EVIDENCE THAT THE MAGNITUDE OF THE TRADE-OFF IN A DICHOTOMOUS TRAIT IS FREQUENCY DEPENDENT.

Many traits, such as wing dimorphism, paedomorphosis, and cyclomorphosis 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 analyzed the evolution of dichotomous traits, assuming that each morph can be characterized by a discrete syndrome of characters. Consideration of the genetic basis of dimorphism suggests that this assumption is incorrect. In this paper, I report a test of this assumption using the wing-dimorphic sand cricket, Gryllus firmus. It is predicted that, rather than remaining constant, the fecundity of macropterous (long-winged, flight-capable) females will decrease as the proportion of macropterous females in the population or family increases. This prediction is supported by fecundity data from lines selected for high and low proportions of macroptery and by sib analysis. Thus, models that seek to predict the evolution of dichotomous traits should take into account the likelihood that values of components of fitness, such as fecundity, may be related to their frequencies in the population.