The classification used for the species (Part III of the Catalog) and for the genera (Part IV of the Catalog) is identical, but the classification has been modified since the publication of Eschmeyer (1990, 1998). The original goal was to provide a framework of orders, families and subfamilies (with occasional use of suborders). The classification portion, however, was secondary to the goal of compiling Parts I and II of the Catalog. Unfortunately, at this time in ichthyology there is no generally accepted classification, and many workers are actively engaged in research on higher-category taxa. Many follow Nelson (1994), and we have used his classification in most areas. At the same time, cladistic studies produce hypotheses of relationships, often based on examination of a very few taxa in each larger taxon, but these hypotheses need to be corroborated by other workers. Cladistic studies offer a rational and logical methodology for studying relationships, but reversals, parallelisms and problems of polarizing characters and outgroup selection for such polarization can be problematic. Often, many trees may be generated from computer programs, sometimes with substantial differences among them. To adopt each new hypothesis as proposed is not warranted in a work such as here presented, where stability for communication to many audiences is desired. The aim of the classification is to group related genera or species together, rather than attempt to reflect relationships evenly. For example, if a group of genera has been recognized as a family but a more recent study shows that these genera are specialized or highly modified members of another family, the genera are moved to the ‘new’ family but may be retained there as a subfamily, thereby keeping those related genera together. In some groups, subfamilies are not used, although they may be used in current literature; these include some small families with only a few genera, but also some large families, such as the Cyprinidae, where some ‘specialized’ subfamilies could be recognized, but the family as a whole has not been divided into subfamilies on which there is general agreement.

Synonyms of family group names are not provided, and the indexes to higher taxa names presented at the end of Parts III and IV include only those names actually mentioned. However, it is possible to determine the current placement of a family or subfamily that is not specifically included in the classification. For example, in the literature one may encounter a species placed in a family that is not included in the Catalog. Since family group names are based on a stem-genus (by dropping the terminal letter or letters and adding -idae for a family or -inae for a subfamily), one may look up the genus in Part II of the Catalog, go to the end of its account, and find where that family is now placed.

Family-group names used in the classification follow current use. Some problems involving family-group names in fishes include currently used family names that are not the oldest for the family and should be replaced by the older names unless a case can be presented to retain the younger name, some family names are being misspelled in the current literature, or two spellings are used (such as Engraulidae or Engraulididae). See Robins et al. (1980:4 [ref. 7111], Steyskal 1980 [ref. 14191] and Géry 1989 [ref. 13422]). These problems are not addressed directly in the Catalog, but some comments regarding family-group names are mentioned under their type genera (e.g., see Phosichthys and Bovicthys).

A few genera or species are not placed within families in the classification. Some are based on mythical specimens, or are indeterminable, or they are names only (without a description); many of these are unavailable names. They are often listed under a class, order or suborder. In the genera listing, unplaced genera appear at the end. The internal name of this table is LINEAGES.