Many traits, such as wing dimorphism, paedomorphosis, and cyclomorphos
is vary dichotomously. Such dimorphisms are maintained in part because
of a trade-off between components of fitness: for example, in insects
, the flightless morph cannot migrate but has a greater fecundity than
the flight-capable morph. Several recent theoretical studies have ana
lyzed the evolution of dichotomous traits, assuming that each morph ca
n be characterized by a discrete syndrome of characters. Consideration
of the genetic basis of dimorphism suggests that this assumption is i
ncorrect. In this paper, I report a test of this assumption using the
wing-dimorphic sand cricket, Gryllus firmus. It is predicted that, rat
her than remaining constant, the fecundity of macropterous (long-winge
d, flight-capable) females will decrease as the proportion of macropte
rous females in the population or family increases. This prediction is
supported by fecundity data from lines selected for high and low prop
ortions of macroptery and by sib analysis. Thus, models that seek to p
redict the evolution of dichotomous traits should take into account th
e likelihood that values of components of fitness, such as fecundity,
may be related to their frequencies in the population.