ian species of small and large body sizes occur. Within the
mammals, rodents are a suitable group of species for an ex-
amination of the fast-slow continuum, in part because they
are primarily herbivores and thus share aspects of their life-
styles. In addition, rodents span only a part of the range of
body sizes of mammals. Thus patterns of life-history traits
that follow the fast-slow continuum should be particularly
evident, with or without statistically controlling possible in-
fluences of body size.
The life cycle of rodents can be adequately approximated
by parameters of partial life-cycle models (Caswell 2001;
Oli and Zinner 2001): juvenile and adult survival (PJand
PA), age at maturity and at the end of the reproductive life-
span (v), and fertility. These key elements of life history
encapsulate the most salient aspects of the life cycle, and
they can be used to predict population growth or evaluate
whether a population is in demographic equilibrium. We
estimated these key life-history traits using data from pop-
ulations of rodents. We followed the approach of Gaillard
et al. (1989) in applying principal component analysis
(PCA) to life-history traits, and examined the outcome for
evidence of a fast-slow continuum. Furthermore, we re-
moved confounded influences of body size and phylogeny
from our data, and again looked for evidence of a fast-
slow continuum. Gaillard et al.’s (1989) analyses suggested
that after body size and the fast-slow continuum, a tertiary
mammalian life-history pattern occurred between altricial
and precocial species. We also looked for evidence of the al-
tricial /precocial pattern of life-history traits among rodent
populations.
Methods
Life-history data and demographic analyses
We compiled life-history data from published sources for
43 populations of 29 species of rodents (Oli and Dobson
2003, Appendix). The database includes recently published
data on Idaho ground squirrels (Spermophilus brunneus
brunneus;Sherman and Runge 2002). For statistical anal-
yses, we transformed variables to improve their fit to a
normal distribution (after Gaillard et al. 1989). Body mass
and fecundity (m) were log-transformed, age at matu-
rity (a), and reproductive life span (v) were calculated in
months and log-transformed. Survival probabilities (PJand
PA) were calculated as monthly rates and square root-arcsin
transformed.
We evaluated the influence of body size by regressing life-
history variables on log-transformed body mass. Residuals
were then analyzed to investigate whether any discernible
patterns in these variables persisted after statistical removal
of body-size effects. We used one-way ANOVA with family
as the main effect to evaluate the effects of phylogeny on
life-history parameters. Alternative methods of evaluating
phylogenetic influences on life-history traits require a well-
corroborated phylogenetic tree, and this was not available
for the Rodentia. We chose to limit our analyses to family
because a substantial proportion of variation in mammal-
ian life-history traits occurs among families (Stearns 1983;
Promislow and Harvey 1990). Residuals were then ana-
lyzed to examine whether the pattern of covariation among
life-history variables persisted after statistical removal of
phylogenetic effects. It is possible that both body size and
phylogeny act synergistically in their effects on life-history
variables. To test this possibility, we simultaneously re-
moved their effects using ANCOVA. Again, we analyzed
residuals to determine whether patterns of covariation of
life-history variables persisted.Quantification of fast-slow continuum
We used three methods for quantifying the fast-slow con-
tinuum in rodent life histories. First, we examined loadings
of life-history variables on the first principal component
(PC1) of a principal components analysis. Relatively high
even loadings for longevity and survival variables would re-
flect the fast-slow continuum. Next, we computed the ratio
of fertility rate to age of first reproduction (F/aratio; after
Oli and Dobson 2003). The F/aratio reflects the magnitude
of reproduction compared to age at maturity, but it also in-
cludes the influence of adult survival (i.e., FPA*m,where
mis fecundity; Oli and Dobson 2005). Because of this, we
also used a similar index, the ratio of fecundity to age at
maturity (viz., m/a). Third, we quantified the fast-slow con-
tinuum based on adult mortality rate. Read and Harvey
(1989) argued that animals at the fast end of the contin-
uum have high mortality rate for their body mass. Thus
we examined the correlation between adult mortality rate
and other life-history variables, after statistically removing
the effect of body mass. We tested the adequacy of differ-
ent measures of the fast-slow continuum for congruence by
looking for significant associations among them.Quantification of precociality
Data on adult female mass and weaning mass were com-
piled from species accounts (Mammalian Species Accounts,
American Society of Mammalogists), and additional body
mass data were compiled from studies reporting demo-
graphic data (Oli and Dobson 2003, Appendix) or from
Silva and Downing (1995). We used the ratios of neonatal
mass and of individual offspring mass at weaning to adult
female body mass as measures of precociality. Larger ratios100 Chapter Eight