<<

The Phylogeny of Atherinomorphs: of a Novel Reproductive System

Lynne R. Parenti

Division of , MRC NHB 159, Department of Zoology, National Museum of Natural History, Smithsonian Institution, Washington, DC, USA.

Lynne R. Parenti • The Phylogeny of Atherinomorphs. Reproductive System 13 Abstract Resumen The Phylogeny of The fishes commonly known in English as Los peces comúnmente conocidos en inglés Atherinomorphs: silversides, , phallostethids, como: silversides, rainbowfishes, phallostethids, , , , , killifishes, ricefishes, halfbeaks, needlefishes, Evolution of a flying fishes, and were combined by flying fishes y sauries fueron integrados por Rosen (1964), into one taxon, now known as Rosen (1964) en un taxon ahora conocido Novel Fish the , largely using osteological como Atherinomorpha, utilizando and reproductive characters. Subsequent esencialmente caracteres osteológicos y Reproductive reviews have supported atherinomorph reproductivos. Revisiones posteriores han , adding characters to the sostenido que Atherinomorpha es un grupo System diagnosis. Today atherinomorphs are monofilético y han agregado otros caracteres a diagnosed as monophyletic by derived la diagnosis. Hoy, los aterinomorfos están characters of the testis, egg, reproductive definidos como monofiléticos por mode, circulatory system, jaw musculature, características derivadas del testículo, huevos, olfactory , and various parts of the formas reproductoras, sistema circulatorio, skeleton including the ethmoid region of the musculatura de la mandíbula, órgano skull, gill arches, pelvic girdle, among others. olfatorio, y varias partes del esqueleto, Support for monophyly of each of the three incluyendo la región etmoidea del cráneo, included taxa, now classified as the orders arcos branquiales, y cintura pélvica, entre , , and otros. El carácter monofilético de cada uno de , and the relationships among los tres taxa incluidos actualmente, clasificados them, varies in quantity and quality. During en los órdenes Cyprinodontiformes, the past twenty years, reproductive and Beloniformes, y Atheriniformes, se define en Lynne R. Parenti molecular data used to infer atherinomorph sus relaciones entre ellos, con variaciones de relationships have grown significantly. In cantidad y calidad. Durante los últimos veinte general, molecular data support hypotheses años, ha aumentado significativamente el uso based on , and, in some cases, de datos sobre reproducción y moleculares provide novel hypotheses and unique para inferir relaciones en los aterinomorfos. challenges to morphological data. A growing En general, datos moleculares apoyan body of data indicates that all atherinomorphs hipótesis basadas en la morfología y, en share a unique testis- that is correlated algunos casos, sustentan hipótesis nuevas y with an array of reproductive modifications únicas referentes a los datos morfológicos. Un such as coupling during mating, relatively número creciente de datos indica que todos long developmental period, sperm-bundle los aterinomorfos tienen un tipo testicular formation, internal fertilization, superfetation, único, correlacionado con una serie de embryo retention, diapause, delayed hatching, modificaciones reproductivas, tales como el hermaphroditism, and live-bearing. apareamiento, un periodo relativamente largo Corroboration of an atherinomorph sister de desarrollo, formación de paquetes de group may include identification of some espermatozoides, fertilización interna, unique aspects of this reproductive system in superfetación, retención de los embriones, other taxa. diapausa, eclosión demorada, hermafroditismo y viviparidad. Corroborar que algún grupo de otro taxa, esté emparentado con los aterinomorfos, puede incluir la identificación de alguno de estos aspectos del sistema reproductor.

• Viviparous Fishes • ViviparousMari Carmen Fishes Uribe and Harry J. Grier, book editors. 14 ,New , Publications, and Homestead, Evolution Florida, 2005. p 13-30. Introduction

n 1964, Donn E. Rosen, then curator in the considered by Gosline to be “higher ” Department of , American Mu- because they have characters such as two dorsal Iseum of Natural History, New York, pub- fins, both with anterior spines or thickened rays, lished a monograph as a Bulletin of the American and an I,5 pelvic-fin ray formula. In contrast, Museum of Natural History, in which he brought the and Exocoetoidei (Cy- together three disparate groups of fishes prinodontiformes and Beloniformes, following in one , the Atheriniformes (Rosen, 1964). Greenwood et al., 1966) were considered by The three suborders in Rosen’s Atheriniformes Gosline to be “intermediate teleosts.” Both have were the (the silversides, rainbow- a single, soft-rayed and may have more fishes, and phallostethids), the - than six pelvic-fin rays, among other characters, toidei (the killifishes and ricefishes), and the that they share with taxa that Gosline thought Exocoetoidei (the sauries, needlefishes, half- to be less advanced (see review by Parenti, 1993). beaks, and flying fishes). Support of this taxon Cyprinodontiformes and Beloniformes were not included evidence largely from two systems, os- considered closely related by Gosline (1971), teology and reproductive biology, which Rosen who postulated that they were derived from, or described in an essay, as was common then, most closely related to the and rather than enumerating putative synapomor- , respectively. phies. Skeletal characters that Rosen considered Monophyly of atherinomorphs and mono- diagnostic of his Atheriniformes included a disc- phyly and relationships of the three included shaped dorsal and ventral ossified mesethmoid taxa was reviewed by Rosen and Parenti (1981) and decoupling of the rostral cartilage from the in conjunction with a phylogenetic analysis of ascending processes of the premaxillae. Repro- Cyprinodontiformes by Parenti (1981). The ductive characters included a large, demersal egg first explicitly cladistic analyses of atherino- with long, adhesive chorionic filaments and morph phylogeny were presented in these two large oil globules (Rosen, 1964:253-255). papers. Atherinomorph monophyly was sup- The new taxon, classified as the series ported by ten characters, again largely those of Atherinomorpha by Greenwood et al. (1966), the skeleton, but also including two reproduc- was not accepted readily by all systematic ich- tive characters (Rosen and Parenti, 1981:20), thyologists. In particular, Gosline (1971, Fig. one of the egg (“a large dermersal egg with long 28B) argued that Rosen had brought together adhesive and short filaments and many lipid fishes from different evolutionary grades that did globules that coalesce at the vegetal pole”), and not share an evolutionary history. Atherinoids one of the testis (“the spermatogonia forming (now classified as the order Atheriniformes fol- only at the blind end of the tubule near the lowing Dyer and Chernoff, 1996; Table 1) were tunica albuginea”).

Lynne R. Parenti • The Phylogeny of Atherinomorphs. Reproductive System 15 Table 1. Annotated classification of Atherinomorph fishes (following Rosen and Parenti, 1981; Collette et al, 1984; Parenti, 1993; Costa, 1998a; Dyer and Chernoff, 1996).

Series Atherinomorpha Greenwood et al., 1966 [= Atheriniformes of Rosen, 1964] Order Atheriniformes Dyer and Chernoff, 1996 [= Atherinoidei of Rosen, 1964; Division I of Rosen and Parenti, 1981] Classification is sequenced. Atherinopsidae Suborder Atherinoidei Family Notocheiridae (including Isonidae) Infraorder Atherines Family Melanotaeniidae (including Bedotiidae, Pseudomugilidae) Family Atherionidae Superfamily Atherinoidea Family (including Dentatherinidae) Family Atherinidae Superorder Cyprinodontea of Dyer and Chernoff, 1996 [Division II of Rosen and Parenti, 1981, order Cyprinodontiformes of Nelson, 1984, not recognized by Rosen, 1964] Order Cyprinodontiformes [= Cyprinodontoidea of Rosen, 1964] Suborder Family Family Suborder Cyprinodontoidei Superfamily Funduloidea Family Family Family Superfamily Valencioidea of Costa, 1998a [= Sept 1 of Parenti, 1981] Family Valenciidae Unranked category including superfamilies Cyprinodontoidea and Poecilioidea Superfamily Cyprinodontoidea Family Cyprinodontidae Superfamily Poecilioidea of Parenti, 1981 [= unnamed of Costa, 1998a] Family Family Order Beloniformes [not recognized by Rosen, 1964] Suborder Adrianichthyoidei [= Adrianichthyoidea of Rosen, 1964] Family Adrianichthyidae (including Horaichthyidae and Oryziidae) Suborder Exocoetoidei Superfamily Exocoetoidea Family Exocoetidae Family Hemiramphidae Superfamily Scomberesocoidea Family Belonidae Family

• Viviparous Fishes 16 Systematics, Biogeography, and Evolution Figure 1. The many, unique characteristics of the Diagrammatic representation of two atherinomorph egg are correlated with reproduc- different testis-types in higher tive and developmental modifications. Filaments teleosts: A: testis lobule in representative atherinomorph, with are derived from the secondary or outer layer of spermatogonia restricted to the the pellucida which is secreted by follicle cells distal end of the lobule; B: testis (Wourms, 1976; Wourms and Sheldon, 1976; lobule in representative non-atherinomorph higher teleost, Loureiro and deSá, 1996). Filaments vary in with spermatogonia distributed number, shape, and relative length (e.g., Able, throughout the length of the testis 1984; Collette et al., 1984; White et al., 1984; lobule (from Grier et al., 1980: Fig. 1) Loureiro and deSá, 1996). Oil globule number ranges from one to over 100 (White et al., 1984: table 93). The relatively long developmental pe- riod in both oviparous and viviparous taxa is cor- related with direct development, that is, loss or reduction of a distinct larval stage, most notably in cyprinodontiforms and beloniforms (Rosen, 1964:253). sence of diapause results from “...developmen- The long developmental period and relatively tal switches between alternative stabilized path- large, desiccation-resistant egg is correlated with ways” (Hrbek and Larson, 1999:1200). This was the evolution of delayed hatching (Martin, said another way by Parenti (1981:364): “...the 1999) or diapause (Wourms, 1972). Delay of annual habit is no more than an exaggeration, hatching of fertilized eggs has been reported in due to extreme environmental fluctuations, of a the atheriniform of the Leu- capability of all cyprinodontiforms to survive resthes, and the cyprinodontiforms stress that involves desiccation”. Now, with our heteroclitus, F. confluentus, and Adinia xenica increased knowledge of delayed hatching pat- (Martin, 1999). Delay of hatching in these taxa terns, I would rewrite that sentence by substitut- is facultative; fertilized, fully-developed eggs can ing “atherinomorphs” for “cyprinodontiforms” hatch, but are stranded in relatively dry habitats (see also Parenti, 1993). Delayed hatching pat- and must await waves, tides, or rains to stimu- terns of atherinomorphs, whether facultative or late hatching (Martin, 1999; Griem and Mar- obligatory, may be homologous and represent tin, 2000). This is in contrast to fertilized eggs another atherinomorph synapomorphy. of annual killifishes, which undergo diapause The testis character proposed as an atherino- and for which delay of hatching is obligatory morph synapomorphy by Rosen and Parenti (Wourms, 1972). Stranding of fertilized eggs out (1981) had been described just the year before of water has been reported also in the Baja Cali- by Harry Grier and colleagues (Grier et al., fornia endemic, Fundulus lima, by Brill (1982). 1980; Fig. 1) who reported this distinctive tes- Facultative delayed hatching of teleost eggs has tis-type in 31 atherinomorph represent- been reported outside atherinomorphs only in ing each of the three orders. Since Grier et al. the lower teleost osmeroid Galaxias maculatus (1980), the atherinomorph testis has been re- (Martin, 1999). ported or confirmed in a total of 79 atherino- The evolutionary relationship between fac- morph species (Parenti and Grier, 2004), ultative and obligatory delay of hatching in including the beloniform adrianichthyid atherinomorphs is unknown. Phylogenetic Horaichthys setnai (Grier, 1984), seven species analyses based on molecules (e.g., Murphy and of atheriniform phallostethids (Grier and Collier, 1997) or morphology (e.g., Costa, 1990, Parenti, 1994), an anablepid, Jenynsia multi- 1998a) have led to various, sometimes conflict- dentata (Martínez and Monasterio de Gonzo, ing, conclusions concerning single or multiple 2002), the internally-fertilizing genus, origins of developmental diapause. A molecular (Grier and Collette, 1987), and phylogenetic analysis of the family Rivulidae by the viviparous halfbeaks, , Hemi- Hrbek and Larson (1999: Fig. 4) supported the rhamphodon and (Downing hypothesis that diapause was present in two dis- and Burns, 1995; Meisner and Burns, 1997a). tantly related groups of South American killi- A possibly similar testis-type in viviparous sur- fishes. Rather than interpret the cladogram fperches, family Embiotocidae, was noted by literally, they considered it unlikely that diapause Grier et al. (1980) but dismissed a year later had originated twice, but that presence or ab- by Grier (1981).

Lynne R. Parenti • The Phylogeny of Atherinomorphs. Reproductive System 17 Figure 2. et al., 1996). Approximately 200 teleost species, Phylogenetic relationships among both freshwater and marine, representing all ma- the three orders of atherinomorph fishes, following Rosen and Parenti Atheriniformes jor teleost lineages, were surveyed by Tsuneki (1981), Collette et al. (1984), Parenti A (1992) for presence or absence of the saccus (1993), Dyer and Chernoff (1996), and vasculosus and extent of its development when this paper. Synapomorphies are: A) Cyprinodontiformes testis a restricted spermatogonial B present. A well-developed saccus vasculosus was type; egg demersal, with several to Beloniformes considered to be a generalized condition, present many chorionic filaments, and in some , Anguilliformes, and several oil globules that coalesce at the vegetal pole; coupling during ostariophysans among lower teleosts, and mating; prolonged developmental Mugiliformes, , Scorpaeni- period; separation of embryonic Evolution of viviparity is considered indepen- formes, , Pleuronectiformes, and afferent and efferent circulation by development of heart in front of dent in atherinomorphs and embiotocids (see among higher teleosts. The head; ossified portion of ethmoid Lydeard, 1993). The phylogenetic significance of saccus vasculosus is reduced or absent in a vari- region of skull highly reduced; reproductive characters in embiotocids, includ- ety of taxa, including atherinomorphs, African infraorbital series represented by the lacrimal, dermosphenotic, and ing both egg and testis, are evaluated relative to , some anabantoids, and a synbranchid two, one or no anterior infraorbital those of atherinomorph fishes elsewhere in this among higher teleosts. bones; lateral process of pelvic bone volume (Grier et al., this volume). The derived The most “clear-cut” result of this survey, and distal end of pleural rib in close association, and, in some taxa, testis and egg are correlated in atherinomorphs according to Tsuneki (1992:74), is that absence connected via a ligament; with a vast array of reproductive modifications of the saccus vasculosus is characteristic of supracleithrum reduced or absent; including coupling during mating, relatively long atherinomorphs, and I agree. Fifteen atherino- dorsal portion of gill arches with a large fourth epibranchial the developmental period, sperm-bundle formation, morphs surveyed, representing all three orders, prominent supporting bone and no internal fertilization, superfetation, embryo re- from both freshwater and un- fourth pharyngobranchial element; tention, diapause, delayed hatching, hermaphro- ambiguously lack the saccus vasculosus. Its ab- medial hooklike projection and ventral flange on the fifth ditism, and live-bearing. Understanding the sence is proposed here as an atherinomorph ceratobranchial bone; supraneural phylogeny of atherinomorph fishes will be en- synapomorphy. bones absent; superficial division of hanced by an understanding of their derived re- A sixteenth atherinomorph synapomorphy, adductor mandibulae with two tendons, one inserting on the productive modifications and evolution of this of the oocyte, was proposed and illustrated by maxilla, a second inserting on the novel reproductive system. Parenti and Grier (2004: Figs. 3,4). Atherino- lacrimal bone; olfactory sensory The purpose of this paper is to review the morph yolk is fluid, rather than granular, epithelium arranged in sensory islets; absence of the saccus current state of our knowledge of atherino- throughout vitellogenesis. vasculosus; B) second infraorbital morph phylogeny, with a focus on live-bearing bone absent; first epibranchial bone taxa, summarizing molecular and morphologi- Atherinomorph with an expanded base and no separate uncinate process; first cal cladistic analyses published during the past pharyngobranchial element absent; two decades. Some phylogenetic analyses of The Series Atherinomorpha was classified as sis- second and third epibranchials solely oviparous taxa are cited but not discussed ter to the Series by Rosen and smaller than the first and fourth; stomach, pyloric caecae and in detail. It is not my goal to summarize all Parenti (1981) without an explicit proposal of pneumatic duct absent cladistic analyses of atherinomorph taxa, which relationship to any particular percomorph ta- is well beyond the scope of this review. xon. In diagnosing atherinomorphs, Rosen (1964) made deliberate comparisons with taxa Atherinomorph Monophyly such as mullets (Mugilidae) or percopsiform fishes that had at one time been proposed as Atherinomorph monophyly was reviewed by closely related to one or another atherinomorph Parenti (1993) who listed 14 diagnostic charac- order (see Parenti, 1993). In discussing ather- ters, five of which were of reproduction and inomorph sister-group relationships, Parenti development (Fig. 2, node A). The atherino- (1993: Fig. 1c) could come to no firm conclu- morph testis-type and associated reproductive sion, arguing that evidence linked atherino- modifications remain among the strongest evi- morphs to paracanthopterygians on one hand dence for monophyly, as argued above. and to percomorphs on the other. Here, I add a fifteenth character to the athe- The close relationship of mullets to ather- rinomorph diagnosis: absence of the saccus inomorphs was proposed by Stiassny (1990) and vasculosus, a hypothalamic circumventricular explored further by Stiassny (1993) who, al- organ of unspecified function (Tsuneki, 1992). though she concluded that there was good evi- The saccus vasculosus has been reported to pro- dence for a sister group relationship, discussed duce a parathyroid hormone-related protein in numerous characters that contradicted that pro- the perciform bream, Sparus aurata (Devlin posal. Of seven characters in support of a mul-

• Viviparous Fishes 18 Systematics, Biogeography, and Evolution let-atherinomorph sister-group relationship morph testis-type (not found in mullets; Grier (Stiassny, 1993: Fig. 1), four were of the pecto- et al., 1980), absence of the saccus vasculosus ral girdle, two of the branchial muscles, and one (well-developed in mullets; Tsuneki, 1992), or of the vertebral column. Three char- innervation of the pectoral and pelvic fin acters were considered to have reversed in ather- muscles by branches of spinal nerve 2 (present inomorphs (viz. Stiassny and Moore, 1992). The and derived in mullets, absent in atherino- study was not intended as an exhaustive review morphs; Parenti and Song, 1996). Detailed of acanthomorph relationships, however, be- molecular analyses of acanthomorph phylogeny cause, for example, another reversal in athe- based on molecular data alone (e.g., Miya et al., rinomorphs would be loss of transforming 2003) confirm atherinomorph monophyly, but ctenoid scales (Roberts, 1993). include other taxa, such as blennioids and go- The hypothesis of a close -atherino- biesocids, along with mullets, as putative ather- morph relationship was taken a step further by inomorph sister taxa. Broader surveys of Johnson and Patterson (1993) who proposed a morphology, as well as molecules, are needed to new taxon, the Smegmamorpha, to include further test monophyly of smegmamorphs and synbranchoid , mastacembeloid eels, the evaluate the hypothesis of a sister group rela- centrarchid Elassoma, gasterosteiforms, mullets, tionship of mullets and atherinomorphs. and atherinomorphs. A single character was pro- posed for smegmamorph monophyly (Johnson Cyprinodontea and Patterson, 1993:572): the first two epineu- ral bones originate at the tip of transverse pro- Relationships among the three groups of ather- cesses or fused parapophyses on the first two inomorph fishes were unspecified by Rosen vertebral centra. In addition, Johnson and (1964). Monophyly of each group and the rela- Patterson (1993:table 2) tabulated selected, tionships among them were considered by Rosen shared derived characters found in some, but and Parenti (1981:21-23) who proposed a sister not all smegmamorphs, such as, for example, group relationship between Cyprinodontiformes infraorbital series with three or fewer bones be- and Beloniformes, together called Division II tween the lacrimal and the dermosphenotic, atherinomorphs. Four characters were proposed present in atherinomorphs, Elassoma, gaster- to support monophyly of Division II (Rosen and osteiforms, and synbranchoids, but absent in Parenti, 1981:21; Fig. 2, node B). A fifth charac- mullets and mastacembeloids. ter may be added from Li (2001:585): absence of Smegmamorph monophyly was one hypo- a stomach, including absence of pyloric caecae thesis tested by Wiley et al. (2000) in a total and a pneumatic duct. evidence analysis of acanthomorph phylogeny The sister group relationship between combining molecular and morphological data. Cyprinodontiformes and Beloniformes has been The strict consensus of 137 equally parsimoni- corroborated (Stiassny, 1990; Saeed et al., 1994; ous based on morphological data recov- Dyer and Chernoff, 1996; Wiley et al., 2000; ered a monophyletic Atherinomorpha with Li, 2001). The relationship was recognized by unresolved relationships to an array of higher Nelson (1984:214) who in his second edition taxa (Wiley et al., 2000: Fig. 8c). A strict con- of Fishes of the World synonymized the two or- sensus of four equally parsimonious trees based ders in an expanded Cyprinodontiformes. This on combined molecular and morphological data decision was reversed in the third edition recovered a group that included the mullet (Nelson, 1994:264, with the three orders Ather- Mugil and the four atherinomorph taxa analyzed iniformes, Beloniformes, and Cyprinodonti- (the atheriniforms and Atherino- formes recognized without inter-ordinal morus, the beloniform , and the relationships expressed). The current edition of cyprinodontiform ), but did not re- Fishes of the World is used worldwide as a stan- cover a monophyletic Smegmamorpha (Wiley dard reference for teleost classification, however, et al., 2000: Fig. 6). Mugil was considered to be and an enlarged order Cyprinodontiformes, in- more closely related to the two atheriniforms cluding beloniforms, was incorporated into nu- than either is to the cyprinodontiform or the merous publications, particularly during the beloniform; that is, atherinomorph monophyly decade between 1984 and 1994 (e.g., Tsuneki, was refuted as well. Morphological characters 1992; Kottelat et al., 1993). The superorder surveyed by Wiley et al. (2000), however, did Cyprinodontea was proposed by Dyer and not include characters such as the atherino- Chernoff (1996) as a formal name for Rosen

Lynne R. Parenti • The Phylogeny of Atherinomorphs. Reproductive System 19 Figure 3. and Parenti’s (1981) Division II, the Cyprino- Phylogenetic relationships among Aplocheilidae families of Cyprinodontiformes as dontiformes and Beloniformes, and is adopted proposed by Parenti (1981) Rivulidae here (Table 1). Cyprinodontiformes

Profundulidae The first explicitly cladistic analysis of cypri- nodontiform fishes was based largely on os- Fundulidae teological characters (Parenti, 1981; Fig. 3). Cyprinodontiform monophyly was corroborated Valenciidae by six complex characters (Parenti, 1981: Fig. 9, node a), and it has been challenged only by Li (2001; see comments below under Beloniformes). Figure 4. One cyprinodontiform character of Parenti Phylogenetic relationships among Anablepidae (1981), prolonged embryonic development, is select genera of Cyprinodontiformes discussed above as an atherinomorph synapo- as proposed by Meyer and Lydeard (1993) based on partial DNA Poeciliidae morphy (see also Parenti, 1993). A major conclu- sequences of the tyrosine kinase sion of Parenti’s (1981) study that refuted earlier gene X-src. Genera in boxes notions of cyprinodontiform relationships was represent families or subfamily groupings of live-bearing taxa and that viviparity had evolved at least three times their close relatives, as also Goodeidae within the order, not once. That is, oviparous sis- proposed by Parenti (1981), from top ter groups of each group of viviparous taxa were to bottom, family Goodeidae, family Anablepidae, and subfamily Cyprinodontidae hypothesized. Also, cyprinodontiforms were di- vided into two monophyletic suborders, the Aplocheiloidei and Cyprinodontoidei. This pro- posal of relationships was tested by Meyer and Lydeard (1993; Fig. 4) using partial DNA se- quences of the tyrosine kinase gene X-src, an oncogene chosen in part because cyprinodonti- Rivulus forms of the genus have long been known to inherit melanomas (see Schartl, 1995). Fundulus No non-cyprinodontiform taxon was used as an outgroup by Meyer and Lydeard, so their analysis was not a test of cyprinodontiform monophyly. Zoogoneticus Also, at least one pivotal taxon, the oviparous Oxyzygonectes, sister to the viviparous Anableps and Jenynsia in the Anablepidae, according to Anableps Parenti (1981), was not included. Nonetheless, Jenynsia the maximum parsimony phylogenetic hypoth- esis of Meyer and Lydeard (1993) corroborated Parenti’s (1981) conclusions that viviparity had evolved at least three times within cyprinodonti- Xiphophorus forms, in anablepids, goodeids, and poeciliids. Xiphophorus Further, the analysis recovered the controversial sister group relationship of the viviparous good- eids and their oviparous relatives, and Crenichthys (see also Grant and Riddle, 1995). Cyprinodontiforms were used as a test taxon Aplocheilichthys by Parker (1997) to evaluate the consequences Aplocheilichthys of combining morphological and molecular data in a phylogenetic analysis. Morphological char- acters described by Parenti (1981, 1984a) are listed and character states coded in a detailed Cyprinodon appendix (Parker, 1997:184-185). Although this Jordanella study was not intended as a thorough re-analy-

• Viviparous Fishes 20 Systematics, Biogeography, and Evolution sis of Parenti’s (1981) hypothesis, some of Anablepidae Parker’s results anticipated those of other stud- ies. For example, Tomeurus, the ovoviviparous Aplocheilichthyinae poeciliine, had been considered primitive to vi- viparous poeciliines since Rosen and Bailey’s (1963) classic review, and this relationship was corroborated by Meyer and Lydeard (1993; Fig. cultratus 4). In contrast, a partitioned X-src dataset of first and second codon positions only, analyzed un- compressa der maximum parsimony (Parker, 1997: Fig. 2), supported the sister group relationship of To- Gambusia affinis meurus and Cnesterodon, a viviparous poeciliine, together considered derived, not , poeci- formosa liines. This relationship was corroborated by an array of synapomorphies in a parsimony analy- latidens sis of morphological characters by Ghedotti (2000; Fig. 5), including neural arch of first ver- metallicus tebra open dorsally, lateral processes on ventral portion of seventh, eighth, and ninth proximal anal-fin radials in adult males present and in contact, and distal tip of the gonopodium with amates elongate, bony processes. These morphological characters were of course known to Rosen and victoriae Bailey (1963) who likely accepted the constraint that an ovoviviparous taxon such as Tomeurus caudimaculatus was basal to a group of viviparous taxa. This constraint was relaxed by Parker (1997) and Cnesterodon decemmaculatus Ghedotti (2000) who independently recovered a novel hypothesis of relationships of Tomeurus. Tomeurus gracilis Division of Cyprinodontiformes into two sub- orders, Aplocheiloidei and Cyprinodontoidei, was corroborated by Costa (1998a; Fig. 6) who Aplocheilidae Figure 5. re-analyzed phylogenetic relationships Phylogenetic relationships among the poecilioid fishes, simplified from using morphology. The hypothesis of Costa Rivulidae Ghedotti (2000: Fig. 20). Species (1998a; Fig. 6) differs from that of Parenti (1981; listed are members of the subfamily Fig. 3) in the placement of Goodeidae as sister to Poeciliinae Profundulidae, not Cyprinodontidae, and the resolution of the relationships of Valenciidae. Profundulidae Both studies used osteology as a principal source of data, but differed in interpretation of signifi- Goodeidae cance of character states of the premaxilla, among others. The sister group relationship of Good- Fundulidae eidae and Profundulidae was recovered also in the molecular hypotheses of Meyer and Lydeard Valenciidae (1993) and Parker (1997). The close relationships of fundulids, goodeids, and profundulids is re- flected in the written classification of Cyprin- odontiformes (Table 1). Anablepidae Monophyly of each of the nine families rec- ognized by Parenti (1981) was corroborated by Poeciliidae Costa (1998a) and the terminal taxa of Meyer and Lydeard’s (1993; Fig. 4) hypothesis are also consistent with Parenti’s hypothesis. Monophyly Figure 6. Phylogenetic relationships among of the Anablepidae and Poeciliidae (sensu Cyprinodontidae families of Cyprinodontiformes as Parenti, 1981) was corroborated in Ghedotti’s proposed by Costa (1998a)

Lynne R. Parenti • The Phylogeny of Atherinomorphs. Reproductive System 21 Figure 7. Phylogenetic relationships among Adrianichthyidae cyprinodontids), and Webb et al. (2004, the families of beloniform fishes, livebearing Goodeidae). Morphological phylo- following Rosen and Parenti (1981) genies or surveys include Parenti (1984a, Ores- and Collette et al. (1984) tias), Chambers (1987, cyprinodontiform Exocoetidae gonopodia; 1990, cnesterodontin gonopodia), Ghedotti (1998, Anablepidae), Rauchenberger Hemiramphidae (1989, Gambusia), Costa (1996b, Fluviphylax; 1997, cyprinodontids), and Rodríguez (1997, ). Conclusions of these studies that bear on higher order relationships of cyprin- Belonidae odontiforms were summarized by Lazara (2001:ix-xiv). Scomberesocidae The genus Xiphophorus, the swordtails and platyfishes, serves as a model taxon in analyses of congruence of phylogenetic pattern with be- havior, development, morphology, and mol- (2000; Fig. 5) morphological analysis, whereas ecules (see, for example, Basolo, 1991; Haas, Meyer and Lydeards’ (1993; Fig. 4) analysis pro- 1993; Meyer et al., 1994, Marcus and McCune, posed a paraphyletic Poeciliidae. The dotted 1999, and Morris et al., 2001). lines in Meyer and Lydeard’s (1993; Fig. 4) cla- dogram indicate relationships for which they Beloniformes found weak support, however. Morphology and molecules do equally well, i.e. agree, in resolv- Beloniform monophyly was supported by seven ing relationships at the tips of the , but vary synapomorphies (Rosen and Parenti, 1981:17), in their ability to recover higher taxa. A mo- including absence of the interhyal bone, reduc- lecular phylogenetic analysis of aplocheiloids by tion or loss of the interarcual cartilage, presence Murphy and Collier (1997) conflicts, in part, of only a single, ventral hypohyal bone, modifi- with the proposals of Parenti (1981) and Costa cations of the gill arch skeleton, as well as a dis- (1998a) based on morphology. A stable classifi- tinctive caudal skeleton characterized by the cation of cyprinodontiforms, at the family level lower caudal lobe with more principal rays than and above, is within reach, however, and is ex- in the upper caudal lobe. pected to include components common to the Beloniform relationships were reviewed by above morphological and molecular analyses. Collette et al. (1984; Fig. 7) who accepted a Additional molecular phylogenies of families monophyletic Beloniformes including Rosen or other subgroups of aplocheiloids include and Parenti’s (1981) proposal that ricefishes those of Murphy and Collier (1999, Aphyo- (family Adrianichthyidae) are more closely re- semion and ), Murphy et al. lated to exocoetoids (families Exocoetidae, (1999a, West African aplocheiloids), and Hemiramphidae, Belonidae, and Scomberesoci- Murphy et al. (1999b, Rivulidae). Morphologi- dae) than to cyprinodontiforms. This proposal cal phylogenies or surveys include Costa (1990, has been criticized recently by Li (2001) who 1998b, Rivulidae; 1995a, ; 1995b, argued that adrianichthyoids are more closely Cynolebiatinae; 1996a, ), related to Cyprinodontiformes than to exo- Loureiro and deSá (1998, Cynolebias), and Aarn coetoids. In particular, Li (2001) claimed that and Shepherd (2001, epiplatines). , some characters proposed as diagnostic of biology, and conservation status of Brazilian Beloniformes sensu lato, although absent in the annual killifishes was reviewed by Costa (2002, primitive cyprinodontiform suborder Aplochei- and references therein). loidei, are present in the derived suborder Molecular phylogenies of families or other Cyprinodontoidei. For example, Li (2001:584) subgroups of cyprinodontoids include Bernardi rejected Rosen and Parenti’s (1981) beloniform (1997, Fundulidae), Breden et al. (1999, Poeci- synapomorphy of a single, ventral hypohyal be- lia), Grady et al. (2001, Fundulus), Hamilton cause cyprinodontoids have a single hypohyal. (2001, ), Hrbek and Meyer, 2003 (Apha- If we accept cyprinodontiform monophyly, nius) Lüssen et al. (2003, ), Lydeard et however, which is supported by a symmetrical al. (1995, Gambusia), Mojica et al. (1997, caudal fin skeleton, absent in adrianichthyoids, ), Parker and Kornfield (1995, and first pleural rib on the second, rather than

• Viviparous Fishes 22 Systematics, Biogeography, and Evolution Figure 8. the third vertebra, among other characters, then matanensis Phylogenetic relationships among similarities between cyprinodontoids and beloniform fishes, simplified from adrianichthyoids must be interpreted as conver- Lovejoy (2000: Fig. 2). The families Exocoetidae Exocoetidae and Scomberesocidae gent rather than indicative of close relationship. were each considered Cyprinodontiforms and beloniforms, sensu hemiramphids monophyletic; the families Rosen and Parenti (1981), are readily distin- Hemiramphidae and Belonidae paraphyletic. “Hemiramphids (IF)” guished by their distinct caudal fin skeletons. hemiramphids refers to the four internally fertilizing Finally, both cyprinodontiforms and adrianich- genera, Zenarchopterus, Nomorhamphus, Dermogenys and thyoids were characterized by Li (2001) as lack- hemiramphids (IF) ing elongate jaws, whereas Parenti (1987, 1989a) argued that the elongate jaws of the belonids large-bodied adrianichthyoids of Sulawesi, some of which have been called ‘duck-billed’ repre- belonids sent additional support for their close relation- ship to exocoetoids. Despite rejection of Li’s belonids (2001) hypothesis, I appreciate his comments as they call for continued critique of the mono- belonids phyly of within the Atherinomorpha as recognized herein. Scomberesocidae Higher-order beloniform sensu lato relation- ships were reviewed by Lovejoy (2000; Fig. 8) who combined data from nuclear and mito- chondrial gene sequences with morphology in a Monophyly of the internally-fertilizing total evidence analysis. The families Exocoetidae halfbeaks, genera Zenarchopterus, Hemirham- (flyingfishes) and Scomberesocidae (sauries) phodon, Dermogenys, and Nomorhamphus, was were considered monophyletic by both Collette supported by the morphological studies of Ander- et al. (1984; Fig. 7) and Lovejoy (2000; Fig. 8). son and Collette (1991), Downing and Burns The families Hemiramphidae (halfbeaks) and (1995), Meisner and Burns (1997b), Meisner and Belonidae (needlefishes), considered monophyl- Collette (1999), and Meisner (2001). Petalichthys etic by Collette et al. (1984), were hypothesized and were included in a reanaly- to be paraphyletic by Lovejoy (2000). sis of beloniform phylogeny by Lovejoy et al. No non-beloniform taxon was included in (2004) and of hemiramphids and Lovejoy’s analysis and a single adrianichthyid spe- belonids corroborated. The last three halfbeak cies (Oryzias matanensis) was an outgroup to the genera are viviparous and together form a mono- ingroup exocoetoids. Therefore, the analysis was phyletic group as corroborated by these analyses. not a test of beloniform or of exocoetoid mono- A fifth genus, the monotypic Tondanichthys, de- phyly. Further, some intriguing taxa were not in- scribed from the type series of ten specimens that cluded in the study, such as the southern African does not include a mature male, is inferred to be , Petalichthys, which remains in a internally fertilizing (Collette, 1995; Meisner and “halfbeak stage” of development for a relatively Collette, 1999). long time before both upper and lower jaws be- Reproductive biology has been used to infer come elongate (Collette et al., 1984:342; phylogenetic relationships among live-bearing Boughton et al., 1991). Also missing was the halfbeaks. Dermogenys is diagnosed by large hemiramphid Oxyporhamphus recently reclassi- sperm bundles and intrafollicular development; fied as a flyingfish, family Exocoetidae, based on whereas, Nomorhamphus is diagnosed by small a re-analysis of morphology (Dasilao et al., 1997). sperm bundles and a long intraluminal devel- Nonetheless, Lovejoy’s (2000; Fig. 8) hypothesis opmental period (Downing and Burns, 1995; offers a novel reinterpretation of traditional Meisner and Burns, 1997b; Meisner and beloniform morphology and invites further study Collette, 1999; Meisner, 2001). These generic of the relationship between morphological and limits are in contrast to those of Brembach molecular data as used in phylogenetic analyses. (1991). Molecular sequences and morphology were com- females are known to carry bundles bined in a total evidence analysis of phylogenetic of fertilized eggs until hatching (Fig. 9), rather relationships of New World needlefishes by than depositing them on over-hanging vegeta- Lovejoy and Collette (2001). tion or the substrate. Aquarium-maintained O.

Lynne R. Parenti • The Phylogeny of Atherinomorphs. Reproductive System 23 Figure 9. , USNM 313908, female, 23.8 mm SL, with cluster of fertilized eggs. Photo by H. H. Tan

Figure 10. sarasinorum, CMK 6557, female, 53.4 mm SL. Above, embryo cluster held to body posterior to pelvic fins. Below, same specimen, chorion measures approximately 2 mm in diameter Fig. 9

nigrimas females reportedly carry large bundles of eggs, but deposit them among or on the substrate soon after spawning (Kottelat, 1990a:54). In spawning in open water, ricefishes are similar to the above mentioned Fundulus lima (Brill, 1982). Embryos in the clusters are relatively well developed, with large, well- formed eyes and pigmented bodies, and appear near hatching (Fig. 10). Because these fertilized eggs may be carried until hatching, Kottelat (1990a:62) proposed that ricefishes be consid- ered a distinct reproductive guild for which he coined the term “pelvic brooders.” This repre- sents one of the few cases of parental care in oviparous atherinomorphs. Females carrying clusters of fertilized eggs, long known in the medaka, Oryzias latipes (Yamamoto, 1975:7), has been reported in at least eight other ricefish species, O. dancena (Fig. 9), O. nigrimas (Kot- telat, 1990a:54) X. oophorus (Kottelat, 1990a: Fig. 6), X. sarasinorum (Fig. 10), O. marmoratus (Kottelat, 1990b: Fig. 5), O. matanensis (Kotte- lat, 1990b:161), O. javanicus (BMNH 1970. 7.22:38-39), O. luzonensis (Blanco, 1947) and is likely to occur in others. The Indian ricefish, Horaichthys setnai, is in- ternally fertilizing and lays fertilized eggs (Kulkarni, 1940). Internal fertilization and em- bryo retention is facultative in some ricefishes. Facultative embryo retention was reported in the medaka, Oryzias latipes, by Amemiya and Murayama (1931). One specimen of Adrianich- thys kruyti, a large, pelagic ricefish from Lake Poso, Sulawesi, was reported to be hermaphro- ditic, having both testis and ovary, by Klie Fig. 10 (1988, in Kottelat, 1990a:57). Much informa- tion available on the reproduction of the large

• Viviparous Fishes 24 Systematics, Biogeography, and Evolution Figure 11. ricefishes is anecdotal (see Weber and de Beau- Atherinopsidae Phylogenetic relationships among fort, 1922; Kottelat, 1990a; Rosen, 1964). More the subfamilies, families or tribes of detailed study of the reproductive morphology atheriniform fishes as proposed by Notocheiridae Dyer and Chernoff (1996), following of the large Sulawesi ricefishes is needed to de- Dyer (1998: Fig. 1d) termine the extent of internal fertilization, em- Bedotiinae bryo retention, and hermaphroditism in the Adrianichthyidae. Melanotaeniinae

Atheriniformes Telmatherinini

Since 1981, there have been several, solely mor- Pseudomugilini phological, phylogenetic analyses of atherini- form fishes (White et al., 1984; Stiassny, 1990; Atherionidae Saeed et al., 1994; and Dyer and Chernoff, 1996; Fig. 11). Monophyly of atheriniforms was Dentatherininae not supported by Rosen and Parenti (1981) or Parenti (1984b), but it has been argued for Phallostethinae strongly in these other studies. Atheriniform monophyly was supported by Atherinomorinae two developmental characters by White et al. (1984:357): short preanal length of flexion lar- Craterocephalinae vae and a single row of melanophores on the dorsal margin of larvae. Eight adult characters Atherininae were added to the diagnosis by Dyer and Chernoff (1996:1): vomerine ventral face con- cave, long A1 muscle tendon to lacrimal, two anterior infraorbital bones, pelvic-rib ligament, cause it is more popular than the vernacular pelvic medial plate not extended to anterior end, “atherinoids.” and second dorsal-fin spine flexible. Phylogenetic analyses of atheriniform sub- This atheriniform diagnosis requires ho- groups include an analysis of the internally moplasy within several characters. For example, fertilizing freshwater and coastal family Phal- newly hatched adrianichthyids also have a single lostethidae, proposed as sister taxon of the ma- row of dorsal melanophores (White et al., rine Dentatherina (Parenti, 1984b, 1989b). This 1984:359). Number of anterior infraorbital bones sister group relationship was corroborated by ranges from one to three in atheriniforms (Dyer Dyer and Chernoff (1996) who classified and Chernoff, 1996). The three anterior infraor- Dentatherina in an expanded Phallostethidae. bital bones of the rainbowfishes Melanotaenia and Atheriniformes are oviparous, although report were considered evidence of their of facultative embryo retention would not be sister-group relationship by Dyer and Chernoff surprising. Phallostethids (sensu Parenti, 1989b) (1996:67), whereas Rosen and Parenti (1981) are internally fertilizing and lay fertilized eggs. considered the character to be primitive for Internal fertilization was reported in the brook atherinomorphs. Outgroups of atheriniforms in silverside, sicculus, by Grier et al. Dyer and Chernoff’s (1996: Table 2) data matrix (1990). include representative cyprinodontiform and Other morphological phylogenies of sub- beloniform taxa and the mullet, Mugil. Given the groups of Atheriniformes include Chernoff strong support for atherinomorph monophyly (1986, menidiines), Saeed et al. (1989, Pseudo- and the tentative support for a mullet-ather- mugilidae), and Dyer (1998, Atherinopsidae). inomorph sister group relationship, as argued above, additional acanthomorph outgroups may Conclusions yield alternate interpretations of phylogeny when included in a parsimony analysis of either mor- The series Atherinomorpha is a well-corrobo- phological or molecular data. rated, monophyletic taxon. As for other such Despite the ambiguity in distribution of char- well-corroborated taxa, the list of characters acters discussed above, I use the formal term diagnostic of the Atherinomorpha continues to order Atheriniformes for the included taxa be- grow (Rosen and Parenti, 1981; Parenti, 1993;

Lynne R. Parenti • The Phylogeny of Atherinomorphs. Reproductive System 25 herein). Each of its included orders, the Cyprin- Acknowledgments and abbreviations odontiformes, Beloniformes, and Atherini- formes is monophyletic, although the quality I am indebted to the organizers, Mari Carmen and quantity of support for each is variable. Uribe, Raúl Pineda, Topiltzin Contreras and Molecular tests of morphological hypotheses Harry J. Grier for inviting me to participate in of atherinomorph relationships agree, in large the II International Symposium on Livebearing part. Where they differ, molecules present novel Fishes. Research on ricefish systematics was sup- hypotheses of relationship, most notably in ported, in part, by NSF grant BSR 87-00351. beloniforms (Lovejoy, 2000), and invite rein- Heok Hui Tan, National University of Singa- terpretations of our traditional understanding pore, photographed the ricefish in figure 8. of morphological characters. Not all molecular Maurice Kottelat, Switzerland, lent specimens in analyses agree, however. Two phylogenetic hy- his care and discussed ricefish systematics. Dar- potheses based on partial sequences of the rell Siebert allowed access to specimens at the oncogene, X-src, yielded different interpretations BMNH. My colleagues, Ralf Britz, Bruce Col- of relationships of the poeciliid Tomeurus. No- lette, Carlos Figueiredo, Dave Johnson, Amy tably, Parker’s (1997: Fig. 2) hypothesis of the Meisner, Joe Nelson, and Melanie Stiassny read sister group relationship of Tomeurus and and provided helpful comments on earlier ver- Cnesterodon was corroborated in a reinterpreta- sions of the manuscript. tion of morphological data by Ghedotti (2000). Institutional abbreviations: BMNH, The Molecular and morphological hypotheses may Natural History Museum, London; USNM, inform each other and point to weak areas in United States National Museum of Natural His- the other analysis. The Cyprinodontiformes has tory, Washington, DC; CMK, Collection of been studied most intensively during the past Maurice Kottelat, Cornol, Switzerland. twenty years; morphological and molecular analyses share repeated statements of relation- ship to such a degree that a stable classification of the order is within reach. Monophyly of the order Atheriniformes has been supported by numerous morphological studies, most recently that of Dyer and Chernoff (1996). Interpretation of polarity of the charac- ters used by Dyer and Chernoff (1996) to diag- nose atheriniforms hinges on the choice of an atherinomorph outgroup, however. Their choice of mullets may be appropriate, but a broader range of outgroup taxa may yield other interpre- tations of atheriniform relationships. A molecu- lar test of the Dyer and Chernoff (1996) hypothesis may reveal novel relationships as well. As for cyprinodontiforms and beloniforms, how- ever, a molecular hypothesis of atheriniforms will not substitute for that based on morphology. Atherinomorph fishes likely will continue to be studied broadly, especially as they include nu- merous model organisms, such as the poeciliids Xiphophorus, Poecilia, and Gambusia, the ricefishes Oryzias, as well as a wide range of live- bearing teleost fish taxa in the families Goodeidae and Anablepidae and other poeciliids. Phyloge- netic hypotheses of atherinomorphs and their subgroups are among the most examined and re-examined within bony fishes, which makes their continued use as model organisms and their unique role in understanding reproduction even more justifiable.

• Viviparous Fishes 26 Systematics, Biogeography, and Evolution References Collette BB. 1995. Tondanichthys kottelati, a new genus and species of freshwater halfbeak (Teleostei: Hemir- Aarn, Shepherd AM. 2001. Descriptive of amphidae) from Sulawesi. Ichthyol Explor Freshwa- sexfasciatus (Cyprinodontiformes: Aplocheili- ters 6:171-174. dae) and a phylogenetic analysis of Epiplatina. Cy- Collette BB, McGowen GE, Parin NV, Mito S. 1984. bium 25:209-225. Beloniformes: development and relationships. In: Able KW. 1984. Cyprinodontiformes: development. In: Moser HG, Richards WJ, Cohen DM, Fahay MP, Moser HG, Richards WJ, Cohen DM, Fahay MP, Kendall AW, Jr., Richardson SL. editors, Ontogeny Kendall AW, Jr., Richardson SL, editors, Ontogeny and systematics of fishes. Special Publication No. 1, and systematics of fishes. Special Publication No. 1 Supplement to Copeia. American Society of Ichthy- Supplement to Copeia. American Society of Ichthy- ologists and Herpetologists. p 335-354. ologists and Herpetologists. p 362-368. Costa WJEM. 1990. Análise filogenética da família Amemiya I, Murayama S. 1931. Some remarks on the Rivulidae (Cyprinodontiformes, Aplocheiloidei). Rev existence of developing embryos in the body of an Brasil Biol 50:65-82. oviparous cyprinodont, Oryzias () latipes Costa WJEM. 1995a. Revision of the neotropical annual (Temminck and Schlegel). Proc Imp Acad Japan fish genus Cynopoecilus (Cyprinodontiformes: Rivuli- 7:176-178. dae). Copeia. 1995:456-465. Anderson WD III, Collette BB. 1991. Revision of the Costa WJEM. 1995b. Pearl Killifishes, the Cynolebia- freshwater viviparous halfbeaks of the genus Hemir- tinae, systematics and biogeography of the neotropical hamphodon (Teleostei: Hemiramphidae). Ichthyol annual fish subfamily (Cyprinodontiformes: Rivuli- Explor Freshwaters 2:151-176. dae). TFH Publications, Inc. Neptune City, NJ. Basolo AL. 1991. Male swords and female preferences. Costa WJEM. 1996a. Phylogenetic and biogeographic Science 253:1426-1427. analysis of the neotropical annual fish genus Simpson- Bernardi G. 1997. Molecular phylogeny of the Funduli- ichthys (Cyprinodontiformes: Rivulidae). J Comp Biol dae (Teleostei, Cyprinodontiformes) based on the cy- 1:129-140. tochrome b gene. In: Molecular Systematics of Fishes. Costa WJEM. 1996b. Relationships, monophyly and Kocher TD, Stepien CA, editors, Academic Press, San three new species of the neotropical miniature poeci- Diego. p 189-197. liid genus Fluviphylax (Cyprinodontiformes: Cyprin- Blanco GJ. 1947. The breeding activities and embryology odontoidei). Ichthyol Explor Freshwaters 7:111-130. of Aplocheilus luzonensis Herre and Ablan. Philippine Costa WJEM. 1997. Phylogeny and classification of the J Sci 77:89-93. Cyprinodontidae revisited (Teleostei: Cyprinodonti- Boughton DA, Collette BB, McCune AR. 1991. Hetero- formes): Are Andean and Anatolian killifishes sister chrony in jaw morphology of needlefishes (Teleostei: taxa? J Comp Biol 2:1-17. Belonidae). Syst Zool 40:329-354. Costa WJEM. 1998a. Phylogeny and classification of the Breden F, Ptacek MB, Rashed M, Taphorn D, Figueiredo Cyprinodontiformes (: Atherinomorpha) a CA. 1999. Molecular phylogeny of the live-bearing reappraisal. In: Malabarba LR, Reis RE, Vari RP, fish genus Poecilia. Mol Phyl Evol 12:95-104. Lucena ZMS, Lucena CAS, editors, Phylogeny and Brembach M. 1991. Systematik und Fortplflanzungsbiologie Classification of Neotropical Fishes, EDIPUCRS der lebendbärenden Halbschnäbler der gattung Der- (Editora Pontifícia Universidade Católica do Rio mogenys und Nomorhamphus (Hemirhamphidae [sic], Grande do Sul), Porto Alegre, . p 537-560. Pisces). Verlag Natur Wissenschaft Solingen. 201 pages. Costa WJEM. 1998b. Phylogeny and classification of Brill, JS, Jr. 1982. Observations on the unique reproduc- Rivulidae revisited: origin and evolution of annualism tive behavior of Fundulus lima Vaillant a killifish from and miniaturization in rivulid fishes (Cyprinodonti- Baja, [sic] California. Freshwater Mar Aquar 5:9-10, formes: Aplocheiloidei). J Comp Biol 3:33-92. 12-15, 74-76, 78, 79, 82, 84-87, 90. Costa WJEM. 2002. Peixes Anuais Brasileiros. Diversi- Chambers J. 1987. The cyprinodontiform gonopodium, dade e Conservação. Série Pesquisa no. 60. Editora da with an atlas of the gonopodia of the fishes of the ge- UFPR, Curitiba. nus Limia. J Fish Biol 30:389-418. Dasilao JC Jr, Sasaki K, Okamura O. 1997. The hemir- Chambers J. 1990. The gonopodia of the fishes of the amphid Oxyporhamphus is a flyingfish (Exocoetidae). tribe Cnesterodontini (Cyprinodontiformes, Poeci- Ichthyol Res 44:101-107. liidae). J Fish Biol 36:903-916. Devlin AJ, Danks JA, Faulkner MK, Power DM, Canario Chernoff B. 1986. Phylogenetic relationships and reclas- AVM, Martin TJ, Ingleton PM. 1996. Immuno- sification of menidiine silverside fishes with emphasis chemical detection of a parathyroid hormone-related on the tribe Membradini. Proc Acad Nat Sci Phila protein in the saccus vasculosus of a teleost fish. Gen 138(1):189-249. Comp Endocrin 101:83-90.

Lynne R. Parenti • The Phylogeny of Atherinomorphs. Reproductive System 2727 Downing AL, Burns JR. 1995. Testis morphology and Grier HJ, Moody DP, Cowell BC. 1990. Internal fertili- spermatozeugma formation in three genera of vivipa- zation and sperm morphology in the , rous halfbeaks: Nomorhamphus, Dermogenys, and Labidesthes sicculus (Cope). Copeia 1990:221-226. Hemirhamphodon. J Morph 225:329-343. Grier HJ, Parenti LR. 1994. Reproductive biology and Dyer BS. 1998. Phylogenetic systematics and historical systematics of phallostethid fishes as revealed by gonad biogeography of the family structure. Environ Biol Fishes 41:287-299. Atherinopsidae (Teleostei: Atheriniformes). In: Mala- Grier HJ, Uribe MC, Parenti LR, De la Rosa-Cruz G. 2005. barba LR, Reis RE, Vari RP, Lucena ZMS, Lucena Fecundity, the germinal epithelium, and folliculogenesis CAS, editors, Phylogeny and classification of Neotro- in viviparous fishes. In: this volume. p 191-216. pical Fishes, EDIPUCRS (Editora Pontifícia Univer- Haas V. 1993. Xiphophorus phylogeny, reviewed on the sidade Católica do Rio Grande do Sul), Porto Alegre, basis of courtship behaviour. In: Schroeder JH, Bauer Brazil. p 519-536. J, Schartl M. editors, Trends in ichthyology. Blackwell, Dyer BS, Chernoff B. 1996. Phylogenetic relationships London. p 279-288. among atheriniform fishes (Teleostei: Atherinomor- Hamilton A. 2001. Phylogeny of Limia (Teleostei: Poeci- pha). Zool J Linn Soc 117:1-69. liidae) based on NADH dehydrogenase subunit 2 se- Ghedotti MJ. 1998. Phylogeny and classification of the quences. Mol Phyl Evol 19:277-289. Anablepidae (Teleostei: Cyprinodontiformes). In: Hrbek T, Larson A. 1999. The evolution of diapause in the Malabarba LR, Reis RE, Vari RP, Lucena ZMS, killifish family Rivulidae (Atherinomorpha, Cyprin- Lucena CAS, editors, Phylogeny and classification of odontiformes): A molecular phylogenetic and biogeo- Neotropical Fishes, EDIPUCRS (Editora Pontifícia graphic perspective. Evolution 53:1200-1216. Universidade Católica do Rio Grande do Sul), Porto Hrbek T, Meyer A.. 2003. Closing of the Tethys Sea and Alegre, Brazil. p 561-582. the phylogeny of Eurasian killifishes (Cyprinodonti- Ghedotti MJ. 2000. Phylogenetic analysis and taxonomy formes: Cyprinodontidae). J Evol Biol 16(1):17-36. of the poecilioid fishes (Teleostei: Cyprinodonti- Johnson GD, Patterson C. 1993. Percomorph phylogeny: formes). Zool J Linn Soc 130:1-53. a survey of acanthomorphs and a new proposal. Bull Gosline WA. 1971. Functional Morphology and classifi- Mar Sci 52:554-626. cation of teleostean fishes. The University of Hawaii Klie K. 1988. Morphologische und histologische Unter- Press, Honolulu. suchungen an neuem -Material – Ein Grady JM, Coykendall DK, Collette BB, Quattro JM. Beitrag zur Systematik und Verwandschaft der Adrian- 2001. Taxonomic diversity, origin, and conservation sta- ichthyidae (Pisces: Atheriniformes). Diplomarbeit, tus of Bermuda killifishes (Fundulus) based on mitochon- Universität, Hamburg. drial cytochrome b phylogenies. Conserv Gen 2:41-52. Kottelat M. 1990a. Synopsis of the endangered Buntingi Grant EC, Riddle BR. 1995. Are the endangered spring- (: Adrianichthyidae and Oryziidae) of fish (Crenichthys Hubbs) and poolfish (Empetrichthys Lake Poso, Central Sulawesi, Indonesia, with a new Gilbert) fundulines or goodeids? A mitochondrial reproductive guild and descriptions of three new spe- DNA assessment. Copeia 1995:209-212. cies. Ichthyol Explor Freshwaters 1:49-67. Greenwood PH, Rosen DE, Weitzman SH, Myers GS. Kottelat M. 1990b. The ricefishes (Oryziidae) of the 1966. Phyletic studies of teleostean fishes, with a pro- Malili lakes, Sulawesi, Indonesia, with description of visional classification of living forms. Bull Amer Mus a new species. Ichthyol Explor Freshwaters 1:151-166. Nat Hist 131:345-455. Kottelat M, Whitten AJ, Kartikasari SN, Wirjoatmodjo Griem JN, Martin KLM. 2000. Wave action: the environ- S. 1993. Freshwater fishes of Western Indonesia and mental trigger for hatching in the California Sulawesi. Hong Kong: Periplus Editions (HK) Ltd. in (Teleostei: Atherinopsidae). Mar Biol collaboration with the Environmental Management 137:177-181. Development in Indonesia (EMDI) Project, Ministry Grier HJ. 1981. Cellular organization of the testis and of State for Population and Environment, Republic of spermatogenesis in fishes. Amer Zool 21:345-357. Indonesia, Jakarta. Grier HJ. 1984. Testis structure and formation of sper- Kulkarni CV. 1940. On the systematic position, structural matophores in the atherinomorph teleost Horaichthys modifications, bionomics and development of a re- setnai. Copeia 1984:833-839. markable new family of cyprinodont fishes from the Grier HJ, Collette BB. 1987. Unique spermatozeugmata province of Bombay. Rec Indian Mus 42:379-423. in testes of halfbeaks of the genus Zenarchopterus Lazara KJ. 2001. The killifishes: an annotated checklist, (Teleostei: Hemiramphidae). Copeia 1987:300-311. synonymy and bibliography of recent oviparous cy- Grier HJ, Linton JR, Leatherland JF, deVlaming VL. prinodontiform fishes: The Killifish Master Index, 1980. Structural evidence for two different testicular Fourth Edition. American Killifish Association, Cin- types in teleost fishes. Amer J Anat 159:331-345. cinnati, OH.

• Viviparous Fishes 28 Systematics, Biogeography, and Evolution Li SZ. 2001. On the position of the suborder Adrian- Meyer A, Lydeard C. 1993. The evolution of copulatory ichthyoidei. Acta Zootax Sin 26:583-588. organs, internal fertilization, placentae and viviparity Loureiro M, deSá RO. 1996. External morphology of the in killifishes (Cyprinodontiformes) inferred from a chorion of the annual fishes Cynolebias (Cyprinodon- DNA phylogeny of the tyrosine kinase gene X-src. Proc tiformes: Rivulidae). Copeia 1996:1016-1022. R Soc Lond B 254:153-162. Loureiro M, deSá RO. 1998. Osteological analysis of the Meyer A, Morrissey JM, Schartl M. 1994. Recurrent ori- killifish genus Cynolebias (Cyprinodontiformes: Ri- gin of a sexually selected trait in Xiphophorus fishes in- vulidae). J Morph 1998:245-262. ferred from molecular phylogeny. Nature 386:539-542. Lovejoy NR. 2000. Reinterpreting recapitulation: system- Miya M, Takeshima H, Endo H, Ishiguro NB, Inoue JG, atics of needlefishes and their allies (Teleostei: Beloni- Mukai T, Satoh TP, Yamaguchi M, Kawaguchi A, formes). Evolution 54:1349-1362. Mabuchi K, Shirai SM, Nishida M. 2003. Major Lovejoy N, Collette BB. 2001. Phylogenetic relationships patterns of higher teleostean phylogenies: a new per- of the new world needlefishes (Teleostei: Belonidae) spective based on 100 complete mitochondrial DNA and the biogeography of transitions between marine sequences. Mol Phyl Evol 26:121-138. and freshwater habitats. Copeia 2001:324-338. Mojica CL, Meyer A, Barlow GW. 1997. Phylogenetic Lovejoy NR, Iranpour M, Collette BB. 2004. Phylogeny relationships of species of the genus Brachyrhaphis and ontogeny of beloniform fishes. Integr Comp Biol (Poeciliidae) inferred from partial mitochondrial 44:366-377. DNA sequences. Copeia 1997:298-305. Lüssen A, Falk TM, Villwock W. 2003. Phylogenetic pat- Morris MR, de Queiroz K, Morizot DC. 2001. Phyloge- terns in populations of Chilean species of the genus netic relationships among populations of northern Orestias (Teleostei: Cyprinodontidae): results of mito- swordtails (Xiphophorus) as inferred from allozyme chondrial DNA analysis. Mol Phyl Evol 29:151-60. data. Copeia 2001:65-81. Lydeard C. 1993. Phylogenetic analysis of species richness: Murphy WJ, Collier GE. 1997. A molecular phylogeny has viviparity increased the diversification of actino- for aplocheiloid fishes (Atherinomorpha, Cyprin- pterygian fishes? Copeia 1993:514-518. odontiformes): the role of vicariance and the origins of Lydeard C, Wooten MC, Meyer A. 1995. Cytochrome b annualism. Mol Biol Evol 14: 790-799. sequence variation and a molecular phylogeny of the Murphy WJ, Collier GE. 1999. Phylogenetic relation- live-bearing fish genus Gambusia (Cyprinodonti- ships of African killifishes in the genera formes: Poeciliidae). Can J Zool 73:213-227. and Fundulopanchax inferred from mitochondrial Marcus JM, McCune AR. 1999. Ontogeny and phylog- DNA sequences. Mol Phyl Evol 11:351-360. eny in the northern swordtail clade of Xiphophorus. Murphy WJ, Nguyen TNP, Taylor EB, Collier GE. Syst Biol 48:491-522. 1999a. Mitochondrial DNA phylogeny of West Afri- Martin KL. 1999. Ready and waiting: delayed hatching can aplocheiloid killifishes (Cyprinodontiformes, and extended incubation of anamniotic vertebrate Aplocheilidae). Mol Phyl Evol 11:343-350. terrestrial eggs. Amer Zool 39:279-288. Murphy WJ, Thomerson JE, Collier GE. 1999b. Phylog- Martínez VH, Monasterio de Gonzo GAM. 2002. Testis eny of the neotropical killifish family Rivulidae (Cy- morphology and spermatozeugma formation in Jenyn- prinodontiformes, Aplocheiloidei). Mol Phyl Evol sia multidentata. In: Meeting book, II International 13:289-301. Symposium on Livebearing Fishes. Abstract. Queré- Nelson JS. 1984. Fishes of the World, Second Edition. taro, Qro., Mexico. John Wiley and Sons. Meisner AD. 2001. Phylogenetic systematics of the Nelson JS. 1994. Fishes of the World, Third Edition. John viviparous halfbeak genera Dermogenys and Nomor- Wiley and Sons. hamphus (Teleostei: Hemiramphidae: Zenarcho- Parenti LR. 1981. A phylogenetic and biogeographic pterinae). Zool J Linn Soc 133 (2): 199-283. analysis of cyprinodontiform fishes (Teleostei, Ather- Meisner AD, Burns JR. 1997a. Testis and andropodial inomorpha). Bull Amer Mus Nat Hist 168:335-557. development in a viviparous halfbeak, Dermogenys sp. Parenti LR. 1984a. A taxonomic revision of the Andean (Teleostei: Hemiramphidae). Copeia 1997:44-52. killifish genus Orestias (Cyprinodontiformes, Cyprin- Meisner AD, Burns JR. 1997b. Viviparity in the halfbeak odontidae). Bull Amer Mus Nat Hist 178: 107-214. genera Dermogenys and Nomorhamphus (Teleostei: Parenti LR. 1984b. On the relationships of phallostethid Hemiramphidae). J Morph 234:295-317. fishes (Atherinomorpha), with notes on the anatomy Meisner AD, Collette BB. 1999. Generic relationships of the of dunckeri Regan, 1913. Amer Mus internally-fertilized southeast Asian halfbeaks (Hemi- Novit No. 2779. 12 p. ramphidae: Zenarchopterinae). In: Séret B, Sire JY, edi- Parenti LR. 1987. Phylogenetic aspects of tooth and jaw tors, Proceedings of the 5th Indo-Pacific Fish Conference, structure of the medaka, Oryzias latipes, and other Nouméa, 1997. Soc Fr Ichtyol, Paris. p 69-76. beloniform fishes. J Zool London 211:561-572.

Lynne R. Parenti • The Phylogeny of Atherinomorphs. Reproductive System 29 Parenti LR. 1989a. Why ricefish are not killifish. J Amer taxonomic revision of the genus (Atherino- Killifish Assoc 22(3):79-84. morpha: Bedotiidae). Amer Mus Novit No. 2979. 33 p. Parenti LR. 1989b. A phylogenetic revision of the phal- Stiassny MLJ. 1993. What are grey mullets? Bull Mar Sci lostethid fishes (Atherinomorpha, Phallostethidae). 52:197-219. Proc Cal Acad Sci 46:243-277. Stiassny MLJ, Moore JA. 1992. A review of the pelvic Parenti LR. 1993. Relationships of atherinomorph fishes girdle of acanthomorph fishes, with comments on (Teleostei). Bull Mar Sci 52:170-196. hypotheses of acanthomorph interrelationships. Zool Parenti LR, Song J. 1996. Phylogenetic significance of the J Linn Soc 104:209-242. pectoral/pelvic fin association in acanthomorph fishes: Tsuneki K. 1992. A systematic survey of the occurrence of A reassessment using comparative neuroanatomy. In: the hypothalamic saccus vasculosus in teleost fish. Stiassny MLJ, Parenti LR, Johnson GD, editors, Inter- Acta Zool 73:67-77. relationships of fishes, Academic Press, San Diego. Webb SA, Graves JA, Macías-Garcia C, Magurran AE, p 427-444. Foighil DO, Ritchie MG. 2004. Molecular phylog- Parenti LR, Grier HJ. 2004. Evolution and phylogeny of eny of the livebearing Goodeidae (Cyprinodonti- gonad morphology in bony fishes. Integr Comp Biol formes). Mol Phyl Evol 30:527-544. 44:333-348. Weber M, de Beaufort LF. 1922. Family Adrianichthyi- Parker A. 1997. Combining molecular and morphologi- dae. In: IV. Heteromi, Solenichthyes, Synentognathi, cal data in fish systematics: examples from the Cyprin- Percesoces, Labyrinthici, Microcyprini. The Fishes of odontiformes. In: Kocher TD, Stepien CA, editors, the Indo-Australian Archipelago. EJ Brill, Leiden. p Molecular Systematics of Fishes, Academic Press, San 376-381. Diego. p 163-188. White BN, Lavenberg RJ, McGowen GE. 1984. Ather- Parker A, Kornfield I. 1995. A molecular perspective on iniformes: development and relationships. In: Moser evolution and zoogeography of cyprinodontid killi- HG, Richards WJ, Cohen DM, Fahay MP, Kendall fishes. Copeia 1995:8-21. AW, Jr., Richardson SL, editors, Ontogeny and sys- Rauchenberger M. 1989. Systematics and biogeography tematics of fishes. Special Publication No. 1, Supple- of the genus Gambusia (Cyprinodontiformes: Poeci- ment to Copeia, American Society of Ichthyologists lidae [sic]). Amer Mus Novit No. 2951. 74 p. and Herpetologists. p 355-362. Roberts CD. 1993. Comparative morphology of spined Wiley EO, Johnson GD, Dimmick WW. 2000. The in- scales and their phylogenetic significance in the terrelationships of acanthomorph fishes: A total evi- Teleostei. Bull Mar Sci 52:60-113. dence approach using molecular and morphological Rodríguez CM. 1997. Phylogenetic analysis of the tribe data. Biochem Syst Ecol 28: 319-350. Poeciliini (Cyprinodontiformes: Poeciliidae). Copeia Wourms JP. 1972. The developmental biology of annual 1997:663-679. fishes. III. Pre-embryonic and embryonic diapause of Rosen DE. 1964. The relationships and taxonomic posi- variable duration in the eggs of annual fishes. J Exp tion of the halfbeaks, killifishes, silversides and their Zool 182:389-414. relatives. Bull Amer Mus Nat Hist 127:217-268. Wourms JP. 1976. Annual fish oogenesis. I. Differentia- Rosen DE, Bailey RM. 1963. The poeciliid fishes (Cy- tion of the mature oocyte and formation of the pri- prinodontiformes), their structure, zoogeography, and mary envelope. Dev Biol 50:338-354. systematics. Bull Amer Mus Nat Hist 126:1-176. Wourms JP, Sheldon H. 1976. Annual fish oogenesis. II. Rosen DE, Parenti LR. 1981. Relationships of Oryzias, Formation of the secondary egg envelope. Dev Biol and the groups of atherinomorph fishes. Amer Mus 50:355-366. Novit No. 2719. 25 p. Yamamoto T. 1975. Medaka (Killifish) Biology and Saeed B, Ivantsoff W, Allen GR. 1989. Taxonomic revi- Strains. Series of Stock Culture in Biological Field. sion of the family Pseudomugilidae (Order Ather- Keigaku Publishing Company, Tokyo. iniformes). Aust J Freshwater Res 40:719-787. Saeed B, Ivantsoff W, Crowley LELM. 1994. Systematic relationships of atheriniform families within Division I of the Series Atherinomorpha () with relevant historical perspectives. Voprosi Ikhtiologii 34(4):1-32. Schartl M. 1995. Platyfish and swordtails: a genetic sys- tem for the analysis of molecular mechanisms in tu- mor formation. Trends Genet 11:185-189. Stiassny MLJ. 1990. Notes on the anatomy and relation- ships of the bedotiid fishes of Madagascar, with a

• Viviparous Fishes 30 Systematics, Biogeography, and Evolution