AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2795, pp. 1-1 8, figs. 1-7 October 9, 1984 On the Relationships of the Triassic-Liassic Redfieldiiform Fishes BOBB SCHAEFFER' ABSTRACT A survey of character distributions among the sumably independent reduction of the branchios- extinct lower actinopterygians has corroborated tegal series. the hypothesis that the freshwater, Triassic-Liassic The sister group of the redfieldiiforms among redfieldiiform fishes form a monophyletic group. the other extinct lower actinopterygians remains This proposal is based mainly on the pattern of unknown, but a survey ofthe dermal skull in these the dermal snout, loss of anterior nares, and pre- fishes has provided a hypothetical "sister" pattern. INTRODUCTION Character analysis in the palaeonisciforms The primary purpose of the present paper and other extinct lower actinopterygian is to attempt a cladistic analysis ofone group groups has lagged well behind that for the ofextinct lower actinopterygians that has long extinct neopterygians. The reasons for this been regarded as a "natural" one, namely, include the absence of close living relatives, the Redfieldiiformes. They are Triassic and the apparent paucity of useful characters in early Jurassic freshwater fishes (fig. 1) whose taxa that are frequently represented by com- remains have been found in continental sed- pressed and inadequately preserved dermal iments in Australia, South Africa, Zambia, skeletons, the fragmentary fossil record, and Morocco, eastern and western United States, the not inconsiderable problems of morpho- and questionably in Madagascar. The design logical interpretation. More or less three di- of the redfieldiiform skeleton has suggested mensional specimens that can be prepared affinity with several other extinct lower ac- chemically or by air abrasion may yield a tinopterygian groups, but there has been no great deal of information, but they represent consensus about relationship. only a small fraction ofthe nearly 200 genera In regard to the living lower actinopteryg- in this paraphyletic assemblage. ians, Patterson (1982) has recently defined I Curator Emeritus, Department of Vertebrate Paleontology, American Museum of Natural History. Copyright © American Museum of Natural History 1984 ISSN 0003-0082 / Price $2.00 2 AMERICAN MUSEUM NOVITATES NO. 2795 FIG. 1. Restorations of three redfieldiiforms. A. Brookvalia. After Hutchinson, 1973 and AMNH 4706, 9254. Flank squamation omitted. B. Daedalichthys. After Brough, 1931 and BMNH 17532-3. C. Cyonichthys. From Schaeffer, 1967. 1 984 SCHAEFFER: REDFIELDIIFORM FISHES 3 two monophyletic groups, the Cladistia for fig. 54), and that the two groups arose from the polypteroids, and the Chondrostei for the a palaeonisciform complex including the acipenseroids and polyodontoids plus their Elonichthyidae, Acrolepidae, and the Rhad- fossil relatives. This restricted use ofthe term inichthyidae (ibid., p. 346). In recent pre- Chondrostei makes it still more imperative cladistic classifications ofthe extinct and liv- that the various taxa ofextinct lower actinop- ing actinopterygians (e.g., Andrews et al., terygians be restudied. The suspected pa- 1967; Romer, 1967) the Redfieldiiformes are raphyly ofnumerous fossil non-neopterygian listed as a suborder or order of the Chon- categories is emphasized by Patterson's (1982, drostei without comment on relationship. fig. 3B) decision to include the palaeonisci- Hutchinson's (1978) last paper on the red- form genus Pteronisculus in a trichotomy with fieldiiforms, which is mainly concerned with the Chondrostei and the Neopterygii on the the genus Helichthys, includes alternative basis of five shared characters. In addition, cladograms for the relationships ofthis genus Gardiner (in press) has concluded that the to the three redfieldiiform families that he Devonian palaeonisciform genera Mimia and defined in 1973 (the Brookvaliidae, Schizur- Moythomasia are successive sister taxa to the ichthyidae, and Redfieldiidae). Prior to his remaining actinopterygians. As such, they death in 1978, Peter Hutchinson and I had should provide us with information on the planned to write a joint paper on redfieldi- primitive state for the actinopterygian brain- iform affinities and interrelationships. case (Mimia is represented by abundant three- The present paper is dedicated to his mem- dimensional material), palate, dermal skull, ory. axial skeleton, fins, and squamation. This in- formation, much of it recently acquired, is ABBREVIATIONS of importance in obviously considerable AMNH, The American Museum of Natural His- working out the patterns of character distri- tory bution within the palaeonisciforms and the BMNH, British Museum (Natural History) redfieldiiforms, and in relation to seeking a UTVPC, University of Texas Vertebrate Paleon- sister group for the latter. tology Collection Discussion of redfieldiiform relationships (see summary in Schaeffer, 1967, p. 329) ef- Anatomical abbreviations are included in the fectively began with Stensio's (1921) opinion legend for figure 4. that the catopterids (=redfieldiids) and the SYSTEMATIC RESUME colobodontids (=perleidids) should be grouped in the family Catopteridae, which, In order to facilitate later discussion, a di- he believed, is closely related to the palaeo- agnosis ofthe Redfieldiiformes is included in niscids. Lehman (1966) placed both fami- this section along with a list of genera ar- lies in the order Perleidiformes, which he ranged, as far as practicable, by character dis- considered to be one of numerous orders of tribution (fig. 6), relative age and source area. fossil, non-teleost actinopterygians (ibid., p. The vertical distribution of the Newark su- 71). Brough (1931, 1934) first considered the pergroup genera is summarized in Olsen, Catopteridae to be "closely related to, and McCune and Thomson (1982, fig. 7). derived from, the Palaeoniscidae," and, in DIAGNOSIS: Lower actinopterygians (non- 1936, to be derived from another palaeon- neopterygians) with a terminal or subtermi- isciform family, the Dicellopygidae. Schaef- nal gape, fusiform body outline and an equi- fer (1967, 1973) has favored derivation from lobate, hemiheterocercal tail. The braincase generalized palaeoniscids, while Lowney (Ms) is typically palaeonisciform with an open oti- has tentatively proposed a sister-group rela- co-occipital fissure and a nearly vertical hyo- tionship with the rhadinichthyid palaeonis- mandibular facet. The parasphenoid is short, ciforms. Hutchinson (1973, p. 345), suggest- with well-developed ascending processes. The ed that the redfieldiiforms and the vomerine area is covered by elongated paired perleidiforms had a common ancestor (ibid., tooth plates (in Ichnolepis). The endoptery- 4 AMERICAN MUSEUM NOVITATES NO. 2795 goids are relatively large and dentigerous. The 8. Molybdichthys Wade, 1935. Middle Triassic dermal snout is composed ofa median rostral (?Ladinian). Same locality as Schizurichthys. and postrostral, paired nasals and paired 9. Daedalichthys Brough, 1931. Lower Triassic "premaxillae" (=premaxillo-antorbitals of (Scythian). Same locality as Atopocephala. some authors). The single nostril is surround- 10. Cionichthys Schaeffer, 1967. Upper Triassic (Middle-Late Carnian) Chinle Group and ed by the nasal, adnasal and the enlarged Dockum Formation, western USA; Newark "premaxilla." The nasal is excluded from the Supergroup, eastern USA. orbit by the adnasal. The maxilla is fixed and 11. Redfieldius Hay, 1899. Lower Jurassic (Het- expanded posteriorly. The preopercular is tangian-Sinemurian), Newark Supergroup, angled; the subopercular is larger than, or eastern USA. about equal to, the size ofthe opercular. Both 12. Dictyopyge Lyell, 1847. Upper Triassic (Mid- a dermohyal and an antopercular are present dle Carnian), Newark Supergroup, eastern in the more generalized taxa. There is a sin- USA. gle, platelike branchiostegal (two in Daedal- 13. Mauritanichthys Martin, 1982. Upper Trias- sic, Argana Valley, Western Atlas, Morocco. ichthys) and a single median gular. The der- 14. Lasalichthys Schaeffer, 1967. Upper Triassic mopterotic is large and rectangular. The (?Carnian), Chinle Group and Dockum For- shoulder girdle includes a clavicle, postclei- mation, western USA. thrum, and a presupracleithrum (observed in 15. Synorichthys Schaeffer, 1967; Schaeffer and Helichthys). Fringing fulcra are probably Mangus, 1970. Upper Triassic (?Carnian), present on all fins. Basal fulcra border the Chinle Group and Dockum Formation, west- unpaired fins, including both lobes ofthe cau- ern USA; Newark Supergroup, eastern USA. dal. The rays of the remote, opposite dorsal Genera 1 through 5 were included by and anal fins are more numerous than the Hutchinson (1973) in his Brookvaliidae and basals. The scales are rhomboidal, with peg- 12, 14 and 15 in the Redfieldiidae. Schizur- and-socket articulation, and with dentine and ichthys was the only genus assigned to the ganoin layers. Family Schizurichthyidae. Hutchinson's (1973, 1978) division of the redfieldiiforms into the Brookvaliidae, Schi- CHARACTER ANALYSIS zurichthyidae and the Redfieldiidae left the The characters discussed below are ar- genus Helichthys (ibid., 1978) in an indeter- ranged as follows: neurocranium, snout, cir- minate status. The potential monophyly of cumorbital series, skull roof, cheek area, these families and the affinities of Helichthys opercular series, palate, mandible, paired