the phylogeny of evolution of a novel fish reproductive...

13Lynne R. Parenti •The Phylogeny of Atherinomorphs. Reproductive System

The Phylogeny ofAtherinomorphs:

Evolution of aNovel Fish

ReproductiveSystem

Lynne R. Parenti

Division of Fishes, MRC NHB 159, Department ofVertebrate Zoology, National Museum of Natural History,

Smithsonian Institution,Washington, DC, USA.

14 • Viviparous Fishes Systematics, Biogeography, and Evolution

The Phylogeny ofAtherinomorphs:

Evolution of aNovel Fish

ReproductiveSystem

Lynne R. Parenti

Abstract

The fishes commonly known in English assilversides, rainbowfishes, phallostethids,killifishes, ricefishes, halfbeaks, needlefishes,flying fishes, and sauries were combined byRosen (1964), into one taxon, now known asthe Atherinomorpha, largely using osteologicaland reproductive characters. Subsequentreviews have supported atherinomorphmonophyly, adding characters to thediagnosis. Today atherinomorphs arediagnosed as monophyletic by derivedcharacters of the testis, egg, reproductivemode, circulatory system, jaw musculature,olfactory organ, and various parts of theskeleton including the ethmoid region of theskull, gill arches, pelvic girdle, among others.Support for monophyly of each of the threeincluded taxa, now classified as the ordersCyprinodontiformes, Beloniformes, andAtheriniformes, and the relationships amongthem, varies in quantity and quality. Duringthe past twenty years, reproductive andmolecular data used to infer atherinomorphrelationships have grown significantly. Ingeneral, molecular data support hypothesesbased on morphology, and, in some cases,provide novel hypotheses and uniquechallenges to morphological data. A growingbody of data indicates that all atherinomorphsshare a unique testis-type that is correlatedwith an array of reproductive modificationssuch as coupling during mating, relativelylong developmental period, sperm-bundleformation, internal fertilization, superfetation,embryo retention, diapause, delayed hatching,hermaphroditism, and live-bearing.Corroboration of an atherinomorph sistergroup may include identification of someunique aspects of this reproductive system inother taxa.

Resumen

Los peces comúnmente conocidos en ingléscomo: silversides, rainbowfishes, phallostethids,killifishes, ricefishes, halfbeaks, needlefishes,flying fishes y sauries fueron integrados porRosen (1964) en un taxon ahora conocidocomo Atherinomorpha, utilizandoesencialmente caracteres osteológicos yreproductivos. Revisiones posteriores hansostenido que Atherinomorpha es un grupomonofilético y han agregado otros caracteres ala diagnosis. Hoy, los aterinomorfos estándefinidos como monofiléticos porcaracterísticas derivadas del testículo, huevos,formas reproductoras, sistema circulatorio,musculatura de la mandíbula, órganoolfatorio, y varias partes del esqueleto,incluyendo la región etmoidea del cráneo,arcos branquiales, y cintura pélvica, entreotros. El carácter monofilético de cada uno delos tres taxa incluidos actualmente, clasificadosen los órdenes Cyprinodontiformes,Beloniformes, y Atheriniformes, se define ensus relaciones entre ellos, con variaciones decantidad y calidad. Durante los últimos veinteaños, ha aumentado significativamente el usode datos sobre reproducción y molecularespara inferir relaciones en los aterinomorfos.En general, datos moleculares apoyanhipótesis basadas en la morfología y, enalgunos casos, sustentan hipótesis nuevas yúnicas referentes a los datos morfológicos. Unnúmero creciente de datos indica que todoslos aterinomorfos tienen un tipo testicularúnico, correlacionado con una serie demodificaciones reproductivas, tales como elapareamiento, un periodo relativamente largode desarrollo, formación de paquetes deespermatozoides, fertilización interna,superfetación, retención de los embriones,diapausa, eclosión demorada, hermafroditismoy viviparidad. Corroborar que algún grupo deotro taxa, esté emparentado con losaterinomorfos, puede incluir la identificaciónde alguno de estos aspectos del sistemareproductor.

• Viviparous FishesMari Carmen Uribe and Harry J. Grier, book editors.New Life Publications, Homestead, Florida, 2005. p 13-30.


Introduction

In 1964, Donn E. Rosen, then curator in theDepartment of Ichthyology, American Mu-seum of Natural History, New York, pub-

lished a monograph as a Bulletin of the AmericanMuseum of Natural History, in which he broughttogether three disparate groups of teleost fishesin one order, the Atheriniformes (Rosen, 1964).The three suborders in Rosen’s Atheriniformeswere the Atherinoidei (the silversides, rainbow-fishes, and phallostethids), the Cyprinodon-toidei (the killifishes and ricefishes), and theExocoetoidei (the sauries, needlefishes, half-beaks, and flying fishes). Support of this taxonincluded evidence largely from two systems, os-teology and reproductive biology, which Rosendescribed in an essay, as was common then,rather than enumerating putative synapomor-phies. Skeletal characters that Rosen considereddiagnostic of his Atheriniformes included a disc-shaped dorsal and ventral ossified mesethmoidand decoupling of the rostral cartilage from theascending processes of the premaxillae. Repro-ductive characters included a large, demersal eggwith long, adhesive chorionic filaments andlarge oil globules (Rosen, 1964:253-255).

The new taxon, classified as the seriesAtherinomorpha by Greenwood et al. (1966),was not accepted readily by all systematic ich-thyologists. In particular, Gosline (1971, Fig.28B) argued that Rosen had brought togetherfishes from different evolutionary grades that didnot share an evolutionary history. Atherinoids(now classified as the order Atheriniformes fol-lowing Dyer and Chernoff, 1996; Table 1) were

considered by Gosline to be “higher teleosts”because they have characters such as two dorsalfins, both with anterior spines or thickened rays,and an I,5 pelvic-fin ray formula. In contrast,the Cyprinodontoidei and Exocoetoidei (Cy-prinodontiformes and Beloniformes, followingGreenwood et al., 1966) were considered byGosline to be “intermediate teleosts.” Both havea single, soft-rayed dorsal fin and may have morethan six pelvic-fin rays, among other characters,that they share with taxa that Gosline thoughtto be less advanced (see review by Parenti, 1993).Cyprinodontiformes and Beloniformes were notconsidered closely related by Gosline (1971),who postulated that they were derived from, ormost closely related to the Beryciformes andMyctophiformes, respectively.

Monophyly of atherinomorphs and mono-phyly and relationships of the three includedtaxa was reviewed by Rosen and Parenti (1981)in conjunction with a phylogenetic analysis ofCyprinodontiformes by Parenti (1981). Thefirst explicitly cladistic analyses of atherino-morph phylogeny were presented in these twopapers. Atherinomorph monophyly was sup-ported by ten characters, again largely those ofthe skeleton, but also including two reproduc-tive characters (Rosen and Parenti, 1981:20),one of the egg (“a large dermersal egg with longadhesive and short filaments and many lipidglobules that coalesce at the vegetal pole”), andone of the testis (“the spermatogonia formingonly at the blind end of the tubule near thetunica albuginea”).


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 sensu Dyer and Chernoff, 1996 [= Atherinoidei of Rosen, 1964; Division I of Rosenand Parenti, 1981] Classification is sequenced.

Family AtherinopsidaeSuborder Atherinoidei

Family Notocheiridae (including Isonidae)Infraorder Atherines

Family Melanotaeniidae (including Bedotiidae, Pseudomugilidae)Family Atherionidae

Superfamily AtherinoideaFamily Phallostethidae (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 AplocheiloideiFamily AplocheilidaeFamily Rivulidae

Suborder CyprinodontoideiSuperfamily Funduloidea

Family ProfundulidaeFamily FundulidaeFamily Goodeidae

Superfamily Valencioidea of Costa, 1998a [= Sept 1 of Parenti, 1981]Family Valenciidae

Unranked category including superfamilies Cyprinodontoidea and PoecilioideaSuperfamily Cyprinodontoidea

Family CyprinodontidaeSuperfamily Poecilioidea of Parenti, 1981 [= unnamed clade of Costa, 1998a]

Family AnablepidaeFamily Poeciliidae

Order Beloniformes [not recognized by Rosen, 1964]Suborder Adrianichthyoidei [= Adrianichthyoidea of Rosen, 1964]

Family Adrianichthyidae (including Horaichthyidae and Oryziidae)Suborder Exocoetoidei

Superfamily ExocoetoideaFamily ExocoetidaeFamily Hemiramphidae

Superfamily ScomberesocoideaFamily BelonidaeFamily Scomberesocidae


The many, unique characteristics of theatherinomorph egg are correlated with reproduc-tive and developmental modifications. Filamentsare derived from the secondary or outer layer ofthe zona pellucida which is secreted by follicle cells(Wourms, 1976; Wourms and Sheldon, 1976;Loureiro and deSá, 1996). Filaments vary innumber, shape, and relative length (e.g., Able,1984; Collette et al., 1984; White et al., 1984;Loureiro and deSá, 1996). Oil globule numberranges 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 orreduction of a distinct larval stage, most notablyin cyprinodontiforms and beloniforms (Rosen,1964:253).

The long developmental period and relativelylarge, desiccation-resistant egg is correlated withthe evolution of delayed hatching (Martin,1999) or diapause (Wourms, 1972). Delay ofhatching of fertilized eggs has been reported inthe atheriniform grunions of the genus Leu-resthes, and the cyprinodontiforms Fundulusheteroclitus, F. confluentus, and Adinia xenica(Martin, 1999). Delay of hatching in these taxais facultative; fertilized, fully-developed eggs canhatch, but are stranded in relatively dry habitatsand must await waves, tides, or rains to stimu-late hatching (Martin, 1999; Griem and Mar-tin, 2000). This is in contrast to fertilized eggsof annual killifishes, which undergo diapauseand for which delay of hatching is obligatory(Wourms, 1972). Stranding of fertilized eggs outof water has been reported also in the Baja Cali-fornia endemic, Fundulus lima, by Brill (1982).Facultative delayed hatching of teleost eggs hasbeen reported outside atherinomorphs only inthe lower teleost osmeroid Galaxias maculatus(Martin, 1999).

The evolutionary relationship between fac-ultative and obligatory delay of hatching inatherinomorphs is unknown. Phylogeneticanalyses based on molecules (e.g., Murphy andCollier, 1997) or morphology (e.g., Costa, 1990,1998a) have led to various, sometimes conflict-ing, conclusions concerning single or multipleorigins of developmental diapause. A molecularphylogenetic analysis of the family Rivulidae byHrbek and Larson (1999: Fig. 4) supported thehypothesis that diapause was present in two dis-tantly related groups of South American killi-fishes. Rather than interpret the cladogramliterally, they considered it unlikely that diapausehad originated twice, but that presence or ab-

sence of diapause results from “...developmen-tal switches between alternative stabilized path-ways” (Hrbek and Larson, 1999:1200). This wassaid another way by Parenti (1981:364): “...theannual habit is no more than an exaggeration,due to extreme environmental fluctuations, of acapability of all cyprinodontiforms to survivestress that involves desiccation”. Now, with ourincreased knowledge of delayed hatching pat-terns, I would rewrite that sentence by substitut-ing “atherinomorphs” for “cyprinodontiforms”(see also Parenti, 1993). Delayed hatching pat-terns of atherinomorphs, whether facultative orobligatory, may be homologous and representanother atherinomorph synapomorphy.

The testis character proposed as an atherino-morph synapomorphy by Rosen and Parenti(1981) had been described just the year beforeby Harry Grier and colleagues (Grier et al.,1980; Fig. 1) who reported this distinctive tes-tis-type in 31 atherinomorph species represent-ing each of the three orders. Since Grier et al.(1980), the atherinomorph testis has been re-ported or confirmed in a total of 79 atherino-morph species (Parenti and Grier, 2004),including the beloniform adrianichthyidHoraichthys setnai (Grier, 1984), seven speciesof atheriniform phallostethids (Grier andParenti, 1994), an anablepid, Jenynsia multi-dentata (Martínez and Monasterio de Gonzo,2002), the internally-fertilizing halfbeak genus,Zenarchopterus (Grier and Collette, 1987), andthe viviparous halfbeaks, Dermogenys, Hemi-rhamphodon and Nomorhamphus (Downingand Burns, 1995; Meisner and Burns, 1997a).A possibly similar testis-type in viviparous sur-fperches, family Embiotocidae, was noted byGrier et al. (1980) but dismissed a year laterby Grier (1981).

Figure 1.Diagrammatic representation of twodifferent testis-types in higherteleosts: A: testis lobule inrepresentative atherinomorph, withspermatogonia restricted to thedistal end of the lobule; B: testislobule in representativenon-atherinomorph higher teleost,with spermatogonia distributedthroughout the length of the testislobule (from Grier et al., 1980: Fig. 1)


Evolution of viviparity is considered indepen-dent in atherinomorphs and embiotocids (seeLydeard, 1993). The phylogenetic significance ofreproductive characters in embiotocids, includ-ing both egg and testis, are evaluated relative tothose of atherinomorph fishes elsewhere in thisvolume (Grier et al., this volume). The derivedtestis and egg are correlated in atherinomorphswith a vast array of reproductive modificationsincluding coupling during mating, relatively longdevelopmental period, sperm-bundle formation,internal fertilization, superfetation, embryo re-tention, diapause, delayed hatching, hermaphro-ditism, and live-bearing. Understanding thephylogeny of atherinomorph fishes will be en-hanced by an understanding of their derived re-productive modifications and evolution of thisnovel reproductive system.

The purpose of this paper is to review thecurrent state of our knowledge of atherino-morph phylogeny, with a focus on live-bearingtaxa, summarizing molecular and morphologi-cal cladistic analyses published during the pasttwo decades. Some phylogenetic analyses ofsolely oviparous taxa are cited but not discussedin detail. It is not my goal to summarize allcladistic analyses of atherinomorph taxa, whichis well beyond the scope of this review.

Atherinomorph Monophyly

Atherinomorph monophyly was reviewed byParenti (1993) who listed 14 diagnostic charac-ters, five of which were of reproduction anddevelopment (Fig. 2, node A). The atherino-morph testis-type and associated reproductivemodifications remain among the strongest evi-dence for monophyly, as argued above.

Here, I add a fifteenth character to the athe-rinomorph diagnosis: absence of the saccusvasculosus, a hypothalamic circumventricularorgan of unspecified function (Tsuneki, 1992).The saccus vasculosus has been reported to pro-duce a parathyroid hormone-related protein inthe perciform sea bream, Sparus aurata (Devlin

et al., 1996). Approximately 200 teleost species,both freshwater and marine, representing all ma-jor teleost lineages, were surveyed by Tsuneki(1992) for presence or absence of the saccusvasculosus and extent of its development whenpresent. A well-developed saccus vasculosus wasconsidered to be a generalized condition, presentin some Osteoglossiformes, Anguilliformes, andostariophysans among lower teleosts, andMugiliformes, Gasterosteiformes, Scorpaeni-formes, Perciformes, Pleuronectiformes, andTetraodontiformes among higher teleosts. Thesaccus vasculosus is reduced or absent in a vari-ety of taxa, including atherinomorphs, Africancichlids, some anabantoids, and a synbranchideel among higher teleosts.

The most “clear-cut” result of this survey,according to Tsuneki (1992:74), is that absenceof the saccus vasculosus is characteristic ofatherinomorphs, and I agree. Fifteen atherino-morphs surveyed, representing all three orders,from both freshwater and marine habitats un-ambiguously lack the saccus vasculosus. Its ab-sence is proposed here as an atherinomorphsynapomorphy.

A sixteenth atherinomorph synapomorphy,of the oocyte, was proposed and illustrated byParenti and Grier (2004: Figs. 3,4). Atherino-morph yolk is fluid, rather than granular,throughout vitellogenesis.

Atherinomorph Sister Group

The Series Atherinomorpha was classified as sis-ter to the Series Percomorpha by Rosen andParenti (1981) without an explicit proposal ofrelationship to any particular percomorph ta-xon. In diagnosing atherinomorphs, Rosen(1964) made deliberate comparisons with taxasuch as mullets (Mugilidae) or percopsiformfishes that had at one time been proposed asclosely related to one or another atherinomorphorder (see Parenti, 1993). In discussing ather-inomorph sister-group relationships, Parenti(1993: Fig. 1c) could come to no firm conclu-sion, arguing that evidence linked atherino-morphs to paracanthopterygians on one handand to percomorphs on the other.

The close relationship of mullets to ather-inomorphs was proposed by Stiassny (1990) andexplored further by Stiassny (1993) who, al-though she concluded that there was good evi-dence for a sister group relationship, discussednumerous characters that contradicted that pro-posal. Of seven characters in support of a mul-

Figure 2.Phylogenetic relationships amongthe three orders of atherinomorphfishes, following Rosen and Parenti(1981), Collette et al. (1984), Parenti(1993), Dyer and Chernoff (1996), andthis paper. Synapomorphies are: A)testis a restricted spermatogonialtype; egg demersal, with several tomany chorionic filaments, andseveral oil globules that coalesce atthe vegetal pole; coupling duringmating; prolonged developmentalperiod; separation of embryonicafferent and efferent circulation bydevelopment of heart in front ofhead; ossified portion of ethmoidregion of skull highly reduced;infraorbital series represented bythe lacrimal, dermosphenotic, andtwo, one or no anterior infraorbitalbones; lateral process of pelvic boneand distal end of pleural rib in closeassociation, and, in some taxa,connected via a ligament;supracleithrum reduced or absent;dorsal portion of gill arches with alarge fourth epibranchial theprominent supporting bone and nofourth pharyngobranchial element;medial hooklike projection andventral flange on the fifthceratobranchial bone; supraneuralbones absent; superficial division ofadductor mandibulae with twotendons, one inserting on themaxilla, a second inserting on thelacrimal bone; olfactory sensoryepithelium arranged in sensoryislets; absence of the saccusvasculosus; B) second infraorbitalbone absent; first epibranchial bonewith an expanded base and noseparate uncinate process; firstpharyngobranchial element absent;second and third epibranchialssmaller than the first and fourth;stomach, pyloric caecae andpneumatic duct absent

Atheriniformes A

Cyprinodontiformes B

Beloniformes


let-atherinomorph sister-group relationship(Stiassny, 1993: Fig. 1), four were of the pecto-ral girdle, two of the branchial muscles, and oneof the vertebral column. Three pelvic fin char-acters were considered to have reversed in ather-inomorphs (viz. Stiassny and Moore, 1992). Thestudy was not intended as an exhaustive reviewof acanthomorph relationships, however, be-cause, for example, another reversal in athe-rinomorphs would be loss of transformingctenoid scales (Roberts, 1993).

The hypothesis of a close mullet-atherino-morph relationship was taken a step further byJohnson and Patterson (1993) who proposed anew taxon, the Smegmamorpha, to includesynbranchoid eels, mastacembeloid eels, thecentrarchid Elassoma, gasterosteiforms, mullets,and atherinomorphs. A single character was pro-posed for smegmamorph monophyly (Johnsonand Patterson, 1993:572): the first two epineu-ral bones originate at the tip of transverse pro-cesses or fused parapophyses on the first twovertebral centra. In addition, Johnson andPatterson (1993:table 2) tabulated selected,shared derived characters found in some, butnot all smegmamorphs, such as, for example,infraorbital series with three or fewer bones be-tween the lacrimal and the dermosphenotic,present in atherinomorphs, Elassoma, gaster-osteiforms, and synbranchoids, but absent inmullets and mastacembeloids.

Smegmamorph monophyly was one hypo-thesis tested by Wiley et al. (2000) in a totalevidence analysis of acanthomorph phylogenycombining molecular and morphological data.The strict consensus of 137 equally parsimoni-ous trees based on morphological data recov-ered a monophyletic Atherinomorpha withunresolved relationships to an array of highertaxa (Wiley et al., 2000: Fig. 8c). A strict con-sensus of four equally parsimonious trees basedon combined molecular and morphological datarecovered a group that included the mulletMugil and the four atherinomorph taxa analyzed(the atheriniforms Melanotaenia and Atherino-morus, the beloniform Strongylura, and thecyprinodontiform Gambusia), but did not re-cover a monophyletic Smegmamorpha (Wileyet al., 2000: Fig. 6). Mugil was considered to bemore closely related to the two atheriniformsthan either is to the cyprinodontiform or thebeloniform; that is, atherinomorph monophylywas refuted as well. Morphological characterssurveyed by Wiley et al. (2000), however, didnot include characters such as the atherino-

morph testis-type (not found in mullets; Grieret al., 1980), absence of the saccus vasculosus(well-developed in mullets; Tsuneki, 1992), orinnervation of the pectoral and pelvic finmuscles by branches of spinal nerve 2 (presentand derived in mullets, absent in atherino-morphs; Parenti and Song, 1996). Detailedmolecular analyses of acanthomorph phylogenybased on molecular data alone (e.g., Miya et al.,2003) confirm atherinomorph monophyly, butinclude other taxa, such as blennioids and go-biesocids, along with mullets, as putative ather-inomorph sister taxa. Broader surveys ofmorphology, as well as molecules, are needed tofurther test monophyly of smegmamorphs andevaluate the hypothesis of a sister group rela-tionship of mullets and atherinomorphs.

Cyprinodontea

Relationships among the three groups of ather-inomorph fishes were unspecified by Rosen(1964). Monophyly of each group and the rela-tionships among them were considered by Rosenand Parenti (1981:21-23) who proposed a sistergroup relationship between Cyprinodontiformesand Beloniformes, together called Division IIatherinomorphs. Four characters were proposedto support monophyly of Division II (Rosen andParenti, 1981:21; Fig. 2, node B). A fifth charac-ter may be added from Li (2001:585): absence ofa stomach, including absence of pyloric caecaeand a pneumatic duct.

The sister group relationship betweenCyprinodontiformes and Beloniformes has beencorroborated (Stiassny, 1990; Saeed et al., 1994;Dyer and Chernoff, 1996; Wiley et al., 2000;Li, 2001). The relationship was recognized byNelson (1984:214) who in his second editionof Fishes of the World synonymized the two or-ders in an expanded Cyprinodontiformes. Thisdecision was reversed in the third edition(Nelson, 1994:264, with the three orders Ather-iniformes, Beloniformes, and Cyprinodonti-formes recognized without inter-ordinalrelationships expressed). The current edition ofFishes of the World is used worldwide as a stan-dard reference for teleost classification, however,and an enlarged order Cyprinodontiformes, in-cluding beloniforms, was incorporated into nu-merous publications, particularly during thedecade between 1984 and 1994 (e.g., Tsuneki,1992; Kottelat et al., 1993). The superorderCyprinodontea was proposed by Dyer andChernoff (1996) as a formal name for Rosen


and Parenti’s (1981) Division II, the Cyprino-dontiformes and Beloniformes, and is adoptedhere (Table 1).

Cyprinodontiformes

The first explicitly cladistic analysis of cypri-nodontiform fishes was based largely on os-teological characters (Parenti, 1981; Fig. 3).Cyprinodontiform monophyly was corroboratedby six complex characters (Parenti, 1981: Fig. 9,node a), and it has been challenged only by Li(2001; see comments below under Beloniformes).One cyprinodontiform character of Parenti(1981), prolonged embryonic development, isdiscussed above as an atherinomorph synapo-morphy (see also Parenti, 1993). A major conclu-sion of Parenti’s (1981) study that refuted earliernotions of cyprinodontiform relationships wasthat viviparity had evolved at least three timeswithin the order, not once. That is, oviparous sis-ter groups of each group of viviparous taxa werehypothesized. Also, cyprinodontiforms were di-vided into two monophyletic suborders, theAplocheiloidei and Cyprinodontoidei. This pro-posal of relationships was tested by Meyer andLydeard (1993; Fig. 4) using partial DNA se-quences of the tyrosine kinase gene X-src, anoncogene chosen in part because cyprinodonti-forms of the genus Xiphophorus have long beenknown to inherit melanomas (see Schartl, 1995).No non-cyprinodontiform taxon was used as anoutgroup by Meyer and Lydeard, so their analysiswas not a test of cyprinodontiform monophyly.Also, at least one pivotal taxon, the oviparousOxyzygonectes, sister to the viviparous Anablepsand Jenynsia in the Anablepidae, according toParenti (1981), was not included. Nonetheless,the maximum parsimony phylogenetic hypoth-esis of Meyer and Lydeard (1993) corroboratedParenti’s (1981) conclusions that viviparity hadevolved at least three times within cyprinodonti-forms, in anablepids, goodeids, and poeciliids.Further, the analysis recovered the controversialsister group relationship of the viviparous good-eids and their oviparous relatives, Empetrichthysand Crenichthys (see also Grant and Riddle, 1995).

Cyprinodontiforms were used as a test taxonby Parker (1997) to evaluate the consequencesof combining morphological and molecular datain a phylogenetic analysis. Morphological char-acters described by Parenti (1981, 1984a) arelisted and character states coded in a detailedappendix (Parker, 1997:184-185). Although thisstudy was not intended as a thorough re-analy-

Figure 3.Phylogenetic relationships amongfamilies of Cyprinodontiformes asproposed by Parenti (1981)

Figure 4.Phylogenetic relationships amongselect genera of Cyprinodontiformesas proposed by Meyer and Lydeard(1993) based on partial DNAsequences of the tyrosine kinasegene X-src. Genera in boxesrepresent families or subfamilygroupings of live-bearing taxa andtheir close relatives, as alsoproposed by Parenti (1981), from topto bottom, family Goodeidae, familyAnablepidae, and subfamilyPoeciliinae

Aplocheilidae

Rivulidae

Profundulidae

Fundulidae

Valenciidae

Anablepidae

Poeciliidae

Goodeidae

Cyprinodontidae

Nothobranchius

CynolebiasRivulusRivulus

FundulusCrenichthysZoogoneticusZoogoneticusXenotocaProfundulusAnablepsJenynsia

CnesterodonPoeciliaXiphophorusXiphophorus

Tomeurus

AplocheilichthysAplocheilichthysFluviphylax

CubanichthysCyprinodonJordanella


sis of Parenti’s (1981) hypothesis, some ofParker’s results anticipated those of other stud-ies. For example, Tomeurus, the ovoviviparouspoeciliine, had been considered primitive to vi-viparous poeciliines since Rosen and Bailey’s(1963) classic review, and this relationship wascorroborated by Meyer and Lydeard (1993; Fig.4). In contrast, a partitioned X-src dataset of firstand second codon positions only, analyzed un-der maximum parsimony (Parker, 1997: Fig. 2),supported the sister group relationship of To-meurus and Cnesterodon, a viviparous poeciliine,together considered derived, not basal, poeci-liines. This relationship was corroborated by anarray of synapomorphies in a parsimony analy-sis of morphological characters by Ghedotti(2000; Fig. 5), including neural arch of first ver-tebra open dorsally, lateral processes on ventralportion of seventh, eighth, and ninth proximalanal-fin radials in adult males present and incontact, and distal tip of the gonopodium withelongate, bony processes. These morphologicalcharacters were of course known to Rosen andBailey (1963) who likely accepted the constraintthat an ovoviviparous taxon such as Tomeuruswas basal to a group of viviparous taxa. Thisconstraint was relaxed by Parker (1997) andGhedotti (2000) who independently recovereda novel hypothesis of relationships of Tomeurus.

Division of Cyprinodontiformes into two sub-orders, Aplocheiloidei and Cyprinodontoidei,was corroborated by Costa (1998a; Fig. 6) whore-analyzed killifish phylogenetic relationshipsusing morphology. The hypothesis of Costa(1998a; Fig. 6) differs from that of Parenti (1981;Fig. 3) in the placement of Goodeidae as sister toProfundulidae, not Cyprinodontidae, and theresolution of the relationships of Valenciidae.Both studies used osteology as a principal sourceof data, but differed in interpretation of signifi-cance of character states of the premaxilla, amongothers. The sister group relationship of Good-eidae and Profundulidae was recovered also in themolecular hypotheses of Meyer and Lydeard(1993) and Parker (1997). The close relationshipsof fundulids, goodeids, and profundulids is re-flected in the written classification of Cyprin-odontiformes (Table 1).

Monophyly of each of the nine families rec-ognized by Parenti (1981) was corroborated byCosta (1998a) and the terminal taxa of Meyerand Lydeard’s (1993; Fig. 4) hypothesis are alsoconsistent with Parenti’s hypothesis. Monophylyof the Anablepidae and Poeciliidae (sensuParenti, 1981) was corroborated in Ghedotti’s

Figure 5.Phylogenetic relationships amongthe poecilioid fishes, simplified fromGhedotti (2000: Fig. 20). Specieslisted are members of the subfamilyPoeciliinae

Figure 6.Phylogenetic relationships amongfamilies of Cyprinodontiformes asproposed by Costa (1998a)

Anablepidae

Aplocheilichthyinae

Procatopodinae

Alfaro cultratus

Priapella compressa

Gambusia affinis

Heterandria formosa

Poeciliopsis latidens

Girardinus metallicus

Poecilia sphenops

Phallichthys amates

Phallotorynus victoriae

Phalloceros caudimaculatus

Cnesterodon decemmaculatus

Tomeurus gracilis

Aplocheilidae

Rivulidae

Profundulidae

Goodeidae

Fundulidae

Valenciidae

Anablepidae

Poeciliidae

Cyprinodontidae


(2000; Fig. 5) morphological analysis, whereasMeyer and Lydeards’ (1993; Fig. 4) analysis pro-posed a paraphyletic Poeciliidae. The dottedlines in Meyer and Lydeard’s (1993; Fig. 4) cla-dogram indicate relationships for which theyfound weak support, however. Morphology andmolecules do equally well, i.e. agree, in resolv-ing relationships at the tips of the tree, but varyin their ability to recover higher taxa. A mo-lecular phylogenetic analysis of aplocheiloids byMurphy and Collier (1997) conflicts, in part,with the proposals of Parenti (1981) and Costa(1998a) based on morphology. A stable classifi-cation of cyprinodontiforms, at the family leveland above, is within reach, however, and is ex-pected to include components common to theabove morphological and molecular analyses.

Additional molecular phylogenies of familiesor other subgroups of aplocheiloids includethose of Murphy and Collier (1999, Aphyo-semion and Fundulopanchax), Murphy et al.(1999a, West African aplocheiloids), andMurphy et al. (1999b, Rivulidae). Morphologi-cal phylogenies or surveys include Costa (1990,1998b, Rivulidae; 1995a, Cynopoecilus; 1995b,Cynolebiatinae; 1996a, Simpsonichthys),Loureiro and deSá (1998, Cynolebias), and Aarnand Shepherd (2001, epiplatines). Taxonomy,biology, and conservation status of Brazilianannual killifishes was reviewed by Costa (2002,and references therein).

Molecular phylogenies of families or othersubgroups of cyprinodontoids include Bernardi(1997, Fundulidae), Breden et al. (1999, Poeci-lia), Grady et al. (2001, Fundulus), Hamilton(2001, Limia), Hrbek and Meyer, 2003 (Apha-nius) Lüssen et al. (2003, Orestias), Lydeard etal. (1995, Gambusia), Mojica et al. (1997,Brachyrhaphis), Parker and Kornfield (1995,

cyprinodontids), and Webb et al. (2004,livebearing Goodeidae). Morphological phylo-genies or surveys include Parenti (1984a, Ores-tias), Chambers (1987, cyprinodontiformgonopodia; 1990, cnesterodontin gonopodia),Ghedotti (1998, Anablepidae), Rauchenberger(1989, Gambusia), Costa (1996b, Fluviphylax;1997, cyprinodontids), and Rodríguez (1997,Poeciliini). Conclusions of these studies thatbear on higher order relationships of cyprin-odontiforms were summarized by Lazara(2001:ix-xiv).

The genus Xiphophorus, the swordtails andplatyfishes, serves as a model taxon in analysesof congruence of phylogenetic pattern with be-havior, development, morphology, and mol-ecules (see, for example, Basolo, 1991; Haas,1993; Meyer et al., 1994, Marcus and McCune,1999, and Morris et al., 2001).

Beloniformes

Beloniform monophyly was supported by sevensynapomorphies (Rosen and Parenti, 1981:17),including absence of the interhyal bone, reduc-tion or loss of the interarcual cartilage, presenceof only a single, ventral hypohyal bone, modifi-cations of the gill arch skeleton, as well as a dis-tinctive caudal skeleton characterized by thelower caudal lobe with more principal rays thanin the upper caudal lobe.

Beloniform relationships were reviewed byCollette et al. (1984; Fig. 7) who accepted amonophyletic Beloniformes including Rosenand Parenti’s (1981) proposal that ricefishes(family Adrianichthyidae) are more closely re-lated to exocoetoids (families Exocoetidae,Hemiramphidae, Belonidae, and Scomberesoci-dae) than to cyprinodontiforms. This proposalhas been criticized recently by Li (2001) whoargued that adrianichthyoids are more closelyrelated to Cyprinodontiformes than to exo-coetoids. In particular, Li (2001) claimed thatsome characters proposed as diagnostic ofBeloniformes sensu lato, although absent in theprimitive cyprinodontiform suborder Aplochei-loidei, are present in the derived suborderCyprinodontoidei. For example, Li (2001:584)rejected Rosen and Parenti’s (1981) beloniformsynapomorphy of a single, ventral hypohyal be-cause cyprinodontoids have a single hypohyal.If we accept cyprinodontiform monophyly,however, which is supported by a symmetricalcaudal fin skeleton, absent in adrianichthyoids,and first pleural rib on the second, rather than

Adrianichthyidae

Exocoetidae

Hemiramphidae

Belonidae

Scomberesocidae

Figure 7.Phylogenetic relationships amongthe families of beloniform fishes,following Rosen and Parenti (1981)and Collette et al. (1984)


the third vertebra, among other characters, thensimilarities between cyprinodontoids andadrianichthyoids must be interpreted as conver-gent rather than indicative of close relationship.Cyprinodontiforms and beloniforms, sensuRosen and Parenti (1981), are readily distin-guished by their distinct caudal fin skeletons.Finally, both cyprinodontiforms and adrianich-thyoids were characterized by Li (2001) as lack-ing elongate jaws, whereas Parenti (1987,1989a) argued that the elongate jaws of thelarge-bodied adrianichthyoids of Sulawesi, someof which have been called ‘duck-billed’ repre-sent additional support for their close relation-ship to exocoetoids. Despite rejection of Li’s(2001) hypothesis, I appreciate his commentsas they call for continued critique of the mono-phyly of clades within the Atherinomorpha asrecognized herein.

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 atotal evidence analysis. The families Exocoetidae(flyingfishes) and Scomberesocidae (sauries)were considered monophyletic by both Colletteet al. (1984; Fig. 7) and Lovejoy (2000; Fig. 8).The families Hemiramphidae (halfbeaks) andBelonidae (needlefishes), considered monophyl-etic by Collette et al. (1984), were hypothesizedto be paraphyletic by Lovejoy (2000).

No non-beloniform taxon was included inLovejoy’s analysis and a single adrianichthyid spe-cies (Oryzias matanensis) was an outgroup to theingroup exocoetoids. Therefore, the analysis wasnot a test of beloniform or of exocoetoid mono-phyly. Further, some intriguing taxa were not in-cluded in the study, such as the southern Africanneedlefish, Petalichthys, which remains in a“halfbeak stage” of development for a relativelylong time before both upper and lower jaws be-come elongate (Collette et al., 1984:342;Boughton et al., 1991). Also missing was thehemiramphid Oxyporhamphus recently reclassi-fied as a flyingfish, family Exocoetidae, based ona re-analysis of morphology (Dasilao et al., 1997).Nonetheless, Lovejoy’s (2000; Fig. 8) hypothesisoffers a novel reinterpretation of traditionalbeloniform morphology and invites further studyof the relationship between morphological andmolecular data as used in phylogenetic analyses.Molecular sequences and morphology were com-bined in a total evidence analysis of phylogeneticrelationships of New World needlefishes byLovejoy and Collette (2001).

Monophyly of the internally-fertilizinghalfbeaks, genera Zenarchopterus, Hemirham-phodon, Dermogenys, and Nomorhamphus, wassupported by the morphological studies of Ander-son and Collette (1991), Downing and Burns(1995), Meisner and Burns (1997b), Meisner andCollette (1999), and Meisner (2001). Petalichthysand Oxyporhamphus were included in a reanaly-sis of beloniform phylogeny by Lovejoy et al.(2004) and paraphyly of hemiramphids andbelonids corroborated. The last three halfbeakgenera are viviparous and together form a mono-phyletic group as corroborated by these analyses.A fifth genus, the monotypic Tondanichthys, de-scribed from the type series of ten specimens thatdoes not include a mature male, is inferred to beinternally fertilizing (Collette, 1995; Meisner andCollette, 1999).

Reproductive biology has been used to inferphylogenetic relationships among live-bearinghalfbeaks. Dermogenys is diagnosed by largesperm bundles and intrafollicular development;whereas, Nomorhamphus is diagnosed by smallsperm bundles and a long intraluminal devel-opmental period (Downing and Burns, 1995;Meisner and Burns, 1997b; Meisner andCollette, 1999; Meisner, 2001). These genericlimits are in contrast to those of Brembach(1991).

Ricefish females are known to carry bundlesof fertilized eggs until hatching (Fig. 9), ratherthan depositing them on over-hanging vegeta-tion or the substrate. Aquarium-maintained O.

Figure 8.Phylogenetic relationships amongbeloniform fishes, simplified fromLovejoy (2000: Fig. 2). The familiesExocoetidae and Scomberesocidaewere each consideredmonophyletic; the familiesHemiramphidae and Belonidaeparaphyletic. “Hemiramphids (IF)”refers to the four internally fertilizinggenera, Zenarchopterus,Nomorhamphus, Dermogenys andHemirhamphodon

Oryzias matanensis

Exocoetidae

hemiramphids

hemiramphids

hemiramphids (IF)

belonids

belonids

belonids

belonids

Scomberesocidae


nigrimas females reportedly carry large bundlesof eggs, but deposit them among plants or onthe substrate soon after spawning (Kottelat,1990a:54). In spawning in open water, ricefishesare similar to the above mentioned Funduluslima (Brill, 1982). Embryos in the clusters arerelatively well developed, with large, well-formed eyes and pigmented bodies, and appearnear hatching (Fig. 10). Because these fertilizedeggs may be carried until hatching, Kottelat(1990a:62) proposed that ricefishes be consid-ered a distinct reproductive guild for which hecoined the term “pelvic brooders.” This repre-sents one of the few cases of parental care inoviparous atherinomorphs. Females carryingclusters of fertilized eggs, long known in themedaka, Oryzias latipes (Yamamoto, 1975:7),has been reported in at least eight other ricefishspecies, 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) andis 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 themedaka, Oryzias latipes, by Amemiya andMurayama (1931). One specimen of Adrianich-thys kruyti, a large, pelagic ricefish from LakePoso, Sulawesi, was reported to be hermaphro-ditic, having both testis and ovary, by Klie(1988, in Kottelat, 1990a:57). Much informa-tion available on the reproduction of the large

Figure 9.Oryzias dancena, USNM 313908,female, 23.8 mm SL, with cluster offertilized eggs. Photo by H. H. Tan

Figure 10.Xenopoecilus sarasinorum, CMK6557, female, 53.4 mm SL. Above,embryo cluster held to bodyposterior to pelvic fins. Below, samespecimen, chorion measuresapproximately 2 mm in diameter Fig. 9

Fig. 10


ricefishes is anecdotal (see Weber and de Beau-fort, 1922; Kottelat, 1990a; Rosen, 1964). Moredetailed study of the reproductive morphologyof the large Sulawesi ricefishes is needed to de-termine the extent of internal fertilization, em-bryo retention, and hermaphroditism in theAdrianichthyidae.

Atheriniformes

Since 1981, there have been several, solely mor-phological, phylogenetic analyses of atherini-form fishes (White et al., 1984; Stiassny, 1990;Saeed et al., 1994; and Dyer and Chernoff,1996; Fig. 11). Monophyly of atheriniforms wasnot supported by Rosen and Parenti (1981) orParenti (1984b), but it has been argued forstrongly in these other studies.

Atheriniform monophyly was supported bytwo developmental characters by White et al.(1984:357): short preanal length of flexion lar-vae and a single row of melanophores on thedorsal margin of larvae. Eight adult characterswere added to the diagnosis by Dyer andChernoff (1996:1): vomerine ventral face con-cave, long A1 muscle tendon to lacrimal, twoanterior infraorbital bones, pelvic-rib ligament,pelvic medial plate not extended to anterior end,and second dorsal-fin spine flexible.

This atheriniform diagnosis requires ho-moplasy within several characters. For example,newly hatched adrianichthyids also have a singlerow of dorsal melanophores (White et al.,1984:359). Number of anterior infraorbital bonesranges from one to three in atheriniforms (Dyerand Chernoff, 1996). The three anterior infraor-bital bones of the rainbowfishes Melanotaenia andChilatherina were considered evidence of theirsister-group relationship by Dyer and Chernoff(1996:67), whereas Rosen and Parenti (1981)considered the character to be primitive foratherinomorphs. Outgroups of atheriniforms inDyer and Chernoff ’s (1996: Table 2) data matrixinclude representative cyprinodontiform andbeloniform taxa and the mullet, Mugil. Given thestrong support for atherinomorph monophylyand the tentative support for a mullet-ather-inomorph sister group relationship, as arguedabove, additional acanthomorph outgroups mayyield alternate interpretations of phylogeny whenincluded in a parsimony analysis of either mor-phological or molecular data.

Despite the ambiguity in distribution of char-acters discussed above, I use the formal termorder Atheriniformes for the included taxa be-

cause it is more popular than the vernacular“atherinoids.”

Phylogenetic analyses of atheriniform sub-groups include an analysis of the internallyfertilizing freshwater and coastal family Phal-lostethidae, proposed as sister taxon of the ma-rine Dentatherina (Parenti, 1984b, 1989b). Thissister group relationship was corroborated byDyer and Chernoff (1996) who classifiedDentatherina in an expanded Phallostethidae.Atheriniformes are oviparous, although reportof facultative embryo retention would not besurprising. Phallostethids (sensu Parenti, 1989b)are internally fertilizing and lay fertilized eggs.Internal fertilization was reported in the brooksilverside, Labidesthes sicculus, by Grier et al.(1990).

Other morphological phylogenies of sub-groups of Atheriniformes include Chernoff(1986, menidiines), Saeed et al. (1989, Pseudo-mugilidae), and Dyer (1998, Atherinopsidae).

Conclusions

The series Atherinomorpha is a well-corrobo-rated, monophyletic taxon. As for other suchwell-corroborated taxa, the list of charactersdiagnostic of the Atherinomorpha continues togrow (Rosen and Parenti, 1981; Parenti, 1993;

Figure 11.Phylogenetic relationships amongthe subfamilies, families or tribes ofatheriniform fishes as proposed byDyer and Chernoff (1996), followingDyer (1998: Fig. 1d)

Atherinopsidae

Notocheiridae

Bedotiinae

Melanotaeniinae

Telmatherinini

Pseudomugilini

Atherionidae

Dentatherininae

Phallostethinae

Atherinomorinae

Craterocephalinae

Atherininae


herein). Each of its included orders, the Cyprin-odontiformes, Beloniformes, and Atherini-formes is monophyletic, although the qualityand quantity of support for each is variable.

Molecular tests of morphological hypothesesof atherinomorph relationships agree, in largepart. Where they differ, molecules present novelhypotheses of relationship, most notably inbeloniforms (Lovejoy, 2000), and invite rein-terpretations of our traditional understandingof morphological characters. Not all molecularanalyses agree, however. Two phylogenetic hy-potheses based on partial sequences of theoncogene, X-src, yielded different interpretationsof relationships of the poeciliid Tomeurus. No-tably, Parker’s (1997: Fig. 2) hypothesis of thesister group relationship of Tomeurus andCnesterodon was corroborated in a reinterpreta-tion of morphological data by Ghedotti (2000).Molecular and morphological hypotheses mayinform each other and point to weak areas inthe other analysis. The Cyprinodontiformes hasbeen studied most intensively during the pasttwenty years; morphological and molecularanalyses share repeated statements of relation-ship to such a degree that a stable classificationof the order is within reach.

Monophyly of the order Atheriniformes hasbeen supported by numerous morphologicalstudies, 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 anatherinomorph outgroup, however. Their choiceof mullets may be appropriate, but a broaderrange 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 willnot substitute for that based on morphology.

Atherinomorph fishes likely will continue tobe studied broadly, especially as they include nu-merous model organisms, such as the poeciliidsXiphophorus, Poecilia, and Gambusia, thericefishes Oryzias, as well as a wide range of live-bearing teleost fish taxa in the families Goodeidaeand Anablepidae and other poeciliids. Phyloge-netic hypotheses of atherinomorphs and theirsubgroups are among the most examined andre-examined within bony fishes, which makestheir continued use as model organisms and theirunique role in understanding evolution of fishreproduction even more justifiable.

Acknowledgments and abbreviations

I am indebted to the organizers, Mari CarmenUribe, Raúl Pineda, Topiltzin Contreras andHarry J. Grier for inviting me to participate inthe II International Symposium on LivebearingFishes. Research on ricefish systematics was sup-ported, in part, by NSF grant BSR 87-00351.Heok Hui Tan, National University of Singa-pore, photographed the ricefish in figure 8.Maurice Kottelat, Switzerland, lent specimens inhis care and discussed ricefish systematics. Dar-rell Siebert allowed access to specimens at theBMNH. My colleagues, Ralf Britz, Bruce Col-lette, Carlos Figueiredo, Dave Johnson, AmyMeisner, Joe Nelson, and Melanie Stiassny readand provided helpful comments on earlier ver-sions of the manuscript.

Institutional abbreviations: BMNH, TheNatural History Museum, London; USNM,United States National Museum of Natural His-tory, Washington, DC; CMK, Collection ofMaurice Kottelat, Cornol, Switzerland.


References

Aarn, Shepherd AM. 2001. Descriptive anatomy ofEpiplatys sexfasciatus (Cyprinodontiformes: Aplocheili-dae) and a phylogenetic analysis of Epiplatina. Cy-bium 25:209-225.

Able KW. 1984. Cyprinodontiformes: development. In:Moser HG, Richards WJ, Cohen DM, Fahay MP,Kendall AW, Jr., Richardson SL, editors, Ontogenyand systematics of fishes. Special Publication No. 1Supplement to Copeia. American Society of Ichthy-ologists and Herpetologists. p 362-368.

Amemiya I, Murayama S. 1931. Some remarks on theexistence of developing embryos in the body of anoviparous cyprinodont, Oryzias (Aplocheilus) latipes(Temminck and Schlegel). Proc Imp Acad Japan7:176-178.

Anderson WD III, Collette BB. 1991. Revision of thefreshwater viviparous halfbeaks of the genus Hemir-hamphodon (Teleostei: Hemiramphidae). IchthyolExplor Freshwaters 2:151-176.

Basolo AL. 1991. Male swords and female preferences.Science 253:1426-1427.

Bernardi G. 1997. Molecular phylogeny of the Funduli-dae (Teleostei, Cyprinodontiformes) based on the cy-tochrome b gene. In: Molecular Systematics of Fishes.Kocher TD, Stepien CA, editors, Academic Press, SanDiego. p 189-197.

Blanco GJ. 1947. The breeding activities and embryologyof Aplocheilus luzonensis Herre and Ablan. PhilippineJ Sci 77:89-93.

Boughton DA, Collette BB, McCune AR. 1991. Hetero-chrony in jaw morphology of needlefishes (Teleostei:Belonidae). Syst Zool 40:329-354.

Breden F, Ptacek MB, Rashed M, Taphorn D, FigueiredoCA. 1999. Molecular phylogeny of the live-bearingfish genus Poecilia. Mol Phyl Evol 12:95-104.

Brembach M. 1991. Systematik und Fortplflanzungsbiologieder lebendbärenden Halbschnäbler der gattung Der-mogenys und Nomorhamphus (Hemirhamphidae [sic],Pisces). Verlag Natur Wissenschaft Solingen. 201 pages.

Brill, JS, Jr. 1982. Observations on the unique reproduc-tive behavior of Fundulus lima Vaillant a killifish fromBaja, [sic] California. Freshwater Mar Aquar 5:9-10,12-15, 74-76, 78, 79, 82, 84-87, 90.

Chambers J. 1987. The cyprinodontiform gonopodium,with an atlas of the gonopodia of the fishes of the ge-nus Limia. J Fish Biol 30:389-418.

Chambers J. 1990. The gonopodia of the fishes of thetribe Cnesterodontini (Cyprinodontiformes, Poeci-liidae). J Fish Biol 36:903-916.

Chernoff B. 1986. Phylogenetic relationships and reclas-sification of menidiine silverside fishes with emphasison the tribe Membradini. Proc Acad Nat Sci Phila138(1):189-249.

Collette BB. 1995. Tondanichthys kottelati, a new genusand species of freshwater halfbeak (Teleostei: Hemir-amphidae) from Sulawesi. Ichthyol Explor Freshwa-ters 6:171-174.

Collette BB, McGowen GE, Parin NV, Mito S. 1984.Beloniformes: development and relationships. In:Moser HG, Richards WJ, Cohen DM, Fahay MP,Kendall AW, Jr., Richardson SL. editors, Ontogenyand systematics of fishes. Special Publication No. 1,Supplement to Copeia. American Society of Ichthy-ologists and Herpetologists. p 335-354.

Costa WJEM. 1990. Análise filogenética da famíliaRivulidae (Cyprinodontiformes, Aplocheiloidei). RevBrasil Biol 50:65-82.

Costa WJEM. 1995a. Revision of the neotropical annualfish genus Cynopoecilus (Cyprinodontiformes: Rivuli-dae). Copeia. 1995:456-465.

Costa WJEM. 1995b. Pearl Killifishes, the Cynolebia-tinae, systematics and biogeography of the neotropicalannual fish subfamily (Cyprinodontiformes: Rivuli-dae). TFH Publications, Inc. Neptune City, NJ.

Costa WJEM. 1996a. Phylogenetic and biogeographicanalysis of the neotropical annual fish genus Simpson-ichthys (Cyprinodontiformes: Rivulidae). J Comp Biol1:129-140.

Costa WJEM. 1996b. Relationships, monophyly andthree new species of the neotropical miniature poeci-liid genus Fluviphylax (Cyprinodontiformes: Cyprin-odontoidei). Ichthyol Explor Freshwaters 7:111-130.

Costa WJEM. 1997. Phylogeny and classification of theCyprinodontidae revisited (Teleostei: Cyprinodonti-formes): Are Andean and Anatolian killifishes sistertaxa? J Comp Biol 2:1-17.

Costa WJEM. 1998a. Phylogeny and classification of theCyprinodontiformes (Euteleostei: Atherinomorpha) areappraisal. In: Malabarba LR, Reis RE, Vari RP,Lucena ZMS, Lucena CAS, editors, Phylogeny andClassification of Neotropical Fishes, EDIPUCRS(Editora Pontifícia Universidade Católica do RioGrande do Sul), Porto Alegre, Brazil. p 537-560.

Costa WJEM. 1998b. Phylogeny and classification ofRivulidae revisited: origin and evolution of annualismand miniaturization in rivulid fishes (Cyprinodonti-formes: Aplocheiloidei). J Comp Biol 3:33-92.

Costa WJEM. 2002. Peixes Anuais Brasileiros. Diversi-dade e Conservação. Série Pesquisa no. 60. Editora daUFPR, Curitiba.

Dasilao JC Jr, Sasaki K, Okamura O. 1997. The hemir-amphid Oxyporhamphus is a flyingfish (Exocoetidae).Ichthyol Res 44:101-107.

Devlin AJ, Danks JA, Faulkner MK, Power DM, CanarioAVM, Martin TJ, Ingleton PM. 1996. Immuno-chemical detection of a parathyroid hormone-relatedprotein in the saccus vasculosus of a teleost fish. GenComp Endocrin 101:83-90.

Lynne R. Parenti •The Phylogeny of Atherinomorphs. Reproductive System 27

Downing AL, Burns JR. 1995. Testis morphology andspermatozeugma formation in three genera of vivipa-rous halfbeaks: Nomorhamphus, Dermogenys, andHemirhamphodon. J Morph 225:329-343.

Dyer BS. 1998. Phylogenetic systematics and historicalbiogeography of the neotropical silverside familyAtherinopsidae (Teleostei: Atheriniformes). In: Mala-barba LR, Reis RE, Vari RP, Lucena ZMS, LucenaCAS, editors, Phylogeny and classification of Neotro-pical Fishes, EDIPUCRS (Editora Pontifícia Univer-sidade Católica do Rio Grande do Sul), Porto Alegre,Brazil. p 519-536.

Dyer BS, Chernoff B. 1996. Phylogenetic relationshipsamong atheriniform fishes (Teleostei: Atherinomor-pha). Zool J Linn Soc 117:1-69.

Ghedotti MJ. 1998. Phylogeny and classification of theAnablepidae (Teleostei: Cyprinodontiformes). In:Malabarba LR, Reis RE, Vari RP, Lucena ZMS,Lucena CAS, editors, Phylogeny and classification ofNeotropical Fishes, EDIPUCRS (Editora PontifíciaUniversidade Católica do Rio Grande do Sul), PortoAlegre, Brazil. p 561-582.

Ghedotti MJ. 2000. Phylogenetic analysis and taxonomyof the poecilioid fishes (Teleostei: Cyprinodonti-formes). Zool J Linn Soc 130:1-53.

Gosline WA. 1971. Functional Morphology and classifi-cation of teleostean fishes. The University of HawaiiPress, Honolulu.

Grady JM, Coykendall DK, Collette BB, Quattro JM.2001. Taxonomic diversity, origin, and conservation sta-tus of Bermuda killifishes (Fundulus) based on mitochon-drial cytochrome b phylogenies. Conserv Gen 2:41-52.

Grant EC, Riddle BR. 1995. Are the endangered spring-fish (Crenichthys Hubbs) and poolfish (EmpetrichthysGilbert) fundulines or goodeids? A mitochondrialDNA assessment. Copeia 1995:209-212.

Greenwood PH, Rosen DE, Weitzman SH, Myers GS.1966. Phyletic studies of teleostean fishes, with a pro-visional classification of living forms. Bull Amer MusNat Hist 131:345-455.

Griem JN, Martin KLM. 2000. Wave action: the environ-mental trigger for hatching in the California grunionLeuresthes tenuis (Teleostei: Atherinopsidae). Mar Biol137:177-181.

Grier HJ. 1981. Cellular organization of the testis andspermatogenesis in fishes. Amer Zool 21:345-357.

Grier HJ. 1984. Testis structure and formation of sper-matophores in the atherinomorph teleost Horaichthyssetnai. Copeia 1984:833-839.

Grier HJ, Collette BB. 1987. Unique spermatozeugmatain testes of halfbeaks of the genus Zenarchopterus(Teleostei: Hemiramphidae). Copeia 1987:300-311.

Grier HJ, Linton JR, Leatherland JF, deVlaming VL.1980. Structural evidence for two different testiculartypes in teleost fishes. Amer J Anat 159:331-345.

Grier HJ, Moody DP, Cowell BC. 1990. Internal fertili-zation and sperm morphology in the brook silverside,Labidesthes sicculus (Cope). Copeia 1990:221-226.

Grier HJ, Parenti LR. 1994. Reproductive biology andsystematics of phallostethid fishes as revealed by gonadstructure. Environ Biol Fishes 41:287-299.

Grier HJ, Uribe MC, Parenti LR, De la Rosa-Cruz G. 2005.Fecundity, the germinal epithelium, and folliculogenesisin viviparous fishes. In: this volume. p 191-216.

Haas V. 1993. Xiphophorus phylogeny, reviewed on thebasis of courtship behaviour. In: Schroeder JH, BauerJ, Schartl M. editors, Trends in ichthyology. Blackwell,London. p 279-288.

Hamilton A. 2001. Phylogeny of Limia (Teleostei: Poeci-liidae) based on NADH dehydrogenase subunit 2 se-quences. Mol Phyl Evol 19:277-289.

Hrbek T, Larson A. 1999. The evolution of diapause in thekillifish family Rivulidae (Atherinomorpha, Cyprin-odontiformes): A molecular phylogenetic and biogeo-graphic perspective. Evolution 53:1200-1216.

Hrbek T, Meyer A.. 2003. Closing of the Tethys Sea andthe phylogeny of Eurasian killifishes (Cyprinodonti-formes: Cyprinodontidae). J Evol Biol 16(1):17-36.

Johnson GD, Patterson C. 1993. Percomorph phylogeny:a survey of acanthomorphs and a new proposal. BullMar Sci 52:554-626.

Klie K. 1988. Morphologische und histologische Unter-suchungen an neuem Adrianichthys-Material – EinBeitrag zur Systematik und Verwandschaft der Adrian-ichthyidae (Pisces: Atheriniformes). Diplomarbeit,Universität, Hamburg.

Kottelat M. 1990a. Synopsis of the endangered Buntingi(Osteichthyes: Adrianichthyidae and Oryziidae) ofLake Poso, Central Sulawesi, Indonesia, with a newreproductive guild and descriptions of three new spe-cies. Ichthyol Explor Freshwaters 1:49-67.

Kottelat M. 1990b. The ricefishes (Oryziidae) of theMalili lakes, Sulawesi, Indonesia, with description ofa new species. Ichthyol Explor Freshwaters 1:151-166.

Kottelat M, Whitten AJ, Kartikasari SN, WirjoatmodjoS. 1993. Freshwater fishes of Western Indonesia andSulawesi. Hong Kong: Periplus Editions (HK) Ltd. incollaboration with the Environmental ManagementDevelopment in Indonesia (EMDI) Project, Ministryof State for Population and Environment, Republic ofIndonesia, Jakarta.

Kulkarni CV. 1940. On the systematic position, structuralmodifications, bionomics and development of a re-markable new family of cyprinodont fishes from theprovince of Bombay. Rec Indian Mus 42:379-423.

Lazara KJ. 2001. The killifishes: an annotated checklist,synonymy and bibliography of recent oviparous cy-prinodontiform fishes: The Killifish Master Index,Fourth Edition. American Killifish Association, Cin-cinnati, OH.

• Viviparous Fishes Systematics, Biogeography, and Evolution28

Li SZ. 2001. On the position of the suborder Adrian-ichthyoidei. Acta Zootax Sin 26:583-588.

Loureiro M, deSá RO. 1996. External morphology of thechorion of the annual fishes Cynolebias (Cyprinodon-tiformes: Rivulidae). Copeia 1996:1016-1022.

Loureiro M, deSá RO. 1998. Osteological analysis of thekillifish genus Cynolebias (Cyprinodontiformes: Ri-vulidae). J Morph 1998:245-262.

Lovejoy NR. 2000. Reinterpreting recapitulation: system-atics of needlefishes and their allies (Teleostei: Beloni-formes). Evolution 54:1349-1362.

Lovejoy N, Collette BB. 2001. Phylogenetic relationshipsof the new world needlefishes (Teleostei: Belonidae)and the biogeography of transitions between marineand freshwater habitats. Copeia 2001:324-338.

Lovejoy NR, Iranpour M, Collette BB. 2004. Phylogenyand ontogeny of beloniform fishes. Integr Comp Biol44:366-377.

Lüssen A, Falk TM, Villwock W. 2003. Phylogenetic pat-terns in populations of Chilean species of the genusOrestias (Teleostei: Cyprinodontidae): results of mito-chondrial DNA analysis. Mol Phyl Evol 29:151-60.

Lydeard C. 1993. Phylogenetic analysis of species richness:has viviparity increased the diversification of actino-pterygian fishes? Copeia 1993:514-518.

Lydeard C, Wooten MC, Meyer A. 1995. Cytochrome bsequence variation and a molecular phylogeny of thelive-bearing fish genus Gambusia (Cyprinodonti-formes: Poeciliidae). Can J Zool 73:213-227.

Marcus JM, McCune AR. 1999. Ontogeny and phylog-eny in the northern swordtail clade of Xiphophorus.Syst Biol 48:491-522.

Martin KL. 1999. Ready and waiting: delayed hatchingand extended incubation of anamniotic vertebrateterrestrial eggs. Amer Zool 39:279-288.

Martínez VH, Monasterio de Gonzo GAM. 2002. Testismorphology and spermatozeugma formation in Jenyn-sia multidentata. In: Meeting book, II InternationalSymposium on Livebearing Fishes. Abstract. Queré-taro, Qro., Mexico.

Meisner AD. 2001. Phylogenetic systematics of theviviparous halfbeak genera Dermogenys and Nomor-hamphus (Teleostei: Hemiramphidae: Zenarcho-pterinae). Zool J Linn Soc 133 (2): 199-283.

Meisner AD, Burns JR. 1997a. Testis and andropodialdevelopment in a viviparous halfbeak, Dermogenys sp.(Teleostei: Hemiramphidae). Copeia 1997:44-52.

Meisner AD, Burns JR. 1997b. Viviparity in the halfbeakgenera Dermogenys and Nomorhamphus (Teleostei:Hemiramphidae). J Morph 234:295-317.

Meisner AD, Collette BB. 1999. Generic relationships of theinternally-fertilized southeast Asian halfbeaks (Hemi-ramphidae: Zenarchopterinae). In: Séret B, Sire JY, edi-tors, Proceedings of the 5th Indo-Pacific Fish Conference,Nouméa, 1997. Soc Fr Ichtyol, Paris. p 69-76.

Meyer A, Lydeard C. 1993. The evolution of copulatoryorgans, internal fertilization, placentae and viviparityin killifishes (Cyprinodontiformes) inferred from aDNA phylogeny of the tyrosine kinase gene X-src. ProcR Soc Lond B 254:153-162.

Meyer A, Morrissey JM, Schartl M. 1994. Recurrent ori-gin of a sexually selected trait in Xiphophorus fishes in-ferred from molecular phylogeny. Nature 386:539-542.

Miya M, Takeshima H, Endo H, Ishiguro NB, Inoue JG,Mukai T, Satoh TP, Yamaguchi M, Kawaguchi A,Mabuchi K, Shirai SM, Nishida M. 2003. Majorpatterns of higher teleostean phylogenies: a new per-spective based on 100 complete mitochondrial DNAsequences. Mol Phyl Evol 26:121-138.

Mojica CL, Meyer A, Barlow GW. 1997. Phylogeneticrelationships of species of the genus Brachyrhaphis(Poeciliidae) inferred from partial mitochondrialDNA sequences. Copeia 1997:298-305.

Morris MR, de Queiroz K, Morizot DC. 2001. Phyloge-netic relationships among populations of northernswordtails (Xiphophorus) as inferred from allozymedata. Copeia 2001:65-81.

Murphy WJ, Collier GE. 1997. A molecular phylogenyfor aplocheiloid fishes (Atherinomorpha, Cyprin-odontiformes): the role of vicariance and the origins ofannualism. Mol Biol Evol 14: 790-799.

Murphy WJ, Collier GE. 1999. Phylogenetic relation-ships of African killifishes in the genera Aphyosemionand Fundulopanchax inferred from mitochondrialDNA sequences. Mol Phyl Evol 11:351-360.

Murphy WJ, Nguyen TNP, Taylor EB, Collier GE.1999a. Mitochondrial DNA phylogeny of West Afri-can aplocheiloid killifishes (Cyprinodontiformes,Aplocheilidae). Mol Phyl Evol 11:343-350.

Murphy WJ, Thomerson JE, Collier GE. 1999b. Phylog-eny of the neotropical killifish family Rivulidae (Cy-prinodontiformes, Aplocheiloidei). Mol Phyl Evol13:289-301.

Nelson JS. 1984. Fishes of the World, Second Edition.John Wiley and Sons.

Nelson JS. 1994. Fishes of the World, Third Edition. JohnWiley and Sons.

Parenti LR. 1981. A phylogenetic and biogeographicanalysis of cyprinodontiform fishes (Teleostei, Ather-inomorpha). Bull Amer Mus Nat Hist 168:335-557.

Parenti LR. 1984a. A taxonomic revision of the Andeankillifish genus Orestias (Cyprinodontiformes, Cyprin-odontidae). Bull Amer Mus Nat Hist 178: 107-214.

Parenti LR. 1984b. On the relationships of phallostethidfishes (Atherinomorpha), with notes on the anatomyof Phallostethus dunckeri Regan, 1913. Amer MusNovit No. 2779. 12 p.

Parenti LR. 1987. Phylogenetic aspects of tooth and jawstructure of the medaka, Oryzias latipes, and otherbeloniform 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 AmerKillifish Assoc 22(3):79-84.

Parenti LR. 1989b. A phylogenetic revision of the phal-lostethid fishes (Atherinomorpha, Phallostethidae).Proc Cal Acad Sci 46:243-277.

Parenti LR. 1993. Relationships of atherinomorph fishes(Teleostei). Bull Mar Sci 52:170-196.

Parenti LR, Song J. 1996. Phylogenetic significance of thepectoral/pelvic fin association in acanthomorph fishes:A reassessment using comparative neuroanatomy. In:Stiassny MLJ, Parenti LR, Johnson GD, editors, Inter-relationships of fishes, Academic Press, San Diego.p 427-444.

Parenti LR, Grier HJ. 2004. Evolution and phylogeny ofgonad morphology in bony fishes. Integr Comp Biol44:333-348.

Parker A. 1997. Combining molecular and morphologi-cal data in fish systematics: examples from the Cyprin-odontiformes. In: Kocher TD, Stepien CA, editors,Molecular Systematics of Fishes, Academic Press, SanDiego. p 163-188.

Parker A, Kornfield I. 1995. A molecular perspective onevolution and zoogeography of cyprinodontid killi-fishes. Copeia 1995:8-21.

Rauchenberger M. 1989. Systematics and biogeographyof the genus Gambusia (Cyprinodontiformes: Poeci-lidae [sic]). Amer Mus Novit No. 2951. 74 p.

Roberts CD. 1993. Comparative morphology of spinedscales and their phylogenetic significance in theTeleostei. Bull Mar Sci 52:60-113.

Rodríguez CM. 1997. Phylogenetic analysis of the tribePoeciliini (Cyprinodontiformes: Poeciliidae). Copeia1997:663-679.

Rosen DE. 1964. The relationships and taxonomic posi-tion of the halfbeaks, killifishes, silversides and theirrelatives. Bull Amer Mus Nat Hist 127:217-268.

Rosen DE, Bailey RM. 1963. The poeciliid fishes (Cy-prinodontiformes), their structure, zoogeography, andsystematics. Bull Amer Mus Nat Hist 126:1-176.

Rosen DE, Parenti LR. 1981. Relationships of Oryzias,and the groups of atherinomorph fishes. Amer MusNovit No. 2719. 25 p.

Saeed B, Ivantsoff W, Allen GR. 1989. Taxonomic revi-sion of the family Pseudomugilidae (Order Ather-iniformes). Aust J Freshwater Res 40:719-787.

Saeed B, Ivantsoff W, Crowley LELM. 1994. Systematicrelationships of atheriniform families within DivisionI of the Series Atherinomorpha (Acanthopterygii) withrelevant historical perspectives. Voprosi Ikhtiologii34(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

taxonomic revision of the genus Rheocles (Atherino-morpha: Bedotiidae). Amer Mus Novit No. 2979. 33 p.

Stiassny MLJ. 1993. What are grey mullets? Bull Mar Sci52:197-219.

Stiassny MLJ, Moore JA. 1992. A review of the pelvicgirdle of acanthomorph fishes, with comments onhypotheses of acanthomorph interrelationships. ZoolJ Linn Soc 104:209-242.

Tsuneki K. 1992. A systematic survey of the occurrence ofthe hypothalamic saccus vasculosus in teleost fish.Acta Zool 73:67-77.

Webb SA, Graves JA, Macías-Garcia C, Magurran AE,Foighil DO, Ritchie MG. 2004. Molecular phylog-eny of the livebearing Goodeidae (Cyprinodonti-formes). Mol Phyl Evol 30:527-544.

Weber M, de Beaufort LF. 1922. Family Adrianichthyi-dae. In: IV. Heteromi, Solenichthyes, Synentognathi,Percesoces, Labyrinthici, Microcyprini. The Fishes ofthe Indo-Australian Archipelago. EJ Brill, Leiden. p376-381.

White BN, Lavenberg RJ, McGowen GE. 1984. Ather-iniformes: development and relationships. In: MoserHG, Richards WJ, Cohen DM, Fahay MP, KendallAW, Jr., Richardson SL, editors, Ontogeny and sys-tematics of fishes. Special Publication No. 1, Supple-ment to Copeia, American Society of Ichthyologistsand Herpetologists. p 355-362.

Wiley EO, Johnson GD, Dimmick WW. 2000. The in-terrelationships of acanthomorph fishes: A total evi-dence approach using molecular and morphologicaldata. Biochem Syst Ecol 28: 319-350.

Wourms JP. 1972. The developmental biology of annualfishes. III. Pre-embryonic and embryonic diapause ofvariable duration in the eggs of annual fishes. J ExpZool 182:389-414.

Wourms JP. 1976. Annual fish oogenesis. I. Differentia-tion of the mature oocyte and formation of the pri-mary envelope. Dev Biol 50:338-354.

Wourms JP, Sheldon H. 1976. Annual fish oogenesis. II.Formation of the secondary egg envelope. Dev Biol50:355-366.

Yamamoto T. 1975. Medaka (Killifish) Biology andStrains. Series of Stock Culture in Biological Field.Keigaku Publishing Company, Tokyo.

• Viviparous Fishes Systematics, Biogeography, and Evolution30

the phylogeny of evolution of a novel fish reproductive...

Documents