Sperm size varies enormously among species, but the reasons for this v
ariation remain obscure. Since it has been suggested that swimming vel
ocity increases with sperm length, earlier studies proposed longer (an
d therefore faster) sperm are advantageous under conditions of intense
sperm competition. Nonetheless, previous work has been equivocal, per
haps because the intensity of sperm competition was measured indirectl
y. DNA profiling now provides a more direct measure of the number of o
ffspring sired by extrapair males, and thus a more direct method of as
sessing the potential for sperm competition. Using a sample of 21 spec
ies of passerine birds for which DNA profiling data were available, we
found a positive relation between sperm length and the degree of extr
apair paternity. A path analysis, however, revealed that this relation
ship arises only indirectly through the positive relationship between
the rate of extrapair paternity and length of sperm storage tubules (S
STs) in the female. As sperm length is correlated positively with SST
length, an increase in the intensity of sperm competition leads to an
increase in sperm length only through its effect on SST length. Why fe
males vary SST length with the intensity of sperm competition is not c
lear, but one possibility is that it increases female control over how
sperm are used in fertilization. Males, in turn, may respond on an ev
olutionary time scale to changes in SST size by increasing sperm lengt
h to prevent displacement from rival sperm. Previous theoretical analy
ses predicting that sperm size should decrease as sperm competition be
comes more intense were not supported by our findings. We suggest that
future models of sperm-size evolution consider not only the role of s
perm competition, but also how female control and manipulation of ejac
ulates after insemination selects for different sperm morphologies.