The larvae of numerous sessile and several motile marine invertebrate species respond to metamorphosis-stimulating cues, usually chemical signals from their benthic habitat. In three series of experiments (IIII), the megalopa larvae of a semi-terrestrial mangrove crab, Sesarma curacaoense de Man, were exposed for differential periods to adult-conditioned water (ACW), which had previously been found an effective cue. In the first experimental series (Exp. I), ACW was added on successively later days of the moulting cycle, while Exp. II comprised treatments in which the cues were provided from the first day (postmoult) and removed on sucessively later days. In Exp. III, the megalopae were exposed to ACW for 3 days at the beginning, in the middle, or near the end of the moulting cycle (days 13, 46, and 79, respectively). All experimental series also comprised a permanently exposed control group (C+) and another one kept in complete absence of the cue (C!), and all experiments were repeated with larvae originating from three different females (F1F3 in Exps. I and II; F4F6 in Exp. III). Both the timing and duration of exposure to ACW had a significant influence on the duration of development to metamorphosis, indicating an interaction between the extrinsic signal and the intrinsic (hormonal) control of the moulting cycle. While continuously exposed control megalopae (C+) required on average 7.6 Å} 0.5 to 7.7 Å} 0.5 d, those in group C! delayed their development by about 34 d (or on average by 48%), 79 d (119%), and 56 d (61%), respectively (F1F3), indicating a great deal of intraspecific variability among hatches. In Exp. I, the cue had to have been added within about the first 67 d to become effective (no significant difference in development time between treatments and control C+), while a later addition of ACW had no more stimulating effect (no difference to C!). In Exp. II, metamorphosis was delayed in all treatments when the cue was removed after an initial exposure of less than 45 d, while later removal did not significantly reduce the stimulating effect of ACW. Together, the results from these two experimental series suggest that an exposure approximately during days 46 of the moulting cycle may be crucial as a temporal window of receptivity for the cue; this period coincides with the transition between intermoult and premoult (stages CD0 of Drachs classification system). Its presumable importance as a critical point was confirmed in Exp. III where development time was consistently shortest in megalopae exposed only during days 46 of the moulting cycle (no significant difference to control C+), while an earlier or later exposure prolonged development to metamorphosis. Intraspecific variability in the responsiveness to ACW was tested by comparing development time in control groups (C+, C!) among megalopae from 14 different hatches (including F1F6). These data showed a significant variability both in the delay due to complete absence of the cue and in the time of development in C+ groups. Twoway ANOVA indicated significant effects of exposure to ACW and of the female, as well as an interaction between these two factors, not only in development time, but also in the extent of individual variability among equally treated sibling larvae (measured as coefficient of variation). Analysis of molt-frequency distributions showed that the cue had a focussing (variabilityreducing) in addition to the accelerating effect. In conclusion, this study showed that the stage of the moulting cycle as well as intraspecific variability within and among hatches must be considered as potentially confounding factors interacting with effects of metamorphosisstimulating cues. The potential evolutionary significance of intraspecific variability in the extent of phenotypic plasticity is discussed in relation to different ecological scenarios.