defend exclusive territories against
other females in mesic grasslands (Perrin et al., 2001; Schradin &
Pillay, 2003b; Willan, 1982). Under such circumstances, males
might not have time or even the possibility to show paternal care,
as they have to adopt an active searching strategy to get access to
receptive females (Ostfeld, 1990).
Adult Offspring Staying in the Group
Offspring of both sexes stay in their natal group even after
reaching adulthood. Group sizes thus increase up to 30 adults at the
end of the breeding season. Offspring remaining in their natal
group can be referred to as helpers when their behavior leads to
increased reproductive success of the breeders (e.g., Taborsky,
1994). In Rhabdomys, there is clear indication that this is likely to
be the case because offspring do not simply benefit from occupy-
ing their parents territory but participate in territorial defense (see
also Schradin, in press). Juveniles also participated in nest con-
struction. In other species, participation in territory defense and
nest building by nonbreeders have been acknowledged as impor-
tant aspects of helping behavior (for a review in cichlids, see
Taborsky, 1994; for birds, see Reyer, 1984; for suricates, see
Doolan & MacDonald, 1996; for callitrichids, see Hubrecht,
1985). Additionally, pups of the group are also very likely to
derive thermoregulatory benefits by the presence of juveniles, as
the night temperatures can fall close to 0 °C. We would therefore
expect that the fitness of breeding pairs would increase because of
help provided by juveniles during the breeding season.
Adult offspring staying in their natal group seem to be repro-
ductively inhibited. During the middle of the breeding season,
many more breeders showed signs of reproductive capability com-
pared with their adult offspring. Adult offspring at this stage
weighed less than old breeders (40–50 g vs. 60–80 g). However,
this weight difference cannot explain a lower percentage of young
adults not being in breeding condition (8%) because about 88% of
SCHRADIN AND PILLAY
breeders at the beginning of the breeding season also weighed
40–50 g. Furthermore, Rhabdomys populations in the mesic grass-
lands can start breeding at below 30 g, and nearly all mice
weighing between 30 and 40 g do breed (Brooks, 1982; Schradin
& Pillay, 2003b). We conclude that adult offspring staying at their
natal nest do not reproduce, although they would be physiologi-
cally capable of reproduction. One reason could be incest avoid-
ance because female Rhabdomys avoid breeding with the father,
and incestuous breeding leads to inbreeding depression (Pillay,
2002). Another important factor might be the short breeding sea-
son of Rhabdomys in the succulent karoo, which is only 3 months
long (Schradin & Pillay, 2003a) compared with 6 months in the
grasslands (Perrin et al., 2001). Whereas offspring of the first litter
born during the breeding season could theoretically start breeding
after the middle of the breeding season when they are 1.5 months
old (Brooks, 1982), their offspring might have a low survival
probability. During the dry summer following the breeding season,
food abundance is low, mice lose about 12% of body weight, and
the mortality rate is 70% during the cold winter (Schradin & Pillay,
2003a). Thus, instead of investing immediately in producing
young with a low survival potential, mice might benefit more by
delaying reproduction and investing in their own survival and
Comparison of Social Organization With Other Muroid
The social organization of Rhabdomys in the succulent karoo
comprises large groups containing several breeding females,
highly amicable relationships between group members but aggres-
sive responses toward strangers, and offspring of both sexes stay-
ing at their natal group even after reaching adulthood and partic-
ipating in nest building and territorial defense. These behavior
patterns could be adaptations to the harsh succulent karoo envi-
ronment. Annual rainfall is only 160 mm, and temperatures regu-
larly fall below 0 °C during winter and spring nights but can be
close to 50 °C during summer days. Rhabdomys is not the only
diurnal small mammal inhabiting the succulent karoo, and the
social systems of the other species there are quite different. The
whistling rat (Parotomys brantsii), although it lives in large colo-
nies with warrens of sometimes hundreds of individuals next to
one another, is a solitary species (Jackson, 1999). Another diurnal
rodent at our field site is the bush karoo rat (Otomys unisulcatus),
whose social system so far is unknown, but it might be either
solitary or living in small groups (personal observation). Another
diurnal small mammal in the same habitat is the round-eared
elephant shrew (Macroscelides proboscideus), which might be
solitary (Sauer, 1973). Round-eared elephant shrews, whistling
rats, bush karoo rats, and Rhabdomys inhabit exactly the same
habitat, are diurnal, and are highly territorial, and the three rodent
species additionally feed to a large extent on the same plant
material (personal observation). Despite these similarities, their
social systems are very different, indicating that the succulent
karoo enables them to develop different strategies to cope with its
harsh environment, providing different ecological niches.
The house mouse (Mus musculus) appears to have the most
similar social system to Rhabdomys in the succulent karoo. Like
Rhabdomys, the house mouse is a group-living solitary forager
(Gerlach, 1998), is polygynous (Lidicker, 1976; Wilkinson &
Baker, 1988), shows communal breeding (Ko ¨nig, 1993, 1994a,
1994b), and is territorial (Hurst, 1987; Lidicker, 1976). When
offspring are weaned, they can stay for some time in their natal
group in both species. However, whereas it is common for Rhab-
domys offspring of both sexes to remain in their natal group
several months after reaching adulthood, male house mice often
disperse when they reach sexual maturity (Gerlach, 1998;
One main difference between the house mouse and Rhabdomys
appears to be in the social relationships of the breeding (also called
dominant) males. The dominant male house mouse lives a rela-
tively solitary life (Gerlach, 1998). In contrast, breeding male
Rhabdomys are highly social, greeting other group members at the
nest by sniffing them and grooming and sitting in body contact
with females and juveniles (see Figure 2). The social nature of
Rhabdomys males is further demonstrated by four behaviors
(Schradin & Pillay, 2003d): (a) Their amicable relationship with
juveniles is similar to that between breeding females and juveniles,
(b) wild males retrieve pups experimentally presented to them, (c)
captive males exhibit high levels of paternal care, and (d) wild
males sleep together with group members. Unpublished results
(Schradin, 2002) from a study that videotaped two natural nests
revealed the breeding male grooming and licking pups in the nest.
In contrast, male house mice do not often sleep in the same nest as
the rest of the group, particularly during communal breeding
(Lidicker, 1976). Whereas male house mice can show paternal care
when kept in captivity in a monogamous situation, they seem to
invest their time mainly in territorial defense under polygynous
conditions, thereby reducing their social interactions with juveniles
(Gandelman, Paschke, Zarrow, & Denenberg, 1970; Gerlach,
1998; Lidicker, 1976).
The striped mouse exhibits a social system in the succulent
karoo, which is very different from its solitary lifestyle in other
parts of South Africa (Brooks, 1974; Perrin, 1980a; Willan, 1982).
In the succulent karoo, Rhabdomys is clearly group living. Why
mice stay together and share the same nest and territory is not yet
understood, but we predict that both ecological constraints and
benefits of philopatry are important factors (Hayes, 2000). Eco-
logical constraints are imposed by high food availability, which
results in high population density and habitat saturation, as dem-
onstrated by a population density of 151 mice/hectare in our field
site (Schradin & Pillay, 2003a), which is several times greater than
that reported from mesic grasslands (10 to 40 mice/hectare; Perrin
et al., 2001). Benefits of philopatry would include benefits of
predator avoidance because of increased vigilance during nest
sharing as well as thermoregulatory benefits (Howard, 1950),
which are important reasons for communal breeding in rodents
(Carter & Roberts, 1997). Communal breeding itself offers a great
advantage when reproductive skew is low (Gerlach & Bartmann,
2002). Whereas Rhabdomys breeds and nests communally in the
succulent karoo, these mice forage alone. Their food (leaves,
flowers, and seeds) is patchily distributed, and foraging in a group
would not assist in food exploitation but rather would increase
predation risk. Mice stay in their natal group even after reaching
adulthood, participating in nest building and territorial defense but
not in reproduction, potentially showing helping behavior. In con-
SOCIAL SYSTEM OF THE STRIPED MOUSE
clusion, Rhabdomys in the succulent karoo is best described as a
group-living solitary forager with communal breeding and helpers
at the nest, one of the most complex social systems found in
Acocks, J. P. H. (1988). Veld types of South Africa. (Available from the
National Botanical Institute, Private Bag X101, Pretoria 0001, South
Braude, S., & Ciszek, D. (1998). Survival of naked mole-rats marked by
implantable transponders and toe-clipping. Journal of Mammalogy, 79,
Brooks, P. M. (1974). The ecology of the four-striped field mouse, Rhab-
domys pumilio (Sparrman, 1784), with particular reference to a popu-
lation on the Van Riebbeeck Nature Reserve, Pretoria. Unpublished
doctoral dissertation, University of Pretoria, Pretoria, South Africa.
Brooks, P. M. (1982). Aspects of the reproduction, growth and develop-
ment of the four-striped mouse, Rhabdomys pumilio (Sparrman, 1784).
Mammalia, 46, 53–64.
Brotherton, P. N. M., & Komers, P. E. (in press). Mate guarding and the
evolution of monogamy in mammals. In U. Reichhard & C. Boesch
(Eds.), Monogamy: Mating strategies and partnership in birds, humans
and other mammals. Cambridge, England: Cambridge University Press.
Carter, C. S., & Roberts, R. L. (1997). The psychobiological basis of
cooperative breeding in rodents. In N. G. Solomon & J. A. French (Eds.),
Cooperative breeding in mammals (pp. 231–266). Cambridge, England:
Cambridge University Press.
Choate, T. S. (1972). Behavioural studies on some Rhodesian rodents.
Zoologica Africana, 7, 103–118.
Crook, J. H. (1964). The evolution of social organisation and visual
communication in the weaver birds (Plocceinae). Behaviour Supplement,
Curtis, B. A., & Perrin, M. R. (1979). Food preferences of the vlei rat
(Otomys irroratus) and the four-striped mouse (Rhabdomys pumilio).
South African Journal of Zoology, 14, 224–229.
David, J. H. M., & Jarvis, J. U. M. (1985). Population fluctuations,
reproduction and survival in the striped field mouse Rhabdomys pumilio
on the Cape Flats, South Africa. Journal of Zoology, London, 207,
Dean, W. R. J., & Milton, S. J. (1999). The karoo. Cambridge, England:
Cambridge University Press.
Dewsbury, D. A. (1985). Paternal behavior in rodents. American Zoologist,
Dobly, A., & Rozenfeld, F. M. (2000). Burrowing by common voles
(Microtus arvalis) in various social environments. Behaviour, 137,
Doolan, S. P., & MacDonald, D. W. (1996). Dispersal and extra-territorial
prospecting by slender-tailed meerkats (Suricata suricatta) in the south-
western Kalahari. Journal of Zoology, 240, 59–73.
Gandelman, R., Paschke, R. E., Zarrow, M. X., & Denenberg, V. H.
(1970). Care of young under communal conditions in the mouse (Mus
musculus). Developmental Psychobiology, 3, 245–250.
Gerlach, G. (1998). Impact of social ties on dispersal, reproduction and
dominance in feral house mice (Mus musculus domesticus). Ethology,
Gerlach, G., & Bartmann, S. (2002). Reproductive skew, costs, and ben-
efits of cooperative breeding in female wood mice (Apodemus sylvati-
cus). Behavioral Ecology, 13, 408–418.
Harper, S. J., & Batzli, G. O. (1996). Monitoring use of runways by voles
with passive integrated transponders. Journal of Mammalogy, 77, 364–
Hayes, L. D. (2000). To nest communally or not to nest communally: A
review of rodent communal nesting and nursing. Animal Behaviour, 59,
Howard, W. E. (1950). Relation between low temperature and available
food to survival of small rodents. Journal of Mammalogy, 32, 300–312.
Hubrecht, R. C. (1985). Home range size, use and territorial behaviour in
the common marmoset Callithrix jacchus at the Tapacura field station
Recife, Brazil. International Journal of Primatology, 6, 533–550.
Hurst, J. L. (1987). Behavioural variation in wild house mice Mus domes-
ticus Rutty: A quantitative assessment of female social organisation.
Animal Behaviour, 35, 1846–1857.
Jackson, T. P. (1999). The social organisation and breeding system of
Brants’ whistling rat (Parotomys brantsii). Journal of Zoology, London,
Jarman, P. J. (1974). The social organisation of antelope in relation to their
ecology. Behaviour, 48, 215–267.
Kenward, R. (1987). Wildlife radio tagging. London: Academic Press.
Kingdom, J. (1974). East African mammals. London: Academic Press.
Komers, P. E., & Brotherton, P. N. M. (1997). Female space use is the best
predictor of monogamy in mammals. Proceedings of the Royal Society
of London, 264, 1251–1270.
Ko ¨nig, B. (1993). Maternal investment of communally nursing female
house mice (Mus musculus domesticus). Behavioral Processes, 30, 61–
Ko ¨nig, B. (1994a). Components of lifetime reproductive success in com-
munally and solitarily nursing house mice—A laboratory study. Behav-
ioral Ecology and Sociobiology, 34, 275–283.
Ko ¨nig, B. (1994b). Fitness effects of communal rearing in house mice: The
role of relatedness versus familiarity. Animal Behaviour, 48, 1449–
Korn, H. (1987). Effects of live-trapping and toe-clipping on body weight
of European and African rodent species. Oecologica, 71, 597–600.
leRoux, A., Schelpe, T., & Wahl, Z. (1997). South African wildflower
guide 1: Namaqualand. Cape Town, South Africa: Botanical Society of
Lidicker, W. Z. (1976). Social behaviour and density regulation in house
mice living in large enclosures. Journal of Animal Ecology, 45, 677–
Lott, D. F. (1991). Intraspecific variation in the social systems of wild
vertebrates. New York: Cambridge University Press.
McGuire, B., Getz, L. L., & Oli, M. K. (2002). Fitness consequences of
sociality in prairie voles, Microtus ochrogaster: Influence of group size
and composition. Animal Behaviour, 64, 645–654.
Myers, N., Mittermeier, R. A., Mittermeier, C. G., Fonseca, G. A. B. D., &
Kent, J. (2000, February 24). Biodiversity hotspots for conservation
priorities. Nature, 403, 853–858.
Nel, J. A. J. (1975). Aspects of the social ethology of some Kalahari
rodents. Zeitschrift fu ¨r Tierpsychologie, 37, 322–331.
Ostfeld, R. S. (1990). The ecology of territoriality in small mammals.
Trends in Ecology and Evolution, 12, 411–415.
Ostfeld, R. S., Miller, M. C., & Schnurr, J. (1993). Ear tagging increases
tick (Ixodes dammini) infestation rate of white-footed mice (Peromyscus
leucopus). Journal of Mammalogy, 74, 651–655.
Parker, K. J., Phillips, K. M., & Lee, T. M. (2001). Development of
selective partner preferences in captive male and female meadow voles,
Microtus pennsylvanicus. Animal Behaviour, 61, 1217–1226.
Perrin, M. R. (1980a). The breeding strategies of two co-existing rodents,
Rhabdomys pumilio (Sparrman, 1784) and Otomys irroratus (Brants,
1827). Acta Oecologica, 1, 383–410.
Perrin, M. R. (1980b). The feeding habits of two co-existing rodents,
Rhabdomys pumilio (Sparrman, 1784) and Otomys irroratus (Brants,
1827), in relation to rainfall and reproduction. Acta Oecologica, 1,
Perrin, M. R., Ercoli, C., & Dempster, E. R. (2001). The role of agonistic
behaviour in the population of two syntopic African grassland rodents,
the striped mouse Rhabdomys pumilio (Sparrman, 1784) and the multi-
SCHRADIN AND PILLAY
mammate mouse Mastomys natalensis (A. Smith 1834) (Mammalia Download full-text
Rodentia). Tropical Zoology, 14, 7–29.
Pillay, N. (2002). Father–daughter recognition and inbreeding avoidance
in the striped mouse, Rhabdomys pumilio. Mammalian Biology, 67,
Rambau, R. V., Stanyon, R., & Robinsom, T. J. (2003). Molecular genetics
of Rhabdomys pumilio subspecies boundaries: mtDNA phylogeography
and karyotypic analysis by fluorescence in situ hybridization (FISH).
Molecular Phylogenetics and Evolution, 28, 564–575.
Reburn, C. J., & Wynne-Edwards, K. E. (1999). Hormonal changes in
males of a naturally biparental and a uniparental mammal. Hormones
and Behavior, 35, 163–176.
Reyer, H.-U. (1984). Investment and relatedness: A cost/benefit analysis of
breeding and helping in the pied kingfisher (Ceryle rudis). Animal
Behaviour, 32, 1163–1178.
Roberts, R. L., Williams, J. R., Wang, A. K., & Carter, C. S. (1998).
Cooperative breeding and monogamy in prairie voles: Influence of the
sire and geographical variation. Animal Behaviour, 55, 1131–1140.
Salvioni, M., & Lidicker, W. Z. (1995). Social organisation and space use
in California voles: Seasonal, sexual, and age-specific strategies. Oeco-
logia, 101, 426–438.
Sauer, E. G. F. (1973). Zum Sozialverhalten der Kurzohrigen Elefanten-
spitzmaus, Macroscelides proboscideus [The social behavior of the
round-eared elephant shrew, Macroscelides probiscideus]. Zeitschrift fu ¨r
Sa ¨ugetierkunde, 38, 65–97.
Schradin, C. (2002). [Namaqualand—In the land of mice]. Unpublished
video (for demonstrations at conferences).
Schradin, C. (2003). Whole day follows of the striped mouse. Manuscript
submitted for publication.
Schradin, C. (in press). Territorial defense in a group living solitary
forager: Who, where against whom? Behavioral Ecology and Sociobi-
Schradin, C., & Pillay, N. (2003a). Demography of the striped mouse
(Rhabdomys pumilio) in the succulent karoo: A unique population in an
extreme environment. Manuscript submitted for publication.
Schradin, C., & Pillay, N. (2003b). Extreme social flexibility in the South
African striped mouse: When to live in groups and when to be alone?
Manuscript submitted for publication.
Schradin, C., & Pillay, N. (2003c). The influence of striped mouse fathers
on pup development under different semi-natural environments. Manu-
script submitted for publication.
Schradin, C., & Pillay, N. (2003d). Paternal care in the social and diurnal
striped mouse (Rhabdomys pumilio): Laboratory and field evidence.
Journal of Comparative Psychology, 117, 317–324.
Schradin, C., Reeder, D., Mendoza, S., & Anzenberger, G. (2003). Pro-
lactin and paternal care: Comparison of three species of monogamous
New World monkeys (Callicebus cupreus, Callithrix jacchus, and Cal-
limico goeldii). Journal of Comparative Psychology, 117, 166–175.
Siegel, S., & Castellan, M. J. (1988). Nonparametric statistics for the
behavioral sciences. New York: McGraw-Hill.
Taborsky, M. (1994). Sneakers, satellites, and helpers: Parasitic and coop-
erative behavior in fish reproduction. Advances in the Study of Behavior,
Waterman, J. M. (1995). The social organization of the Cape ground
squirrel (Xerus inauris; Rodentia: Sciuridae). Ethology, 101, 130–147.
Webster, A. B., & Brooks, R. J. (1981). Social behavior of Microtus
pennsylvanicus in relation to seasonal changes in demography. Journal
of Mammalogy, 62, 738–751.
Wilkinson, G. S., & Baker, A. E. M. (1988). Communal nesting among
genetically similar house mice. Ethology, 77, 103–114.
Willan, K. B. R. (1982). Social ecology of Otomys irroparatus, Rhabdomys
pumilio and Mastomys natalensis. Unpublished doctoral dissertation,
University of Natal, Pietermaritzburg, South Africa.
Willan, K., & Meester, J. (1989). Life-history styles of southern African
Mastomys natalensis, Otomys irroratus and Rhabdomys pumilio (Mam-
malia, Rodentia). In M. N. Bruton (Ed.), Alternative life-history styles of
animals (pp. 421–439). Dordrecht, the Netherlands: Kluwer Academic.
Wirminghaus, J. O., & Perrin, M. R. (1993). Seasonal-changes in density,
demography and body-composition of small mammals in a Southern
temperate forest. Journal of Zoology, 229, 303–318.
Wood, M. D., & Slade, N. A. (1990). Comparison of ear-tagging and
toe-clipping in prairie voles, Microtus ochrogaster. Journal of Mammal-
ogy, 71, 252–255.
Received March 24, 2003
Revision received May 24, 2003
Accepted June 1, 2003 ?
SOCIAL SYSTEM OF THE STRIPED MOUSE