Systematics and biogeography of the genus Phalloptychus Eigenmann , 1907 ( Cyprinodontiformes : Poeciliidae : Poeciliinae )

The genus Phalloptychus is revised. Phalloptychus iheringii is removed from the synonym of P. januarius. Three species are recognized: P. eigenmanni, P. januarius, and P. iheringii. Phalloptychus eigenmanni can be distinguished from its congeners by the number of pelvic-fin rays in females, the number of predorsal scales, and the predorsal distance in females. Phalloptychus iheringii and P. januarius can be distinguished by the number of epipleural ribs, number of gonopodial rays, and a significantly different number of vertical bars along body side in females. Lectotypes are designated for Girardinus iheringii and G. januarius. Redescriptions and known distribution ranges are provided for each species as well as an identification key. Derived features supportting the monophyly of the genus and infrageneric clades are presented and discussed. The biogeographic implications of the geographic distribution of Phalloptychus species are discussed.


Introduction
The genus Phalloptychus Eigenmann, 1907 contains poeciliids endemic to South America occurring along the coastal drainages from Bahia to Rio Grande do Sul states of Brazil (Fig. 1).Males are typically smaller than females; the gonopodium is long and asymmetrical and bears spines along its length.Papers concerning Phalloptychus are extremely rare in systematic literature.These are mostly confined to original descriptions.Even from a biological standpoint Phalloptychus species are poorly studied.
The taxonomic history of Phalloptychus began with the first described species currently in the genus: Girardinus januarius Hensel, 1868, based on specimens from Rio de Janeiro.A second species, G. iheringii Boulenger, 1889, was described from Rio Grande do Sul on the basis of specimens collected by Hermann von Ihering. Garman (1895) synonymized G. iheringii and G. caudimaculatus Hensel, 1868 [= Phalloceros caudimaculatus] with G. januarius, since the later had page priority over the second.Garman (1895) erected the genus Glaridodon on the basis of tooth and jaw morphology, with G. uninotatus Poey, 1854 as type species, and placed G. januarius in his new genus Glaridodon.The name Glaridodon is preoccupied in Therapsida, and Garman (1896) replaced it with Glaridichthys in the subsequent year.It is quite evident from Garman's (1895) figure of a Gl.januarius male (Plate VIII, fig.15), showing a distal appendix on gonopodium, that he had only Phalloceros Eigenmann, 1907 specimens at hand.That is the reason why he considered both genera to be synonyms.Schneider & Ribeiro's (1903) examination of the fish collection in the Museu Nacional do Rio de Janeiro, identified and recognized Gl. januarius and described G. zonatus based on specimens with uncertain locality.Steindachner (1907) did not treat Glaridodon as a preoccupied name and recognized Glaridodon as a subgenus of Girardinus Poey, 1854 based on his examination of material

P R O O F S
from Santa Catarina state.Eigenmann (1907) recognized Glaridichthys januarius as generically distinct from G. caudimaculatus on the basis of teeth and gonopodium structure.Therefore, Eigenmann (1907) created the genus Phalloptychus for Glaridichthys januarius, and erected the genus Phalloceros for G. caudimaculatus.Philippi (1908) considered G. caudimaculatus a junior synonym of Glaridichthys januarius, for he believed that Eigenmann created Phalloptychus and Phalloceros based respectively on juvenile and adult male specimens of Glaridichthys januarius.Moreover, judging from his figures he had solely examined specimens of G. caudimaculatus.Langer (1913) followed Philippi further confusing the situation.Regan (1913) in his revision of the Poeciliinae synonymized G. iheringii and G. zonatus with P. januarius.Henn (1916) following Eigenmann, recognized the distinctness of genera Phalloceros and Phalloptychus and described P. eigenmanni from specimens collected by Haseman in the rio Catu at Alagoinhas, Bahia.
Henn also provided a redescription of P. januarius from specimens collected by Hermann von Ihering in the Rio Grande do Sul and cited Santa Catarina, Rio Grande do Sul, and Uruguay, as part of distribution range of the species.Rudolph von Ihering (1931) considered Phalloptychus composed by two subspecies: P. januarius januarius and P. januarius eigenmanni.More recently, Lucinda (2003) recognized three valid species in the genus: P. eigenmanni, P. iheringii, and P. januarius.Lucinda (2003) regarded Girardinus zonatus as Incertae Sedis in Poeciliinae, since syntypes are missing and the type locality is uncertain.
Lucinda & Reis (2005: fig.2a) presented (but not discussed) a phylogenetic hypothesis of relationships among Phalloptychus species.This hypothesis (Fig. 2) is part of a more inclusive phylogenetic study on the relationships among poeciliine genera.Thus, the phylogenetic position of the genus Phalloptychus in the subfamily Poeciliinae was discussed by Lucinda & Reis (2005: fig.1).Although the transformation series analysis was provided and intrageneric relationships of Phalloptychus were depicted by Lucinda & Reis (2005), these authors did not present the diagnoses of intrageneric clades.These clade diagnoses are provided herein.
Thus, this paper has the following aims: (1) to diagnose the species of Phallotorynus and their geographic distribution; (2) present the diagnoses of the genus and its intrageneric clades and (3) and to propose a hypothesis of biogeographic relationships among those species.

Material and Methods
The entries under examined material for each species follow the sequence: country, state, museum acronym, catalogue number, total number of examined specimens in the lot (in parentheses; number of cleared and stained specimens is indicated by an asterisk and separated from total number by a bar), type status, collection locality, date, collector.

O F S
counts, respectively.All counts were made on the left side of adult specimens whenever possible, except the number of teeth, which have been counted on both sides.In species descriptions, numbers in square brackets following the counts indicate number of specimens for each count.Fourteen measurements were obtained as distances between 13 homologous landmarks on the lateral left surface of head and body, as described in Lucinda (2005).Tables of descriptive morphometrics were elaborated with "Datax, version 4.2" by Roberto Reis and Nelson Fontoura.Measurements, other than SL, are expressed as percents of SL, except those that are subunits of the head, which are expressed as percents of HL.
Nonparametric statistical tests of null hypotheses of character similarity were performed using the software "Sigma Stat for Windows (Jandel Scientific), in order to evaluate meristic data that failed to pass tests for normality and equal variance.This methodology has been extensively described and commented by Weitzman & Malabarba (1999).
Number and disposition of cephalic pores followed the nomenclature of Rosen & Mendelson (1960), Gosline (1949), andParenti (1981).Only adult individuals have been examined to avoid undesirable ontogenetic variation.Nomenclature of the gonopodium follows Rosen & Gordon (1953) and Lucinda & Reis (2005).Descriptions of gonopodium morphology are based on fully developed gonopodia of large adult males.Osteological nomenclature adopted followed Rosen & Bailey (1963) and Parenti (1981).Clearing and staining followed the method of Taylor & Van Dyke (1985).Cladistic procedures, character state assignments, transformation series and clade numbers follow the phylogenetic analysis performed by Lucinda & Reis (2005).Character states illustrations are provided or referred to by Lucinda & Reis (2005).

Results
Three Phalloptychus species are herein recognized: P. eigenmanni, P. januarius, and P. iheringii, whose descriptions are provided below.Autapomorphies and synapomorphies supporting the monophyly of the genus and the infrageneric clade are presented below.Two asterisks indicate uniquely derived and unreversed features.
Distribution.Phalloptychus eigenmanni is solely known from the type locality (Fig. 1).1908.Following Haseman's itinerary during his expedition in South America (Eigenmann, 1911) it is evident that he has only collected in rio Catu, Alagoinhas at that date.Then, the aforementioned evidence leads to the conclusion that the three specimens of lot AMNH 22657 are actually the missing paratypes of P. eigenmanni from FMNH.

Remarks
The sole mature male known from this species has been damaged during clearing and staining for the monograph of Rosen & Bailey (1963).This specimen is a male that is now housed in the American Museum of Natural History as AMNH 22657, from which only the head is still preserved.Since P. eigenmanni is probably extinct (Rosa & Menezes, 1996;MMA, 2004), the only information about its gonopodium is that given in Rosen & Bailey (1963: fig.31E).Phalloptychus iheringii (Boulenger) Fig. 5; Tables 1-2 Girardinus iheringii Boulenger, 1889: 266.Type locality: Rio Grande do Sul.Restricted by Ihering (1893: 29)  eigenmanni by the number of pelvic-fin rays in females (5 vs. 6, respectively), by the number of predorsal scales (10-12 vs. 13, respectively), and the shorther predorsal distance in females (55.6-61.3 vs. 64.0-68.6 % SL, respectively).P. iheringii is distinguished from P. januarius by the number of epipleural ribs (10-11 vs. 12-13 [one specimen had 11 in one side and 12 in the other], respectively), by the number of gonopodial rays (8 vs. 9, respectively), and by a significantly greater number of vertical bars along body sides of females (range = 8-18, median = 12 vs.range = 5-12, median = 7, respectively).Although some overlap occurs, Mann-Whitney rank sum tests (Fig. 6) indicate significant differences (P < 0.001).Preorbital ramus of cephalic sensory system represented by two to five grooved neuromasts.Preorbital canal absent.Anterior portion of supraorbital ramus (pores 1 and 2a) parallel to upper lip with three inconspicuous neuromasts on each side.Posterior portion of supraorbital ramus (pores 2b, 3, 4a) composed of two grooved neuromasts.Posterior remnants of infra-orbital ramus represented by three neuromasts (pores 4b, 5, 6a) and by one short canal (pores 6b and 7).Preopercular ramus represented by large canal (sometimes completely open, forming groove) along preopercular posterolateral border and by prolonged canal along preopercle ventral border opened by four pores.Opercular canal absent.Mandibular ramus composed of two or three superficial neuromasts (pores Z, Ya, and Yb) on anterior border of ventral surface of mandible and by one superficial neuromast near maxillary distal end (pore W).
Distribution.Phalloptychus iheringii occurs in coastal drainages in Santa Catarina and Rio Grande do Sul states of Brazil (Fig. 1).
Remarks.Boulenger (1889) described Girardinus iheringii in honor of Hermann von Ihering who collected and sent him the specimens on which the description has been based (Ihering, 1893).This name has subsequently been used by Eigenmann & Eigenmann (1891) referring to Rio Grande do Sul specimens.Boulenger originally cited the type locality as "Rio Grande do Sul", probably because it was the only collection information he had.Ihering (1893) states that type specimens have been collected "an der Mündung des Rio Camaquam" [= in the mouth of rio Camaquã], thus restricting the type locality.In the same paper, however, Ihering claimed that G. iheringii should be regarded as a junior synonym of G. januarius, based on information provided in letter by Hingeldorf.Presumably, for this reason, subsequent authors have considered both names synonyms.However, the study of color pattern has revealed that the number of vertical bars along body sides of females is significantly greater in southern populations of Phalloptychus from Santa Catarina and Rio Grande do Sul (range = 8-18, median = 12) than for northern populations from Rio de Janeiro, São Paulo and Paraná (range = 5-12, median = 7).See Fig. 6 for a nonparametric statistical expression of differences and simi-larities among Phalloptychus populations.These differences along with differences in number of epipleural ribs and gonopodial rays give support to the resurrection of P. iheringii as a valid species.Phalloptychus januarius (Hensel) Fig. 7; Tables 1-2 Girardinus januarius Hensel, 1868: 360.Type locality: "aus den Pfützen und Gräben um Rio de Janeiro" [= from the puddles and ditches around Rio de Janeiro].
Diagnosis.Phalloptychus januarius can be autapomorphically diagnosed by the possession of nine anal-fin rays in males .Furthermore, Phalloptychus januarius can be distinguished from P. eigenmanni by the number of pelvicfin rays in females (5 vs. 6, respectively), by the number of predorsal scales (10-12 vs. 13, respectively), and the shorter predorsal distance in females (55.respectively).Phalloptychus januarius is distinguished from P. iheringii by the number of epipleural ribs (12-13 [one specimen had 11 in one side and 12 in the other] vs. 10-11, respectively), by the number of gonopodial rays (9 vs. 8, respectively), and by a significantly lower number of vertical bars along body sides of females (range = 5-12, median = 7, vs. range = 8-18, median = 12, respectively).Although some overlap occurs, Mann-Whitney rank sum tests (Fig. 6) indicate significant differences (P < 0.001).Gonopodial complex composed of three functional gonapophyses and nine gonactinosts.Gonactinosts 2, 3, 4 fused.Gonactinost 4 with wing-like expansions.Ligastyle present.Gonopodium sinistrally asymmetrical.Eight gonopodial rays.R1 and R2 unbranched and short, with 8 segments.R3 with dorsal convexity located near base ranging from second to tenth or eleventh segments.Tip of R3, R4a, and R4p ventrally bent and joint.Dorsal convexity located between segments 14 to 21 (rarely 12 to 18) of R4p.Twelve to 17 spines on distal segments of R4p.Spines retrorse except two or three last spines, directed forwards or upwards.R6, R7, R8 branched.Five to 8 segments before bifurcation of ray 6. Anterior and posterior branches of R6 almost or fully ankylosed.Distal end of R6 modified in arrow-shaped expansion.R7 with six or seven segments anterior to bifurcation.Anterior and posterior branches of R7 moderately expanded and partially to completely ankylosed.Anterior branch of R7 greater than posterior.R8 with seven or eight segments anterior to bifurcation.Anterior and posterior branches of R8 normal.Anterior branch of R8 greater than posterior.Four to six segments on anterior branch of R8.Four segments on posterior branch of R8.R9 minute not attached to pterygiophore.Remarks.Besides the types examined, the type series includes the lots ZMB 7423 (4) and CAS-SU 1132 (1).It is not absolutely sure, whether the lot ZMB 31497 belongs to the type series, since this number is based secondarily on an old label from the anatomical collection: "No.25204, R. Janeir [sic]" -the remainder of the label is illegible.This is the collection of the former "Institute of Anatomy of the Humboldt-University" in the 19th century.And this belonged to the Medical Faculty.This is due to the fact that human-anatomists were traditionally doing much research work on comparative anatomy of vertebrates.Thus, collected material was often divided between Zoological (Philosophical) Faculty with the Zoological Collection and the Medical Faculty.Only from approximately 1890 onwards the collections were reunited.But, apparently also R. Virchow, the famous pathologist working at the Charity Hospital at Humboldt-University took over some of the original anatomical collection for study (Paepke & Seegers 1986;P. Bartsch in litt., 2000).As there is no apparent evidence that Hensel has not examined specimens from lot ZMB 31497, it is advisable to label the lot paralectotypes.

Discussion
The genus Phalloptychus is a well-diagnosed, monophyletic group (Lucinda & Reis, 2005).The genus is traditionally allocated in the tribe Cnesterodontini Hubbs.This tribe, as Type localities are listed as in original descriptions.Museum acronyms are: BMNH -Natural History Museum, London; CM -Carnegie Museum, Pittsburgh; FMNH -Field Museum of Natural History, Chicago; FURG -Fundação Universidade de Rio Grande, Rio Grande; MCP -Museu de Ciências e Tecnologia da Pontifícia Universidade Católica do Rio Grande do Sul, Porto Alegre; MHNCI -Museu de História Natural do Capão da Imbúia, Curitiba; MZUSP -Museu de Zoologia da Universidade de São Paulo, São Paulo; CAS-SU -Stanford University Collection at the California Academy of Sciences, San Francisco; UFPB -Universidade Federal da Paraíba, João Pessoa; UMMZ -University of Michigan Museum of Zoology, Ann Arbor; USNM -National Museum of Natural History, Smithsonian Institution, Washington, DC; and ZMB -Museum für Naturkunde, Humboldt-Universität, Berlin.Counts of scales, fin rays and vertebrae were taken according to Lucinda (2005).The three numbers in parentheses separated by bars following the vertebrae counts indicate respectively: number of vertebrae anterior to first dorsal pterygiophore, number of vertebrae between first and last dorsal pterygiophore, and number of vertebrae posterior to last dorsal pterygiophore.Other counts include (a) number of teeth on outer premaxillary row; (b) number of teeth on outer dentary row; (c) number of branchiostegal rays; (l) number of caudal-fin rays attached to hypural plate; (d) number of upper accessory cartilages; and (e) number of lower accessory cartilages.All counts, except scales, were performed in cleared and stained specimens.The minute and incompletely ossified ray attached to the last normal gonopodial ray has been considered in gonopodial ray-counts.Rudimentary and procurrent rays were included in pectoral and caudal fin

Fig. 1 .
Fig. 1.Coastal drainages from Bahia to Rio Grande do Sul, and Uruguay showing distribution of Phalloptychus eigenmanni (triangle), P. januarius (squares), and P. iheringii (circles).Each symbol may represent more than one lot and/ or locality.T = type locality.

Fig. 2 .
Fig. 2. Intrageneric relationships of Phalloptychus.The numbers on the branches refer to the character state transformations series listed by Lucinda & Reis (2005).
. Originally, Henn cited ten paratypes of P. eigenmanni.In 1956, Donn E. Rosen borrowed several lots from FMNH (loan #Z-7445), including three paratypes of P. eigenmanni and possibly never returned them, although his loan was closed in 1970.Since then they have been considered missing in the FMNH.There is correspondence to FMNH (between Loren P. Woods and Donn E. Rosen) that suggests they might have remained at AMNH (with Woods' permission) and that Rosen might have turned them into skeleton preps (M. A. Rogers, pers.comm.).Although neither the original invoice nor the correspondence seem to indicate clearly that these are paratypes, they might be the specimens now housed in the American Museum of Natural History and catalogued as "Phalloptychus eigenmanni Henn, paratypes 3 spec.(1 missing), AMNH 22657 (FMNH 55877) from rio Catu, Alagoinhas" (B.Brown, pers.comm.).Collection data, number of specimens, number of male and female specimens of lots FMNH 55877 and AMNH 22657 suggest that they represent the original paratypic series of Henn.John D. Haseman collected both lots FMNH 55877 and AMNH 22657 in March 4 th ,
Fins hyaline.Fin rays with two rows of brown chromatophores along each side, along extension of ray.Pale brown band on dorsal fin near its base (sometimes not easily seen).Eight to 18 narrow brown bars (median = 12) along body sides of female specimens.Dark brown spot at base of R3.

Fig. 6 .
Fig. 6.Tukey box plot of number of bars along flanks in female specimens of Phalloptychus januarius and P. iheringii.Significant differences were found between P. januarius (Rio de Janeiro, São Paulo and Paraná) and P. iheringii populations (Santa Catarina and Rio Grande do Sul).* Type locality of P. januarius, ** type locality of P. iheringii.

Table 1 .
Descriptive morphometrics of males of Phalloptychus species.Measurements 1-10 are percents of standard length and measurements 11-13 are percents of head length.