represent offspring that are closer to adult body size at birth
and at the end of the major period of maternal investment,
and these species might be considered more precocial, as-
suming that the degree of development correlates with body
mass of offspring. Using this assumption, we applied corre-
lation analysis to ask if the degree of precociality was asso-
ciated with principal components of life-history variables.
Another variable that might reflect an aspect of the “shape”
of life histories is the length of the pre-adult period com-
pared to the length of life. We estimated this by taking the
ratio of age at maturity to reproductive lifespan (viz., a/v),
and compared this ratio to principal components of the life
history. Ratios that reflected the degree of development of
offspring and the timing of reproductive onset within life
cycles were compared to other variables using Spearman
rank correlations.
Results
We analyzed life-history data for 43 populations from
29 species and 9 families of rodents. Sciuridae was the most
represented family (26 populations), followed by Hetero-
myidae and Muridae (4 populations each), and other fami-
lies were represented by 1 or 2 populations. Our sample
included species from a wide range of body masses (0.019 –
23.8 kg) and life-history characteristics (Oli and Dobson
2003, Appendix). So our sample, although not large, in-
cluded rodents with diverse patterns of life history and
population dynamics.
The pattern of life histories of a population are well-
described by five variables: a, v, PJ, PA, and m. We used
fecundity (m) for this analysis rather than fertility (F), be-
cause the latter incorporates the influence of adult survival.
These life-history traits all varied significantly with mean
adult body mass, with between about 20% and 40% of the
variation in life histories statistically accounted for by body
mass (table 8.1). We used the family taxonomic level as
classes in an analysis of variance to estimate the potential
influence of phylogeny on the life-history traits and body
mass. Phylogeny could account for about 60% to 90% of
the variation in life histories, and 90% of the variation in
body mass. The latter result indicates that body mass and
phylogeny were strongly confounded for the rodent pop-
ulations that we examined. Thus we adjusted life-history
traits statistically for both body mass and phylogeny using
the general linear model, and did not consider these influ-
ences separately in further analyses. Combined influences
of body mass and phylogeny were not greatly different from
the same analysis with phylogeny alone, explaining be-
tween about 60% and 90% of the variation among life-
history variables.
We looked for evidence of a fast-slow continuum among
the life-history traits (transformed, but not adjusted for
body mass and phylogeny), using a principal component
analysis (table 8.2). The first two principal component (PC)
axes statistically explained 84% and another 8% of the
variation in traits, respectively. The first PC had strong pos-
itive loadings for a, v, PJ, and PA, reflecting a fast-slow con-
tinuum. Thus timing variables and survival increased or de-
creased together along a fast-slow continuum. A trade-off
was evident with reproduction, as reflected by a negative
PC1 loading for m. The second PC axis reflected changes
in juvenile survival that were somewhat associated with
changes in mand a trade-off with adult survival. The same
general pattern was evident when adult body mass was in-
cluded in the analysis, except that body mass was also
strongly associated with the fast-slow continuum reflected
by the first PC axis. The first three PC axes explained 77%,
another 11%, and another 7%, respectively, of the varia-
tion in life-history traits and body mass. The second PC
axis reflected changes in body mass that were statistically
independent of PC1, and changes in juvenile survival and its
associations with fecundity and adult survival occurred onFast and Slow Life Histories of Rodents 101Table 8.1 Regressions (for the influence of adult body mass), analyses of variance (for the influence of family-level phylogeny),
and analyses of covariance using the general linear model (for the influence of body mass and phylogeny)
Influence of adult body mass Influence of family-level phylogeny Influence of mass and familyTrait R^2 FP R^2 FP R^2 FP
a 0.403 27.6 0.0001 0.816 18.9 0.0001 0.845 20.0 0.0001
v 0.395 26.7 0.0001 0.805 17.5 0.0001 0.807 15.4 0.0001
PJ 0.197 10.0 0.003 0.585 6.0 0.0001 0.590 5.2 0.0003
PA 0.284 16.2 0.0002 0.879 30.9 0.0001 0.902 33.9 0.0001
m 0.180 9.0 0.005 0.818 19.1 0.0001 0.839 19.1 0.0001
Mass 0.897 37.1 0.0001
NOTE: Sample size in all cases is 43 populations.