Cross-linked walled protein microcapsules have previously been developed and shown to be ingested, digested and assimilated by suspension-feeding bivalves. In this study, we examined the effect of adding carbohydrate to protein capsules on the utilization of encapsulated protein by mussels (Mytilus trossulus) in a series of in vitro enzyme incubations and an in vivo feeding experiment. Two types of carbohydrate (amylose and maltodextrin) were separately encapsulated at a carbohydrate to protein ratio of 1:1 by weight. In vitro carbohydrate and protein leakage rates from mixed carbohydrate/protein capsules were low [e.g. <4% (24h)(-1)] compared to previously reported leakage rates for protein capsules [e.g. 5 to 20% (24h)(-1)]. Proteolytic breakdown of mixed carbohydrate/protein capsules by mussel-style extracts was similar to that for protein capsules, when the encapsulated carbohydrate was amylose, a straight-chain (alpha-1,4-linkage) polysaccharide; however; capsules containing maltodextrin, which is partially comprised of alpha-1,6-linkages, were not as readily digestible in vitro. In contrast to results from in vitro enzyme assays, maltodextrin/protein capsules were utilized in vivo by mussels with similar efficiencies to those for protein and amylose/protein capsules. These findings demonstrate that utilization of microencapsulated protein by mussels cannot be accurately predicted from results of in vitro enzyme assays. Importantly, up to 50% of the dry weight of cross-linked walled protein microcapsules can be replaced with carbohydrate without significantly affecting encapsulated protein digestion and assimilation efficiencies for M. trossulus.