Do the highly diverse fish faunas that associate with coral reefs have
distinguishing taxonomic and ecological characteristics, as proposed
by Cheat and Bellwood (1991) and Bellwood (1996)? Does a 50 my old (Eo
cene) fossil fish fauna from Italy represent a coral-reef fish assembl
age that provides unique information about the evolution of such assem
blages, as claimed by Bellwood (1996)? I compared the structure of the
reef fish faunas of adjacent tropical regions rich and poor in coral
reefs, in both America and Polynesia, and found that they exhibit no s
ubstantive differences in relative species richness among families of
typical ''coral-reef'' fishes. While coral-rich regions have larger re
ef fish faunas, a variety of factors probably contribute to such diffe
rences. Thus coral-reef fish faunas may lack a distinctive taxonomic s
tructure. A similar comparative approach would be useful for assessing
whether assemblages of fishes on coral reefs have distinctive ecologi
cal characteristics. Based on patterns of habitat use by modern tropic
al shorefishes, the Italian Eocene fauna includes few definite reef fi
shes, and may well consist primarily of non-reef fishes preserved in a
non-reef habitat. Until we know more about the environment in which t
hose fossils were preserved, that fauna can contribute little to under
standing how coral reef fish assemblages have evolved.