Abstract During four successive experiments spanning 24 months, the bivalve Protothaca staminea (Conrad) grew significantly more at all sizes inside cages that excluded large (> 6 mm) predators and croppers than inside cage controls and uncaged controls in a clean-sand site in Mugu Lagoon, California. For a clam of 2.5 cm initial length, predator and cropper exclusion increased average linear growth to 2.2 times control levels and volumetric growth to 2.5 times control Caging over the same periods had no repeatable effect on Protothaca growth in a muddy-sand site. Laboratory tests of the feeding behavior of Protothaca with and without each of four common consumers suggest that the presence of consumers did not cause sufficient disruption of feeding to produce the growth differences observed in clean sand. However, monthly seining data over 16 months revealed the persistent abundance of three fishes, Leptocottus armatus Girard, Hypsopsetta guttulata (Girard), and Paralichthys californicus (Ayres), containing abundant Protothaca staminea and Macoma spp. siphons in their stomachs. Estimates of habitat-specific intensity of siphon nipping suggest that the rate of cropping per individual Protothaca was 30 to 92% lower in the muddy sand, in part because of a switch to the more abundant deposit-feeding bivalve, Macoma nasuta (Conrad). Thus, habitat-specific cropping differences can explain the between-habitat difference in the response of Protothaca to caging.