with regard to their feeding habits, and are generally will-
ing to try new foods (Southern 1954; Meehan 1984); how-
ever, under certain conditions they appear to avoid novel
foods (Misslin 1982; Humphries et al. 2000). Neophilia,
like neophobia, incurs costs and benefits. Benefits could in-
clude the ability to find new foods readily, which would en-
hance the capacity to exploit new environments. Neverthe-
less the extent of the difference with rats remains somewhat
of a puzzle.
Aggression and social status
Rats and mice differ in the ways that agonistic interactions
are used to establish and maintain social relations, partly
reflecting the greater range of the rats’ behavioral reper-
toire. Overall, both male rats and mice display more ag-
gression than females, although rats appear to be able to
establish dominance hierarchies within their living groups
that perpetuate themselves despite changes in relative phys-
ical prowess of males in that group. In contrast, male mice
constantly compete for dominance among each other. Com-
paratively few data exist for aggression in female rats, al-
though aggression in female mice is well documented in a
variety of contexts.
Most studies on aggression have been conducted in the
laboratory, although observations of wild rats confirm that
aggression, while not very common, is highest among males
(see descriptive drawings, including “play fighting,” in Pel-
lis and Pellis 1987; and Berdoy 2002 for film footage of ag-
onistic interactions). Agonistic interactions between sexes
and between females are relatively infrequent; a dominance
hierarchy among females is less readily apparent (Calhoun
1962a; Robitaille and Bovet 1976; Adams and Boice 1983,
1989; Berdoy et al. 1995; for laboratory rats see also Blan-
chard and Blanchard 1977, Rampaud 1984; and Blanchard
et al. 1988).
Predictably, larger rats are more likely to win contests
upon first encounter with smaller individuals. However,
observations of wild rats in stable groupsalso show stable
linear dominance hierarchies among males, where social
status can become “settled” as the outcome of previous en-
counters and determines the outcome of future ones (Ber-
doy et al. 1995; see also Smith et al. 1994 for laboratory
rats). In these conditions, age becomes a better predictor of
social status than weight, as dominant individuals maintain
their social status even though initial body weight asymme-
tries with younger (and at the time lighter) individuals have
disappeared or been reversed. It is noteworthy that the exis-
tence of such “settled dominance” (Berdoy et al. 1995) may
mitigate against an easy demonstration of the role of com-
petitive abilities in maintaining social status because, once
dominance has been established, changes in competitive
abilities may not be immediately reflected by a change in so-
cial status. This relationship may explain why, even in lab-
oratory settings, the evidence that dominance is correlated
with body weight has been conflicting, with only about half
the studies reporting a correlation (reviewed in Berdoy et al.
1995). It is not known to what extent such settled domi-
nance is also found in rats living at low densities (and there-
fore with less frequent contacts), but its existence among
wild rats in stable groups raises the question as to why
young individuals with greater competitive abilities should
accept the status quo. First, the benefits of living in groups
may outweigh the cost of being relegated to a subordinate
position, particularly if the costs of escalated aggression are
great relative to the value of the contested resource. Second,
as a consequence of some of the features of rat mating be-
havior (see the following), dominant males often cannot mo-
nopolize access to estrous females or to food resources (Cal-
houn 1962a; Berdoy and Macdonald 1991; Berdoy 1994;
Berdoy et al. 1995). Finally, dominance hierarchies, while
settled, are not static, and can change substantially between
six monthly periods, thus allowing competitive subordi-
nates to rise to prominent social status eventually (M. Ber-
doy, unpublished data). Clearly, the relative tolerance of
such social dynamics allows rats to live at high densities.
Aggression has also been studied in a variety of strains
of house mice, including wild house mice (Lagerspetz 1964;
Brain et al. 1989). Similar to rat aggression tendencies,
male-male aggression is important for the establishment
and maintenance of territories, in some instances for domi-Comparative Social Organization and Life History of Rattusand Mus 385Figure 32.4 Neophobia: grain consumption (in g) of a “stable” enclosed col-
ony of wild rats at familiar and novel containers (but same food) introduced at
day 0 (X axis). The numbers of rats were constant throughout the observations
and the containers were accessible in equal measure to all colony rats (thus
controlling for possible spatial factors). The dotted line indicates the average
amount expected to be consumed if both types of containers were used in
equal measure. Note the lack of consumption for the first 10 days, the relatively
rapid change over the next three days (days 11–13) and the still very reduced
(although slowly increasing) consumption after about 1 month (Berdoy, unpub-
lished data).