Rodent Societies: An Ecological & Evolutionary Perspective

the third PC axis. For both of the previous PCAs, however,
only the first principal component explained substantial
variation in the life-history data.
We conducted a principal component analysis for life-
history traits that were statistically adjusted for influences
of body mass and family-level phylogeny (table 8.2). The
first three PC axes statistically explained 42%, another
23%, and another 20%, respectively, of the variation in life-
history traits. Loadings on the first PC axis revealed strong
positive loadings for a, v, PJ, and PA, reflecting a fast-slow
continuum. A trade-off of the continuum and reproduction,
however, was not evident. Rather, variation in fecundity
was reflected by a high positive loading for mon the second
PC axis. The third PC axis reflected changes in adult sur-
vival that were somewhat associated with changes in m,
and a trade-off with juvenile survival. From the statistical
definition of principal components, patterns on the second

and third PC axes were statistically independent of the fast-

slow continuum as reflected by the first PC.

When life-history data were not adjusted for body size

and phylogeny, scores on the first PC were strongly associ-

ated with some possible measures (e.g., F/aand m/a) of the

fast-slow continuum (table 8.3), but in a negative direction.

Thus as PC1 scores increased and F/aand m/adecreased,

life histories moved from fast to slow. Scores on PC1 were

also strongly positively associated with adult mortality, as

reflected by PA. Mortality equaled 1 minus PA, so as mor-

tality increased, life histories moved from slow to fast. Sur-

prisingly, the strongest single association with PC1 was

with the ratio of PA/a. Scores on the first PC were signifi-

cantly negatively associated with our indices of precociality,

with precocial species more likely to fall at the fast end of

the continuum. For this PCA, the scores on the second PC

showed no significant associations.

102 Chapter Eight

Table 8.2 Factor loadings from principal component analyses

Adjusted for adult body mass

Unadjusted Unadjusted and family-level phylogeny

Traits PC1 PC2 PC1 PC2 PC3 PC1 PC2 PC3

a 0.46 0.17 0.45 0.02 0.16 0.58 0.28 0.08

v 0.46 0.03 0.44 0.01 0.04 0.56 0.28 0.06

PJ 0.42 0.76 0.39 0.20 0.79 0.50 0.28 0.45

PA 0.46 0.36 0.43 0.17 0.35 0.28 0.16 0.85

m 0.44 0.50 0.41 0.35 0.46 0.11 0.86 0.26

Mass 0.32 0.90 0.07

NOTE: Sample size in all cases is 43 populations.

Table 8.3 Spearman rank correlations between principal component scores that represent the fast-slow continuum

(viz., PC1), some possible proxies for the continuum, and variables that reflect developmental patterns

Adjusted for adult body mass

Unadjusted and family-level phylogeny

Index PC1 PC2 PC1 PC2 PC3

F/a 0.763*** 0.191 0.244 0.379 0.064

m/a 0.773*** 0.188 0.257 0.367 0.0048

PA 0.879*** 0.184 0.303 0.003 0.466*

PA/a 0.928*** 0.009 0.365 0.080 0.414*

Neonate mass /

Adult mass 0.746** 0.108 0.046 0.261 0.057

Weaning mass /

Adult mass 0.589** 0.137 0.071 0.009 0.049

a/v 0.102 0.367 0.129 0.449* 0.048

NOTES: Sample size was 27 for the ratios of neonate mass and of weaning mass to adult body mass; it was 43 for all other

variables.

***P 0.0001

**P 0.001

*P 0.01.