dition or ability to invest in offspring was gauged as a cor-
relate of maternal mass or age (McShea and Madison 1986;
Armitage 1987a; Krackow 1997), immediate reproductive
history or current litter composition in which male off-
spring were either assumed to be or were demonstrably
more costly to produce than daughters (Gosling 1986;
Krackow and Hoeck 1989; Clark et al. 1990; Millesi et al.
1999), or where resources available to the mother were al-
tered in various ways including direct food restriction (Riv-
ers and Crawford 1974; McClure 1981; Labov et al. 1986;
Meikle and Drickamer 1986; König 1989b; Meikle and
Thornton 1995; Sikes 1995, 1996a, 1996b; Moses et al.
1998; Wright et al 1988; Lamb and Aarde 2001), food aug-
mentation (Duquette and Millar 1998, Trombulak 1991),
density reduction to reduce competition (Wright et al.
1988), or separation of young (Clark et al. 1991). Fifteen of
these 23 studies report either skewed sex ratios or differen-
tial investment in offspring, and hence are consistent with
the predictions of the model. The remaining eight show no
evidence of differential investment. Of the 15 studies con-
sistent with predictions, only one (Gosling 1986) meets
the necessary second criterion of demonstrating facultative
adjustment. Consequently, the patterns of differential mor-
tality or biased growth in the remaining 14 studies are
equally consistent with nonfacultative explanations to be
discussed later. As a number of these studies are of partic-
ular importance in evaluating the likelihood of Trivers and
Willard-like manipulations, I will consider them in greater
detail.
McClure’s (1981) study of eastern woodrats (Neotoma
floridana), in which she reported differential survival of
nursing male offspring, is often cited as the only clear ex-
ample of sex-biased allocation leading to differential mor-
tality (Clutton-Brock 1991). This study is potentially very
important because the experimental treatment (food re-
striction) was imposed postnatally, so the resulting biased
sex ratios presumably resulted from postnatal maternal ad-
justment. Further, the sex ratio bias was coupled with a re-
duction in litter size in an apparent attempt by the mother
to reduce overall cost of reproduction. In an effort to rep-
licate McClure’s (1981) findings, Sikes (1995, 1996b) fol-
lowed the same experimental protocol of postnatal food
restriction with eastern woodrats and with northern grass-
hopper mice (Onychomys leucogaster), but obtained dis-
similar results. In neither of the studies by Sikes was there
any evidence of sex-biased mortality or growth rates as a
function of food restriction. Further, although these stud-
ies validated the first assumption of the T-W hypothesis
(that offspring condition at weaning is dependent on ma-
ternal condition during parental investment), a follow-up
study (Sikes 1996a) showed that the compromised offspring
condition produced by maternal food restriction was not
maintained after the young were able to feed independently.
McClure’s results also prompted an exceptionally well-
done study of bushy-tailed woodrats (Neotoma cinerea) by
Richard Moses and colleagues. The major criticism of Mc-
Clure’s work by Moses et al. (1998) was that, although she
stated that male offspring of restricted mothers were often
outside the nest and not nursing, no direct evidence was
provided that their deaths were due to facultative adjust-
ment. This same point had been made previously by vari-
ous authors concerning tests of the T-W hypothesis (Myers
1978; Charnov 1982; Clutton-Brock 1991). To address this
shortcoming, Moses et al. (1998) followed an experimen-
tal protocol of postnatal food restriction similar to that of
McClure. Although these authors documented male-biased
mortality among offspring of food-restricted mothers, as
reported by McClure (1981), when they sought evidence
that these differences were due to facultative adjustment by
mothers, they found no hint of support. Instead, Moses
et al. (1998) concluded that male-biased mortality occurred
in the absence of maternal intervention and was most likely
due to the greater energetic demand of larger male offspring
in this sexually dimorphic species. Similar circumstances
may explain sex-biased mortality in other sexually dimor-
phic species, such as horses (Equus caballus;Monard et al.
1997). Based on their results, Moses et al. concluded that
their study failed to support the T-W hypothesis.
The descriptive study by Gosling (1986) offers the best
support for the T-W hypothesis in rodents. This was a long-
term study involving dissection of 5,853 female coypus
(Myocastor coypus) from control operations in England.
Of the sample, nearly 1,500 females contained litters old
enough to sex and were included in subsequent analyses.
Gosling documented sex-biased investment in that small lit-
ters were significantly more likely to be aborted if they were
female biased than if male biased. Abortion was of entire
litters, and clearly was facultative rather than simply a re-
sponse to litter size, because male-biased litters tended to
be retained even if small. To be sure, there was a tendency
toward male-biased litters among younger mothers and
among those mothers giving birth in the summer, but ow-
ing to the power afforded by an enormous sample size, age
and season could be excluded as potential explanatory vari-
ables; Gosling concluded that biased investment in subse-
quent litters was indeed facultative and under maternal con-
trol. Gosling’s documentation of facultative adjustment of
maternal investment to favor one sex over the other makes
this the only study of rodents to satisfy both criteria cen-
tral to the T-W hypothesis. As has been pointed out re-
peatedly (Myers 1978; Charnov 1982; Clutton-Brock 1991;
Moses et al. 1998), without solid evidence of facultativeFacultative Sex Ratio Adjustment 131