Costs of reproduction measured under laboratory conditions may have li
ttle bearing on actual costs observed in the field, because both envir
onmental conditions and behavior can modify the manifestation of these
costs. I studied survival, growth, and reproduction in a wild populat
ion of southern water skinks (Eulamprus tympanum) over 3 yr to measure
costs of reproduction in females, to examine how the environment infl
uences these costs, and to assess the extent to which females are able
to reduce costs. The proportion of mature females that reproduced eac
h year varied from 30-60%. Frequent reproduction appeared costly in te
rms of both survival rates and growth. Growth rates and survival rates
were negatively correlated with reproduction during the 3-yr study. I
n the coolest and wettest year of the study, when food intake by femal
es was relatively low, gravid females and females that had reproduced
in the previous year survived less well than nonreproducing females. T
here was no evidence of a survival cost of reproduction in the other t
wo years of the study. Reproduction also imposed a fecundity cost, as
overall, mean growth rates of reproducing females were reduced relativ
e to those of nonreproducing females of the same body size. Reduced gr
owth rate translates to reduced litter size in these skinks, because f
ecundity increases with body size. These results suggest that reproduc
tion can be costly in terms of both survival and future reproduction i
n these skinks, but that survival costs are not consistent from year t
o year and are mediated by environmental conditions. Females appear to
adjust their frequency of reproduction to minimize these costs, as th
ose females that skipped opportunities for reproduction were those exp
ected to gain the greatest growth and fecundity benefit by skipping.