Abstract
THE species problem in freshwater fishes, in which allopatry is the norm, is discussed. Biological species criteria break down because of the general occurrence of allopatry, and the implications for the taxonomy of freshwater fishes are explored. Essentially morphological criteria must be used, but their use must be subject to careful scrutiny, and patterns of variation must be studied. Galaxias maculatus (Jenyns) is shown to occur in Australia, New Zealand, Chatham Islands, Chile, Tierra del Fuego, and Falkland Islands in the typical diadromous form. Subspecies are not recognised. Landlocked or freshwater-limited populations occur through much of the range of the species, and in these populations vertebral number is reduced to about 55, from about 62 in diadromous stocks. These populations are not given separate taxonomic recognition, variation in vertebral number being regarded as related to the transition from a diadromous to a wholly freshwater life history; G. usitatus McDowall and G. parrishi Stokell become synonyms of G. maculatus. Further populations in small lakes in North Auckland, New Zealand, have only about 49 vertebrae. These are regarded as specifically distinct and are placed in G. gracilis McDowall.