on reproduction of young females (Wolff et al. 2002). None
of these studies involved specific tests of hypotheses about
the possible roles of urinary chemosignals mediating pos-
sible delays of puberty or suppression of breeding among
females.
Among actual field investigations, those on house mice
involve the use of highway cloverleaf islands and structural
enclosures. Urine from females in high-density populations
on cloverleaf islands successfully delayed the onset of pu-
berty in females maintained in the laboratory (Massey and
Vandenbergh 1980). Urine from females in low-density
populations housed at cloverleaf islands did not alter the on-
set of puberty. When female density was artificially inflated
by adding females to cloverleaf islands, urine collected from
resident females was capable of delaying puberty in young
females maintained in the laboratory (Coppola and Van-
denbergh 1987). Male urine collected in both high- and
low-density conditions successfully accelerated puberty in
female mice maintained in the laboratory (Massey and Van-
denbergh 1981). These studies demonstrate that the urinary
chemosignals that accelerate or delay puberty are produced
in wild house mice living under natural conditions. They
cannot, however, be used to indicate the possible popula-
tion consequences of these signals.
A series of 0.1 ha outdoor enclosures was used to test
both acceleration and delay chemosignals in a seminatural
setting using introduced populations of house mice (Drick-
amer and Mikesic 1990). Urine from mice maintained in
the laboratory was sprayed weekly on food sources and nest
sites. Two enclosures were treated with urine from grouped
females, two were treated with adult male urine, and two
enclosures were treated with water as a control. Popula-
tions in the enclosures with male urine added attained
higher densities, exhibited earlier puberty for females and
more reproduction by females than mice in the control en-
closures. In contrast, mice in the enclosures treated with
urine collected from grouped females exhibited less repro-
duction, lower overall growth and densities, and less repro-
duction than mice in the control enclosures. In this instance,
field effects on puberty and reproduction were recorded,
though the urine supplies used to treat the different popu-
lations came from wild mice housed in the laboratory. Thus
these two sets of investigations provide strong evidence for
urinary chemosignal effects on reproduction in wild house
mice, but no study has yet been conducted that involves
both donors and recipients in a field setting.
Seminatural enclosures have been used effectively to
study numerous aspects of rodent behavioral ecology with
mixed results when compared with laboratory studies (see
the following). Similar enclosures could be used to test the
acceleration and delay effects of chemosignals in a field set-
ting in vole species. In an open 10-ha grassland, using ex-
perimental manipulations involving removing adults of one
sex or the other, Rodd and Boonstra (1988) found that
adult females play a critical role in terms of delaying the
sexual maturation of young meadow voles at higher densi-
ties. In contrast, experimental manipulations of gray-tailed
voles in which adult sex ratios were skewed toward males
or females in large 0.2 ha enclosures produced only mini-
mal and short-term (one week) effects of delayed sexual
maturation (Wolff et al. 2002).Laboratory Phenomena —Natural Phenomena?One question that arises after considering the numerous
laboratory studies and the available field-test data concerns
whether some of the observed effects of urinary chemosig-
nals on aspects of reproduction in rodents are laboratory
artifacts (table 9.1). Do these phenomena translate to natu-
ral settings? How many of the effects are characteristic of
natural populations of rodents?
Drickamer (1989) tested whether female behavior might
influence pregnancy termination (the Bruce effect) in house
mice. Wild female house mice were given choices between
soiled bedding containing odors from the stud male and a
strange male. For the first half of the 19 –20 day gestation
period the inseminated females exhibited a strong prefer-
ence for the stud male, avoiding the odor of a strange male
and the possible negative consequences. These preferences
were absent in the second half of gestation, but by then,
the embryos had implanted and fetuses were developing.
This short study suggests that, given a chance under natu-
ral rather than caged conditions, recently inseminated fe-
males would choose to associate with the area of the stud
male and could, thereby, avoid any possible pregnancy ter-
mination effects from a strange male.
More recently, Wolff (2003c) summarized studies on
several rodent species that suggest that further testing of the
chemosignal effects on rodent reproduction is warranted
before we jump to conclusions about laboratory phenom-
ena and their possible importance in natural populations.
Field tests of the Bruce effect conducted in large seminatural
enclosures failed to provide any evidence of pregnancy dis-
ruption in gray-tailed voles (de la Maza et al. 1999) or prai-
rie voles (Mahady and Wolff 2002). An interesting side note
to these two studies is that the gray-tailed vole has a pro-
miscuous mating system, while the prairie vole has a mo-
nogamous mating system.
One hypothesis about suppression or delays in repro-
duction (phenomena similar to the Whitten effect) involves
negative influences from the presence of predators. The110 Chapter Nine