Protein phylogenies were used to test the hypothesis that aspects of the innate immune system of vertebrates have been conserved since the last common ancestor of vertebrates and arthropods. The phylogeny of lysozymes showed evidence of conservation of function, but phylogenies of seven other protein families did not. Natural resistance-associated macrophage protein, nitric oxide synthetase, and serine protease families all showed a pattern of gene duplication within vertebrates after their divergence from arthropods, giving rise to immune system-expressed genes in vertebrates. Insect hemolin, a member of the immunoglobulin superfamily, was found not to be closely related to members of that family having an immune system role in vertebrates; rather, it appeared most closely related to both arthropod and vertebrate molecules expressed in the nervous system. Thus, hemolin seems to have evolved its role independently in insects, probably through duplication of a neuroglian-like ancestor. Furthermore, vertebrate immune system-expressed serpins, chitinases, and pentraxins were found to lack orthologous relationships with arthropod members of the same families also functioning in immunity. Therefore members of these families have evolved immune system functions independently in the two phyla. It is now widely recognized that the specific immune system of vertebrates has no counterpart in invertebrates; these phylogenetic analyses suggest that there is a similar evolutionary discontinuity with respect to innate immunity as well.