Two methods are used to analyze latitudinal ranges of fish species: means of ranges of all species with a midpoint in the same 5⚬ latitude band plotted against latitude, and means of ranges of all species occurring in the same 5⚬ latitude band plotted against latitude. Application of the first method to data on 894 teleost species from the Indo-Pacific and Atlantic oceans showed that species with midpoints of latitudinal ranges near the equator have, on the average, much greater latitudinal ranges than species with midpoints at higher latitudes. Ranges of species were more or less the same when the second method was used, because of larger ranges and greater species numbers of low-latitude species. Rapoport's rule, according to which high-latitude species have greater latitudinal range than low-latitude species, therefore does not generally apply to marine teleosts. In contrast, the rule applies to North American (720 species) and North European (61 species) freshwater fish above a latitude of approximately 40⚬N as demonstrated by both methods. The rule has not been shown to apply to any group of organisms below a latitude of approximately 25⚬-35⚬N. Attempts to explain greater species numbers of tropical species by narrower environmental tolerances of tropical species, as indicated by Rapoport's rule, are therefore premature, and such an explanation does not hold for marine teleosts.