Systematics of the family Ariidae (Ostariophysi, Siluriformes), with a redefinition of the genera. Author Alexandre P. Marceniuk Author Naércio A. Menezes text Zootaxa 2007 1416 1 126 http://www.zoobank.org/urn:lsid:zoobank.org:pub:FFC65592-D8DB-41BE-AEAC-A41EAB6C6185 journal article z01416p001 [[ Family Ariidae Bleeker ]] Introduction The Ariidae are widely distributed, most of its species occurring along tropical and temperate areas of the world in coastal waters of the continents, estuarine regions and lower portions of coastal rivers. A restricted number of species is either entirely confined to marine waters where they can be found at depths of 150 meters or to fresh waters in the upper courses of rivers 500 kilometers away from their mouths. The group was established by Bleeker (1862) as Phalanx Arii and formally defined by Regan (1911). Since then the inaccurate description of new taxa sometimes based on ontogenetic phases of the same species or on sexual differences added to the proliferation of names. The genera presently recognized were primarily defined on the basis of traditional morphological characters, such as shape of tooth plates, eye position on head, presence or absence of skin over the eye, extension of the branchial opening, type of ornamentation of skull bones, number and shape of barbels, etc., presently considered inconsistent or of limited information (pers. obs.). Lack of adequate comparative material in studies of systematics and taxonomy of the Ariidae led to the recognition of species from widely separated regions of the world under the same generic name and this is reflected in classifications proposed during the 19th century (Cuvier & Valenciennes, 1840a, 1840b; Bleeker, 1858, 1863; Guenther , 1864). Later on new genera and classifications were established based on geographic distribution of species involving faunas of different regions such as: Africa and Asia (Chaudhuri, 1916; Herre, 1926; Fowler, 1936, 1941; Chandy, 1953; Misra, 1959; Tobor, 1969; Jayaram & Dhanze, 1978; Jayaram, 1982, 1984; Taylor, 1986); Australia and Papua New Guinea (Weber & Beaufort, 1913; Hardenberg, 1941, 1948; Roberts, 1978; Kailola, 1999) and Americas (Jordan & Gilbert, 1883; Eigenmann & Eigenmann, 1890; Jordan & Evermann, 1898; Meek & Hildebrand, 1923; Gosline, 1945; Fowler, 1951; Taylor & Menezes, 1977; Figueiredo & Menezes, 1978; Kailola & Bussing, 1995; Acero, 2003). More recently attempts have been made to redefine genera and to establish generic relationships, but primarily based on similarity of external characters not on presence of shared derived characters as proposed by the cladistic method (Hennig, 1950, 1966). Tilak (1965) primarily using characters associated with skull bones and the Weberian apparatus of species from India redefined some of the ariid genera and suggested that the characters he found could be used for species and genera worldwide. Higuchi (1982) however, studying the osteology of five species from southeastern and southern Brazil concluded that Tilak’s characters were not useful to diagnose genera from the Atlantic Ocean. Jayaram & Dhanze (1986) tried to define relationships among ariid genera from India also based on osteological and other morphological characters. They recognized two main “evolutionary lines”, one including the apparently closely related genera Ketengus and Batrachocephalus and the other the genus Arius , more specialized. Osteogeneiosus was considered the basal unit. However, no evidence that he was dealing with monophyletic groups was presented. The major difficulties in undertaking a comprehensive study on the systematics of the Ariidae in order to provide a better definition of the included genera and to study their relationships is the great diversity and wide distribution of the group. Many of the species are rare in museum collections and are not available for anatomical studies. Recent studies of the phylogenetic relationships of ariid genera (Kailola, 1990a, 2004; Betancur-R., 2003; Betancur-R. & Acero, 2004; Betancur-R. et al., 2004) are geographically restricted and do not take into account species of many of the known ariid genera (see Discussion and comparison with previous classifications below). In spite of the array of controversies concerning definition of genera and their relationships within the Ariidae , the monophyletic condition of the family was never seriously questioned. First considered as relatively primitive or “generalized in form and structure” during pre-cladistic studies (Regan, 1911; Berg, 1940), Ariidae was considered more primitive than Doradidae , Plotosidae , Schilbeidae and Bagridae an opinion not shared by Shelden (1937), Tilak (1963, 1965, 1967), Greenwood et al. (1966) and Chardon (1968). Using cladistic methodology, Mo (1991), Lundberg (1993) and de Pinna (1993) confirmed Ariidae as a monophyletic group. The first two suggested Doradoidea as the ariid sister group, but de Pinna (1993) attributed this condition to family Claroteidae and both this family and Ariidae would be the sister group of Schilbeidae and Pangasiidae . A similar conclusion was reached by de Britto (2002). Later on de Pinna (1998) also recognized Doradoidea as the sister group of the Ariidae . The main objective in the present work is to revise and redefine the taxonomic status of ariid nominal genera using exclusive internal and external morphological characters and a combination of morphological characters and to propose a new species arrangement within the valid genera, examining the largest possible number of ariid representatives. The phylogenetic analysis is the subject of a future publication.