If, as some philosophers maintain, there are no composites, we do not have to ask whether, as others hold, composition is identity. Here I argue that both groups are wrong: there are composites, and composition is not identity. I examine one argument for excluding composites from our ontology, based on their alleged causal redundancy. I give reason to think that composites are ineliminable in causal explanations of macroscopic effects. I go on to argue that the relation between composites and their (...) components is not one of identity. (shrink)