« OP ILLINOIS LIBRARY AT URBANA-CHAMPAIQN Biaopv DEC 161996 590.5 FI n.s. Phyiogeny and Paedomorphosi African Family of Freshwater ^ (Gonorynchiformes: Kneriidae) Terry Grande October 31, 1994 Publication 1459 PUBLISHED BY FIELD MUSEUM OF ^ : avvj.i\»My&j?f:^ {fHm for Contributors to i^ieldiair. ;!:^fr infwbers ano rcscaa ■K of at Icnsi 5\ submitted .inipu(er .cnals iirc j;o, Iliinoi:- ,h wide tn a Hit r.fSfvlr (0(* ;«;//« ap,<.J luld foMov> crsity Press, vSfani'ord Figures n- •e must b- ISPRSNTEDOW • •-■:■ ^'cff PAPFR- FIELDIANA Zoology ^JEW SERIES, NO. 78 Phylogeny and Paedomorphosis in an African Family of Freshwater Fishes [Gonorynchiformes: Kneriidae) rerry Grande Research Associate Department of Geology Field Museum of Natural History Roosevelt Road at Lake Shore Drive Chicago, Illinois 60605-2496 Assistant Professor Department of Biology Loyola University 5525 North Sheridan Road Chicago, Illinois 60626 Accepted May 26, 1994 Published October 31, 1994 Publication 1459 PUBLISHED BY FIELD MUSEUM OF NATURAL HISTORY © 1994 Field Museum of Natural History ISSN 0015-0754 PRINTED IN THE UNITED STATES OF AMERICA Table of Contents Abstract 1 Introduction 1 Methods 2 Materials Examined 2 List of Abbreviations 4 Results and Systematic Descriptions 4 Unnamed Kneriid Subgroup A 5 Grasseichthys Gery, 1 964 6 Cromeria Boulenger, 1901 8 Unnamed Kneriid Subgroup B 13 Unnamed Kneriid Subgroup C 13 Parakneiua Poll, 1 965 13 Knerl\ Steindachner, 1 866 16 Discussion 17 Acknowledgments 19 Literature Cited 20 List of Illustrations 1 . Phylogenetic hypothesis of gonorynchi- form relationships 3 2. Phylogenetic hypothesis of kneriid rela- tionships showing two monophyletic groups: Grasseichthys + Cromeria and Parakneria + Kneria 5 3. Coronoid processes of Chanos chanos and Kneria wittei 6 4. Skull of Grasseichthys gabonensis and ventral skull of Chanos chanos 7 5. Anterior vertebrae and neural arches of Parakneria tanzaniae 8 6. Skull, anterior vertebrae, and neural arches of Cromeria nilotica 8 7. Specimen of Grasseichthys gabonensis and Cromeria nilotica occidentalis 9 8. Lateral view of suspensorium of Gras- seichthys gabonensis 10 9. Anterior vertebrae and neural arches of Grasseichthys gabonensis 10 10. Skull of Cromeria nilotica 11 1 1 . Lateral view of suspensorium of Crome- ria nilotica 11 12. Ventral gill arches of Parakneria tanza- niae and Chanos cyprinella 12 13. Specimens of Parakneria tanzaniae and Kneria wittei; specimen of K. wittei showing opercular apparatus 13 14. Dorsal and ventral views of skull of Parakneria tanzaniae 14 15. Lateral and medial views of suspensoria of Parakneria tanzaniae 15 16. Dorsal and ventral views of skull of Kneria wittei 17 17. Lateral and medial views of suspensoria of Kneria wittei 18 18. Anterior vertebrae and neural arches of Kneria wittei 19 List of Tables Data matrix from which the phylogenetic hypothesis for kneriid fishes was con- structed 4 Meristic counts of representative sp)ecies of Kneria, Parakneria, Cromeria, and Grasseichthys for comparisons 9 f [ Phylogeny and Paedomorphosis in an African Family of Freshwater Fishes (Gonorynchiformes: Kneriidae) Terry Grande i Abstract The freshwater gonorynchiform family, Kneriidae, lives in west and central Africa and con- tains some of the smallest known vertebrates. The extremely small size and paedomorphic nature of some of these species pose an interesting theoretical problem for phylogenetic analyses. The Kneriidae is defined as monophyletic in part by the presence of lateral extensions of the mesethmoids and the articulation of the first neural arch with both the exoccipitals and su- praoccipital. The family is frequently divided into two clades, one consisting of the genera Cromeria and Grasseichthys, and the other consisting oi Parakneria and Kneria. Cromeria and Grasseichthys are unusually small fish that mature at only 30 and 16 mm standard length, respectively, whereas Parakneria and Kneria are substantially larger by the time they reach maturity. Three characters supporting the Cromeria + Grasseichthys clade are the lack of scales, incomplete cranial ossification, and the lack of median contact between the frontal bones. Incongruent with those three characters is a peculiarly shaped vomer shared by Cromeria, Parakneria, and Kneria. At first look, parsimony would seem to indicate that the Cromeria + Grasseichthys group is monophyletic. The three characters supporting that group, however, are problematic, because they all are either absence or reductive characters and could be the result of paedomorphic developmental patterns. In contrast, the vomer character suggesting a Crome- ria + Parakneria + Kneria group is neither a loss nor a reductive character. Additional non- reductive characters are needed to better resolve the family interrelationships. Introduction convenience. Gosline (1960) stated that "each family within the Gonorynchoidei is so widely dif- The teleostean order Gonorynchiformes is ferent that any relationship between them is dif- geographically widespread, is morphologically di- ficult to comprehend" (p. 353). verse, and has a fossil record dating back to the Not until Greenwood et al. (1966) reexamined Early Cretaceous. For many years the order was gonorynchiform relationships was any significant poorly defined and its monophyly in doubt (Gayet, progress made. They proposed the order Gono- 1986). Various gonorynchiform taxa have in the rynchiformes with two suborders: the Gonoryn- past been included in the Salmoniformes, Clupeo- choidei (including Gonorynchidae) and the Cha- morpha and Elopomorpha (Regan, 1929). Gosline noidei (including Chanidae, Kneriidae, and (1960) recognized the suborder Gonorynchoidei Phractolaemidae). Kneriidae 5en5M Greenwood et in which he included the Chanidae, Gonorynchi- al. (1966) included two tiny (probably paedo- dae, and Kneriidae, but he grouped these gono- morphic) genera, Cromeria and Grasseichthys. rynchoids within the Cypriniformes. Gosline Gonorynchiformes was diagnosed as having an (1960) was not convinced of the monophyly of epibranchial organ, upper and lower intermuscu- Gonorynchoidei and gave no diagnostic characters lar bones, a loosely articulated suspensorium, and for the taxon. He simply used it as a group of one or more pairs of cephalic ribs. In 1970 Rosen FIELDIANA: ZOOLOGY, N.S., NO. 78, OCTOBER 31, 1994, PP. 1-20 1 and Greenwood included the Gonorynchiformes within the superorder Ostariophysi as the sister group to the Otophysi (fishes possessing a Weber- ian apparatus and including all non-gonorynchi- form ostariophysans). Lenglet (1974) agreed that the Gonorynchiformes and Otophysi are closely related but considered Rosen and Greenwood's (1970) findings to be premature. He considered the interrelationships within the order uncertain, and the placement of Cromeria and Grasseichthys in the Gonorynchiformes problematic because of their probable paedomorphic nature. Lenglet (1974) found that Cromeria and Grasseichthys do not share all the characters of the Gonorynchi- formes and proposed that the two genera be placed within a new suborder, Cromeroidei, on the basis of a reduction in cranial osteology, lack of scales, and lack of suprapreopercle. Roberts (1973) crit- icized Rosen and Greenwood's (1970) classifica- tion and stated that the inclusion of the Gonoryn- chiformes within the Ostariophysi would render the latter paraphyletic. But later. Fink and Fink (1981) corroborated Rosen and Greenwood's placement of the Gonorynchiformes within the Ostariophysi and further diagnosed the order. Al- though today few ichthyologists question the monophyly of the group, many debate the rela- tionships among its taxa. Howes (1 985), in a study based on muscle characters, was unable to resolve relationships of taxa within the order. In a study of the phylogenetic interrelationships of both Recent and fossil gonorynchiform fishes, Grande (1992) confirmed the monophyly of the order. This was based on several characters, in- cluding the loss of the orbitosphenoid bone, a re- duction and separation of the pterosphenoid bones, a reduction of the parietals along with expansion of the frontal bones posteriorly, an extension of the supraorbital canal ending on the frontals in- stead of the parietals, the loss of teeth on the fifth ceratobranchial, lateral expansion of the first neu- ral arch, the presence of a strong ascending process of the parasphenoid that rises high and passes in front of the prootics, the presence of cephalic ribs, and the presence of an epibranchial organ. Grande (1992) also showed that Chanos (a marine Indo- Pacific form) is the sister group to a group con- taining the Gonorynchidae plus an African fresh- water clade consisting of the Phractolaemidae {Phractolaemus) and the Kneriidae (Grasseich- thys, Cromeria, Parakneria, Knerid) (Fig. 1). This paper will focus on the phylogenetic inter- relationships of the family Kneriidae. The Knerii- dae has long been a systematically problematic group. Myers (1938) suggested that Cromeria (a monotypic genus) was only a larval stage of some species of Kneria. Howes (1985) tentatively syn- onymized Parakneria and Kneria, stating that "characters separating the genera Kneria and Parakneria appear to be ones of degree and on the basis of synapomorphies of the skeleton and jaw musculature the two genera should probably be considered a single genus" (p. 299). The family is thus in need of review. Methods Skeletal material was prepared using a modified version of Dingerkus and Uhler's (1977) method of clearing and double staining. Measurements and meristic counts were taken from all specimens ex- amined as outlined by Hubbs and Lagler (1949). Cranial and postcranial osteological illustrations were made using a Wild M5 or M8 dissecting mi- croscope equipped with a camera lucida. Comparative morphological data were analyzed to assess the relationships among kneriid fishes by means of phylogenetic analysis, or cladistics (Nel- son & Platnick, 1981). Character polarity was de- termined by outgroup comparison. Outgroups chosen for this study included Chanos (a gener- alized gonorynchiform), Phractolaemus (the hy- pothesized sister group to the Kneriidae), and Op- sariichthys (a generalized cypriniform). Derived character information (character ma- trix. Table 1 ) was processed using outgroup root- ing, Deltran optimization, and the branch-and- bound version of PAUP version 3.0 (Swofford, 1990). These data were also analyzed using the Hennig 86 program to test the stability of my data matrix and the reliability of PAUP. Characters were optimized using a character description pro- gram, D3, version 2, by Buckup (1991). This pro- gram generates lists of character state transfor- mations from the output produced by Hennig 86. Materials Examined Chanos chanos: 1 6 specimens (SL = 30-220 mm): AMNH 87984 (cleared & stained); ansp 122260, 122264 (alcoholics); fmnh 971 10, 3981, 101 19 (cleared & stained). Chanos cyprinella: 58 specimens (SL = 70-105 FIELDIANA: ZOOLOGY ^ a ^ ^ -O O O »<- Co / O ^^ O