Considerable attention has recently been focused on using levels of develop
mental instability among members of a population to detect environmental or
genetic stresses on animals or plants. It is not yet clear, however, that
high developmental instability in a sample of individuals always indicates
environmental stress or poor genetic quality. We studied 13 fragmented popu
lations of prairie phlox (Phlox pilosa L.) to test the hypothesis that deve
lopmental instability should decrease with increasing population size - as
expected if small populations suffer genetic problems associated with inbre
eding or are exposed to more environmental stress than larger populations.
We used two different measures of developmental instability, each calculate
d for two different traits: radial asymmetry of flowers (for petal width an
d petal length) and modular fluctuating asymmetry of leaves (for leaf width
s at two points along the leaf). There were weak but significant correlatio
ns among individuals for four of six pairwise combinations of these measure
s. Surprisingly, three of our four measures of developmental instability sh
owed strong population size effects that were opposite to those expected: d
evelopmental instability increased with population size. We conclude that m
easures of developmental instability cannot be applied uncritically for bio
monitoring without considerable knowledge of developmental mechanisms, natu
ral history, and population biology of the species in question.