Several recent studies using natural populations of Drosophila show that on
e must be very careful when sorting among the targe number of molecular pol
ymorphisms found at most loci to identify the nucleotide changes responsibl
e far phenotypic variation in complex traits. Indeed, several mutations wit
hin a single allele can interact to generate the overall observed effect. T
he results are instructive both for those interested in the genetics of evo
lutionary change and for those attempting to ferret out the genetic basis o
f complex human diseases.