not accelerate reproduction during winter and urine from
grouped females does not delay reproduction during sum-
mer. Taken together, these studies suggest that for female
house mice, variation in the timing of puberty and repro-
duction is strongly influenced by a suite of factors that con-
vey information about the adequacy of conditions for repro-
duction. If, as seems probable from the field tests on house
mice, there are fitness consequences stemming from the in-
fluence of urinary chemosignals, then investigations of the
underlying mechanisms are warranted.
A possible adaptive significance of the Bruce effect has
been proposed in several reviews (e.g., Schwagmeyer 1979;
Labov 1981b; Storey 1986a). The initial idea was that ter-
minating a pregnancy because of the presence of a strange
male (who has presumably displaced the stud male) is ad-
vantageous for the female in that she does not carry a lit-
ter to term that could be destroyed by infanticide. This may
be true for some species. However, another idea that has
gained credence is that pregnancy termination is a form
of postfertilization mate selection (Dawkins 1976). Females
could gain in terms of the fitness of progeny if they mate
with a more aggressive and dominant male (the one who has
displaced the stud male). This hypothesis may have some
validity for some species with early termination of preg-
nancy, but for vole species, in which the Bruce effect can oc-
cur as late as day 17 of gestation, loss of a pregnancy would
be a huge cost to the female. The processes involved in preg-
nancy termination can also be, in part, functions of other
factors, such as the aggressiveness of both the female and
the male(s) involved (Storey 1990). Alternatively, if the two
field studies to test this phenomenon are correct and the
Bruce effect does not occur in natural populations, then hy-
pothesizing some functional significance to this response is
superfluous. As de la Maza et al. (1999) and Mahady and
Wolff (2002) argued, a rodent with a short lifespan can-
not afford to waste pregnancies every time a strange male
passes by.
It is noteworthy that pregnancy termination in house
mice occurs only in the period prior to implantation of the
fetuses. Since implantation occurs by about day 5 or 6 of
the 19 –20 day gestation period, it is evident that there is a
roughly 67–75% chance that the embryos have already im-
planted if a male takeover effect occurs at random with re-
spect to the stage of gestation for that female. It would be
quite interesting to determine, for those species of small ro-
dents where there is a process by which males replace each
other in terms of holding particular spatial areas, whether
these shifts in males take place randomly or within particu-
lar periods of time after insemination. Further, one could
postulate that in those species that are polygynous, there
is some likelihood that a male will depart an area in search
of additional mates after a given female is inseminated. In
most species of voles, males wander and do not guard their
mates; consequently, strange males are not deterred from
entering females’ territories. It seems unlikely that females
in these species would abort pregnancies following expo-
sure to strange males. The association between mate guard-
ing and the Bruce effect warrants further theoretical and
empirical testing.Future Explorations of Rodent Chemosignals
and ReproductionHaving examined what is known about three general types
of urinary chemosignals influencing rodent reproduction, I
now provide some thoughts and speculations on what sorts
of investigations should occur in the future. What types of
findings will enable us to better understand the ecological
and mechanistic aspects of these phenomena?Comparative approach
Much of what we know in detail concerns only the house
mouse, with bits and pieces of information from several
other taxa, primarily Microtusand Peromyscus.Even for
house mice, the complete picture regarding release and pos-
sible effects of urinary chemosignals on puberty and repro-
duction and /or how a reproductive response affects lifetime
reproductive success has not been obtained. No tests have
been conducted under any type of field conditions of uri-
nary chemosignals from estrous, pregnant, or lactating fe-
males, all of which accelerate puberty in laboratory tests. We
have no real information on the possible occurrence of the
Bruce effect under natural conditions. In fact, no compara-
tive studies have been conducted of commensal house mice
that have been living in association with humans in dwell-
ings and other human-made structures, or with feral house
mice that are free-living in fields and other natural settings.
All of these areas need further investigation in the species
in which we know the most about the urinary chemosignals
that influence reproduction. Gaps in our current under-
standing of these phenomena are obvious from table 9.1.
An important avenue of research would be to systemati-
cally test the effects of several urinary chemosignals on re-
production under a set of standard conditions in a labora-
tory, and then in natural settings consistent with the normal
socioecology of each species. This experimental regime
could be accomplished with a range of vole or deer mouse
taxa. Large-scale studies involving several species with dif-
ferent social systems and from a variety of ecological set-
tings is necessary to provide sufficient data to test the vari-112 Chapter Nine