Genetic and developmental constraints have often been invoked to explain patterns of existing morphologies. Yet, empirical tests addressing this issue directly are still scarce. We here set out to investigate the importance of maternal body size as an evolutionary constraint on egg size in the tropical butterfly Bicyclus anynana, employing an artificial two-trait selection experiment on simultaneous changes in body and egg size (synergistic and antagonistic selection). Selection on maternal body size and egg size was successful in both the synergistic and the antagonistic selection direction. Yet, responses to selection and realized heritabilities varied across selection regimes: the most extreme values for pupal mass were found in the synergistic selection directions, whereas in the antagonistic selection direction realized heritabilities were low and nonsignificant in three of four cases. In contrast, for egg size the highest values were obtained in the lines selected for low pupal mass. Thus, selection on body size yielded a stronger correlated response in egg size than vice versa, which is likely to bias (i.e., constrain), if weakly, evolutionary change in body size. However, it seems questionable whether this will prevent evolution toward novel phenotypes, given enough time and that natural selection is strong. Correlated responses to selection were overall weak. Egg and larval development times tended to be associated with changes in maternal size, whereas variation in pupal development times weakly tended to follow variation in egg size. Lifetime fecundity was similar across selection regimes, except for females simultaneously selected for large body mass and small egg size, exhibiting increased fecundity. Multiple regressions showed that lifetime fecundity and concomitantly reproductive investment were primarily determined by longevity, as expected for an income breeder, whereas egg size was primarily determined by pupal mass. Evidence for a phenotypic trade-off between egg size and number was weak.