ABSTRACT
Feeding periodicities and rhythms of suspension-feeding bivalves in coastal waters are closely related to diel and tidal cyclic environmental changes, such as food availability and predation risk. Although such feeding periodicities of bivalves are important for understanding how their feeding strategies adapt to localized habitats, the feeding periodicity in bivalves in the field remains to be clarified, especially in infaunal species. The present study focused on the morphological condition of the crystalline style as an indicator of the feeding activity of the infaunal bivalve Manila clam, Ruditapes philippinarum, and estimated their feeding periodicity in the field. First, the effects of feeding (siphon-extending) behaviour, food and the light/dark cycle on the condition of the crystalline style (defined based on its morphology and hardness) were investigated in the laboratory. Results of the ordered logistic models and model selection showed that clams extending their siphons kept their crystalline styles hard. Furthermore, the effects of the light/dark cycle and tidal phase on the condition of the crystalline style were investigated in natural intertidal and subtidal zones and assessed using the results of the ordered logistic models. In the intertidal zone, the effect of the tidal phase was significant, whereas the effect of the light/dark cycle was not, and feeding was estimated to be active at ebb and flood tides, when the density of food particles was likely to be high. On the other hand, in the subtidal zone, the results of the model selection were not clear because of significant effects of interactions between the light/dark cycle and tidal phase. However, when the interactions were removed from the models, the effect of the light/dark cycle was significant, and feeding was estimated to be active at night when the predation risk from visual predators would be relatively low. These results confirm that R. philippinarum possesses behavioural plasticity in feeding periodicity, which is adaptable to both intertidal and subtidal zones.
On the Complexity of Logistic Regression Models
