PROCEEDINGS OF THE UNITED STATES NATIONAL MUSEUM issued H^?(Vv4 O^MI ^H the SMITHSONIAN INSTITUTION U. S. NATIONAL MUSEUM ^o\. 107 Washington : 1957 No. 3383 THE FROGFISHES OF THE FAMILY ANTENNARIIDAE By Leonard P. Schultz In my attempt to identify fishes for inclusion in the report on "The Fishes of the Marshall and Marianas Islands," I find it impossible to place confidence in the literature describing marine reef fishes of the tropical Indo-Pacific regions unless the genus or the family in which bhey belong has been revised. Ichthyologists or fishery biologists svho have not attempted a revision of a genus of widely ranging marine ishes cannot possibly assess the untrustworthiness of most current faunal lists. To correctly identify as to genus and to species, tropical [ndo-Pacific fishes must be considered on a world-wide basis. Those tew individuals who have attempted revisions of genera have found a disturbing percentage of the scientific names currently applied to be imreliable. Even a serious attempt at revision of a genus may not clarify the nomenclatorial status of more than 95 percent of the species. The doubtfully identified 5 percent results mostly from the inaccessibility to the author of those types of species preserved in the scattered museums of the world. The problems of classification and analysis of species and genera sncountered in one family are the same as found in many other fami-ies that I have studied. In general, there are three such problems. (1) The characters used in distinguishing species or genera in each family must be carefully evaluated. Only in a general way are the characters of one family reliable for use in another family. It is im-portant to determine for each character its variability, and what reliance can be placed on it for each genus. Frequently this cannot be estabhshed until aU species have been studied in the family. Another problem encountered is, (2) what constitutes a genus in the Family being revised? Those who have studied many complex fish families know that the limits of a genus are variable. The characters 47
BioStor
