Aquapress ISSN

Transcription

1 aqua International Journal of Ichthyology Vol. 18 (4), 15 October 2012 Aquapress ISSN

2 aqua - International Journal of Ichthyology Managing Editor: Heiko Bleher Via G. Falcone 11, Miradolo Terme (PV), Italy Tel. & Fax: heiko@aquapress-bleher.com Scientific Editor: Frank Pezold College of Science & Engineering Texas A&M University Corpus Christi 6300 Ocean Drive Corpus Christi, TX Tel frank.pezold@tamucc.edu Editorial Board: Gerald R. Allen Department of Aquatic Zoology, Western Australian Museum, Perth, Australia Nina G. Bogutskaya Zoological Institute of the Russian Academy of Sciences, St. Petersburg, Russia Friedhelm Krupp Curator of Fishes, Senckenberg Research Institute and Natural History Museum, Frankfurt am Main, Germany Flávio C. T. Lima Museu de Zoologia da Universidade de São Paulo São Paulo, Brasil Axel Meyer Lehrstuhl für Zoologie und Evolutions biologie, Universität Konstanz, Germany Paolo Parenti Department of Enviromental Sciences, University of Milano-Bicocca, Milan, Italy Mário de Pinna Museu de Zoologia da USP, São Paulo, Brazil John E. Randall Bishop Museum, Honolulu, Hawaii, U.S.A. Richard Winterbottom Centre of Biodiversity & Conservation Biology, Royal Ontario Museum, Toronto, Canada Helen K. Larson Curator Emeritus, Fishes - Museum and Art Gallery of the Northern Territory. Darwin, Australia Scope aqua is an international journal which publishes original scientific articles in the fields of systematics, taxonomy, bio geography, ethology, ecology, and general biology of fishes. Papers on freshwater, brackish, and marine fishes will be considered. aqua is fully refereed and aims at publishing manuscripts within 2-4 months of acceptance. In view of the importance of color patterns in species identi - fication and animal ethology, authors are encouraged to submit color illustrations in addition to descriptions of coloration. It is our aim to provide the international scientific community with an efficiently published journal meeting high scientific and technical standards. Call for papers The editors welcome the submission of original manuscripts which should be sent in digital format to the scientific editor. Full length research papers and short notes will be considered for publication. There are no page charges and color illustrations will be published free of charge. Authors will receive one free copy of the issue in which their paper is published and an e-print in PDF format. Subscription Notice At least one volume (4 issues) of aqua is being published per year, each issue comprising pages (incl. cover). The subscription rate (for one volume = 4 issues): Personal subscription: Euro 75,00 (online edition) - Euro 100,00 (online edition + print edition incl. priority mail); Institutional subscription: Euro 800,00 (online edition) - Euro 1000,00 (online edition + print edition incl. priority mail). Subscription enquires should be sent to the publisher at the address given below or by to: aqua@aquapress-bleher.it - aqua@aquapress-bleher.com Special Publication Since 2003 Aquapress publishes a series of Special Publications, which are produced at irregular intervals. All Special Publications have about 100 or more pages and are available separately from regular issues of aqua. Enquiries about subscriptions and prices should be sent to the publisher at the address given here above or by to: aqua@aquapress-bleher.it - aqua@aquapress-bleher.com ISSN Publisher: Aquapress, Redazione aqua, I Miradolo Terme (Pavia), Italy Printer: Pronto Stampa Srl Bergamo Italy Copyediting and layout: Rossella Bulla 2012 aqua, International Journal of Ichthyology

3 aqua, International Journal of Ichthyology Laimosemion paryagi (Cyprinodontiformes: Aplocheiloidei: Rivulidae), a new species from the upper Mazaruni river drainage of Western Guyana F. B. M. Vermeulen¹, W. H. Suijker² & G. E. Collier³ 1) Tanki Leendert 194c, Aruba. vermeulen@setarnet.aw 2) Straatweg 50 Rotterdam, The Netherlands. w.suijker@upcmail.nl 3) Department of Biological Science, The University of Tulsa, S. Tucker Dr. Tulsa, OK glen-collier@utulsa.edu Received: 16 March 2012 Accepted: 27 September 2012 Abstract Laimosemion paryagi, new species, is described from the upper Mazaruni river system, a tributary of the Essequibo River. It is a member of the Laimosemion breviceps group former known as the Rivulus breviceps group, and shares a robust body and deep caudal peduncle with Laimosemion breviceps (Eigenmann, 1909) and Laimosemion lyricauda (Thomerson et al., 1991) and to a lesser degree with L. gransabanae (Lasso et al., 1992) and Laimosemion torrenticola (Vermeulen & Isbrücker, 2000). These species are all endemic to the Guiana Highlands in western Guyana and the neighbouring Gran Sabana in Eastern Venezuela. It is distinguished from other species in the L. breviceps group by morphology and its remarkable male color pattern of red blotches on a turquoise ground color on the flanks and in the unpaired fins and having a rounded caudal fin as opposed to one having extension or being spade shaped. Analysis of mitochondrial DNA sequence reveals that it is genetically distinct from all other members of this group and that inhabitants of the Guyana highlands diverged from each other early in the history of the genus, commensurate with the geological age of the Guiana Shield. Zusammenfassung Laimosemion paryagi nov. spec. wird vom Flusssystem des oberen Mazaruni beschrieben, einem Zuflusssystem des Essequibo. Diese neue Art gehört zur Laimosemion-breviceps-Gruppe, die früher Rivulus-breviceps-Gruppe genannt wurde, und hat einen robusten Körper und einen tiefen Schwanzstiel mit Laimosemion breviceps (Eigenmann, 1909) und Laimosemion lyricauda (Thomerson et al., 1991) gemeinsam, in geringerem Grade auch mit L. gransabanae (Lasso et al., 1992) and Laimosemion torrenticola (Vermeulen & Isbrücker, 2000). All diese Arten sind endemisch im Guyana-Hochland in West-Guyana sowie im benachbarten Gran Sabana in Ost-Venezuela. Die neue Art unterscheidet sich von den anderen Arten der L.-breviceps-Gruppe durch die Körpergestalt und das auffällige Farbmuster der Männchen: rote Flecken auf türkisfarbenem Grund an den Flanken und den unpaaren Flossen sowie eine abgerundete Schwanzflosse, während die anderen Arten mit Fortsatz oder spatenförmigen Schwanzflossen ausgestattet sind. Eine Analyse der mitochondrialen DNA-Sequenz zeigt, dass die neue Art sich genetisch von allen anderen Angehörigen der Gruppe unterscheidet und dass die Bewohner des Guyana- Hochlandes als Divergenz zu allen anderen Arten in der Frühzeit der Gattungsgeschichte entstanden sind, gemessen am geologischen Alter des Guyana-Schildes. Résumé Laimosemion paryagi, nouvelle espèce, est décrit comme originaire du système du haut Mazaruni, un tributaire de l Essequibo River. C est un membre du groupe Laimosemion breviceps, jadis groupe Rivulus breviceps, et partage un corps robuste et un large pédoncule caudal avec Laimosemion breviceps (Eigenmann, 1909) et Laimosemion lyricauda (Thomerson et al., 1991) et, à un moindre degré, avec L. gransabanae (Lasso et al., 1992) et Laimosemion torrenticola (Vermeulen & Isbrücker, 2000). Ces espèces sont toutes endémiques des hauteurs de l ouest de la Guyana et du Gran Sabana voisin, à l est du Venezuela. Il se distingue des autres espèces du groupe L. breviceps par la morphologie et par le remarquable patron de coloration du mâle fait de taches rouges sur fond turquoise sur les flancs et les nageoires impaires et par une caudale arrondie au contraire de celle de forme étendue ou en pelle. L analyse de séquence mitochondriale d ADN révèle que l espèce se distingue génétiquement de tous les autres membres de ce groupe et que les espèces des hauteurs de Guyana diffèrent l une de l autre tôt dans l histoire du genre, proportionnellement à l âge géologique du Bouclier guyanais. 181 Sommario Laimosemion paryagi, nuova specie, è descritto dal sistema superiore del fiume Mazaruni, un affluente del fiume Essequibo. Si tratta di un membro del gruppo Laimosemion breviceps un tempo noto come gruppo Rivulus breviceps, e condivide con L. breviceps (Eigenmann, 1909), L. lyricau - da (Thomerson et al., 1991) e, in misura minore, con L. gran sabanae (Lasso et al, 1992) e L. torrenticola (Veraqua vol. 18 no October 2012

4 Laimosemion paryagi (Cyprinodontiformes: Aplocheiloidei: Rivulidae), a new species from the upper Mazaruni river drainage of Western Guyana meulen & Isbrücker, 2000), un corpo robusto e un altrettanto ro bu sto peduncolo caudale. Queste specie sono tutte en de miche del le Guiana Highlands nella parte occidentale della Guyana e della vicina Gran Sabana nel Venezuela orientale. Si distingue dalle altre specie del gruppo L. breviceps per la notevole colorazione del maschio, fatta di chiazze rosse su un fondo di colore turchese sui fianchi e sulle pinne impa ri, e per la morfologia tra cui una pinna caudale arrotonda ta rispetto ad una avente estensione o a forma di vanga. L'analisi della sequenza di DNA mitocondriale rivela che è ge neticamente distinta da tutti gli altri membri di questo grup po e che le specie che abitano gli altopiani della Guya na si allontanarono uno dall'altro piuttosto presto nella sto ria del genere, commisurato con l'età geologica dello scu do della Guiana. INTRODUCTION The recently erected genus Laimosemion (Cyprinodontiformes: Aplocheiloidei: Rivulidae) currently includes twenty four species (Costa, 2011). It represents a molecularly defined clade (Murphy et al., 1999; Hrbek and Larson, 1999; Hrbek et al., 2004; Suijker and Collier, 2006) within a large, diverse assemblage of species (formerly assigned to the genus Rivulus) that are distributed throughout Northwestern South America; central and eastern Brazilian Amazon; river basins of Guianas; upper Orinoco River basin in Venezuela and Colombia; Rio Negro basin in Brazil; and lower Peruvian Amazon. They are found in coastal lowland areas as well as mountainous terrain occasionally up to altitudes of 1300 meters above sea level. Within these areas they are found in small streams, creeks, swampy or wet places adjacent to creeks and rain pools. Past authors have attempted to arrange related forms of this larger group into a number of complexes or species groups or subgenera (Hoedeman, 1958 and 1961; Huber, 1992 and 1999; Costa, 2006). Hoedeman (1958) coined the term breviceps complex for small forms, coarsely scaled which included breviceps, frenatus, agilae and geayi. Huber (1999) grouped these species into the subgenus Laimosemion based on small size, distinctive male patterns, a red opercular shield and that two had been reported to have bifid epipleural ribs, a unique character among the Rivulidae. Costa (2006) refined the morphological and osteological definition of this subgenus and erected a new subgenus Owiyeye for a distinctive set of related species from west of the Guiana Shield. Subsequently, Costa (2011) elevated the subgenus Laimosemion to genus and redefined it to also include those species he formerly included in Owiyeye primarily on the basis of previously published molecular phylogenies (Hrbek and Larson, 1999; Murphy et al. 1999: Hrbek et al., 2004; and Suijker and Collier, 2006). The first and second authors have visited Guyana several times since These series of expeditions were the first serious attempts to relocate all species of the former genus Rivulus described by Eigenmann (1909 and 1912) based upon material collected during his expedition of During these and other expeditions to the neighboring areas of Venezuela, five additional species have been described that can be included in this group of species (Thomerson at al., 1991; Lasso et al., 1992; Vermeulen and Isbrücker, 2000; Lasso- Alcala et al., 2006: Suijker and Collier, 2006). The new species presented in this paper was first collected during a survey of the Guyana plateau by W. H. Suijker and Y. Suijker, accompanied by S. C. Paryag on April 4, MATERIAL AND METHODS Specimens analysed: Wild-caught specimens were used as holotype and paratype while first generation aquarium bred descendents of wild-caught fish were used for other paratypes, preparations of karyotypes and DNA extractions. Specimens of Laimosemion gransabanae from San Rafael de Kamoirán, Venezuela, Laimosemion breviceps from Kaieteur Falls, Guyana and Laimosemion torrenticola from Kamarang, Guyana, were also used for DNA extractions. Molecular analyses: DNA was extracted from ethanol preserved specimens using a DNAeasy kit (Qiagen). Amplification of portions of the mitochondrial genes for 12S rrna, 16S rrna, and cytochrome B was done as previously described (Murphy and Collier, 1996, and GenBank accession numbers for the taxa used are: L. geayi (AF002433, AF002537, AF002483), L. strigatus (AF002434, AF002538, AF002484), L. agilae (AF002432, AF002536, AF002482), L. frenatus (AF002435, AF002539, AF002485), L. xiphidius (AF002436, AF002540, AF002486), L. lyricauda (AF002439, AF002543, AF2489), L. rectocaudatus(af002440, AF002544, AF002490), L. sp. Tobogan (AF002437, AF002541, AF002487), L. sp. DCT (AF002438, AF002542, AF002488), L. mahdianensis (DQ501248, DQ501249, DQ501250), R. cylindraceus (U41799, AF002533, U41781) and R. roloffi aqua vol. 18 no October

5 F. B. M. Vermeulen, W. H. Suijker and G. E. Collier (U41798, AF002434, U41780). The GenBank accession numbers for L. breviceps and L. paryagi are JX through JX Sequences were aligned by Clustal W and analyzed using MEGA 4.1 (Tamura et al., 2007). Karyotype: Metaphase preparations of chromosomes were prepared from gill epithelium of a single wild-caught male as described by Kligerman and Bloom (1977). Morphometrics: Measurements and counts follow Huber (1992). All visible and minute rays from the anal and dorsal fins were counted. Measurements are taken with a Mitutoyo Dial Calliper to the nearest tenth of a millimetre and are presented in percentages of the standard length. Clearing and staining of a specimen was performed as described by Taylor and Van Dyke (1985). Laimosemion paryagi, n. sp. Figs 1-2, Table I Holotype: ZMA (Zoölogisch Museum Amsterdam) , male, 39,2 mm SL: Guyana, Upper Mazaruni District, upper Mazaruni river about 7 Fig. 1. Laimosemion paryagi n. sp. (not preserved). Male from the terra typica. Photo by F. Vermeulen. Fig. 2. Laimosemion paryagi n. sp. (not preserved). Female from the terra typica. Photo by F. Vermeulen. 183 aqua vol. 18 no October 2012

6 Laimosemion paryagi (Cyprinodontiformes: Aplocheiloidei: Rivulidae), a new species from the upper Mazaruni river drainage of Western Guyana miles downstream from Kamarang, a little creek about 15 minutes walk on a bush trail in eastern direction starting at the right bank near a rapid locally called Sand Landing (Figures 3 and 4). Coordinates: N W, altitude approx. 495 meters above sea level. April 4, Paratypes: ZMA , one female, 36,9 mm SL: Collected same locality and date as holotype and 6 males F1 generation, original from same location. Diagnosis: Laimosemion paryagi is a member of the L. breviceps species group as indicated by its relative small size, short snout length, bright coloration and relatively low number of scales (LL+C). It is distinguished from lowland close relatives (L. agilae, L. xiphidius, L. cladophorus, L. geayi and L. mahdiaensis) by having a deeper caudal peduncle ( % of SL, vs % Fig. 3. The habitat with the tea colored water lighting up in the sun. L. paryagi was not seen in the creek itself but in swampy places beside the creek. Photo by S. Sladkowski. of SL). It is distinguished from the remaining species in the group by having a rounded caudal fin (versus a spade-shaped caudal fin in L. breviceps, L. gransabanae and L. torrenticola, and a lyre-shaped caudal fin in L. lyricauda). Its remarkable coloration of red blotches on a turquoise background on the flanks and in all unpaired fins in males (versus absence of turquoise color and red blotches) is unique within the group. Description: Morphometric data are presented in Table I. Largest specimen 39.2 mm SL, male. The dorsal outline of the body has a weak convex profile, straight from anterior of the dorsal fin along the caudal peduncle. Ventral profile convex, to the caudal peduncle more straight. Body cylindrical. Greatest body depth at level of pelvic fin insertion. Dorsal fin short triangular shaped, no filaments in either males or females, somewhat larger in males. Dorsal fin origin directly above 6 th or 7 th anal finray. Anal fin rectangular in males, more rounded or slightly rectangular in females, no fin rays extended as filaments. Caudal fin rounded in both sexes. Pelvic fins rounded, in males reaching the second ray of anal fin, in females smaller not reaching anal fin. Pectoral fins in males somewhat larger than in females but both not reaching base of pelvic fin. Dorsal fin-rays 9; anal fin-rays 11-12; pelvic finrays 6. Scales large and cycloid, head scaled except the ventral part of cheeks. Longitudinal series of scales upon the end of lateral plate, some scales (3-5) on the caudal fin base, no scales on base of other fins. Frontal scales are circularly arranged around the central A-scale. The squamation pattern is of the E type in most samples and of the E-D type in 2 of the 8 samples. Scales along the mid- lateral line have single minute sensory organs. Epipleural ribs are not bifid (Fig. 8). Coloration in life: Male: Dorsum brown to dark brown, laterally the body is light brown with aubergine red spots arranged to form weak and broken lines, more intense on the anterior portion. These clear, visible red spots are shown only on the upper half of the sides and are only weakly present on the lower half. The lower half of the sides is iridescent turquoise from the lower edge of the eye, over the gill-cover, to the end of the caudal peduncle. Venter and cheeks are whitish to grey. Iris of the eye is golden. No opercular or humeral markings are present. The side of the head is orange to yellow. The dorsal fin is turquoise with large aubergine- red blotches that tend to form aqua vol. 18 no October

7 F. B. M. Vermeulen, W. H. Suijker and G. E. Collier Table I. Morphometric data of the holotype and paratypes of Laimosemion paryagi. Standard length is given in mm, other morphometric data are presented as percentages of standard length. The abbreviation dev = deviation. Holotype Paratypes 8 samples male female male male male male male male lowest highest average dev Standard length in mm In % of SL Total length Pre-dorsal length Pre-anal length Body depth at anal level Head length Snout length Interorbital space Eye diameter Depth of caudal peduncle Snout to dorsal Snout to anal Meristic data Number of dorsal fin rays Number of anal fin rays Dorsal insertion to anal insertion Longest dorsal fin ray Longest anal fin ray Scalation Number scales LL + C Number scales to D insertion Head scalation type E E E E E damaged D-E E-D lines along the rays, while the distal portion of the dorsal fin fades to more red than turquoise to form a marginal aubergine- red zone. Anal fin is bluish to turquoise near the base with some weak red blotches. The distal two thirds of the anal fin becomes more red than turquoise to form a marginal aubergine- red zone. Caudal fin basis is turquoise, with the center more iridescent and the edges of which fade to form an aubergine-red margin around the outer edge. The central part of the caudal has rows of numerous red dots that form 3-4 vertical bars. The pelvic fin is almost completely aubergine-red with a weak bluish at the base. Pectoral fin is hyaline orange. Female: Dorsum light brown, pale brown on the sides and yellow brown to the vent. The anterior part of the flanks show hyaline red markings. Dorsal fin is hyaline orange with weak darker dots forming horizontal lines. Anal fin is hyaline orange with more orange toward the distal end. Caudal fin is hyaline orange without any markings or marginal band. Pelvic fins hyaline. Pectoral fins hyaline orange. Eye with golden iris. No Rivulus spot or supracaudal ocellus is present in either sex or at any stage of life. There is a dark (black) stripe that appears below the midline from behind the pectoral fin to nearly the base of the caudal. This is seen in both sexes when they are stressed. 185 aqua vol. 18 no October 2012

8 Laimosemion paryagi (Cyprinodontiformes: Aplocheiloidei: Rivulidae), a new species from the upper Mazaruni river drainage of Western Guyana Behaviour: In general, species that were formerly placed in the genus Rivulus (now divided into Anablepsoides, Atlantirivulus, Cynodonichthys, Kryptolebias, Laimosemion, Melanorivulus and Rivulus by the classification as proposed by Costa, 2011), are non-schooling, solitary fish that are found in very small streams and swamps. In our collecting experience, they are found in the dense vegetation of swamps, under rocks or between logs and leaf litter in cascading mountain creeks. Often they can be found in the wet leaf litter adjacent to small streams. As a result they are often missed in general Fig. 4. Main river systems of Guyana. aqua vol. 18 no October

9 F. B. M. Vermeulen, W. H. Suijker and G. E. Collier ichthyological surveys of larger fishes or larger bodies of water. They are most often discovered by coincidence during expeditions with other objectives. Collectors of Laimosemion almost never see the species that they are looking for until they see the fish in their nets after blind collecting. Members of the L. breviceps species group display behaviour that is an exception to the general pattern described above for the larger group of related genera. Although they do not school, they are clearly seen not hiding and in open water. This behaviour, witnessed in the field by authors in many occasions, is also shared by Anablepsoides waimacui and Anablepsoides amphoreus. All members of the Laimosemion breviceps group, as well as the two species mentioned above, do not share their habitat with any other fish species. The lack of fish predators may be responsible for this remarkable behaviour. Molecular phylogeny: Sequences of portions of the mitochondrial genes for 12S and 16S RNA and cytochrome b for L. paryagi were added to those previously determined (Murphy et al., 1999; Suijker and Collier, 2006). The resulting phylogeny places the new species as a member of this group with the closest relationship being to L. breviceps. Partial DNA sequences for L. gransabanae and L. torrenticola place them as close relatives of L. breviceps as well. Karyotype: Metaphase spreads (Figure 7) reveal 2n=26. The haploid complement consists of one large metacentric and one medium size metacentric element and a graded series of smaller metacentric and submetacentric elements. This is the lowest number recorded to date for a species of the genus Laimosemion. Low diploid numbers also characterize other members of the group shown in Figure 6 (Suijker and Collier, 2006). Distribution: Known only from the type locality, a small black- water creek, about 15 minutes walk on a hunting trail starting at the right bank of the upper Mazaruni river about 10 km downstream from Kamerang near rapids locally called Sand Landing. Etymology: The name paryagi is chosen to honour Mr. Subhas Chand Paryag, from Georgetown, Guyana, co-collector of the new species and local helper during most of the expeditions in Guyana made by first and second author. Discussion: Laimosemion paryagi is a member of the L. breviceps species group as indicated by its relative small size, short snout length and bright Fig. 5. Upper Mazaruni River at Kamarang. 187 aqua vol. 18 no October 2012

10 Laimosemion paryagi (Cyprinodontiformes: Aplocheiloidei: Rivulidae), a new species from the upper Mazaruni river drainage of Western Guyana L. gaeyi L. strigatus L. agilae L. frenatus L. xiphidius L. lyricauda L. paryagi L. breviceps L. mahdiaensis L. rectocaudatus L. sp. Tobogan L. sp. DCT R. cylindraceus R. roloffi Fig. 6. Molecular phylogeny for L. paryagi and selected members of the genus Laimosemion. Neighbor-joining tree based upon 1048 base pairs of portions of the cytochrome b, 12S, and 16S mitochondrial genes. The values below the nodes are bootstrap values. Rivulus cylindraceus and Rivulus roloffi were used as outgroups to root the tree. Previous subgeneric groupings (Costa, 2006) and current generic assignment (Costa, 2011) are indicated by brackets. coloration. It is distinguished from other members of the group principally by the shape of caudal fin and a unique color pattern. This placement is also supported by the molecular and cytogenetic data. The dark line displayed below the midline when stressed is a character that is shared with L. breviceps, L. agilae and L. lyricauda. This unique, shared color pattern may be another character that unites the group. This group has also been assigned to the subgenus Laimosemion (Huber, 1999; Costa, 2006). However, the primary osteological diagnostic characteristic of this subgenus was given as bifid epipleural ribs (Costa, 2006). The subgenus Owiyeye was erected (Costa, 2006) to include species from west of the Guiana Shield. In Figure 6 this subgenus is represented by L. rectocaudatus, L. sp. Tobogán and L. sp. DCT (Maroa). This subgenus is characterized by, among other characters, frontal scales transversely arranged and squamation S-patterned. None of these characters are found in L. pa - ryagi. Thus, L. paryagi cannot be assigned to either of these subgenera as defined by Costa (2006). Fig. 7. Metaphase chromosomes from gill epithelia of Laimosemion paryagi. aqua vol. 18 no October

11 F. B. M. Vermeulen, W. H. Suijker and G. E. Collier Costa (2011) conceded that the morphological characters had not been fully examined in all species of the subgenus Laimosemion and recognized the monophyly of the set of species included in the subgenera Laimosemion and Owiyeye revealed by previous published molecular phylogenies. Accordingly Laimosemion was raised to generic rank to include all species of the two former subgenera. However, the species of Owiyeye differ from the other species in the new genus in their size, behaviour and, most importantly, geographical distribution. Unfortunately, most species formerly assigned to Owiyeye have not been made available for molecular analysis. Until there can be a comprehensive analysis of both morphological and molecular characters, the taxonomic status of these species will remain ambiguous. For these reasons, we prefer to informally use the term breviceps group to refer to those small species found along the eastern edges of the Guiana Shield and adjacent lowlands where they are found in creeks and wetland areas with stony or sandy bottoms. The Guiana Shield is an ancient geological formation that forms one of the three cratons of the South American plate. The Guiana highlands are the higher elevations of this formation found in west central Guyana. Some of the most spectacular waterfalls in world flow off these highlands to the surrounding lowlands. These highlands were presumably never inundated by epicontinental seas. Hence species of freshwater fishes endemic to this region likely diversified over a longer period of time than species from the surrounding lowlands. The molecular phylogeny is consistent with this notion in that all members from the interior and highland areas of the Guiana Shield are deeply diverged from one another. Even though these species are each other s closest relatives, they represent a diversity of morphology and coloration unparalleled in other parts of the former genus Rivulus. Whether this is simply a function of the age of the Guiana Shield or other factors remain to be seen. ACKNOWLEDGEMENTS Authors are grateful to the Ministry of Fisheries and Wildlife of the Co-operative Republic of Guyana providing permits, to Subhas Chand Paryag and his wife Jenny for their hospitality and help during most expeditions of authors and to Mr. Vishnu Mesir, at that time head of the Upper Mazaruni district, for his help and hospitality in Kamarang. They also thank Mr. Siegmund Sladkowski, Germany for the use of his image from the Terra Typica and his assistance in the field. Fig. 8. Fifth epipleural rib dissected from the right side of a cleared and stained specimen of L. paryagi. 189 aqua vol. 18 no October 2012

12 Laimosemion paryagi (Cyprinodontiformes: Aplocheiloidei: Rivulidae), a new species from the upper Mazaruni river drainage of Western Guyana REFERENCES COSTA, W. J. E. M Relationships and taxonomy of the killifish genus Rivulus (Cyprinodontiformes: Aplocheiloidei: Rivulidae) from the Brazilian Amazonas river basin, with notes on historical ecology. aqua, Journal of Ichthyology and Aquatic Biology 11 (4): COSTA, W. J. E. M Phylogenetic position and taxonomic status of Anablepsoides, Atlantirivulus, Cynodonichthys, Laimosemion and Melanorivulus (Cyprinodontiformes: Rivulidae). Ichthyol. Explor. Freshwaters, Vol. 22, No. 3, pp , 1 fig., 1 tab., EIGENMANN, C. H Some new genera and species of fishes from British Guiana. Annals of the Carnegie Museum, vol. VI (1): EIGENMANN, C. H The freshwater fishes of British Guiana, including a study of the ecological grouping of the species, and the relation of the fauna of the plateau to that of the lowlands. Mem. Carnegie Mus., V, pp , 103 pls. HOEDEMAN, J. J Studies on Cyprinodontiform Fishes [4]. The frontal scalation pattern in some groups of tooth carps (Pisces, Cyprinodontiformes). Bulletin of Aquatic Biology, 1 (3): HOEDEMAN, J. J Studies on Cyprinodontiform Fishes [8]: Preliminary key to the species and subspecies of the genus Rivulus. Bulletin of Aquatic Biology 2 (18): HRBEK, T. & LARSON, A The evolution of diapause in the killifish family Rivulidae (Antherinomorpha, Cyprinodontiformes): A molecular phylogenetic and biogeographic perspective. Evolution 53, HRBEK, T., PEREIRA DE DEUS, C. & PIRES FARIAS, I Rivulus duckensis (Teleostei; Cyprinodontiformes): new species from the Tarumã Basin of Manaus, Amazonas, Brazil, and its relationship to other neotropical Rivulidae. Copeia 2004 (3), HUBER, J. H Review of Rivulus, Ecobiogeography Relationships Cybium, Societe francais d Ichthyologie, 572 pp. HUBER, J. H Updates to the phylogeny and systematics of the Neotropical cyprinodont genus Rivulus and its allied (Cyprinodontiformes: Rivulinae). Cybium, 23: KLIGERMAN, A. & BLOOM S Rapid chromosome preparations from solid tissues of fishes. Journal of the Fisheries Research Board of Canada, 34, LASSO, C., TAPHORN, D. & THOMERSON, J Rivulus gransabanae, a new species of killifishes from Venezuela (Cyprinodontiformes: Rivulidae). Ichthyological Exploration of Freshwaters 2 (4): LASSO-ALCALA, O. M., TAPHORN, D. C., LASSO, C. A. & LEON-MATA, O Rivulus sape, a new species of killifish (Cyprinodontiformes: Rivulidae) from the Paragua River system, Caroní River drainage, Guyana Shield, Venezuela. Zootaxa 1275: MURPHY, W. J. & COLLIER, G.E Phylogenetic relationships within the aplocheiloid fish genus Rivulus (Cyprinodontiformes, Rivulidae): Implications for Caribbean and Central American biogeography. Molecular Biology and Evolution 13: MURPHY, W. J. & COLLIER, G. E A molecular phylogeny for aplocheiloid fishes (Atherinomorpha, Cyprinodontiformes): The role of vicariance and the origins of annualism. Molecular Biology and Evolution 14: MURPHY, W. J., THOMERSON, J. E. & COLLIER, G. E Phylogeny of the neotropical killifish family Rivulidae (Cyprinodontiformes, Aplocheiloidei) inferred from mitochondrial DNA sequences. Molecular Phylogenetics and Evolution 13: SUIJKER, W. H. & COLLIER, G. E Rivulus mahdiaensis, a new killifish from central Guyana, (Cyprinodontiformes: Rivulidae). Zootaxa 1246: TAMURA K, DUDLEY J, NEI M. & KUMAR, S MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Molecular Biology and Evolution 24: TAYLOR, W. R. & VAN DYKE, G. C Revised procedures for staining and clearing small fishes and other vertebrates for bone and cartilage study. Cybium 9: THOMERSON J. E., BERKENKAMP, H. O. & TAPHORN, D Rivulus lyricauda, a new species from the Guyana shield in Eastern Venezuela (Cyprinodontiformes, Rivulidae) Ichthyol. Explor. Freshwaters, 1 (4): , 4 fig., 1 tab. VERMEULEN F. B. M. & ISBRÜCKER, I. J. H Rivulus torrenticola n. sp. (Actinopterygii: Cyprinodontiformes: Rivulidae), a new killifish from highlands in the Guyana Shield. Beaufortia, Bul. Zool. Mus. University of Amsterdam, Vol. 50, No. 10 VERMEULEN, F. B. M. & HRBEK, T Kryptolebias sepia n. sp. (Actinopterygii: Cyprinodontiformes: Rivulidae), a new killifish from the Tapanahoni River drainage in southeast Surinam. Zootaxa, 928: aqua vol. 18 no October

13 aqua, International Journal of Ichthyology Nothobranchius kardashevi and Nothobranchius ivanovae (Cyprinodontiformes: Nothobranchiidae): two new annual killifishes from the Katuma River drainage, western Tanzania Stefano Valdesalici Via Cà Bertacchi 5, Viano (RE), Italy. or Received: 28 March 2012 Accepted: 29 June 2012 Abstract Two new annual killifish species, Nothobranchius kardashevi, new species, and N. ivanovae, new species, are described based on specimens collected in ephemeral pools in the Katuma River drainage system, western Tanzania. Nothobranchius kardashevi, new species, belongs to the N. ugandensis species group and N. ivanovae to the N. taeniopygus species group. Both are distinguished from the respective other group members by a diagnostic combination of male colouration and morphological characters. Zusammenfassung Die beiden neuen Killifisch-Arten, Nothobranchius kardashevi nov. spec. und N. ivanovae nov. spec., werden auf der Grundlage von Exemplaren beschrieben, die in saisonalen Tümpeln im Einzugsgebiet des Katuma-Flusses in West-Tansania gesammelt wurden. Nothobranchius kardashevi nov. spec. gehört zur N.-ugandensis-Artengruppe, N. ivanovae hingegen zur N.-taeniopygus-Artengruppe. Sie unterscheiden sich jeweils von den anderen Gruppenangehörigen durch eine Kombination aus Farbmerkmalen bei den Männchen und morphologischen Kennzeichen. Résumé Deux nouvelles espèces de killys annuels, Notobranchius kardashevi, nouvelle espèce, et N. ivanovae, nouvelle espèce, sont décrites sur base de spécimens collectés dans des mares éphémères dans le système de drainage de la Katuma River, à l ouest de la Tanzanie. Notobranchius kardashevi, nouvelle espèce, appartient au groupe N. ugandensis et N. ivanovae, au groupe N. taeniopygus. Tous deux se distinguent des autres membres respectifs des groupes par une combinaison diagnostique de la coloration du mâle et des caractéristiques morphologiques. Sommario Due nuove specie di killifish annuali, Nothobranchius kardashevi, nuova specie, e N. ivanovae, nuova specie, sono descritte sulla base di esemplari raccolti in pozze effimere nel sistema di drenaggio del fiume Katuma, Tanzania occidentale. Nothobranchius kardashevi appartiene al gruppo di specie N. ugandensis, mentre N. ivanovae al gruppo di specie N. taeniopygus. Entrambi sono distinti dai rispettivi altri membri del gruppo da una combinazione diagnostica comprendente la colorazione del maschio e caratteristiche morfologiche. INTRODUCTION The killifish genus Nothobranchius occurs in the subtropical and tropical parts of eastern Africa, from Sudan to South Africa, and from Chad to Zanzibar and Mafia islands in Tanzania. All known species are annual fishes, living in temporary pools and swamps formed during the rainy season (Wildekamp 2004). Six Nothobranchius species are currently known from central Tanzania: N. neumanni from the Great Ruaha, Bubu, Wembere, and Malagarasi River basins, and Lakes Manyara and Victoria drainage systems; N. robustus from the Lake Victoria drainage system; N. seegersi from Malagarasi River basin; N. taeniopygus from the Wembere, Malagarasi, and Bubu River basins and the Lakes Victoria and Rukwa drainages systems; and at least two undescribed species: N. sp. aff. neumanni Mbeya Type, and N. sp. Lake Victoria (De Vos et al. 2001; Seegers 1997; Shidlovsky 2010; Valdesalici & Kardashev 2011; Wildekamp 1990, 2004). In May 2011, Kiril Kardashev and Iva Ivanova collected specimens of two Nothobranchius species from ephemeral pools in the Katuma River drainage system, western Tanzania, during a Nothobranchius species survey. As a result of the present study, these specimens are herein described as N. kardashevi, new species, and N. ivanovae, new species. 191 aqua vol. 18 no October 2012

14 Nothobranchius kardashevi and Nothobranchius ivanovae, two new annual killifishes from the Katuma River drainage, western Tanzania MATERIAL AND METHODS Measurements and counts were taken as described in Amiet (1987), Huber (2000), and Valdesalici (2010). Measurements were made with a digital calliper, partly under a dissecting microscope, and rounded to the nearest 0.1 mm. If not stated otherwise, measurements are presented as percentages of standard length (SL), except for eye diameter and snout length, which are given as percentage of head length (HL). Terminology for the cephalic neuromast series follows Scheel (1968) and Huber (2000) and for the frontal squamation Hoedeman (1958). Osteological preparations (clearing and staining, below: C&S) were made according to Taylor and Van Dyke (1985), but not stained for cartilages. Morphological data from Seegers (1996), Valdesalici et al. (2009) and Wildekamp (1990, 1994) were used for comparisons. Type material is deposited in the following institutions: Royal Museum for Central Africa (MRAC), Tervuren, Belgium and Museo Civico di Storia Naturale Giacomo Doria (MSNG), Genova, Italy. Nothobranchius kardashevi, n. sp. (Figs 1-2, Table I) Fig. 1. Nothobranchius kardashevi, adult male, about 35 mm SL, not preserved, Karira stream, western Tanzania. Photo by I. Ivanova. Fig. 2. Nothobranchius kardashevi, adult female, about 30 mm SL, not preserved, Karira stream, western Tanzania. Photo by I. Ivanova. Holotype: MRAC B2-09-P-1, 1 male, 37.7 mm SL, pond on left side of the road from Kabungo to Iloba, Karira stream, Katuma river drainage, altitude 1023 m, Tanzania, S, E, 21 May 2011, Kiril Kardashev and Iva Ivanova. Paratypes: MRAC B2-09-P-2-5, 1 male, 36.1 mm SL C&S, 3 females, mm SL; MSNG 56889, 1 male, 31.2 mm SL; same data as holotype. Diagnosis: Nothobranchius kardashevi males similar to other members of N. ugandensis species group, differing from all other species of the genus by comaqua vol. 18 no October

15 Stefano Valdesalici Table I. Morphometric data for Nothobranchius kardashevi. Measurements were given as percentages of standard length in mm. Eye diameter and snout length were given as percentages of head length. Cleared and stained material is included. Holotype Males (n = 3) Females (n = 3) Standard length (mm) Body depth at pelvic fin Predorsal length Preanal length Prepelvic length Caudal peduncle length Caudal peduncle depth Dorsal fin base length Anal fin base length Head length Snout length Eye diameter bination of male colouration characters consisting of light blue scales with a broad irregularly reticulated pattern on body forming oblique bars in bright red, vivid bright red colouration on the head and dorsum, yellow or blue anal fin, large vivid red spot pattern on dorsal and anal fins; rounded head, dorsal profile of head slightly concave to nearly straight, convex from nape to end of dorsal fin base; variable cephalic squamation. Differs from other members of the N. ugandensis species group by caudal fin spotted (vs. caudal fin plain). Distinguished from N. nubaensis by reduced red pigmentation, having irregular broad margin on scales (vs. regular reticulation), by hyaline pectoral fins (vs. red), and having blue anal fin and pelvic fins (vs. yellow). Distinguished from N. ugandensis by absence of dorsal, anal and caudal fin margin (vs. present, sometimes one or more margins absent), by broad irregular blotch pattern on dorsal fin in two colours (vs. only one colour and more regular pattern often reduced as a series of dots), by anal fin pattern with large irregular spots in two colours (vs. without any spots or if present in only one colour and reduced to a few dots), by broad spots on caudal fin (vs. if present, faint dots on yellow morph). In addition, N. kardashevi males have a relatively longer head compared to N. nubaensis ( vs % SL), relatively higher number of scales in longitudinal series (28-32 vs ), fewer anal fin rays (16-17 vs ), more slender caudal peduncle compared to N. ugandensis ( vs. 12,5-16,2% SL), reduced number of circumpeduncular scales (12 vs. 14 in N. nubaensis and in N. ugandensis, and a different posterior supraorbital neuromast series (with two neuromasts vs. three in N. ugandensis and N. nubaensis). Description: See Figs 1-2 for overall appearance and Table I for morphometric data for the type series. Robust Nothobranchius with rounded body, maximum length recorded in males 37.7 mm SL. Dorsal profile slightly concave to nearly straight on head, convex from nape to end of dorsal fin base, ventral profile convex, slightly concave on caudal peduncle posterior to dorsal and anal fin. Snout slightly pointed, mouth directed upwards, lower jaw longer than upper jaw, posterior end of rictus at same level as or slightly above centre of eye. Branchiostegal membrane projecting posteriorly from opercle. Dorsal and anal fins located posterior to mid-body, tips rounded with short filamentous rays. Both fins with papillate contact organs on fin rays. Dorsal fin tip reaching caudal fin. Number of dorsal fin rays 17-18, anal fin rays Pectoral fin (18-20 rays) approximately triangular; in some specimens pectoral fins reach pelvic fins, and pelvic fins (6 rays) reach the anus. Caudal fin rounded (24-28 rays). Scales cycloid, body and head entirely scaled, except for ventral surface of head. Scales in median lateral series on caudal fin base. Transverse row of scales 12. Circumpeduncular scales 12. Cephalic squamation pattern variable. Frontal neuromast series of the open type. Anterior supraorbital series with one or two neuromast. Posterior 193 aqua vol. 18 no October 2012

16 Nothobranchius kardashevi and Nothobranchius ivanovae, two new annual killifishes from the Katuma River drainage, western Tanzania supraorbital series in shallow groove, with two neuromasts. One neuromast on each scale of median longitudinal series. Six branchiostegal rays. Vomerine teeth present in a large patch. Lateral process of post-temporal present. Second pharyngobranchial with one tooth. Single short ante ro dor sal process of urohyal. Number of vertebrae 30. Premaxilla and dentary with many irregularly distributed unicuspid, slightly curved teeth of different size, a small number of larger ones on the outer row of upper and lower jaw. Females smaller than males, maximum observed size 33.6 mm SL. In females, dorsal fin rounded, anal fin triangular with rounded tip, branchiostegal membrane not projecting from opercle. Colour in life: Males (Fig. 1). Body and Fig. 3. Geographic distribution of Nothobranchius kardashevi (open black rhomboid), N. ivanovae (black rhomboids, open black rhomboid: type locality), N. ugandensis (violet rhomboids), N. taeniopygus Ipati River (blue rhomboid), N. taeniopygus Wogo River (red rhomboid). Map prepared by Béla Nagy. aqua vol. 18 no October

17 Stefano Valdesalici head scales light blue, dark red margin of some scales on dorsal part of body and caudal peduncle broader. Snout, frontal, and dorsal portion of head red. Operculum with 3 oblique stripes extending from eye to dorsal part of the head; branchiostegal membrane whitish. Dorsal fin light blue with irregular rows of dark red spots and dots; proximally somewhat fused forming an irregular blotch, medially elongated on fin rays, distally denser and smaller. Anal fin light blue with few dark red proximal spots and dots. Irregular violet spots are present over entire dorsal fin and anal fin. Pelvic fins light blue, red dots near the base. Pectoral fins hyaline with light blue margins. Caudal fin light blue with dark red spots and dots, proximally fused forming an irregular blotch more dotted distally. Iris golden. Female (Fig. 2). Body and head scales pale brown, with silver iridescence on scale centre. Opercular region silver. Abdomen silver. All fins hyaline. Iris golden. Etymology: The species name is dedicated to my friend, aquarist, collector, and discoverer of the species, Kiril Kardashev from Dupnitsa, Bulgaria. Distribution and habitat: (Figs 3-4) Nothobranchius kardashevi is currently only known from a residual pool in the temporary Karira stream, Fig. 4. Type locality of Nothobranchius kardashevi and N. ivanovae. Residual pool in Karira stream, western Tanzania. Photo by K. Kardashev. Katuma River drainage, western Tanzania. The other fish present were N. ivanovae, new species, and unidentified Barbus species. Nothobranchius ivanovae, n. sp. (Figs 5-6, Table II) 195 Holotype: MRAC B2-09-P-6, 1 male, 44.7 mm SL, pond on left side of the road from Kabungo to Iloba, Karira stream, Katuma river drainage, altitude 1023 m, Tanzania, S, E, 21 May 2011, Kiril Kardashev and Iva Ivanova. Paratypes: MRAC B2-09-P-7-12, 4 males mm SL, 2 females mm SL; MRAC B2-09-P-13-15, 3 males mm SL C&S; same data as holotype; MSNG 56890, 1 male 32.1 mm SL; 2 females mm SL; series of ponds on right side of the road from Kabungo to Iloba, Katuma river drainage, altitude 1010 m, Tanzania, S, E, 22 May 2011, Kiril Kardashev and Iva Ivanova. Diagnosis: Nothobranchius ivanovae males similar to other members of N. taeniopygus species group, differing from all other species of the genus by caudal and anal fin with a broad light sub-marginal band and dark margin. Differs from the similar and geographically closest N. taeniopygus from Ipati River ( phenotype 4 sensu Wildekamp 2004) by the following combination of characters: caudal fin red (vs. caudal fin pale brownish red) with subdistal band (vs. without band), anal fin with a creamy yellow broad submarginal band (vs. thin white band), and from N. taeniopygus from Wogo river ( phenotype 2 sensu Wildekamp 2004) by the following combination of characters: blue-grey scales (vs. light blue), caudal fin red (vs. pale brownish grey) with a subdistal and distal band (vs. without any band). In addition, N. ivanovae males have a relatively shorter head compared to N. taeniopygus Ipati river ( vs % SL), smaller eye ( vs % HL) more scales on lateral line (31 vs ), more dorsal and anal fin rays (17-18 vs and vs , respectively), shorter head compared to N. taeniopygus Wogo river ( vs % SL), more scales on lateral line (31 vs. 30), and more dorsal fin rays (17-18 vs. 15). Description: See Figs 3-4 for overall appearance and Table II for morphometric data for the type series. Deep bodied Nothobranchius, laterally comaqua vol. 18 no October 2012

18 Nothobranchius kardashevi and Nothobranchius ivanovae, two new annual killifishes from the Katuma River drainage, western Tanzania pressed, maximum length recorded in males 44.7 mm SL. Dorsal profile slightly concave to strongly concave on head, convex from nape to end of dorsal fin base, ventral profile convex, slightly concave on caudal peduncle posterior to dorsal and anal fin. Snout slightly pointed, mouth directed upwards, lower jaw longer than upper jaw, posterior end of rictus at same level as or slightly above centre of eye. Branchiostegal membrane projecting posteriorly from opercle. Dorsal and anal fins located posterior to mid-body, tips rounded with short filamentous rays. Both fins with papillate contact organs on fin rays. Dorsal fin tip not reaching caudal fin. Number of dorsal fin rays 17-18, anal fin rays Pectoral fin (20 rays) approximately triangular; in some specimens pectoral fins reach origin of pelvic fins, and pelvic fins (6-8 fin rays) reach the anal fin origin. Caudal fin rounded (24-26 rays). Scales cycloid, body and head entirely scaled, except for ventral surface of head. Scales in median lateral series on caudal fin base. Transverse row of scales 12. Circumpeduncular scales 14. Cephalic squamation pattern variable. Frontal neuromast series of the open type. Anterior supraorbital series in shallow grooves, each with two or three neuromasts. Posterior supraorbital series with three neuromasts. One neuromast on each scale of median longitudinal series. Six branchiostegal rays. Vomerine teeth present in a small patch. Lateral process of post-temporal not present. Second pharyngobranchial with one tooth. Bifid antero - dorsal process of urohyal. Number of vertebrae 28. Premaxilla and dentary with many irregularly distributed unicuspid, slightly curved teeth of different size, a small number of larger ones on the outer row of upper and lower jaw. Females smaller than Fig. 5. Nothobranchius ivanovae, adult male, about 35 mm SL, not preserved, Katuma River drainage, western Tanzania. Photo by I. Ivanova. Fig. 6. Nothobranchius ivanovae, adult female, about 30 mm SL, not preserved, Katuma River drainage, western Tanzania. Photo by I. Ivanova. aqua vol. 18 no October

19 Stefano Valdesalici Table II. Morphometric data of Nothobranchius ivanovae. Measurements were given as percentages of standard length in mm. Eye diameter and snout length were given as percentages of head length. Cleared and stained material is included. Holotype Males (n = 6) Females (n = 4) Standard length (mm) Body depth at pelvic fin Predorsal length Preanal length Prepelvic length Caudal peduncle length Caudal peduncle depth Dorsal fin base length Anal fin base length Head length Snout length Eye diameter males, maximum observed size 33.1 mm SL. In females, dorsal fin rounded, anal fin triangular with rounded tip, branchiostegal membrane not projecting from opercle. Colour in life: Males (Fig. 5). Body and head scales light blue-grey with thin dark red margin, forming a reticulated pattern on body and head. Snout, frontal, and dorsal portion of head pale brown. Branchiostegal membrane whitish. Dorsal fin grey-green with irregular rows of dark red small dots; proximally larger, distally denser and smaller, dark red spots on upper anterior margin. Anal fin cream with a proximal red line and broad dark red-brown margin. Pelvic fins cream, red-brown near the base and with a brown rim. Pectoral fins hyaline with light blue margins. Caudal fin red, narrow submarginal orange band, with a broad black margin. Iris silver, with faint black vertical bar through centre of eye. Female (Fig. 6). Body and head scales grey brown with brown scale margin, forming a reticulated pattern. Opercular region silver to golden. Abdomen silver to golden. All fins hyaline. Iris silver. Etymology: The species name is dedicated to the aquarist, collector, and discoverer of the species, Iva Ivanova from Dupnitsa, Bulgaria. Distribution and habitat: (Figs 3-4) Nothobranchius ivanovae is currently known from a large swampy area that covers about 50 km from Kabungo to Sibwesa, and belongs to the Katuma river drainage, western Tanzania. The type locality was at the time of collection a residual pool in the dry bed of Karira stream. Other fish species present were N. kardashevi and unidentified Barbus species. Discussion: Nothobranchius kardashevi clearly belongs to the N. ugandensis species group (Valdesalici 2009). The N. ugandensis species group has a wide distribution, ranging from Sudan (seasonal rivers around Nuba mountains) and Ethiopia (Alvero River drainage) in the north to the Bubu River drainage and Lake Rukwa drainage systems in Tanzania in the south. Nothobranchius kardashevi seems to be related to N. sp. Lake Victoria, both have large red spots on the anal fin (vs. usually absent or with few small spots in N. ugandensis and the anal fin completely spotted in N. nubaensis). The latter species is still not formally described, but seems, based on distribution and colouration, an artificial assemblage of several related species. Populations are known from the Lake Victoria and Rukwa drainage systems, and the Malagarasi River basin. The closest geographically known related species are the populations of N. sp. Lake Victoria (known as N. sp. Limba Limba or N. sp. TSTS 10-5 ) found in the Malagarasi drainage, and N. sp. Piti River in the Lake Rukwa drainage (Seegers 1997; Shidlovsky 2010, Wildekamp 1990). Both are clearly separated from N. kardashevi due to their distinctive anal fin coloration (yellow vs. blue), caudal fin pattern (plain or with proximal part spotted vs. completely spotted) and spot coloration on the dorsal and anal fins 197 aqua vol. 18 no October 2012

20 Nothobranchius kardashevi and Nothobranchius ivanovae, two new annual killifishes from the Katuma River drainage, western Tanzania (one colour vs. two colours). The light blue with red spots caudal fin pattern can be found in other species, which belong to different subgenera or species groups: N. eggersi (Adiniops), N. jubbi (Nothobranchius), and in Zononothobranchius: N. malaissei, and N. symoensi, both belonging to the N. taeniopygus species group and not to the N. ugandensis species group like N. kardashevi. Nothobranchius ivanovae belongs to the N. taeniopygus species group, an assemblage of species, that has a wide distribution in Tanzania and Zambia (Valdesalici & Amato 2011).Within this group, N. ivanovae shows similarities with some populations of N. taeniopygus. This latter species was first described by Hilgendorf in 1891, and was redescribed by Wildekamp in The neotype, designated from Kapatu Creek on Wembere drainage, received a proper description. In the central Tanzanian shield (Giddelo et al. 2002) Nothobranchius taeniopygus is widespread from the drainage of the Bubu River in the east, to Lake Victoria and Eyasi drainage systems in the north, the Malagarasi River drainage in the west, and Lake Rukwa drainage in the south. Seegers (1996) was first to report on different, geographically restricted forms within this species, and recently Wildekamp (2004) recognized the different known forms of N. taeniopygus as four distinct phenotypes. Nothobranchius ivanovae is clearly separated from the known phenotypes by its unique combination of characters. ACKNOWLEDGEMENTS I want to thank Francesca Fontana for the assistance with the first draft of the manuscript, Kiril Kardashev and Iva Ivanova for donating specimens, information, and fish pictures, Stefan van der Voort for grammar suggestions, and Béla Nagy for preparing the distribution map. I am grateful to Miguël Parrent (MRAC) and Giuliano Doria (MSNG) for access to material under their care. REFERENCES AMIET, J. L Faune du Cameroun. Fauna of Cameroon. Vol. 2. Le genre Aphyosemion Myers (Pisces: Teleostei: Cyprinodontiformes). Sciences Naturales, Compiègne, 262 pp. DE VOS, L., SEEGERS, L., TAVERNE, L. & THYS VAN DER AUDENAERDE, D L ichtyofaune du basin de la Malagarasi (Système du Lac Tanganyka) une synthèse de la connaissance actuelle. Annales, Sciences Zoologiques, Musée Royal de l Afrique Centrale 285: GIDDELO, C. S., ARNDT, A. D. & VOLCKAERT, F. A. M Impact of rifting and hydrography on the genetic structure of Clarias gariepinus in eastern Africa. Journal of Fish Biology 60: HILGENDORF, F. M Eine Aufzählung der von Emin Pascha und Dr. Stuhlmann gesammelten Fische und Krebse. Sitzungs-Berichte der Gesellschaft Naturforschender Freunde zu Berlin 1: HOEDEMAN, J. J The frontal scalation pattern in some groups of tooth carps (Pisces, Cyprinodontiformes). Bulletin of Aquatic Biology 1: HUBER, J. H Killi-Data Updated checklist of taxonomic names, collecting localities and bibliographic references of oviparous Cyprinodont fishes (Cyprinodontiformes). Société Française d Ichtyologie, Paris, 538 pp. SCHEEL, J. J Rivulins of the Old World. Tropical Fish Hobbyist Publication, Neptune City, 480 pp. SEEGERS, L The fishes of the Lake Rukwa drainage. Annales, Sciences Zoologiques, Musée Royal de l Afrique Centrale 278: SEEGERS, L Killifishes of the World. Old World Killis II. Verlag A.C.S. Mörfelden-Walldorf, 112 pp. SHIDLOVSKY, K Collecting Nothobranchius in South- Western Tanzania. Nothobranchius archives 1: TAYLOR, W. R. & VAN DYKE G. C Revised procedures for staining and clearing small fishes and other vertebrates for bone and cartilage study. Cybium 9: VALDESALICI, S Nothobranchius boklundi (Cyprinodontiformes: Nothobranchiidae): a new annual killifish with two male colour morphs from the Luangwa River basin, Zambia. aqua, International Journal of Ichthyology 16 (2): VALDESALICI, S. & AMATO, G Nothobranchius oestergaardi (Cyprinodontiformes: Nothobranchiidae), a new annual killifish from Mweru Wantipa Lake drainage basin, northern Zambia. aqua, International Journal of Ichthyology 17 (2): VALDESALICI, S., BELLEMANS, M., KARDASHEV, K. & GOL- UBTSOV, A Nothobranchius nubaensis (Cyprinodontiformes: Nothobranchiidae) a new annual killifish from Sudan and Ethiopia. aqua, International Journal of Ichthyology 15 (3): VALDESALICI, S. & KARDASHEV, K Nothobranchius seegersi (Cyprinodontiformes: Nothobranchiidae), a new annual killifish from the Malagarasi River drainage, Tanzania. Bonn zoological Bulletin 60: WILDEKAMP, R. H Redescription of two lesser known Nothobranchius from Central Tanzania, N. taeniopygus and N. neumanni (Cyprinodontiformes: Aplocheilidae). Ichthyological Exploration of Freshwaters 1: WILDEKAMP, R. H The Nothobranchius species from Uganda, with description of a new polymorphic species (Cyprinodontiformes: Aplocheilidae). Ichthyological Exploration of Freshwaters 5: WILDEKAMP, R. H A world of killies. Atlas of the oviparous cyprinodontiform fishes of the world. Volume 4. The American Killifish Association, Elyria, 398 pp. aqua vol. 18 no October

21 aqua, International Journal of Ichthyology Male Grijalva mosquitofish (Heterophallus milleri Radda, 1987) increase individual mating preferences in front of an audience David Bierbach 1*, Madlen Ziege 1, Claudia Zimmer 1, Rüdiger Riesch 2, Ingo Schlupp 3, Lenin Arias-Rodriguez 4 & Martin Plath 1 1) Johann-Wolfgang-Goethe University of Frankfurt, Evolutionary Ecology Group, Max-von-Laue-Str. 13, D-60438, Frankfurt am Main, Germany. s: MZ madlen.ziege@googl .com; CZ cla.zim@web.de; MP mplath@bio.uni-frankfurt.de 2) North Carolina State University, Department of Biology & W. M. Keck Center for Behavioral Biology, 127 David Clark Labs, Raleigh, NC , USA. rwriesch.evolutionarybiology@gmail.com 3) University of Oklahoma, Department of Biology, 730 Van Vleet Oval, Norman, OK 73019, USA. Schlupp@ou.edu 4) División Académica de Ciencias Biológicas, Universidad Juárez Autónoma de Tabasco (UJAT), C.P Villahermosa, Tabasco, México. leninariasrodriguez@hotmail.com *corresponding author s address: Johann-Wolfgang-Goethe University of Frankfurt, Evolutionary Ecology Group, Max-von-Laue-Str. 13, D 60438, Frankfurt am Main, Germany Tel: ; Fax: David.Bierbach@gmx.de Received: 27 February 2012 Accepted: 9 June 2012 Abstract Socially influenced mate choice behavior is currently a growing field in the study of sexual selection and evolution. Here, we provide the first description of male Grijalva mosquitofish (Heterophallus milleri) courtship behavior, and further report on an unparalleled audience effect in that species. Lab-reared male Grijalva mosquitofish significantly increased their preference for an initially preferred female in a full contact design in the presence of another male. This is somewhat unexpected as previous studies found males of other members of the family Poeciliidae to interact more evenly with stimulus females when faced with an audience, and thus decreasing their preference for the initially preferred female. As those audience effects are assumed to represent male tactics to prevent sperm competition risk arising through male mate choice copying, we argue that male mate choice copying might not play a major role in the reproductive biology of H. milleri. Zusammenfassung Der Einfluss des sozialen Umfeldes auf das Partnerwahlverhalten ist als Fokusbereich bei der Erforschung von sexueller Selektion und Evolution von zentralem Interesse. In dieser Studie beschreiben wir erstmalig das Balzverhalten des Grijalva-Moskitofischs (Heterophallus milleri) und zeigen weiterhin einen bisher beispiellosen Publikums - effekt in dieser lebendgebärenden Fischart. Laborgehaltene Grijalva-Moskitofisch-Männchen steigerten ihre sexuelle Präferenz für ein zuvor bevorzugtes Weibchen signifikant, wenn sie von einem anderen Männchen beobachtet wurden. Dies ist ungewöhnlich, da Männchen anderer Mitglieder der Familie Poeciliidae in früheren Studien gleichmäßiger mit den Stimulus-Weibchen interagierten und somit ihre eingangs gezeigte Präferenz abschwächten. Da solche Publikumseffekte als Strategien zur Vermeidung von Spermienkonkurrenz im Zuge von männlichem Partnerwahlkopieren angesehen werden, vermuten wir, dass Partnerwahlkopieren im männlichen Geschlecht von H. milleri kaum eine Rolle spielt. Résumé Le comportement de choix du partenaire sous influence sociale est actuellement un champ d investigation croissant dans l étude de la sélection et de l évolution sexuelles. Nous proposons ici la première description du comportement de la parade de reproduction du mâle Heterophallus milleri. Des mâles de cette espèce élevés en laboratoire augmentaient leur préférence pour une femelle déjà choisie par un comportement de plein contact en présence d un autre mâle. C est chose plutôt inattendue dans la mesure où d autres études ont montré que les mâles d autres membres des Poeciliidae interagissent même davantage avec des femelles stimulus en présence d un groupe et donc qu ils diminuent leur intérêt pour la femelle initialement préférée. Comme ces «effets de groupe» sont sensés 199 aqua vol. 18 no October 2012

22 Male Grijalva mosquitofish (Heterophallus milleri Radda, 1987) increase individual mating preferences in front of an audience représenter des tactiques mâles pour prévenir ce risque de compétition de sperme survenant par l imitation du choix de la partenaire par le mâle, nous soutenons la thèse que l imitation susdite pourrait ne pas jouer un rôle majeur dans la biologie reproductrice de H. milleri. Sommario Il comportamento socialmente influenzato della scelta del compagno è attualmente un settore in crescita nello studio della selezione sessuale e dell'evoluzione. Qui for nia mo la prima descrizione del comportamento di cor teg giamento del maschio di gambusia Grijalva (Heterophallus milleri) e un resoconto su un "effetto spettatore" senza pre cedenti in questa specie. I maschi di gambusia Grijalva alle vati in labo - ratorio aumentano significativamente la loro preferenza per una femmina inizialmente prescelta in un disegno sperimentale a pieno contatto in presenza di un altro maschio. Questo è un risultato inaspettato, poiché stu di precedenti hanno evidenziato che maschi di altri membri della famiglia Poeciliidae interagiscono di fronte a un pubblico in modo più uniforme con le femmine stimo lo, diminuendo l inte - resse per la femmina inizialmente preferita. Poiché si presume che tale effetto pubblico rappresenti una tattica del sesso maschile per prevenire il ri schio di competizione da spermatozoi derivante dal copia re la scelta della compagna da parte di un maschio, si so stiene che questo fenomeno di imitazione non svolgerebbe un ruolo importante nella biologia riproduttiva di H. milleri. INTRODUCTION Gregarious animals often choose their mating partners within social aggregations, i.e. in front of conspecifics, so mate choice decisions and any other communicatory interactions between choosing individuals, their potential mates, and bystanding individuals are usually part of extensive communication networks (Matos & Schlupp 2005; McGregor & Peake 2000; Peake 2005; Earley & Dugatkin 2005). By-standing individuals may extract and subsequently use information from observed (sexual) interactions to modify their own mating decisions ( social eavesdropping ; Naguib et al. 2004; Dabelsteen 2005; Earley 2010); for instance, male fiddler crabs (Uca mjoebergi Rathbun, 1924) use rivals courtship behavior to detect receptive females (Milner et al. 2010). Another form of social eavesdropping is mate choice copying, during which a female s propensity to mate with a given male increases after it had observed that male interact sexually with another female (Pruett-Jones 1992, Galef 2008; Vakirtzis 2011). Mate choice copying appears to be widespread throughout the Animal Kingdom (teleost fishes: Dugatkin 1992; Schlupp et al. 1994; Witte 2006; birds: Freed-Brown and White 2009; mammals: Galef et al. 2008, including humans: Wayneforth 2007, Place et al. 2010). Most studies reported on female mate choice copying, but males, too, were found to copy other males mating decisions (poeciliid fishes: Schlupp & Ryan 1997; Bierbach et al. 2011a; pipefish: Widemo 2006; humans: Waynforth 2007; Place et al. 2010). However, at least in internally fertilizing species, copulating with a previously mated female clearly elevates sperm competition risk for both, the copier and the copied male (Plath & Schlupp 2008; Plath & Bierbach 2011). Males that are at risk of being copied by another male are predicted to respond by trying to prevent the rival from copying. Indeed, males may alter their mating behavior strategically when observed by a sexually active rival (i.e., by a potential copier; Plath et al. 2008a,b,c; Ziege et al. 2009; Dosen & Montgomerie 2004; Bierbach et al. 2011b, 2012; Wronski et al. 2012). Such audience effects have been examined most thoroughly in male Atlantic mollies (Poecilia mexicana Steindacher 1863, family Poeciliidae), which decrease sexual activity and cease expressing mating preferences when a rival male is nearby (Plath et al. 2008a,b,c, 2010; Ziege et al. 2009; Bierbach et al. 2011b, 2012). Moreover, males interacted first with the previously nonpreferred female when an audience male was presented, thereby potentially exploiting male mate choice copying behavior so as to lead the rival away from the preferred mate ( deceptive behavior, Plath et al. 2008c, 2010; Bierbach et al. 2012). These behavioral responses to an audience male were confirmed in a comparative study by Bierbach et al. (2012) for 9 different species of poeciliid fishes (P. mexicana, P. orri Fowler, 1943, P. lati - pinna (LeSueur, 1821), P. latipunctata Meek, 1904, P. reticulata Peters, 1859, P. picta (Regan, 1913), Limia tridens (Hilgendorf, 1889), L. sulphurophilia Rivas, 1980, Gambusia sexradiata Hubbs, 1936) and also found in male sand gazelles (Gazella marica, Wronski et al. 2012). The only exception from this general pattern appears to be the Grijalva mosquitofish (Heterophallus milleri Radda, 1987), a close relative to the genus Gambusia (Hrbek et al lists H. milleri in the genus Gambusia) that is endemic to the Río Grijalva/Usumacinta drainage in southern Mexico (Miller et al. 2005). Interestingly, males in this species appear to increase their mating preferences when confronted with a rival (Ziege et al. 2008). However, in contrast to the comparative study by aqua vol. 18 no October

23 David Bierbach, Madlen Ziege, Claudia Zimmer, Rüdiger Riesch, Ingo Schlupp, Lenin Arias-Rodriguez and Martin Plath Bierbach et al. (2012) as well as the study first describing male deceptive mating behavior (Plath et al. 2008c), which both established male mating preferences through measuring direct sexual interactions, the study by Ziege et al. (2008) used association times instead. Furthermore, wild-caught fish were used by Ziege et al. (2008) instead of labreared individuals that were used to being held in glass tanks and had previously been exposed to handling stress. Thus comparability of the seemingly contradictory results from Ziege et al. (2008) with the results of other studies is difficult. We, thus, hypothesize that laboratory-reared H. milleri males would show similar behavioral responses to the presence of an audience male reduced sexual activity and cessation in the expression of mating preferences as reported for other lab-reared poeciliids. Our present study, therefore, made use of laboratory stocks of H. milleri that were established during the past three years to better control for effects related to handling stress while we further measured direct sexual interactions to establish male mating preferences, which increases comparability with the data from other poeciliid species. Thus, the current study (a) describes the behavioral repertoire of male H. milleri in terms of courtship and mating behavior, and (b) tests for audience effects in the controlled environment of a lab-based experimental setup. MATERIAL AND METHODS Origin and maintenance of the study organism: All fish used in this study were lab-reared descendants of wild-caught fish collected in 2006 in a small creek, Arroyo Bonita, a tributary to the Río Oxolotán near the village of Tapijulapa (for details see Riesch et al. 2011). Test fish were maintained as randomly outbred mixed-sex stocks in 1000-L tanks in a greenhouse of the Aquatic Research Facility of the University of Oklahoma in Norman, U.S.A. Experiments were conducted in November Test fish were kept at a temperature of 25 C, under a natural light: dark regime and fed daily ad libitum with commercially available flake food (TetraMin ). We isolated focal (N = 30) males in 38-l tanks for four days prior to the tests to ensure that they were motivated to mate (Schlupp & Plath 2005). Each focal male was tested only once; however, due to the limited number of males available from our stocks, some males were also used as audience males after they were used as a focal male, but never on the same day. All fish used in a trial were taken from different stock tanks to avoid confounding effects of familiarity (Bierbach et al. 2011b) Experimental design: We used a full contact design in which focal males mating behavior towards two types of females was videotaped by a Canon 3CCDF Digital Video Camcorder and videos were used for a qualitative description of courtship displays (a sample tape is provided in the online supplement). To analyze males change in mating behavior when presented with a by-standing rival, sexual behaviors towards the two stimulus females were scored as a measure of mating preferences. Each focal male was tested for its baseline mating preference and was then retested either without an audience (control) or with an audience male present (audience treatment). This allowed us to examine the change in focal males behavior from the 1 st to the 2 nd part of the tests while we could differentiate between effects induced by an audience and changes that would occur over the course of the experiment even without an audience present (Plath et al. 2008c;). The test tank (50 cm length x 30 cm width x 30 cm height) was filled to 20 cm with aged tap water of C. Black plastic covered all sides except the front. Prior to the tests, we selected two different-sized stimulus females [mean (± SE) SL; large: 33.3 ± 0.5 mm; small: 25.0 ± 0.7 mm] from the stock tanks and introduced them into the test tank. We exchanged stimulus females after each trial and retransferred them to the stock tanks, so that some females may have been used more than once (but never on the same day). Poeciliid males typically prefer to mate with larger females (e.g., Bisazza et al. 1989; Herdman et al. 2004) as fecundity usually correlates positively with body size (e.g., Pires et al. 2011) a pattern that was also confirmed for H. milleri (Riesch et al. 2011). Afterwards, we introduced the focal male (22.2 ± 0.5 mm, N = 30) into a transparent Plexiglas cylinder (10 cm diameter) in the center of the tank and left the fish undisturbed for 5 minutes. After this habituation period, we gently lifted the cylinder. During a 10- min observation period, male sexual behaviors (nipping, gonopodial thrusts, and courtship displays towards either type of females) were scored, and we noted with which female the male interacted first. We decided a priori to terminate a trial if the male did not show any sexual behavior during the 1 st part of the test; N = 6 trials (out of 30) were discarded based on this criterion so that data 201 aqua vol. 18 no October 2012

24 Male Grijalva mosquitofish (Heterophallus milleri Radda, 1987) increase individual mating preferences in front of an audience from N = 24 males (12 per treatment) could be analyzed further. Nipping is a typical pre-copulatory behavior in poeciliids, whereby the male approaches the female from behind and touches her genital region with his snout (Liley 1966; Parzefall 1969). Thrusting was defined as attempts to introduce the male s gonopodium into the female s genital opening. Immediately after the first preference test, we repeated measurement of male mating preferences, Fig. 1. Videotaped courtship sequence of H. milleri. For a description see main text. aqua vol. 18 no October

25 David Bierbach, Madlen Ziege, Claudia Zimmer, Rüdiger Riesch, Ingo Schlupp, Lenin Arias-Rodriguez and Martin Plath Fig. 2. Influence of a by-standing male on male mate choice behavior in Grijalva mosquitofish. During the 1 st Part, the focal male could interact with two stimulus females [large (black bars) or small (gray bars)] without an audience present. During the second part of the trials, half of the males were visually presented with an audience male (2 nd Part with Audience), while another half of trials were repeated without an audience (2 nd Part Control). Depicted are means (± SE) for (A) nipping and (B) thrusting, along with results from paired t-tests. Also given are (C) fractions of males that interacted with either the large or small female first when released from their acclimatization cylinder (Binomial tests). 203 aqua vol. 18 no October 2012

26 Male Grijalva mosquitofish (Heterophallus milleri Radda, 1987) increase individual mating preferences in front of an audience but in one half of the trials, an audience male was presented, while the other half was repeated without an audience (control). To initiate this 2 nd part of a trial, we reintroduced the focal male into the acclimatization cylinder. An audience male (21.5 ± 0.5 mm) was placed in another transparent cylinder in the central back of the tank, while for the control an empty cylinder was presented. The audience male was confined in his cylinder throughout the experiment. After another 5 minutes for habituation (during which all four fish could interact visually), measurement of male preferences was carried out as described above. Interactions between males were not quantified; qualitatively, we often observed synchronized swimming up and down of the two males. After a trial was completed, all fish were measured for standard length (SL) to the closest millimeter. Statistical analysis: All data were tested for normal distribution using Shapiro-Wilk tests and, throughout, all relative data were arcsine (square root)-transformed prior to the statistical analyses. However, as courtship displays were performed by only 16% (4 out of 24) of males in both parts of the test, 40%, (10 out of 24) showed courtship in at least one part, we did not quantitatively analyze this behavior. Poeciliid males tend to reduce sexual activity (determined herein as the sum of nipping and thrusting with both stimulus females) upon exposure to a rival (Bierbach et al. 2012). In our first analysis, we therefore compared the change of male sexual activity from the 1 st to 2 nd part of the tests. We used the relative sexual activity during the second part of the tests (as a fraction of the sexual activity during both parts) as dependent variable in a univariate GLM (general linear model) including treatment (with or without audience) as factor. As covariates we included focal males SL, as well as the size difference between the large and small stimulus female (SL large SL small female), but removed both as they had no significant effects ( focal male Fig. 3. (A) Change in male sexual activity during the 2 nd part of the tests (as fraction of the sum of sexual behaviors during both parts). During the 2 nd part males were randomly assigned to either a control treatment (control, black bars) or a treatment during which their mate choice was observed by a rival male (audience, gray bars). Values above 0.5 indicate increased sexual activity during the 2 nd part. (B) Changes in individual focal males mating preferences induced by the presence of an audience male. Depicted are preference scores (see main text), whereby negative values indicate that male preferences decreased in strength. Data in (A) and (B) are means ± SE. aqua vol. 18 no October

27 David Bierbach, Madlen Ziege, Claudia Zimmer, Rüdiger Riesch, Ingo Schlupp, Lenin Arias-Rodriguez and Martin Plath body size : F 1,22 = 0.09, P = 0.76; female body size difference : F 1,21 = 0.01, P = 0.94). To explore the direction of male preferences, we compared numbers of sexual behaviors with the large and the small stimulus female during either of the two test parts using paired t-tests; nipping and gonopodial thrusting were analyzed separately in this analysis. Moreover, males were predicted to alter their individual mating preferences under the influence of an audience (Plath & Schlupp 2008; Plath & Bierbach 2011). To explore this effect, a preference score, based on focal males individual preferences, was calculated as: [(fraction of nipping and thrusting directed toward the initially preferred female during the 2 nd part of a trial) (fraction of nipping and thrusting with the same female during the 1 st part)]. Hence, the score expressed the change between the two parts of the tests, whereby negative values would indicate that preferences decreased in strength while positive values would indicate an increase in strength of preference (Plath et al. 2008a; Ziege et al. 2009; Bierbach et al. 2011b, 2012). Data were compared in a GLM using treatment (with or without audience) as a factor. In the initial model, we included focal male body size (F 1,22 = 1.93, P = 0.18) and female body size difference (F 1,21 = 0.18, P = 0.68) but removed both covariates from the final model as they had no significant effect. Finally, males of other poeciliids, including the closely related G. sexradiata, initially approach the previously non-preferred female upon presentation of a rival (Bierbach et al. 2012). We asked whether a similar pattern would be uncovered in H. milleri and first compared the numbers of males that directed their first sexual behavior towards the large or small female using Binomial tests. Data from the 1 st and 2 nd part of the tests were analyzed separately. Second, we compared the proportion of males that altered their individual preferences from the 1 st to 2 nd part between the control group and the audience treatment using a chi²-test. All statistical analyses were run using SPSS RESULTS Description of male courtship displays: Heterophallus milleri males showed conspicuous courtship displays during our mate choice tests. During courtship, males usually start to swim calmly behind the female s anal fin and follow the female in exactly this position for up to one minute. From this position males move to nip at the female s gonopore similar to other poeciliids. Males then suddenly start to circle around the female, which typically remains stationary throughout (Fig. 1 A-F). During these circles males rest for a short period in front of the female s head and bend their bodies to an S-shape (Fig. 1 G-I). In this position males raise their tails and vibrate them while folding all unpaired fins. Additionally, males heavily swing their gonopodia at this stage (Fig. 1 G-H). After this episode, they either quickly swim behind the female and repeat the circling performance or they slowly swim backwards with all unpaired fins folded (Fig. 1 J-L) until both the male s and the female s heads are at the same level (Fig. 1 L). They then quickly turn around and try to thrust the female, which now typically swims away (Fig. 1 M-O). We were also able to observe those courtship circles from above in natural habitats of H. milleri, as this species often swims directly below the water surface (see Riesch et al. 2011). Changes in sexual activity induced by an audience male: Even though visual inspection of our data suggests that males slightly increased sexual activity when observed by an audience (Fig. 2a), no significant difference was detected in the GLM (mean square = 0.04, F 1,23 = 0.94, P = 0.34). Audience-induced changes in the overall direction of male mating preferences: Males tended to direct more nipping and thrusting towards the larger female during the 1 st part of the tests, however, this effect was not significant (see Fig. 3a, b for results from paired t-tests), and males in both treatments also did not show a significant preference during the 2 nd part (Fig. 3a-b). Audience-induced changes in individual male mating preferences: When comparing the strength of change in individual male mating preferences from the 1 st to 2 nd part of the tests using GLM (i.e., preference scores) irrespective of whether the larger or smaller female was preferred a significant difference between treatments was uncovered (mean square = 0.49, F 1,23 = 5.49, P = 0.028), and males increased their initial preference in response to the presence of an audience (Fig. 2b). First sexual interaction: When released from the acclimatization cylinder most focal males first interacted with the larger stimulus female during the 1 st part of the tests (see Fig. 2c for P-values from Binomial tests). This pattern remained virtually unchanged during the 2 nd part of the tests in both treatments (Fig 2c), and even though the effect was not significant, qualitatively, most males 205 aqua vol. 18 no October 2012

28 Male Grijalva mosquitofish (Heterophallus milleri Radda, 1987) increase individual mating preferences in front of an audience still approached the larger stimulus female first during the second test part. Comparing the fraction of males that switched to another female during the 2 nd test part, we did not find a significant difference between the control and audience treatments (chi²-test; ² = 1.15, df = 1, P = 0.28). DISCUSSION Surprisingly, some males showed vigorous courtship behavior, e.g., circled around a female and performed a swinging movement of their gonopodium directly in front of a female. This was somewhat unexpected since courtship is reduced and rare in the closely related genus Gambusia (e.g. Rosen & Tucker 1961; Hughes 1985; Pilastro et al. 1997). However, if put into a larger phylogenetic context, the tribe Gambusiini also comprises genera and species (e.g. the genus Xiphophorus or the pike killifish, Belonesox belizanus Kner 1860; Hrbek et al. 2007) that are known to often vigorously court their females (Farr 1989). Furthermore, females of at least some members of the genus Gambusia were found to exhibit a mating preference for males with large gonopodia (see Langerhans et al for discussion, but see Gasparini et al for absence of such a preference in P. reticulata), so we suspect courtship displays of male H. milleri to be an advertising strategy that might exploit such a female preference. Our mate choice tests confirmed previous results from Ziege et al. (2008) in which the authors established male mating preferences and the influence of an audience by measuring association times in a dichotomous design. Males increased their preferences for the initially preferred female when observed by an audience. Nevertheless, as wildcaught individuals were used by Ziege et al. (2008) we initially hypothesized that those effects might have been an artifact of testing wild caught fish rather than representing the species normal behavioral repertoire; wild H. milleri were clearly not accustomed to the handling stress. While being stressed, wild-caught H. milleri males may not have perceived a fourth fish (the audience male) as a rival, but safety-in-numbers due to the addition of another shoal mate may have had a calming effect (Queiroz & Magurran 2005), resulting in an increased motivation to mate. However, after confirming these results in the present study, we now ask why male H. milleri behave differently compared to other poeciliids which typically decrease their sexual activity and interact more with the initially non-preferred female when presented with an audience male especially compared to members of the closely related genus Gambusia (see Bierbach et al. 2012). One hypothesis that has been put forward to explain the unusual behavior of H. milleri males was that males could try to monopolize preferred females and defend them from other males (Ziege et al. 2008). Preliminary data from a comparison of male aggressive behavior in several poeciliids, however, identify H. milleri as being little aggressive (D. Bierbach et al., unpublished data) while high aggression would be predicted if overt competition with rivals played a major role. Generally, it was argued that audience effects (i.e., reduced sexual behavior, reduced mating preferences, and approach of the usually non-preferred female first) can be interpreted as tactics to reduce sperm competition risk arising from male mate choice copying (see Schlupp & Ryan 1997; Plath & Schlupp 2008; Plath & Bierbach 2011). However, little is known about male mate choice copying in the genus Heterophallus as well as in the related genus Gambusia. Possibly, male mate choice copying does not play a major role in the reproductive biology of H. milleri (for a similar argument see Ziege et al. 2008). In this context, it will be important to not only establish whether or not males copy other males choices under laboratory conditions, but especially knowledge about the probability for male mate choice copying to occur under natural conditions is vital to our understanding of the evolutionary forces shaping male behavioral responses to rivals (Plath & Bierbach 2011). In summary, we are as yet lacking a compelling answer as to the question of why H. milleri males differ in their response to an audience from any other poeciliid investigated so far. Our study, therefore, identifies H. milleri as being of central importance for future studies looking into adaptive explanations of audience-induced behavioral modification in male poeciliid fishes, especially when trying to answer the question why closely related species differ in their behavioral repertoire. ACKNOWLEDGEMENTS We thank Amber Makowicz for her help in the lab. Financial support came from the Deutsche Forschungsgemeinschaft (PL 470/3-1). This research was conducted under the Mexican authorization of the Municipio de Tacotalpa in Tabasco and SEMARNAT. aqua vol. 18 no October

29 David Bierbach, Madlen Ziege, Claudia Zimmer, Rüdiger Riesch, Ingo Schlupp, Lenin Arias-Rodriguez and Martin Plath REFERENCES BIERBACH, D., KRONMARCK, C., HENNIGE-SCHULZ, C., STADLER, S. & PLATH, M. 2011a. Sperm competition risk affects male mate choice copying. Behavioral Ecology and Sociobiology 65: BIERBACH, D., GIRNDT, A., HAMFLER, S., KLEIN, M., MUCKSCH, F., PENSHORN, M., SCHWINN, M., ZIMMER, C., SCHLUPP, I., STREIT, B. & PLATH, M. 2011b. Male fish use prior knowledge about rivals to adjust their mate choice. Biology Letters 7: BIERBACH, D., KONSTANTIN, Y., STADLER, S., SCHENKEL, D., MAKOWICZ, M. E., NIGL, H., GEUPEL, H., RIESCH, R., SCHLUPP, I. & PLATH, M Audience effects alters male mate choice in poecilids. In: Social learning theory: Phylogenetic considerations across animal, plant, and microbial taxa (Ed. K. B. Clark). Nova Science Publishers, Hauppauge, NY (in press). BISAZZA, A., MARCONATO, A. & MARIN, G Male mate preference in the mosquitofish Gambusia holbrooki. Ethology 83: DABELSTEEN, T Public, private or anonymous? Facilitating and countering eavesdropping. In: Animal communication networks (Ed. P. K. McGregor): Cambridge University Press, Cambridge. DOSEN, L. D. & MONTGOMERIE, R Mate preferences by male guppies (Poecilia reticulata) in relation to the risk of sperm competition. Behavioral Ecology and Sociobiology 55: EARLEY, R. L Social eavesdropping and the evolution of conditional cooperation and cheating strategies. Philosophical Transactions of the Royal Society B: Biological Sciences 365: EARLEY, R. L. & DUGATKIN, L.A Fighting, mating and networking: pillars of poeciliid sociality. In: Animal communication networks (Ed. P. K. McGregor): Cambridge University Press, Cambridge. FARR, J. A Sexual selection and secondary sexual differentiation in poeciliids: Determinants of male mating success and the evolution of female choice. In: Ecology & Evolution of Livebearing Fishes (Poeciliidae) (Eds. G. K. Meffe & F. F. Snelson Jr.): Prentice Hall, Englewood Cliffs, NJ. FREED-BROWN, G. & WHITE, D. J Acoustic mate copying: female cowbirds attend to other females vocalizations to modify their song preferences. Philosophical Transactions of the Royal Society B: Biological Sciences 276: GALEF, B. G Social influences on the mate choices of male and female Japanese quail. Comparative Cognition and Behavior Reviews 3: GALEF, B. G., LIM, T. C. W. & GILBERT, G. S Evidence of mate choice copying in Norway rats, Rattus norvegicus. Animal Behaviour 75: GASPARINI, C., PILASTRO, A. & EVANS, J.P Male genital morphology and its influence on female mating preferences and paternity success in guppies. PLoS ONE 6:e HERDMAN, E. J. E., KELLY, C. D. & GODIN, J.-G. J Male mate choice in the guppy (Poecilia reticulata): do males prefer larger females as mates? Ethology 110: HRBEK, T., SECKINGER, J. & MEYER, A A phylogenetic and biogeographic perspective on the evolution of poeciliid fishes. Molecular Phylogenetics and Evolution 43: HUGHES, A. L Male size, mating success, and mating strategy in the mosquitofish Gambusia affinis (Poeciliidae). Behavioral Ecology and Sociobiology 17: LANGERHANS, R. B., LAYMAN, C. A. & DEWITT, T. J Male genital size reflects a tradeoff between attracting mates and avoiding predators in two live-bearing fish species. Proceedings of the National Academy of Sciences of the United States of America 102: LILEY, N. R Ethological isolating mechanisms in four sympatric species of poeciliid fishes. Behaviour (Suppl.) 13: MCGREGOR, P. K. & PEAKE, T Communication networks: social environments for receiving and signaling behaviour. Acta Ethologica 2: MILLER, R. R., MINCKLEY, W. L. & NORRIS, S. M Freshwater Fishes of México. University of Chicago Press, Chicago, 419 pp. MILNER, R. N. C., JENNIONS, M. D. & BACKWELL, P. R. Y Eavesdropping in crabs: an agency for lady detection. Biology Letters 6: NAGUIB, M., AMRHEIN, V. & KUNC, H. P Effects of territorial intrusions on eavesdropping neighbors: communication networks in nightingales. Behavioral Ecology 15: PARZEFALL, J Zur vergleichenden Ethologie verschiedener Mollienesia-Arten einschließlich einer Höhlenform von Mollienesia sphenops. Behaviour 33: PILASTRO, A., GIACOMELLO, E. & BISAZZA, A Sexual selection for small size in male mosquitofish (Gambusia holbrooki). Philosophical Transactions of the Royal Society B: Biological Sciences 264: PEAKE, T. M Eavesdropping in communication networks. In: Animal communication networks (Ed. P. K. McGregor): Cambridge University Press, Cambridge. PIRES, M. N., BANET, A., POLLUX, B. L. A. & REZNICK, D. N Variation and evolution of reproductive strategies in poeciliid fishes. In: Ecology & Evolution of Poeciliid Fishes (Eds. J. Evans, A. Pilastro & I. Schlupp): University of Chicago Press, Chicago. PLACE, S. S., TODD, P. M., PENKE, L. & ASENDORPF, J. B Humans show mate copying after observing real mate choices. Evolution and Human Behavior 31: PLATH, M. & BIERBACH, D Sex and the publicsocial eavesdropping, sperm competition risk, and male mate choice. Communicative and Integrative Biology 4: PLATH, M. & SCHLUPP, I Misleading mollies the effect of an audience on the expression of mating 207 aqua vol. 18 no October 2012

30 Male Grijalva mosquitofish (Heterophallus milleri Radda, 1987) increase individual mating preferences in front of an audience preferences. Communicative and Integrative Biology 1: PLATH, M., BLUM, D., SCHLUPP, I. & TIEDEMANN, R. 2008a. Audience effect alters mating preferences in Atlantic molly (Poecilia mexicana) males. Animal Behaviour 75: PLATH, M., BLUM, D., TIEDEMANN, R. & SCHLUPP, I. 2008b. A visual audience effect in a cavefish. Behaviour 145: PLATH, M., RICHTER, S., TIEDEMANN, R. & SCHLUPP, I. 2008c. Male fish deceive competitors about mating preferences. Current Biology 18: PLATH, M., RICHTER, S., SCHLUPP, I. & TIEDEMANN, R Misleading mollies: surface but not cave dwelling Poecilia mexicana males deceive competitors about mating preferences. Acta Ethologica 13: PRUETT-JONES, S Independent versus non-independent mate choice: do females copy each other? American Naturalist 140: QUEIROZ, H. & MAGURRAN, A. E Safety in numbers? Shoaling behaviour of the Amazonian red-bellied piranha. Biology Letters 1: RIESCH, R., COLSTON, T. J., JOACHIM, B. L. & SCHLUPP, I Natural history and life history of the Grijalva gambusia Heterophallus milleri Radda, 1987 (Teleostei: Poeciliidae). Aqua 17: ROSEN, D. E. & TUCKER, A Evolution of secondary sexual characters and sexual behavior patterns in a family of viviparous fishes (Cyprinodontiformes - Poeciliidae). Copeia 2: SCHLUPP, I., MARLER, C. & RYAN, M. J Benefit to male sailfin mollies of mating with heterospecific females. Science 263 (5145): SCHLUPP, I. & PLATH, M Male mate choice and sperm allocation in a sexual/asexual mating complex (Poecilia mexicana, P. formosa, Poeciliidae, Teleostei). Biology Letters 1: SCHLUPP, I. & RYAN, M. J Male sailfin mollies (Poecilia latipinna) copy the mate choice of other males. Behavioral Ecology 8: VAKIRTZIS, A Mate choice copying and nonindependent mate choice: a critical review. Annales Zoologici Fennici, 48: WAYNEFORTH, D Mate choice copying in humans. Human Nature 18: WIDEMO, M. S Male but not female pipefish copy mate choice. Behavioral Ecology 17: WITTE, K Learning and mate choice. In: Fish cognition and behavior (Eds. C. Brown, K. N. Laland & J. Krause): Blackwell, Oxford. WRONSKI, T., BIERBACH, D., CZUPALLA, L.-M., LERP, H., ZIEGE, M., CUNNINGHAM, P. L. & PLATH, M Rival presence leads to reversible changes in male mate choice of a desert dwelling ungulate. Behavioral Ecology 23: ZIEGE, M., PADUR, L., DUWE, V., RAMM, A., SCHARNWE- BER, K., RIESCH, R. & PLATH, M Audience effect alters mate choice in male Heterophallus milleri (Poeciliidae). Bulletin of Fish Biology 10: ZIEGE, M., MAHLOW, K., HENNIGE SCHULZ, C., KRON- MARCK, C., TIEDEMANN, R., STREIT, B. & PLATH, M Audience effects in the Atlantic molly (Poecilia mexicana) prudent male mate choice in response to perceived sperm competition risk? Frontiers in Zoology 6: 17. aqua vol. 18 no October

31 aqua, International Journal of Ichthyology Uranoscopus rosette, a new species of stargazer (Uranoscopidae: Trachinoidei) from the Red Sea John E. Randall 1 and Rachel J. Arnold 2 1) Bishop Museum, 1525 Bernice St., Honolulu, HI , USA. jackr@hawaii.rr.com 2) School of Aquatic and Fishery Sciences, University of Washington, 1122 NE Boat St., Box , Seattle, WA 98195, USA. schoenrj@uw.edu Received: 28 July 2012 Accepted: 3 September 2012 Abstract Uranoscopus rosette is described as a new species of stargazer (family Uranoscopidae) from 11 specimens, mm SL, collected from the Gulf of Aqaba, Red Sea, and formerly misidentified as U. fuscomaculatus Kner, U. sulphureus Valenciennes, or U. dollfusi Brüss. It is distinct in having two narrowly separated dorsal fins with IV spines and 13 soft rays, no scales dorsally on body anterior to origin of second dorsal fin; oblique scale rows midlaterally on side of body; a strong cleithral spine; supracleithrum without a distinct posterior spine; 5-7 short spines ventrally on preopercle; body depth at anal-fin origin in SL; head width in SL; lingual lure flat and broadly triangular at base with long cirri on edges, ending in a ribbon-like filament with short cirri; and a unique color pattern dorsally on the body of irregular rows of dark gray-brown spots, many rosette-like, the head densely dark-dotted dorsally. Zusammenfassung Uranoscopus rosette wird als neue Art der Himmelsgucker (Familie Uranoscopidae) auf der Grundlage von 11 Exemplaren mit mm SL beschrieben, die im Golf von Aqaba im Roten Meer gefangen wurden und zunächst fälschlicherweise als U. fuscomaculatus Kner, U. sulphureus Valenciennes oder U. dollfusi Brüss bestimmt worden waren. Die neue Art lässt sich durch folgende Merkmale unterscheiden: zwei schmale, getrennte Rückenflossen mit IV Hartstrahlen und 13 weichen Strahlen; fehlende Schuppen am Rücken im vorderen Rumpfteil bis zum Ursprung der zweiten Rückenflosse; schräge Schuppenreihen in der Mitte an der Seite; ein starker Cleithrumdorn; Supracleithrum ohne deutlichen hinteren Dorn; 5-7 kurze Dornen bauchwärts auf dem Präoperculum; Körpertiefe am Ansatz der Afterflosse 3,6-4,05 in SL; Kopfbreite 2,5-2,8 in SL; Köderfortsatz flach und breit dreieckig an der Basis mit langen Cirrhen an den Rändern, er endet in einem bandähnlichen Fortsatz mit kurzen Cirrhen; und ein unverkennbares Farbmuster rückseitig auf dem Rumpf mit unregelmäßign Reihen dunkler graubrauner Flecken, viele von ihnen rosettenähnlich, am Kopf rückseitig dicht gesät dunkle Flecken. Résumé Uranoscopus rosette est décrit comme une nouvelle espèce d Uranoscope (famille Uranoscopidae) à partir de 11 spécimens, mm de LS, collectés dans le golfe d Aqaba, mer Rouge, et jadis mal identifié comme U. fuscomaculatus Kner, U. sulphureus Valenciennes ou U. dollfusi Brüss. Elle se distingue par deux nageoires dorsales à peine séparées, avec IV épines et 13 rayons mous, l absence d écailles dorsalement sur l avant du corps antérieur à la seconde dorsale; rangées obliques d écailles sur le côté du corps; une forte épine cleithrale; un supracleithrum sans épine postérieure distincte; 5-7 épines ventralement sur le préopercule; la hauteur du corps à l origine de l anale 3,6-4,05 en LS; la largeur de la tête 2,5-2,8 en LS; un leurre lingual aplati et quasi triangulaire à la base avec de longs cirres sur les bords se terminant en un filament en forme de ruban avec des cirres courts; et un patron de coloration unique, dorsalement sur le corps, de taches grisbrun sombre en rangées irrégulières, beaucoup en forme de rosette et la tête densément couverte de points foncés dorsalement. Sommario Uranoscopus rosette è descritto come una nuova specie di pesce prete (famiglia Uranoscopidae) sulla base di 11 esemplari di mm SL, raccolti nel Golfo di Aqaba, Mar Rosso, e in precedenza erroneamente identificato come U. fuscomaculatus Kner, U. sulphureus Valenciennes o U. dollfusi Brüss. Si distingue per le seguenti caratteristiche: due pinne dorsali di poco separate con IV spine e 13 raggi molli, assenza di scaglie dorsalmente sul corpo, anteriormente all origine della seconda pinna dorsale; serie di scaglie oblique medio-laterali sul lato del corpo, una robusta spina cleitrale; supracleitrum senza una distinta spina posteriore; 5-7 corte spine ventralmente sul preopercolo; altezza del corpo a livello dell origine della pinna anale in SL; larghezza della testa in SL; esca linguale piatta e sostanzialmente triangolare alla base con cirri lungo i bordi, terminante in un filamento nastriforme con corti cirri; e, dorsalmente sul corpo, un motivo particolare costituito da righe irregolari di macchie di 209 aqua vol. 18 no October 2012

32 Uranoscopus rosette, a new species of stargazer (Uranoscopidae: Trachinoidei) from the Red Sea colore grigio scuro-marrone, molti motivi a rosetta e la testa densamente punteggiata di scuro dorsalmente. INTRODUCTION The fishes of the family Uranoscopidae are aptly named stargazers because their eyes are placed dorsally or dorsolaterally on the cuboid head and directed upward. They are benthic on sedimentary substrata, usually buried by day with only their eyes and fringed mouth visible. Most species have a fleshy lure developed from tissue of the oral valve inside the lower jaw that can be wriggled outside the mouth, worm-like, to attract their usual prey of small fishes. Pietsch (1989) published an extensive study of the osteology and myology of the family Uranoscopidae, provided comparison with the nine other families of the suborder Trachinoidei, and a key to the seven genera of the family. He listed 23 species of Uranoscopus that he expected to be valid. Twenty-three years later, Eschmeyer s updated Catalog of Fishes has the same 23 valid species in the genus, and no others. Dor (1970) published the first record of a species of Uranoscopus for the Red Sea, U. fuscomaculatus Kner (1868), type locality Fiji. In his checklist of Red Sea fishes, Dor (1984) added two more species. Brüss (1986; 1987a, 1987b) described four new Red Sea species of Uranoscopus and provided a key. Goren & Dor (1994) raised the species total to eight, one of which, U. scaber Linnaeus, is an immigrant from the Mediterranean via the Suez Canal. Khalaf & Disi (1997: 178) illustrated a stargazer from the Gulf of Aqaba as U. fuscomaculatus. In the most recent checklist of Red Sea fishes, Golani & Bogorodsky (2010) eliminated the Red Sea record of U. oligolepis Bleeker as a locality error, and followed Randall et al. (1990) in treating U. fuscomaculatus Kner from Fiji as a synonym of U. sulphureus Valenciennes in Cuvier & Valenciennes (1832), type locality Tonga. Photographs of the holotype of U. fuscomaculatus Figs 1-2. Holotype of Uranoscopus fuscomaculatus Kner, NMW 59954, mm SL, Kandavu Island, Fiji. Photos by H. Wellendorf. aqua vol. 18 no October

33 John E. Randall and Rachel J. Arnold in the Naturhistorische Museum in Vienna (Figs 1-2) and the badly damaged holotype of Uranoscopus sulphureus in the Muséum National d Histoire Naturelle in Paris (Fig. 3) have revealed both as valid species, neither of which is represented in the Red Sea. We describe here the Red Sea stargazer previously misidentified as U. fuscomaculatus or U. sulphureus as a new species. MATERIAL AND METHODS Type specimens of the new species have been deposited in the Bernice P. Bishop Museum, Hono lulu (BPBM); Hebrew University, Jerusalem (HUJ); Senckenberg Museum, Frankfurt (SMF); U. S. National Museum of Natural History, Washington, D.C. (USNM); and the University of Washington, Seattle, WA (UW). Length of specimens are given as standard length (SL), measured in millimeters from the median anterior margin of the upper lip (disregarding the cirri) to the base of the caudal fin (posterior end of hypural plate); head length is measured from the same anterior point to the posterior end of the opercular membrane; body depth is taken from the origin of the first dorsal fin; head width is the maximum width; snout length is from the median anterior point of the upper lip to the nearest edge of the orbit; orbit diameter and interorbital width are the least measurements; length of the interorbital fossa is measured from the median point at the base of the upper lip to the posterior edge of the fossa; upper-jaw length is the maximum length from the median anterior point of the upper lip; distance between gill openings is taken dorsally; cleithral spine length is measured on the dorsal edge; the distance between basipterygial processes is taken from the spine tips; fin-ray lengths are those of the longest rays. Proportional measurements are rounded to the nearest Oblique rows of scales on the body are counted (as given by Kishimoto, 1984) from the most anterior row crossing the body midlaterally (a point well beneath the pectoral fin) to the most posterior row at the centre of the caudal-fin base. Lateral-line pores were counted only on subadult paratype (obscure in adults). Fig. 3. Holotype of Uranoscopus sulphureus, MNHN 5152, Tonga. Photo by R. Causse. 211 aqua vol. 18 no October 2012

34 Uranoscopus rosette, a new species of stargazer (Uranoscopidae: Trachinoidei) from the Red Sea Uranoscopus rosette n. sp. (Figs 4-9; Table I) Uranoscopus fuscomaculatus (non Kner) Dor 1970: 23, fig. (Eilat). Uranoscopus fuscomaculatus (non Kner) Khalaf & Disi 1997: 170 (Gulf of Aqaba). Uranoscopus dollfusi (non Brüss) Debelius 1998: 169, lower fig. (Sabargad, Egypt). Uranoscopus dollfusi (non Brüss) Lieske & Myers 2004: 176, lower fig. (El Quadim Bay, Egypt). Holotype: BPBM 20844, female, 203 mm SL, Red Sea, Gulf of Aqaba, off Coral World pier and undersea observatory, Eilat, Israel, sand patch in reef with sparse sea grass, 15 m, quinaldine and hand net, J. E. Randall, 10 August Paratypes: USNM , 2: mm, Red Sea, Egypt, off ancient site about 5 km north of Koseir (26 8 N, E), in cove, edge of precipitous reef with numerous caves, 0-6 m, rotenone, H. Fehlmann et al., 8 Jan 1965; HUJ 7178, 168 mm, Red Sea, Gulf of Aqaba, Eilat, M. Dor, 23 June 1965; SMF 14284, 92 mm, Gulf of Aqaba, Eilat, W. Poper, 7 Jul 1969; USNM , 110 mm, Red Sea, Gulf of Aqaba, Israel, Eilat, V. G. Springer; 8 Jul 1969; USNM , 211 mm, Red Sea, Gulf of Aqaba, bay at El Hamira, rock and coral ledge, grading into sand, m, rotenone, V. G. Springer et al., 9 Sep 1969; HUJ 8534, 216 mm, Egypt, Gulf of Aqaba, east coast of Sinai Peninsula, El Hamira, M. Tsurnamal, 4 Mar 1975; USNM , mm, female, Red Sea, Gulf of Aqaba, Jordan, Aqaba, Marine Science Station, N, E, T. Paulus, May 1990; SMF 24189, mm, Red Sea, Gulf of Aqaba, Jordan, Aqaba, Marine Science Station, N, , E. T. Paulus, May 1990; UW , mm, same data as preceding. Diagnosis: Dorsal rays IV + 13; pectoral rays 18; oblique scale rows on midside of body 37-41; no scales on body above lateral line anterior to second dorsal fin; body depth of adult at anal-fin origin in SL; head broad and flat dorsally, the width in SL; bones dorsally on head rugose; occipital lobes not developed as knobs; cleithral spine large, slightly angled dorsally; in head length; supracleithrum not forming a distinct spine posteriorly; 5-7 small spines ventrally on preopercle; 3-5 blunt preorbital spines; eyes small, the orbit diameter in head length; lingual lure broadly triangular and flat at base, with long cirri on each side, narrowing to a slender ribbon-like filament with short cirri on edges, its length half the distance separating cleithral spines; dorsal fins narrowly separated; margin of second dorsal fin straight; caudal fin slightly rounded; color in life light gray-brown, grading to white ventrally, the dorsal half of body with irregular longitudinal rows of dark brown spots of about eye size, many rosette-like, and scattered dark brown dots, especially posteriorly; dark dots dorsally on head; first dorsal fin largely black; second dorsal and anal fins with translucent membranes, the dorsal rays and posterior anal rays dotted with dark brown, the margin white; caudal fin dark gray with black dots on rays, the posterior margin white; largest specimen 216 mm SL. Fig. 4. Holotype of Uranoscopus rosette in life, Gulf of Aqaba, Red Sea, 15 m. Photo by J. E. Randall. aqua vol. 18 no October

35 John E. Randall and Rachel J. Arnold Description: Dorsal rays IV + 13; anal rays 13; principal caudal rays 12, the median 10 branched; upper and lower procurrent caudal rays 5, the posterior 2 segmented; pectoral rays 18; pelvic rays I,5; scales cycloid, embedded, nonimbricate, in oblique rows, those in longitudinal series on midside of body 40 (37-41); no scales on body above lateral line anterior to second dorsal fin; lateral-line pores of subadult 49 to base of caudal fin (scales obscure in adults); preopercular spines 6 (5-7); preorbital spines blunt, 3 (3-5); branchiostegal rays 6; vertebrae 25 (Fig. 9). Body subquadrangular, strongly tapering, the depth at origin of anal fin 4.05 ( ) in SL; caudal-peduncle depth 8.7 ( ) in SL; head very flat dorsally and coarsely granular; occipital lobes not developed as knobs; supracleithrum Fig. 5. Holotype of Uranoscopus rosette when fresh, BPBM 20844, 203 mm. Photo by J. E. Randall. Fig. 6. Holotype of Uranoscopus rosette in alcohol, dorsal view. Photo by H. A. Randall. 213 without a distinct posterior spine; head length 2.65 ( ) in SL; head wider than body, the width 2.65 ( ) in SL; snout length 5.75 ( ) in head length; eye small, the orbit diameter 7.1 ( ) in head length; no cirrus on eye; interorbital width 4.85 ( ) in head length; maximum interorbital depth 3.3 ( ) in head length; length of interorbital fossa 5.2 ( ) in head length; distance between gill openings 4.0 ( ) in SL; distance between basipterygial processes 4.45 ( ) in head length; cleithral spine large, nearly covered by skin, angling dorsally about 10, the length 4.55 ( ) in head length; dorsal margin of opercle ascending, then broadly rounded posteriorly; a distinct respiratory groove between dorsal margin of opercle and ventral margin of supracleithrum and base of cleiaqua vol. 18 no October 2012

36 Uranoscopus rosette, a new species of stargazer (Uranoscopidae: Trachinoidei) from the Red Sea thral spine; opercular membrane broad (as wide as orbit diameter in holotype), with a conspicuous fringe of cirri; lateral line commencing above suprascapular spine, passing to below origin of soft dorsal fin, continuing along base of fin, curving ventrally on caudal peduncle to midbase of caudal fin, and continuing to posterior end of fin. Mouth forming an angle of about 100 to horizontal axis of body, the opening strongly curved when viewed dorsally; upper-jaw length 2.55 ( ) in head length; width of maxilla equal to orbit diameter in holotype; width of mouth 2.2 ( ) in head length; margin of lips with well-formed cirri, 15 in upper lip and 21 in lower lip of holotype; teeth in jaws within a zone of fleshy papillae, small, conical, and recurved; upper jaw with a maximum of six rows medially, narrowing to one or two laterally; teeth in lower jaw about twice as large, in two rows medially, narrowing to one laterally, about 20 in outer row of holotype; villiform teeth in a broad patch on palatines; tongue broadly rounded; lingual lure damaged in holotype, described here from paratypes: very simi - lar to Figure 3c of Kishimoto in Carpenter & Niem (2001: 3521), flat and broadly triangular at base, about as long as width of mouth, with a fringe of long cirri along each edge; cirri progressively longer basally, the longest about equal to Fig. 7. Smallest paratype of Uranoscopus rosette, SMF 14284, 92 mm, Gulf of Aqaba. Photo by S. Tränkner. aqua vol. 18 no October

37 John E. Randall and Rachel J. Arnold Table I. Proportional measurements of type specimens of Uranoscopus rosette as percentages of the standard length. Holotype Paratypes BPBM SMF HUJ UW USNM HUJ Standard Length (mm) Sex female male female male female female Body depth Caudal-peduncle depth Head length Head width Snout length Orbit diameter Interorbital width Least depth of infraorbitals Greatest depth of infraorbitals Length of interorbital fossa Width of interorbital fossa Distance between gill openings Between basipterygial processes Upper-jaw length Width of mouth Cleithral-spine length Predorsal length Base of first dorsal fin Longest dorsal spine Longest dorsal ray Longest anal ray Caudal-fin length Pectoral-fin length Pelvic-fin length interorbital width; distal part of lure a slender ribbon-like filament with short cirri on edges, its length one-half distance separating cleithral spines; anterior nostril a small tapering conical tubule; posterior nostril a longitudinal slit in a slightly elevated base. Dorsal fins narrowly separated; origin of first dorsal fin above tip of cleithral spine; predorsal length 2.5 ( ) in SL; first dorsal spine longest, 4.25 ( ) in head length; margin of second dorsal fin posterior to third ray straight; third dorsal soft ray longest, 2.2 ( ) in head length; twelfth anal soft ray longest, 2.5 ( ) in head length; caudal fin slightly rounded, 3.3 ( ) in SL; posterior margin of pectoral fins strongly angular, the seventh or eighth pectoral ray longest, reaching posterior to a vertical at base of third anal ray, 3.5 ( ) in SL; dorsal margin of pectoral fin straight; fifth pelvic ray longest, 1.95 ( ) in head length. 215 Color of holotype in alcohol pale yellowish brown dorsally, grading to yellowish white ventrally, with about three irregular longitudinal rows of rosette-like brown spots on about dorsal half of body; dorsal part of head to interorbitals finely dotted with dark brown; first dorsal fin black except for a small pale triangular part anteriorly at base and a small pale spot at rear base of fin; second dorsal fin with translucent membranes and light brown rays, dotted with brown, pale at tips of last ten rays; anal fin similar, but only last four rays with a few small dark spots; caudal fin with translucent membranes and brown rays finely spotted with dark brown, the ray tips paler; pectoral fins with translucent membranes and brown rays, paler at tips; pelvic fins pale yellowish brown. Color in life of holotype (Fig. 4); color when fresh (Fig. 5); color in dorsal view in alcohol (Fig. 6). Figure 7 provides three views of the smallest paratype in alcohol; the oblique white area doraqua vol. 18 no October 2012

38 Uranoscopus rosette, a new species of stargazer (Uranoscopidae: Trachinoidei) from the Red Sea sal to the tip of the pectoral fin is from damage to the specimen. The underwater photograph of Fig. 8 is provided to show color variation. Etymology: The species name rosette is in reference to the diagnostic circular clusters of spots, many rose-like, dorsally on the body. Remarks: Uranoscopus rosette has long been misidentified as Uranoscopus fuscomaculatus Kner, type locality, Fiji. Photographs of Kner s holotype of U. fuscomaculatus, NMW (Figs 1 & 2) reveal a more elongate fish, the body depth at analfin origin about 4.5 in SL, compared to in U. rosette. Also it is not as broad as U. rosette, the head width about 3 in SL, compared to Other noteworthy differences from U. rosette are the long slender lingual filament without a fringe of cirri on the edges, the horizontal cleithral spine (axis of spine of U. rosette angles dorsally), the colofr pattern of straight rows of dark spots dorsally on the body, compared to clusters of dark spots or irregular circles in approximate rows, and the caudal fin lacking a white posterior margin. As noted above, Uranoscopus fuscomaculatus is not a senior synonym of U. sulphureus Valenciennes, nor of U. rosette. The holotype of U. sulphureus is badly damaged, but the photographs (Fig. 3) shows very large eyes, narrow interorbital space, and a distinctive U-shaped interorbital fossa extending to the posterior margin of the eyes that clearly differentiate the species. Underwater photographs of Uranoscopus rosette have also been misidentified as U. dollfusi Brüss, a species previously reported from mud or silty sand substrata in the Gulf of Suez, Gulf of Oman, and the Persian Gulf (Randall, 1995). It is readily distinguished by lacking dark spots dorsally on the body, having a white band along the base of the black first dorsal fin, suprascapular spines, and a slender, blade-like, tapering, lingual lure, without cirri on the edges. Fig. 8. Uranoscopus rosette, Hurghada, Egypt, 12 m. Photo by A. Golubev. aqua vol. 18 no October

39 John E. Randall and Rachel J. Arnold The holotype of Uranoscopus rosette was discovered buried in a sand patch within a reef area at a depth of 15 m, with just the eyes and mouth visible. It was anesthetized, carefully dug from the sediment to avoid the cleithral spine, and photographed A second specimen of Uranoscopus (BPBM 41103, 213 mm SL) was collected at the same time and place as the holotype of U. rosette. It differs in color pattern in having about the upper half of the body speckled with small dark brown spots (none larger than 2 mm). The counts and measurements are remarkably like those of U. rosette, and we identify it as this species, but we do not make it a paratype. ACKNOWLEDGEMENTS We thank Dr. Tilman J. Alpermann of the Senckenberg Museum in Frankfurt, Dr. Daniel Golani of the Hebrew University in Jerusalem, and Shirleen Smith of the U.S. National Museum of Natural History for the loan of specimens; Sandra Raredon of the USNM for x-rays of Uranoscopus rosette; Romain Causse of the Muséum National d Histoire Naturelle in Paris for the photograph of the holotype of U. sulphureus; Sven Tränkner of the Senckenberg Museum for the photograph of the smallest paratype of U. rosette; Helmut Wellendorf of the Naturhistorisches Museum in Vienna for the photographs of the holotype of U. fuscomaculatus; and Andrey Golubev for the undersea photograph of U. rosette in Egypt. We are also grateful to Loreen R. O Hara and Arnold Y. Suzumoto of the Bishop Museum for curatorial assistance, and Helen A. Randall for the dorsal-view photograph of the holotype of U. rosette. Sergey V. Bogorodsky, Theodore W. Pietsch, and Helen A. Randall reviewed the manuscript. Fig. 9. X-ray of paratype of Uranoscopus rosette, USNM , 168 mm, N of Koseir, Egypt. Photo by S. Raredon. 217 aqua vol. 18 no October 2012

40 Uranoscopus rosette, a new species of stargazer (Uranoscopidae: Trachinoidei) from the Red Sea REFERENCES BRÜSS, R Two new species of Uranoscopus Linnaeus, 1758, from the Red Sea: U. dollfusi n. sp. and U. bauchotae n. sp. Bulletin Muséum. National d Histoire. Naturelle ser. 4, sect. A, 8 (4): BRÜSS, R. 1987a. Tiefenwasser- und Tiefseefische aus dem Roten Meer. XIII. Uranoscopus marisrubri n. sp. aus dem zentralen und nördlichen Roten Meer (Pisces: Perciformes: Uranoscopidae). Senckenbergiana Biologica 68 (1/3): BRÜSS, R. 1987b. Uranoscopus dahlakensis n. sp. und Neunachweis von U. scaber Linnaeus 1758 im Roten Meer (Pisces: Perciformes: Uranoscopidae). Senckenbergiana Biologica 68 (1/3): CARPENTER, K. E. & NIEM, V. H The Living Marine Resources of the Western Central Pacific, vol. 6: v pp. Food and Agriculture Organization of the United Nations, Rome. CUVIER, G. & VALENCIENNES, A Histoire Naturelle de Poissons, vol. 8. F. G. Levrault, Paris, xix pp. DEBELIUS, H Red Sea Reef Guide. IKAN, Frankfurt. 321 pp. ESCHMEYER, W. N. & COLLABORATORS Catalog of Fishes. html. DOR, M Nouveaux poissons pour la faune de la Mer Rouge. Bulletin Sea Fisheries Research Station Haifa, Israel 54: DOR, M CLOFRES: Checklist of the fishes of the Red Sea. Israel Academy of Sciences and Humanities, Jerusalem. xxii pp GOLANI, D. & BOGORODSKY, S. V Fishes of the Red Sea reappraisal and updated checklist. Zootaxa 2463: GOREN, M. & DOR, M An Updated Checklist of the Fishes of the Red Sea CLOFRES II. Israel Academy of Sciences and Humanities, Jerusalem and Interuniversity Institute for Marine Sciences, Elat, xii pp. KHALAF M. A. & DISI, A. M Fishes of the Gulf of Aqaba. Marine Science Station, Aqaba, Jordan, 252 pp. KISHIMOTO, H Redescription and lectotype designation of the stargazer, Uranoscopus kaianus Günther. Copeia 1984 (4): KNER, R Ueber neue Fische aus dem Museum der Herren Johann Cäsar Goeffroy & Sohn in Hamburg. Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften Mathematisch-Naturwissenschaftliche. 58 (1-2): LIESKE, E. & MYERS, R. F Coral Reef Guide Red Sea. Harper Collins Publishers Ltd., London, 384 pp. PIETSCH, T. W Phylogenetic relationships of trachinoid fishes of the family Uranoscopidae. Copeia 1989 (2): RANDALL, J. E., ALLEN, G. R. & STEENE, R. C Fishes of the Great Barrier Reef and Coral Sea. Crawford House Press, Bathurst, New South Wales, 507 pp. RANDALL, J. E Coastal Fishes of Oman. University of Hawai i Press, Honolulu, xiii pp. aqua vol. 18 no October

41 aqua, International Journal of Ichthyology On the validity of the cirrhitid fish genus Itycirrhitus Michelle R. Gaither 1-2 and John E. Randall 3 1) California Academy of Sciences, 55 Music Concourse Dr., San Francisco, CA 94118, USA. michellergaither@gmail.com 2) Hawai i Institute of Marine Biology, P.O. Box 1346, Kaneohe, HI 96744, USA. 3) Bishop Museum, 1525 Bernice St., Honolulu, HI , USA. jackr@hawaii.rr.com. Received: 6 June 2012 Accepted: 17 September 2012 Abstract The hawkfish Cirrhitus wilhelmi Lavenberg & Yañez was described from Easter Island in 1972, reclassified in Amblycirrhitus by Pequeño (1989), and the range extended to the Pitcairn Islands by Randall (1999). Randall (2001) described the new genus Itycirrhitus for this species. The similarity in general morphology and color to the Hawaiian Cirrhitops fasciatus and C. mascarenensis, type locality Mauritius, prompted the present molecular study to determine the validity of Itycirrhitus. Combining mitochondrial cytochrome oxidase I (COI) and cytochrome b (cyt b) sequence data from specimens of I. wilhelmi with comparable data from the two species of Cirrhitops, we found high levels of divergence between the two genera (COI, d = 15%; cyt b, d = 20-21%) that are similar to comparisons between Itycirrhitus and other genera of the family (COI, d = 8-18%; cyt b, d = 12-20%). This level of divergence is typical of other generic comparisons within Cirrhitidae (COI, d = 15-20%; cyt b, d = 15-21%). Zusammenfassung Der Büschelbarsch (Korallenwächter) Cirrhitus wilhelmi Lavenberg & Yañez wurde 1972 von den Osterinseln beschrieben und später von Pequeño (1989) in die Gattung Amblycirrhitus eingeordnet, hinzu kam eine Erweiterung des Verbreitungsgebietes bis zu den Pitcairn-Inseln von Randall (1999). Bald danach wurde von Randall (2001) für diese Art die neue Gattung Itycirrhitus beschrieben. Die Ähnlichkeit nach Färbung und allgemeiner Morphologie zu den hawaiianischen Cirrhitops fasciatus und C. mascarenensis, Typuslokalität Mauritius, gab den Anlass zu der vorliegenden molekularen Studie mit dem Ziel, die Gültigkeit von Itycirrhitus zu prüfen. Auf der Grundlage der Sequenzdaten der mitochondrialen Cytochromoxidase I (COI) sowie des Cytochrom b (cyt b) wurden Exemplare von I. wilhelmi mit solchen der beiden Cirrhitops-Arten verglichen; es zeigte sich ein hoher Grad an Divergenz zwischen den beiden Gattungen (COI, d = 15%; cyt b, d = 20-21%), ähnliche Werte wie beim Vergleich zwischen Itycirrhitus und anderen Gattungen der Familie (COI, d = 8-18%; cyt b, d = 12-20%). Dieser Grad der Divergenz ist typisch für andere exemplarische Vergleiche innerhalb der Cirrhitidae (COI, d = 15-20%; cyt b, d = 15-21%). Résumé Le poisson-faucon Cirrhitus wilhelmi Ravenberg & Yañez a été décrit en provenance de l île de Pâques, en 1972, et reclassé comme Amblycirrhitus par Pequeño (1989), et la distribution étendue jusqu aux îles Pitcairn par Randall (1999). Randall (2001) a décrit le nouveau genre Itycirrhitus pour cette espèce. L analogie en morphologie générale et en couleur avec Cirrhitops fasciatus d Hawaï et C. mascarenensis, localité-type l île Maurice, a déterminé la présente étude moléculaire pour tester la validité d Itycirrhitus. En combinant les données des séquences de la cytochrome oxydase I (COI) et du cytochrome b (cyt b) de spécimens de I. wilhelmi avec des données comparables de deux espèces de Cirrhitops, nous avons trouvé des niveaux élevés de divergences entre les deux genres (COI, d = 15%; cyt b, d = 20-21%) similaires à des comparaisons entre Itycirrhitus et d autres genres de la famille (COI, d = 8-18%; cyt b, d = 12-0%). Ce niveau de divergence est typique d autres comparaisons génétiques parmi les Cirrhitidae (COI, d = 15-20%; cyt b, d = 15-21%). Sommario Il pesce falco Cirrhitus wilhelmi Lavenberg & Yañez è stato descritto dall Isola di Pasqua nel 1972, riclassificato in Amblycirrhitus da Pequeño (1989) e la sua area di presenza estesa alle isole Pitcairn da Randall (1999). Randall (2001) ha descritto il nuovo genere Itycirrhitus per questa specie. La somiglianza nella morfologia generale e nel colo - re alla specie hawaiana Cirrhitops fasciatus e a C. mascarenensis, località tipo Mauritius, ha incoraggiato il presente studio molecolare per determinare la validità di Itycirrhitus. Combinando i dati di sequenza della citocromo ossidasi I (COI) e del citocromo b (cit b) mitocondriali da campioni di I. wilhelmi con dati comparabili delle due specie di 219 aqua vol. 18 no October 2012

42 On the validity of the cirrhitid fish genus Itycirrhitus Cirrhitops, abbiamo trovato alti livelli di divergenza tra i due generi (COI, d = 15%, cit b, d = 20-21%), simili a confronti tra Itycirrhitus e altri generi della famiglia (COI, d = 8-18%, cit b, d = 12-20%). Questo livello di divergenza è tipica di altri confronti generici all'interno Cirrhitidae (COI, d = 15-20%; cyt b, d = 15-21%). INTRODUCTION The 33 species of fishes of the perciform family Cirrhitidae, popularly known as hawkfishes, are found in the tropical and subtropical Indo-Pacific region, except for three in the Atlantic and three in the eastern Pacific (two of which, Cirrhitichthys oxycephalus and Oxycirrhites typus, are range extensions from the Indo-Pacific). Cirrhitid fishes all have X dorsal spines, III anal spines, 14 pectoral rays (the lower five to seven rays unbranched and thickened), two flat opercular spines, a serrate preopercle, cycloid scales, one to several cirri at the tip of each membrane of the dorsal spines, and no swim bladder. Most species occur in shallow water on coral reefs or rocky substrata, often in areas exposed to wave action. When in a surge zone, they use their thickened lower pectoral rays to wedge themselves in cracks in the reef or within branches of coral. Hawkfishes feed mainly on benthic crustaceans and occasionally on small fishes. Exceptions are Cyprinocirrhites polyactis that feeds well above the substratum on zooplankton, and O. typus that often makes short forays from the bottom to prey on the larger animals of the demersal plankton. At least some of the species of the family are protogynous hermaphrodites (Sadovy & Donaldson 1995). The generic classification of the Cirrhitidae has a long and confused history, Günther (1860) recognized eight genera in the family. A century and 11 cirrhitid publications later, Schultz in Schultz & collaborators (1960) listed 13 genera, including Isobuna and Serranocirrhitus, now known to be serranids. Randall (1963) included 10 genera and 34 species in his revision of the family. Randall (2001) again revised the genera of cirrhitids, adding three new monotypic genera from species formerly classified in Cirrhitus. One of these, Itycirrhitus, was described for a small species, Cirrhitus wilhelmi, from Easter Island. This species was first described by Lavenberg & Yañez (1972) and later reclassified in Amblycirrhitus by Pequeño (1989). Randall (1999) extended the range to the Pitcairn Islands. The key to genera of Randall (2001) is reproduced here. Key to the Genera of Cirrhitidae 1a. Snout not elongate, its length about in head length; body not slender, the depth in SL; canine teeth in jaws markedly longer than inner villiform teeth, those at front of upper jaw and side of lower jaw enlarged b. Snout elongate, its length in head length; body slender, the depth in SL; canine teeth in jaws only slightly longer than inner villiform teeth and nearly uniform in size... Oxycirrhites 2a. Caudal fin rounded, truncate, or slightly emarginate; dorsal soft rays 11-15; snout not short, its length SL b. Caudal fin lunate; dorsal soft rays 16 or 17; snout short, its length in head length... Cyprinocirrhites 3a. Small scales on cheek in more than 12 rows a. Rows of large scales on cheek 4-6 (small scales also usually present) a. Lower 7 pectoral rays unbranched and thickened; first 2 supraneural bones in space before second neural spine; more than 40 cirri in 2 series on posterior flap of anterior nostril b. Lower 6 pectoral rays unbranched and thickened; first 3 supraneural bones in space before second neuralspine; fewer than 15 cirri on posterior flap of anterior nostril a. Supraorbital ridge high, continuing more than half eye diameter posterior to orbit; lower opercular spine acute, forming an angle of 45 or less; no scales in interorbital space; pectoral fins reaching slightly beyond a vertical at tips of pelvic fins; body depth in SL... Cristacirrhitus 5b. Supraorbital ridge low and not continuing posterior to eye; lower opercular spine forming an angle of 90 ; a V-shaped band of scales in posterior half of interorbital space; pectoral fins short, not reaching a vertical at tips of pelvic fins; body depth in SL Cirrhitus 6a. Dorsal soft rays 13; three-fourths or more of preopercular margin coarsely serrate; palatine teeth absent; body very deep, the depth in SL, and very compressed, the width in depth; longest pectoral rays not extending beyond a vertical at pelvic-fin tips.... Neocirrhites aqua vol. 18 no October

43 Michelle R. Gaither and John E. Randall 6b. Dorsal soft rays 11-12; about upper half of preopercular margin finely or coarsely serrate; palatine teeth present; body not very deep, the depth in SL, and not very compressed, the width in depth; longest pectoral rays extending beyond a vertical at pelvic-fin tips a. Upper margin of preopercle finely serrate (25 or more serrae); exposed end of posttemporal finely serrate; first (most medial) branchiostegal ray strongly curved and nearly parallel with second ray; gill membrane across throat naked; scales on cheek separated from serrate edge of preopercle by a broad naked zone crossed by irregular sensory channels (may show as ridges); no scales on snout; scales ventrally on chest extremely small; 3 rows of large scales above lateral line in middle of body; dorsal soft rays Notocirrhitus 7b. Upper margin of preopercle coarsely serrate (fewer than 13 serrae); exposed end of posttemporal with 3-5 serrae; first (most medial) branchiostegal ray nearly straight and not parallel to second ray; gill membrane across throat scaled; scales on cheek extending to base of preopercular serrae; scales on snout extending to below anterior nostrils; scales ventrally on chest one-half or more size of scales on side of body; 4 rows of large scales above lateral line in middle of body; dorsal soft rays (rarely 12 or 14) Itycirrhitus 8a. Rows of large scales above lateral line to base of spinous portion of dorsal fin 5; a single cirrus from membrane near tip of each spine of dorsal fin; membranes between longest dorsal spines incised at most one-fifth spine length; palatine teeth absent... Paracirrhites 8b. Rows of large scales above lateral line to base of spinous portion of dorsal fin 3 or 4; a tuft of cirri from membrane near tip of each spine of dorsal fin; membranes between longest dorsal spines incised one-third or more of spine length; palatine teeth present or absent a. Palatine teeth absent; maxilla reaching to or beyond a vertical through middle of eye; dorsal spines short, the longest in head length; dorsal profile of head convex; lower 5 pectoral rays unbranched; interorbital fully scaled; snout not pointed; preorbital without a free posterior margin... Isocirrhitus 9b. Palatine teeth present; maxilla not reaching a vertical through middle of eye; dorsal spines in head length; dorsal profile of head straight to slightly convex; lower 5 to 7 pectoral rays unbranched; interorbital scaled or naked; snout pointed or not pointed; preorbital with or without a free posterior margin a. Dorsal soft rays (rarely 15); first 2 pectoral rays unbranched; snout not pointed, the dorsal profile from interorbital to upper lip convex; interorbital naked... Cirrhitops 10b. Dorsal rays (rarely 13); first pectoral ray unbranched, second branched; snout pointed, the dorsal profile straight; interorbital scaled or naked a. Preopercular margin finely serrate; first 2 supraneural bones in space before second neural spine; preorbital without a free hind margin; interorbital scaled; first dorsal soft ray not produced into a filament; lower 5 (rarely 6) pectoral rays unbranched Amblycirrhitus 11b. Preopercular margin coarsely serrate; all 3 supraneural bones in space before second neural spine; preorbital with a free hind margin for one-fourth to one-half distance from lower edge to eye; interorbital not scaled; first dorsal soft ray usually produced into a filament; lower 6 or 7 pectoral rays unbranched... Cirrhitichthys After the separation of Oxycirrhites in the key, mainly by its snout length, and Cyprinocirrhites by its lunate caudal fin, the characters leading to the other genera are less incisive, in particular those leading to Itycirrhitus. This study was initiated to determine if an analysis of DNA sequence data from species of cirrhitid fishes will support the generic classification. Specifically we collected mitochondrial sequence data to compare Itycirrhitus wilhelmi (Fig. 1) with the similarly colored species of Cirrhitops, C. fasciatus (Fig. 2) and C. mascarenensis. Additionally, we investigated the level of genetic divergence between the two monotypic genera Notocirrhitus and Neocirrhites which are closely grouped in the key. 221 MATERIALS AND METHODS For genetic analysis a 1cm fin clip was obtained from three specimens of Itycirrhitus wilhelmi from Easter Island. DNA was isolated using the modiaqua vol. 18 no October 2012

44 On the validity of the cirrhitid fish genus Itycirrhitus Fig. 1. Itycirrhitus wilhelmi, Easter Island. Photo by J. E. Randall. Fig. 2. Cirrhitops fasciatus, Oah u, Hawaiian Islands. Photo by J. E. Randall. aqua vol. 18 no October

45 Michelle R. Gaither and John E. Randall Fig. 3. Maximum Likelihood (ML) phylogenetic tree of the cytochrome oxidase I gene for 10 species (7 genera) of hawkfish (Family Cirrhitidae). Numbers above branches reflect bootstrap support for ML tree and Bayesian posterior probabilities, respectively. Only major nodes with bootstrap support of >60 are labeled. 223 aqua vol. 18 no October 2012

46 On the validity of the cirrhitid fish genus Itycirrhitus Table I. Mitochondrial cytochrome oxidase I (COI) and cytochrome b (cyt b) sequences used in this study. Species name, Barcode of Life Data systems ( Ratnasingham & Hebert 2007) sequence identification numbers and GenBank accession numbers (in italics; and museum voucher IDs (in parentheses) are listed. Sequences generated in this study are in bold. Number of individuals (if >1) in which haplotype was detected is listed in parentheses. Species COI cyt b Amblycirrhitus pinos TZAIB (HLC-12147) A. bimacula JX EU684136, JX Cirrhitops fasciatus EU684132, EU (BPBM 40485, 40888) EU684137, EU (BPBM 40485, 40888) TZAIB (HLC-12324) GBGC , GBGC C. mascarenensis EU684134, EU (BPBM ) EU684139, EU (BPBM ) Cirrhitus pinnulatus JX645649, JX (2) JX645661, JX645662, JX Itycirrhitus wilhelmi JX645651, JX (2) JX645658, JX645659, JX Neocirrhites armatus TZAIC (HLC-10878) JX645664, JX645665, JX TZAIB (HLC-13128) TZAIC (HLC-11757) JX645653, JX645654, JX Notocirrhitus splendens Paracirrhites arcatus JX645656(2) (MA655017, MA655018) TZAIB (HLC-12066) TZAIB (HLC-12067) TZAIB (HLC-12068) TZAIB (HLC-12070) TZAIB (HLC-15197) P. forsteri DSFSG451-11(ADC11_214.7 #2) DSFSG566-11(ADC11_214.7 #4) Epinephelus lanceolatus HQ174825, HQ HQ174838, HQ Plectropomus leopardus GBGC , GBGC , FSCS AY963555, AY fied HotSHOT method (Meeker et al. 2007; Truett et al. 2000). We amplified approximately 660 bp the mitochondrial cytochrome c oxidase subunit I (COI) gene and approximately 750 base pairs of cytochrome b (cyt b) gene using the primers employed in Randall and Schultz (2009). Polymerase chain reactions (PCR) were carried out in a 20 µl volume containing 5-20 ng of template DNA, 0.4 µm of each primer, 10 µl of the premixed PCR solution BioMix Red (Bioline Inc., Springfield, NJ, USA), and deionized water to volume. After an initial 7 min denaturation at 95 C, each of 35 cycles consisted of denaturation for 30 s at 94 C, annealing at 56 C for COI and 58 C for cyt b for 30 s, and extension at 72 C for 45 s with a final 10 min extension at 72 C. Amplification products were purified using 0.75 units of Exonuclease I: 0.5 units of Shrimp Alkaline Phosphatase (ExoSAP, USB, Cleveland, OH, USA) per 7.5 µl PCR products at 37 C for 60 minutes, followed by deactivation at 80 C for 15 minutes. DNA sequencing was performed with fluorescentlyaqua vol. 18 no October

47 Michelle R. Gaither and John E. Randall labeled dideoxy terminators on an ABI 3730XL Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) at the University of Hawai i Advanced Studies of Genomics, Proteomics and Bioinformatics sequencing facility. Sequences for each locus were aligned, edited, and trimmed to a common length using the DNA sequence assembly and analysis software Geneious Pro 5.0 (Biomatters, LTD, Auckland, NZ). In all cases, alignment was unambiguous with no indels or frameshift mutations. Unique sequences were deposited in Gen- Bank ( see Table I for accession numbers). Maximum Likelihood (ML) trees were constructed using the default settings implemented in the program MEGA 5.05 (Tamura et al. 2007). COI and cyt b sequences of Cirrhitops fasciatus (N = 5, Hawai i) and Cirrhitops mascarenensis (N = 2, Mauritius), were obtained from GenBank (Randall & Schultz 2009, Table I). For inclusion in the phylogenetic tree COI sequences from representatives of three genera in the family Cirrhitidae were downloaded from the Barcode of Life Data (BOLD) Systems 2.5 website ( (Table I). Corresponding cyt b sequences were available from GenBank for only Amblycirrhitus bimacula. Additionally, tissue samples of Amblycirrhitus bimacula (N = 1, Hawai i), Cirrhitus pinnulatus (N = 3, Guam), Neocirrhites armatus (N = 3, Guam), and Notocirrhitus splendens (N = 2, Kermadec Islands, New Zealand) were obtained and sequenced at both loci and deposited in GenBank (Table I). Available voucher specimens and tissues (Table I) are deposited in the: Biodiversity Institute of Ontario (HLC), Guelph, Canada; Bernice P. Bishop Museum (BPBM), Honolulu, USA; South African Institute of Aquatic Biodiversity (ADC), Grahamstown, South Africa; and the Auckland Museum (MA), Auckland, New Zealand. Phylogenetic trees were rooted with two species from the family Serranidae (Table I, Epinephelus lanceolatus and Plectropomus leopardus). For analysis, all COI and cyt b sequences were trimmed to 557 bp and 671 bp, respectively. Bootstrap support values were calculated using default settings with 1000 replicates. For comparison we ran a Bayesian Markov Chain Monte Carlo (MCMC) analysis as implemented in the program MRBAYES (Huelsenbeck & Ronquist 2001). We employed default settings and ran simulations for 1,000,000 generations until the standard deviation of split frequencies were below 0.01 as recommended in the MRBAYES manual. Average percent divergence (d) between genera was calculated in ARLEQUIN 3.5 (Excoffier & Lischer 2010). RESULTS Our comparisons of Itycirrhitus wilhelmi and the two species of the genus Cirrhitops (C. fasciatus and C. mascarenensis) at two mitochondrial genes confirmed the distinction of these two genera. Across all COI sequences analyzed we detected 197 variable nucleotide sites. There were two haplotypes in three I. wilhelmi specimens. The average pairwise difference between C. fasciatus and C. mascarenensis was 39 bp (d = 7%), while I. wilhelmi differs from the two species of Cirrhitops by an average of 85 bp (d = 15%). All generic level comparisons resulted in average base pair differences that range between 82 bp (d = 15%; Cirrhitops vs. Neocirrhites) and 113 (d = 20%; Neocirrhites vs. Amblycirrhitus). The exception is I. wilhelmi and Neocirrhites armatus which differ by an average of only 47 bp (d = 8%). Analyses of cyt b sequences were similarly robust. We detected 256 variable nucleotide sites. There were three haplotypes in three I. wilhelmi specimens. The average pairwise difference between C. fasciatus and C. mascarenensis was 75 bp (d = 11%). In contrast, the average pairwise differences between I. wilhelmi and the two species of Cirrhitops were 140 and 132 respectively (d = 21% and 20%, respectively). Generic level comparisons resulted in average base pair differences that ranged between 84 bp (d = 13%; Neocirrhites vs. Itycirrhitus) and 138 bp (d = 21%; Amblycirrhitus vs. Cirrhitops). After repeated attempts, we were able to amplify the Notocirrhitus splendens specimens at COI but not cyt b. When comparing N. splendens with Neocirrhites armatus, which are closely grouped in the key, we found an average pairwise difference between these monotypic genera of 84 bp (d = 15%). A similar level of divergence was detected between N. splendens and I. wilhelmi (87 bp, d = 16%), which are also closely grouped in the key. Phylogenetic analyses revealed a single tree topology for both mitochondrial markers with only the relative positions of the outgroups varying (Fig. 3). Both Maximum Likelihood and Bayesian analyses indicated that N. armatus is more closed related to I. wilhelmi than I. wilhelmi is to the members of the genus Cirrhitops. 225 aqua vol. 18 no October 2012

48 On the validity of the cirrhitid fish genus Itycirrhitus DISCUSSION Phylogenetic analyses indicate strong concordance between our molecular data and generic level classifications within the family Cirrhitidae (see key above). Species within the same genus always grouped together with intragenetic divergences ranging from 7-14% at COI. The genus Itycirrhitus was described by Randall (2001) based on seemingly small morphological differences. Here we report sequence data that support the generic level designation for Itycirrhitus wilhelmi. This species is 15% divergent at COI from either of the species of the genus Cirrhitops; a level of divergence common among the other genera of the family (8 to 20%). Our finding of a similar level of divergence between Neocirrhites armatus and Notocirrhitus splendens supports the designation of these monotypic genera. Interestingly, I. wilhelmi did not cluster in the phylogenetic tree with the morphologically similar Cirrhitops (Figs 1-2). Instead, despite significant morphological differences, Neocirrhites armatus was found to be only 8% divergent from I. wilhelmi, a similar level of divergence between the two species of Cirrhitops (7%). In this case we found surprising incongruence between the morphological characters used to delineate these species and the level of molecular divergence detected. ACKNOWLEDGEMENTS We thank the following for providing tissue samples for this study: Dr. Gerald R. Allen, Dr. Alfredo Cea, Dr. Mark V. Erdmann, Daniel Pelicier, and Dr. Gordon W. Tribble. Support for the first author was provided by a grant/cooperative agreement from the National Oceanic and Atmospheric Administration, Project R/HE-1, which is sponsored by the University of Hawaii Sea Grant College Program, SOEST, under Institutional Grant No. NA09OAR The views expressed herein are those of the authors and do not necessarily reflect the views of NOAA or any of its subagencies. UNIHI-SEAGRANT-JC This is contribution #1516 from the Hawai i Institute of Marine Biology and #8739 from the School of Ocean and Earth Science and Technology. REFERENCES EXCOFFIER, L. & LISCHER, H. E. L ARLEQUIN suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources 10: GÜNTHER, A Catalogue of the Acanthopterygian Fishes in the Collection of the British Museum. British Museum, London, 548 pp. HUELSENBECK, J. P. & RONQUIST, F MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17: LAVENBERG, R. J. & YAÑEZ L. A A new species of Cirrhitus from Easter Island (Pisces, Cirrhitidae), Gayana, Zoologia, 21: MEEKER, N. D., HUTCHINSON, S. A., HO, L., & TREDE, N. S Method for isolation of PCR-ready genomic DNA from zebrafish tissues. BioTechniques 43: PEQUEÑO, G Peces de Chile. Lista sistemática revisada y comentada. Revista de Biología Marina, Valparaíso 24: RANDALL, J. E Review of the hawkfishes (family Cirrhitidae). Proceedings of the United States National Museum 114: RANDALL, J. E Report on fish collections from the Pitcairn Islands. Atoll Research Bulletin, 461, 53 pp. RANDALL, J. E Revision of the generic classification of the hawkfishes (Cirrhitidae), with descriptions of three new genera. Zootaxa 12: RANDALL, J. E. & SCHULTZ, J. K Cirrhitops mascarenensis, a new species of hawkfish from the Mascarene Islands, southwestern Indian Ocean. Smithiana Bulletin 9: RATNASINGHAM, S. & HEBERT, P. D. N BOLD: The Barcode of Life Data System ( Molecular Ecology Notes 7: SADOVY, Y. & DONALDSON, T. J Sexual pattern of Neocirrhites armatus (Cirrhitidae) with notes on other hawkfish species. Environmental Biology of Fishes 42: SCHULTZ, L. P., CHAPMAN, W. M., LACHNER, E. A. & WOODS, L. P Fishes of the Marshall and Marianas Islands. Smithsonian Institution United States National Museum Bulletin, 202 (2), 438pp. TAMURA, K., DUDLEY, J., NEI, M. & KAMAR, S MEGA4: molecular evolutionary genetics analysis (MEGA) software version 4.0. Molecular Biology and Evolution 24: TRUETT, G. E., MYNATT, R. L., TRUETT, A. A., WALKER, J. A. & WARMAN, M. L Preparation of PCR-quality mouse genomic DNA with hot sodium hydroxide and Tris (HotSHOT). BioTechniques 29: aqua vol. 18 no October

49 aqua, International Journal of Ichthyology A new species of Rainbowfish (Chilatherina: Melanotaeniidae), from the Sepik River System of Papua New Guinea Gerald R. Allen 1 and Peter J. Unmack 2 1) Western Australian Museum, Locked Bag 49, Welshpool DC, Perth, Western Australia tropical_reef@bigpond.com 2) Department of Biology, WIDB 401, Brigham Young University, Provo UT 84602, U.S.A. Received: 9 July 2012 Accepted: 3 September 2012 Abstract A new species of rainbowfish, Chilatherina pagwiensis, is described on the basis of 31 specimens, mm SL, collected from a tributary of the Sepik River near Pagwi village, East Sepik Province, Papua New Guinea. Morphologically it is closely allied to C. campsi, an inhabitant of mainly hilly or mountainous terrain in the Markham, Ramu, and Sepik river systems of northern Papua New Guinea and the Purari system, south of the Central Dividing Range. Both species are characterised by the absence of vomerine and palatine teeth, and males and females are relatively slender (greatest body depth % SL) compared to most other family members. However, the two species exhibit pronounced modal differences in soft dorsal and anal rays. Chilatherina pagwiensis usually has soft dorsal rays compared to usual counts of in C. campsi. Likewise, the former species most frequently has soft anal rays compared to (most frequently 22-23) in C. campsi. Analysis of genetic relationships based on mitochondrial cytochrome b sequences indicates a close relationship between C. pagwiensis and the ancestor to C. axelrodi and C. campsi. The mean p-distance of C. pagwiensis to its relatives was 2.8 and 3.8% respectively, which are similar to values recorded between other rainbowfish sister species. Zusammenfassung Beschrieben wird eine neue Art der Regenbogenfische: Chilatherina pagwiensis, auf der Grundlage von 31 Exemplaren mit 20,5-76,1 mm SL, die in einem Nebenfluss des Sepik River nahe dem Dorf Pagwi, Provinz Ost-Sepik, Papua- Neuguinea, gesammelt wurden. Morphologisch ist sie ganz ähnlich C. campsi, einem Bewohner des hauptsächlich hügeligen oder bergigen Geländes in den Systemen der Flüsse Markham, Ramu, and Sepik in Papua-Neuguinea sowie im Purari-System, südlich der Bergkette Central Dividing Range. Die beiden Arten sind gekennzeichnet durch das Fehlen von Vomer- und Gaumenzähnen und eine relativ schlanke Gestalt bei Männchen wie Weibchen (größte Körpertiefe 23,7-32,3% SL), verglichen mit den meisten anderen Angehörigen der Familie. Nach dem Maßstab der weichen Rücken- und Afterflossenstrahlen zeigen die beiden Arten jedoch deutliche Unterschiede: Chilatherina pagwiensis hat gewöhnlich weiche Rückenflossenstrahlen im Vergleich zu bei C. campsi. Entsprechend zeigt die erstgenannte Art weiche Afterflossenstrahlen, C. campsi hingegen (meistens 22-23). Die Analyse der genetischen Verwandtschaft auf der Grundlage des mitochondrialen Cytochrom b spricht für eine nahe Verwandtschaft zwischen C. pagwiensis und dem Vorfahren von C. axelrodi und C. campsi. Die mittlere p-distanz von C. pagwiensis zu seinen beiden Verwandten betrug 2,8 bzw. 3,8%, was etwa den Werten entspricht, die man bei anderen Regenbogenfisch- Zwillingsarten festgestellt hat. Résumé Une nouvelle espèce de poisson arc-en-ciel, Chilatherina pagwiensis, est décrite sur base de 31 spécimens, 20,5-76,1 de LS, collectés dans un tributaire du Sepik, près de Pagwi village, East Sepik Province, Papouasie-Nouvelle-Guinée. Morphologiquement, elle est très proche de C. campsi qui occupe surtout des régions de collines ou de montagnes dans les systèmes du Markham, du Ramu et du Sepik, au nord de la Papouasie-Nouvelle-Guinée et dans le système du Purari, au sud de la chaîne de montagnes centrale. Les deux espèces se caractérisent par l absence de dents vomerines et palatines et mâles et femelles sont relativement allongés (hauteur maximale du corps 23,7-32,3% de la LS) comparé à la plupart des autres membres de la famille. Néanmoins, les deux espèces montrent de sérieuses divergences modales en rayons mous de la dorsale et en rayons de l anale. Chilatherina pagwiensis a généralement rayons mous à la dorsale contre en moyenne pour C. campsi. De même, la nouvelle espèce a le plus souvent rayons mous dans l anale contre (le plus souvent 22-23) pour C. campsi. L analyse des relations génétiques basée sur des séquences du cytochrome mitochondrial b révèle une relation étroite entre C. pagwiensis et l ancêtre de C. axelrodi et de C. campsi. Le principal écart p entre C. pagwiensis et ses proches était respectivement de 2,2 et 227 aqua vol. 18 no October 2012

50 A new species of Rainbowfish (Chilatherina: Melanotaeniidae), from the Sepik River System of Papua New Guinea de 3,8%, valeurs similaires à celles mentionnées pour d autres espèces proches de poissons arc-en-ciel. Sommario Una nuova specie di pesce arcobaleno, Chilatherina pagwiensis, è descritta sulla base di 31 esemplari di mm SL, raccolti da un affluente del fiume Sepik vicino al villaggio Pagwi, East Sepik, Papua Nuova Guinea. Morfologicamente è molto vicina a C. campsi, una specie che abita i territori prevalentemente collinari o montuosi dei sistemi fluviali del Markham, Ramu e Sepik del nord di Papua Nuova Guinea e il sistema Purari, a sud del Central Dividing Range. Entrambe le specie sono caratterizzate dall'assenza di denti vomerini e palatini, e maschi e femmine sono relativamente affusolati (altezza massima del corpo % SL) rispetto alla maggior parte degli altri membri della famiglia. Tuttavia, le due specie presentano marcate differenze modali nel numero dei raggi molli della dorsale e della anale. Chilatherina pagwiensis ha di solito raggi molli dorsali rispetto ai conteggi usuali di in C. campsi. Allo stesso modo, la prima ha più frequentemente raggi molli anali rispetto ai (più spesso 22-23) di C. campsi. Analisi delle relazioni genetiche mitocondriali basate su sequenze del citocromo b indicano una relazione stretta tra C. pagwiensis e l'antenato di C. axelrodi e C. campsi. La distanza media p di C. pagwiensis dalle forme parentali era del 2,8 e del 3,8% rispettivamente, valori simili a quelli registrati tra le altre specie sorelle dei pesci arcobaleno. INTRODUCTION Rainbowfishes (Melanotaeniidae) are among the most common inhabitants of fresh waters of the Australia/New Guinea region, both in terms of species and number of individuals. They occur in a huge variety of lotic and lentic habitats. Allen (1995) presented an overview of the family, recognising 55 species, including 13 from Australia, 40 from New Guinea, and three, which are shared. Recent taxonomic studies resulted in the description 24 additional species, including two Chilatherina (Allen & Renyaan, 1996a; Price, 1997), two Glossolepis (Allen, 2001; Allen & Renyaan, 1998), 19 Melanotaenia (Allen, 1996; Allen, 1997; Allen & Hadiaty, 2011; Allen & Renyaan, 1996b and 1998; Allen & Unmack, 2008; Allen et al., 2008; Kadarusman et al. 2010, 2011 and 2012; McGuigan, 2001), and a new monotypic genus, Pelangia (Allen, 1998). In addition, several species that were previously considered as synonyms or subspecies, including Glossolepis kabia (Herre, 1935), Melanotaenia australis (Castelnau, 1875) and M. rubrostriatus (Ramsay & Ogilby, 1886) were elevated to full specific status (Allen et al., 2002; Graf & Ohee, 2009). Therefore, the current species total (excluding the species described below) for the family is 80. Seven genera are currently recognised including Cairnsichthys Allen, 1980 (Australia), Chilatherina Regan, 1914 (northern New Guinea), Glossolepis Weber, 1907 (northern New Guinea), Iriatherina Meinken, 1974 (Australia and southern New Guinea), Melanotaenia Gill, 1862 (Australia and New Guinea), Pelangia Allen, 1998 (southern New Guinea), and Rhadinocentrus Regan, 1914 (Australia). However, an extensive phylogenetic study currently in progress by the authors reveals there is a critical need to reassess current generic assignments, particularly regarding species of Chilatherina and Glossolepis. Moreover, several species with widespread distributions, e.g. Melanotaenia goldiei (Macleay, 1883) of southern New Guinea and M. trifasciata (Rendahl, 1922) of northern Australia, are likely divisible into numerous taxa. The present paper describes a new species of Chilatherina from the Sepik River System of Papua New Guinea that was initially collected by the first author in 1982, although at the time it was incorrectly identified as C. fasciata (Weber, 1913), a species with which it co-occurs. It was subsequently collected by the second author and colleagues in 2007, who recognised its unique status, which was confirmed by genetic comparisons with other members of the genus. Pending the outcome of our current generic study, we provisionally assign the new species to Chilatherina. The genus was reviewed by Allen (1981), who recognised six species: C. axelrodi Allen, 1979, C. campsi (Whitley, 1957), C.crassispinosa (Weber, 1913), C. fasciata, C. lorentzi (Weber, 1907), and C. sentaniensis (Weber, 1907). Allen (1983) subsequently recognised C. bulolo (Whitley, 1938) as an additional valid species, and more recently three additional species were described, including C. alleni Price, 1997, C. bleheri Allen, 1985, and C. pricei Allen & Renyaan, MATERIALS AND METHODS Counts and measurements that appear in parentheses refer to the range for paratypes if different from the holotype. Type specimens are deposited at Auburn University Museum, Auburn, Alabama, USA (AUM) and the Western Australian Museum, Perth (WAM). Specimens from the following institutions were also examined: Australian Museum, Sydney (AMS), National Museum of Natural History, Washington, D.C. (USNM), and Zoologisch Museum, Amsterdam (ZMA). aqua vol. 18 no October

51 Gerald R. Allen and Peter J. Unmack Supplementary Table I. GenBank numbers for the individuals sequenced. Under locality, ANGFA, RSG and IRG refer to rainbowfish hobbyist groups in Australia, North America and Europe respectively. Species River basin Locality Individual GenBank # C. alleni Wapoya Tiawiwa R 2 JX C. axelrodi Nemayer ANGFA 1 JX C. bulolo Ramu Whege R 1 JX C. bulolo Ramu Whege R 2 JX C. bulolo Ramu Whege R 3 same as #1 C. campsi Ramu Whege R 1 JX C. campsi unknown RSG 1 same as Whege C. campsi unknown IRG 1 same as Whege C. campsi Markham Uruf Ck 1 JX C. crassispinosa Sepik Amuku Ck 1 JX C. crassispinosa Sepik Amuku Ck 2 same as #1 C. crassispinosa Sepik Nimbrim Ck 1 same as #1 C. crassispinosa Sepik Nimbrim Ck 2 same as #1 C. pagwiensis Sepik Kuins Ck 3 JX C. pagwiensis Sepik Kuins Ck 5 JX C. fasciata Tami Hewa R 1 JX C. fasciata Sepik Kuins Ck 1 JX The methods of counting and measuring are as follows: dorsal and anal rays the last ray of the anal and second dorsal fins is divided at the base and counted as a single ray; lateral scales number of scales in horizontal row from upper edge of pectoral-fin base to caudal-fin base, excluding the small scales posterior to the hypural junction; transverse scales number of scales in vertical row between anal fin origin and base of first dorsal fin; predorsal scales - number of scales along midline of nape in front of first dorsal fin; cheek scales - total number of scales covering suborbital and preoperculum; standard length (SL) measured from tip of upper lip to caudal-fin base; head length (HL) measured from tip of the upper lip to upper rear edge of gill opening; caudal peduncle depth is least depth and caudal peduncle length imeasured between two vertical lines, one passing through base of the last anal ray and the other through caudal-fin base; caudal concavity horizontal distance between verticals at tips of shortest and longest rays. DNA material (see Appendix Table): the following Chilatherina species were included for DNA sequence (museum numbers are for formalin preserved specimens collected at the same time): C. pagwiensis (from paratype locality, see below, AUM 47592), C. alleni (Tiawiwa River, Wapoga River system, Papua, Indonesia, S, E, WAM unregistered); C. axelrodi (aquarium strain from Australia, originally from the Pual River, Papua New Guinea, approximately S, E); C. bulolo (Whege River, Ramu River system, Papua New Guinea, S, E, WAM unregistered); C. campsi (Whege River, Ramu River system, Papua New Guinea, S, E, WAM unregistered; Uruf Creek, Markham River system, Papua New Guinea, S, E; aquarium strain from Europe and North America of unknown origin); C. crassispinosa (Amuku Creek, tributary of Screw River, Sepik River system, Papua New Guinea, S, E, AUM 47601; Nimbrim Creek, tributary of Nina River, Sepik River system, on Maprik-Lumi Road, Papua New Guinea, S, E, AUM 47533); C. fasciata (Hewa River bridge crossing at kilometre 60 on Genyem-Lereh Road, Papua, Indonesia, S, E, WAM unregistered; Kuins Creek (same as the paratype locality of C. pagwiensis, see below, AUM 47591). We extracted genomic DNA from muscle tissue for each specimen using DNeasy Tissue Kits (QIA- GEN Inc., Chatsworth, CA). We amplified the mtdna cytochrome b (cytb) gene using two primers that flanked the gene, Glu18 TAACCAGGACTAATGRCTTGAA and RF.Thr.48 GCAGTAGGAGGGAATTTAAC- CTTCG. Final concentrations for polymerase chain reaction (PCR) components per 25 µl reaction were as follows: 25 ng template DNA, aqua vol. 18 no October 2012

52 A new species of Rainbowfish (Chilatherina: Melanotaeniidae), from the Sepik River System of Papua New Guinea µm of each primer, units of Taq DNA polymerase, 0.1 mm of each dntp, 2.5 µl of 10X reaction buffer and 2.5mM MgCl 2. Amplification parameters were as follows: 94 C for 3 min followed by 35 cycles of 94 C for 30 s, 48 C for 30 s, and 72 C for 90 s, and 72 C for 7 min. We examined PCR products on 1% agarose gels using SYBR safe DNA gel stain (Invitrogen, Eugene, OR, USA). We purified PCR products using a Montage PCR 96 plate (Millipore, Billerica, MA, USA). Sequences were also obtained via cycle sequencing with Big Dye 3.0 dye terminator ready reaction kits using 1/16th reaction size (Applied Biosystems, Foster City, CA). Sequencing reactions were run with an annealing temperature of 52 C following the ABI manufacturer s protocol. We purified sequenced products using SephadexTM columns (G.E. Healthcare, Piscataway, NJ). Sequences were obtained using an Applied Biosystems 3730 XL automated sequencer at the Brigham Young University DNA Sequencing Center. All unique sequences obtained in this study were deposited in GenBank, accession numbers JX JX (Supplementary Table I). Sequences were edited using Chromas Lite 2.0 (Technelysium, Tewantin, Queensland, Australia) and imported into BioEdit (Hall, 1999). Sequences were checked for unexpected frameshift errors or stop codons in MEGA 5.05 (Tamura et al., 2011). For Maximum likelihood (ML) analysis we identified the best fitting model of molecular evolution using the Akaike Information Criterion in ModelTest 3.7 (Posada & Crandall, 1998) as implemented in PAUP* 4.0b10 (Swofford, 2003). The best model of evolution found by ModelTest was GTR+I. We employed traditional tree-based ML phylogenetic analyses using RAxML (Stamatakis, 2006; Stamatakis et al., 2008) with bootstrapping for 1,000 pseudoreplicates, and the final best ML tree was calculated using the GTRGAMMA model on the CIPRES cluster at the San Diego Supercomputer Center (Miller et al., 2010). Among species variation was calculated using mean between group p-distances in MEGA. Chilatherina pagwiensis, n. sp. Pagwi Rainbowfish (Figs 1-2; Tables I-II) Holotype: AUM 58007, male, 76.1 mm SL, upper Kuins Creek, S, E, on Maprik-Pagwi road, East Sepik Province, Papua New Guinea, seine, J. W. Armbruster, P. J. Unmack, and A.Y. Ko ou, 12 October Paratypes: AUM 47649, 4 specimens, mm SL, collected with holotype; WAM P , 26 specimens, mm SL, lower Huins Creek, approximately S, E, about 10 km north of Pagwi on Maprik road, East Sepik Province, Papua New Guinea, rotenone, G. Allen, 19 October Comparative material examined (Papua New Guinea unless stated otherwise): Chilatherina axelrodi USNM , 3 specimens, mm SL, Bewani district, Papua New Guinea; WAM P , holotype, 81.1 mm SL, Bewani district, Papua New Guinea; WAM P , 22 specimens, mm SL, Bewani district, Papua New Guinea; Chilatherina campsi AMNH 20211, 50 mm SL, holotype of Centratherina tenius Nichols, 1956, Wahgi Valley; AMS IB.3337, 56 mm SL, holotype of C. campsi, tributary of middle Jimmi River; AMS IB.3342, 50 mm SL, paratype of C. campsi, same location as previous specimen; USNM , 23 specimens, mm SL, tributary of Markham River, about 105 km northwest of Lae; WAM P , 43 specimens, mm SL, Wahgi Valley near Minj; WAM P , 5 specimens, mm SL, Pima River, upper Purari system, 78 km southwest of Goroka; WAM P , 7 specimens, mm SL, Omsis River, 22 km west of Lae; WAM P , 10 specimens, mm SL, small tributary of Omsis River; Chilatherina fasciata USNM , 11 specimens, mm SL, tributary of Markham River about 25 km west of Lae; WAM P , 4 specimens, mm SL, Lake Wanam, 25 km west of Lae; WAM P , 4 specimens, mm SL, small stream 15 km west of Lae; WAM P , 6 specimens, mm SL, Clearwater Creek, 65 km west of Lae; WAM P , 78 specimens, mm SL, stream on Angoram Road, 73 km east of Wewak; WAM P , 10 specimens, mm SL, stream 5 km beyond Maprik on Dreikikir road; WAM P , 42 specimens, mm SL, Bagi Creek, about 10 km west of Angoram; WAM P , 81 mm SL, stream on Maprik road 113 km west of Wewak; WAM P , 7 specimens, mm SL, Trubum Creek on Maprik road, 79 km west of Wewak; ZMA , 104 mm SL, lectotype of Rhombatractus fasciata (Weber), Boearin River, tributary of Upper Sermowai River, Papua aqua vol. 18 no October

53 Gerald R. Allen and Peter J. Unmack Province, Indonesia; ZMA , 11 specimens, mm SL, Mamberamo River at Pioniersbivak, Papua Province, Indonesia. Diagnosis: A species of Chilathernia distinguished by the following combination of characters: dorsal rays V or VI-I,10 to 13 (usually 10 to 12); anal rays I,19-24 (usually I,20 or 21); pectoral rays 12 to 15 (usually 13 or 14); lateral scales (usually 39-40), predorsal scales 19 to 21; total gill rakers on first arch 14 to 16; vomerine and palatine teeth absent; colour when freshly collected generally pale grey on upper back grading to white over most of body; diffuse greyish stripe on each scale row of back; fins generally reddish, except pectorals semi-translucent whitish. Description: Dorsal rays V-I,12 (V or VI,10 to 13); anal rays I,21 (I,19 to 24); pectoral rays 13 (12 to 15); pelvic rays I,5; branched caudal rays 15; lateral scales 40 (38 or 41); transverse scales 11 (10 or 11); predorsal scales 20 (19-21); cheek scales 17 (15-20); total gill rakers on first arch 14 (14-16). Body depth 3.2 ( ) in SL; greatest body depth by sex and size class as follows: males mm SL, % SL (x = 27.5, N = 3), males mm SL, % SL (x = 30.5, N = 10), male 70+ mm SL, 30.9 % SL (N = 1), females mm SL, % SL (x = 27.2, N = 6), females mm SL, % SL (x = 28.1, N = 5); head length 4.5 ( ) in SL; greatest width of body 2.9 ( ) in greatest body depth; snout length 3.1 ( ) in HL; eye diameter 3.1 ( ) in HL; interorbital width 3.3 ( ) in HL; depth of caudal peduncle 2.3 ( ) in HL; length of caudal peduncle 1.7 ( ) in HL. Jaws about equal, oblique, premaxilla more or less Fig. 1. Freshly collected specimen of Chilatherina pagwiensis, male paratype, 64.3 mm SL, Kuins Creek, East Sepik Province, Papua New Guinea. Photo by P. J. Unmack. Fig. 2. Chilatherina pagwiensis, preserved male, holotype, 76.1 mm SL, Kuins Creek, East Sepik Province, Papua New Guinea. Photo by G. R. Allen. 231 aqua vol. 18 no October 2012

54 A new species of Rainbowfish (Chilatherina: Melanotaeniidae), from the Sepik River System of Papua New Guinea Table I. Proportional measurements of selected type specimens of Chilatherina pagwiensis expressed as percentage of the standard length. Holotype Paratype Paratype Paratype Paratype Paratype Paratype AUM WAM AUM WAM WAM AUM AUM P P P Sex male male male female female female male Standard length (mm) Body depth Body width Head length Snout length Eye diameter Bony interorbital width Maxillary length Depth of caudal peduncle Length of caudal peduncle Predorsal distance Preanal distance Prepelvic distance nd dorsal fin base Anal fin base Pectoral fin length Pelvic fin length Longest ray 1st dorsal fin Longest ray 2nd dorsal fin Longest anal ray Caudal fin length Caudal concavity Fig. 3. Map of New Guinea with arrow indicating approximate collection locations of Chilatherina pagwiensis. aqua vol. 18 no October

55 Gerald R. Allen and Peter J. Unmack Table II. Summary of dorsal, anal, and pectoral fin-ray counts for Chilatherina campsi, C. fasciata, and C. pagwiensis. Soft Dorsal Rays C. campsi C. fasciata C. pagwiensis Soft Anal Rays C. campsi C. fasciata C. pagwiensis Pectoral Rays C. campsi C. fasciata C. pagwiensis Fig. 4. Habitat of Chilatherina pagwiensis at type locality, upper Kuins Creek, East Sepik Province, Papua New Guinea. Photo by P. J. Unmack. 233 aqua vol. 18 no October 2012

56 A new species of Rainbowfish (Chilatherina: Melanotaeniidae), from the Sepik River System of Papua New Guinea straight, without abrupt bend between anterior horizontal portion and lateral part; maxilla ends below anterior edge of eye or slightly anterior to this point; maxillary length 3.6 ( ) in HL; lips thin; jaw teeth slender, conical, slightly movable, and with slightly curved tips, extending onto outer surface of lips; teeth of upper jaw in 5-6 irregular rows anteriorly, reduced to 2-3 rows posteriorly, where they are exposed when mouth is closed; teeth in lower jaw smaller, in about 2-4 irregular rows anteriorly, reduced to 1 or 2 rows posteriorly; vomerine and palatine teeth absent. Scales of body cycloid, relatively large, and arranged in regular horizontal rows; scale margins weakly crenulate; predorsal scales extending forward to about middle of interorbital space or slightly posterior; preopercle with 2-3 scale rows between its posterior angle and eye. First dorsal fin origin about level with anal fin origin; Predorsal length 2.1 ( ) in SL; longest spine (usually first in female and third in male) of first dorsal fin 1.3 ( ) in HL, its depressed tip reaching spine of second dorsal fin in females and reaching to about base of first or second soft ray in mature males; length of second-dorsal fin base 4.3 ( ) in SL; longest rays (generally anterior ones in females and posterior ones in males) of second dorsal fin 1.3 ( ) in HL, the depressed posterior rays extending about one half length of caudal Table III. Mean p-distances between Chilatherina species for cytochrome b C. alleni 1 C. axelrodi C. bulolo C. campsi C. crassispinosa C. pagwiensis C. fasciata peduncle or less in females and nearly full length of caudal peduncle in mature males; preanal length 2.1 ( ) in SL; length of anal-fin base 2.4 ( ) in SL; longest (anterior rays in females and posterior rays in males) anal rays 1.5 ( ) in HL; prepelvic length 3.2 ( ) in SL; pelvic fins relatively short, 1.4 ( ) in HL, depressed tips barely reaching to base of anal spine in adults; length of pectoral fins 1.3 ( ) in HL; length of caudal fin 1.1 ( ) in HL; caudal fin forked, caudal concavity 3.5 ( ) in HL. Colour when freshly collected (Fig. 1): pale grey on upper back grading to white over most of body; diffuse greyish stripe on each scale row of back; fins generally reddish, except pectorals semitranslucent whitish. Colour in alcohol (Fig. 2): brown on upper Fig. 5. Maximum likelihood tree for Chilatherina species based on analysis of cytochrome b sequences (1,141 bp). Bootstrap values were obtained from 1,000 replicates. aqua vol. 18 no October

57 Gerald R. Allen and Peter J. Unmack half and yellowish tan on lower half of head and body; diffuse blackish mid-lateral stripe from rear edge of gill cover to caudal-fin base, slightly more than one scale wide for most of its length, and 6-7 intense black spots along middle portion of lower edge; two pairs of short curved dark grey bars on lower side, just above anteriormost anal rays; fins generally dusky grey. Remarks: Genetic analysis (below) indicates a close sister-species relationship between the new taxon and the ancestral lineage to both C. axelrodi and C. campsi. Morphologically, it is most closely allied to C. campsi, an inhabitant of mainly hilly or mountainous terrain in the Markham, Ramu, and Sepik river systems of northern Papua New Guinea, and south of the Central Dividing Range in the Purari system (Allen, 1981). Both species are characterised by the absence of vomerine and palatine teeth, and males and females are relatively slender (greatest body depth % SL) compared to most other family members. However, the two species exhibit pronounced modal differences in soft dorsal and anal rays (Table II). Chilatherina pagwiensis usually has soft dorsal rays compared to usual counts of in C. campsi. Likewise, the former species most frequently has soft anal ray compared to (most frequently 22-23) in C. campsi. The new species was found in the same habitat with C. fasciata (Weber, 1913), which is widely distributed in northern New Guinea. The two species are easily confused without microscopic examination of the dentition, especially specimens under about 60 mm SL. However, C. fasciata differs in having both vomerine and palatine teeth, and although counts of dorsal and anal soft rays overlap with those of C. pagwiensis, it tends to have higher counts, especially for the anal rays (Table II). There is also an apparent difference in maximum SL with C. fasciata attaining 105 mm versus less than 80 mm SL for C. pagwiensis. Although genetic analysis reveals a close relationship with C. axelrodi, that species differs in having both palatine and vomerine teeth (both edentate in C. pagwiensis) and mature males (> 50 mm SL) generally have a much deeper body (depth averages more than 37% of the SL, n = 28 versus less than 32% SL, n = 12, for C. pagwiensis). Finally, the colour pattern of C. axelrodi is significantly different, consisting of a broad, broken black mid-lateral stripe and 6-10 black bars anteriorly on the lower sides. The species is known only from a single creek in the vicinity of Pagwi village, Sepik River system, Papua New Guinea (Fig. 3). Both authors were given different names for the same creek, Kuin and Huin and it remains unresolved as to which name is correct. It is the only creek which crosses the southern section of the Maprik-Pagwi road, once in its upper reaches and once in the lower reaches. Given it seems widespread within this creek it is likely to be found in similar habitats elsewhere in the Sepik drainage. The type locality (Fig. 4) consisted of a quiet pool in a small creek, flowing through secondary forest. The paratypes were collected to a depth of 1.8 m over a rock and mud bottom in a slow-flowing stream in closed-canopy forest. The latter site is situated approximately 400 km upstream from the sea at an elevation of about 30 m above sea level. DNA analysis of Chilatherina: A total of 17 individuals representing seven species were sequenced for 1,141 bp of cytb. Of 1,141 bp, 1,001 were constant, 27 variable characters were parsimony uninformative, and 113 characters were parsimony informative. ML recovered a single tree with a -ln score of (Fig. 5). All but two nodes had high ML bootstrap support (>97, Fig. 5). Between species the mean p-distances varied between 1.1 and 6.5% (Table III). The Chilatherina species examined formed three lineages. One contained the new species, C. pagwiensis which was most closely related to the ancestor to C. axelrodi and C. campsi, differing to those species by a p-distance of 2.8 and 3.8% respectively. The next closest lineage contained C. bulolo and C. crassispinosa which were closely related (p-distance of 1.1%). The third lineage consisted of C. fasciata and C. alleni (Fig. 5). These levels of genetic divergence between species are similar to those exhibited by many other rainbowfishes (Allen et al., 2008). ACKNOWLEDGEMENTS We thank J. Armbruster and A. Ko ou, who assisted with various aspects of the fieldwork. We also thank the following individuals who provided additional material for genetic analysis: G. Briggs, J. Graf, G. Lange, and R. Smith and his staff from Hydrobiology. DNA sequencing was conducted in the laboratory of Jerald Johnson (Brigham Young University) whose help is gratefully acknowledged. This study was supported in part by the US NSF award (DEB ) for the All Catfish Species Inventory. We also thank the former Department of Fisheries and Marine Resources of Papua New Guinea for providing sponsorship of collecting activities during 235 aqua vol. 18 no October 2012

58 A new species of Rainbowfish (Chilatherina: Melanotaeniidae), from the Sepik River System of Papua New Guinea , and particularly the Papua New Guinea Biological Foundation, which provided financial support for Allen s 1982 visit to the Sepik. REFERENCES ALLEN, G. R Chilatherina axelrodi, a new species of rainbowfish (Melanotaeniidae) from Papua New Guinea. Tropical Fish Hobbyist 28 (5): 48-49, ALLEN, G. R A generic classification of the rainbowfishes (family Melanotaeniidae). Records of the Western Australian Museum 8 (3): Allen, G. R A revision of the rainbowfish genus Chilatherina (Melanotaeniidae). Records of the Western Australian Museum 9 (3): Allen, G. R Chilatherina bulolo, a valid species of rainbowfish (Melanotaeniidae) from northern New Guinea. Fishes of Sahul, 1 (2): ALLEN, G. R Three new rainbowfishes (Melanotaeniidae) from Irian Jaya and Papua New Guinea. Revue françasise d Aquariologe 12 (2): ALLEN, G. R Rainbowfishes in Nature and in the Aquarium. Tetra Verlag, Melle, Germany. ALLEN, G. R Two new species of rainbowfishes (Melanotaenia: Melanotaeniidae), from the Kikori River system, Papua New Guinea. Revue française d Aquariologie Herpétologie 23 (1-2): ALLEN, G. R A new species of rainbowfish (Melanotaenia: Melanotaeniidae), from the Lakekamu Basin, Papua New Guinea. Revue française d Aquariologie Herpétologie 24 (1-2): ALLEN, G. R A new genus and species of rainbowfish (Melanotaeniidae) from fresh waters of Irian Jaya, Indonesia. Revue française d Aquariologie Herpétologie 25 (1-2): ALLEN, G. R A new species of rainbowfish (Glossolepis: Melanotaeniidae) from Irian Jaya, Indonesia. Journal of the Australia New Guinea Fishes Association 15 (3): ALLEN, G. R. & CROSS, N. J Rainbowfishes of Australia and Papua New Guinea. T.F.H. Publications, Neptune, New Jersey, 141 pp. ALLEN, G. R. & HADIATY, R. K A new species of rainbowfish (Melanotaeniidae), from western New Guinea (West Papua Province, Indonesia). Fishes of Sahul 25 (1): ALLEN, G. R., MIDGLEY, H. & ALLEN, M Field guide to the freshwater fishes of Australia. Western Australian Museum, Perth, 394 pp. ALLEN, G. R. & RENYAAN, S. J. 1996a. Chilatherina pricei, a new species of rainbowfish (Melanotaeniidae) from Irian Jaya. Revue française d Aquariologie Herpétologie 23 (1-2): 5-8. ALLEN, G. R. & RENYAAN, S. J. 1996b. Three new species of rainbowfishes (Melanotaeniidae) from the Triton Lakes, Irian Jaya, New Guinea. aqua, Journal of Ichthyology and Aquatic Biology 2 (2): ALLEN, G. R. & RENYAAN, S. J Three new species of rainbowfishes (Melanotaeniidae) from Irian Jaya, Indonesia. aqua, Journal of Ichthyology and Aquatic Biology 3 (2): ALLEN, G. R. & UNMACK, P. J A new species of rainbowfish (Melanotaeniidae: Melanotaenia) from Batanta Island, western New Guinea. Aqua, International Journal of Ichthyology 13 (3-4): ALLEN, G. R., UNMACK, P. J. & HADIATY, R. K Two new species of rainbowfishes (Melanotaenia: Melanotaeniidae), from western New Guinea (Papua, Barat Province, Indonesia). aqua, International Journal of Ichthyology 14 (4): CASTELNAU, F. L Researches on the fishes of Australia. Philadelphia Centennial Expedition of Intercolonial Exhibition Essays, : GRAF, J. & H. L. OHEE, H. L Glossolepis from northern New Guinea. Fishes of Sahul. Journal of the Australian New Guinea Fishes Association 23 (2): HALL, T. A BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symposium Series 41: HERRE, A. W. C. T New fishes obtained by the Crane Pacific expedition. Field Museum of Natural History, Publications, Zoölogical Series 18 (12): KADARUSMAN, SUDARTO, PARADIS, E. & POUYAUD, L Description of Melanotaenia fasinensis, a new species of rainbowfish (Melanotaeniidae) from West Papua, Indonesia with comments on the rediscovery of M. ajamaruensis and the endangered status of M. parva. Cybium 34 (2): KADARUSMAN, SUDARTO, SLEMBROUCK, J. & POUYAUD, L Description of Melanotaenia salawati, a new species of rainbowfish (Melanotaeniidae) from Salawati Island, West Papua, Indonesia. Cybium 35 (3): KADARUSMAN, HADIATY, R.K., SEGURA, G., SETIAWIBAWA, G., CARUSO, D. & POUYAUD L Four new species of rainbowfishes (Melanotaeniidae) from Arguni Bay, West Papua, Indonesia. Cybium 36 (2), MACLEAY, W Contribution to a knowledge of the fishes of New Guinea, No. 4. Proceedings of the Linnean Society of New South Wales 8 (2): MCGUIGAN, K. L An addition to the rainbowfish (Melanotaeniidae) fauna of north Queensland. Memoirs of the Queensland Museum 46 (2): MILLER, M. A., PFEIFFER, W. & SCHWARTZ, T Creating the CIPRES Science Gateway for inference of large phylogenetic trees. In, Proceedings of the Gateway Computing Environments Workshop (GCE), 14 Nov. 2010, New Orleans, LA pp 1-8. NICHOLS, J. T A new melanotaeniid fish from New Guinea. American Museum Novitates 1802: 1-2. POSADA, D. & CRANDALL, K. A ModelTest: testing the model of DNA substitution. Bioinformatics 14: PRICE, D. S Chilatherina alleni, a new species of rainbowfish (Melanotaeniidae) from Irian Jaya. Revue française d Aquariologie Herpétologie 24 (3-4): RAMSAY, E. P. & OGILBY, J. D A contribution to the knowledge of the fish-fauna of New Guinea. Proceedings of aqua vol. 18 no October

59 Gerald R. Allen and Peter J. Unmack the Linnean Society of New South Wales (Series 2) 1 (1): REGAN, C. T Report on the freshwater fishes collected by the British Ornithologists Union expedition and the Wollaston expedition in Dutch New Guinea. Transactions of the Zoological Society of London 20 (part 6, no. 1): RENDAHL, H Meddelelser fra det Zoologiske Museum, Kristiania. Nr. 5. A contribution to the ichthyology of north-west Australia. Nyt Magazin for Naturvidenskaberne 60: STAMATAKIS, A RAxML-VI-HPC: Maximum Likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22: STAMATAKIS, A., HOOVER, P. & ROUGEMONT, J A rapid bootstrap algorithm for the RAxML web-servers. Systematic Biology 75: SWOFFORD, D. L PAUP. Phylogenetic analysis using parsimony (and other methods), version 4.0b10. Sinauer, Sunderland. TAMURA, K., PETERSON, D., PETERSON, N., STECHER, G., NEI, M. & KUMAR, S MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28: WEBER, M Süsswasserfische von Neu-Guinea ein Bei - trag zur Frage nach dem früheren Zusammenhang von Neu Guinea und Australien. In: Nova Guinea. Résultats de l expédition scientifique Néerlandaise à la Nouvelle-Guinée. Süsswasserfische von Neu-Guinea v. 5 (Zool.) part 2: WEBER, M Süsswasserfische aus Niederländisch Südund Nord-Neu-Guinea. In: Nova Guinea. Résultats de l expédition scientifique Néerlandaise à la Nouvelle-Guinée. Zoo - logie 9 (4), Leiden: WHITLEY, G. P Fishes from inland New Guinea. Records of the Australian Museum 24 (3): aqua vol. 18 no October 2012

60 Book review SHORE FISHES OF EASTER ISLAND John E. Randall & Alfred Cea University of Hawaii Press pp. cloth bound Price: USD $35 Easter Island, a tiny remote outpost in the South Pacific, is perhaps best known for its colossal stone statues, and the controversy surrounding their origin. However, it is one of the world s most significant ichthyological locations due to its highly unique faunal community. Particularly significant, is the remarkably high percentage of endemism among its shorefishes, 21.7 % - a figure exceeded only by the Hawaiian Islands (25 % endemism) for the entire Indo-Pacific region. Although the fauna is small, reflecting its small size and isolated nature, as well as its subtropical location and lack of habitat diversity, it contains a highly interesting blend of zoogeographic elements from southern subtropical Oceania, the anti-equatorial Pacific, eastern Pacific, and the broader tropical Indo-west and central Pacific. In addition to the large percentage of endemics, about 14.5 % of the shorefishes are cosmopolitan forms. Shore Fishes of Easter Island presents a concise, well-written account of the island s fish community. The 57 families are presented in phylogenetic order, each briefly introduced with a section that includes its salient morphometric and meristic features, distribution, and pertinent references. Species coverage (in alphabetical order by genus and species within each family) follows a similar format, accompanied by insightful remarks dealing with abundance, habitat, feeding habits, and local common names. Each species is well illustrated with a diagnostic colour photograph, many of them taken in the natural habitat. Both sexes are illustrated for species that exhibit sexual di-chromatism, such as wrasses, parrotfishes, and triggerfishes. The authors are well qualified for this task. John ( Jack ) Randall is a living ichthyological legend with nearly 800 publications to his credit. Alfredo Cea is a medical doctor from Santiago, Chile, who has made many visits to the island for the study of coastal ecology and anthropological research. The author of this review accompanied Randall on his first visit to the island in There was very little information about the fish fauna at that time and no comprehensive collections had been made. Now, largely thanks to the dedication of Randall and Cea, the current fauna of the island stands at 169 species of which 139 can be classified as shorefishes and 26 are epipelagic inhabitants. In addition to its encyclopaedic coverage of the families and species, this volume includes a historic review of Easter Island ichthyology as well as valuable sections on zoogeography and marine conservation. There is also a comprehensive reference section. An added bonus is a checklist of the shore and epipelagic fishes of Easter Island at the end of the book, which also includes the distribution category (e.g. tropical Indo-Pacific, anti-tropical, eastern Pacific, etc.) for each species. This book is highly recommended for serious ichthyologists, divers, and students of natural history. Gerald R. Allen aqua vol. 18 no October

61 Book review THE BIOLOGY OF GOBIES Patzner, R. A., J. L. Van Tassell, M. Kovačié, and B. G. Kappor, (editors). Science Publishers, Channel Islands, U.K., Enfield, New Hampshire, USA pp. Gobies are a group of charismatic fishes that live in tropical and temperate freshwater, estuarine, and fully marine habitats to a recorded depth of over 1,100 meters. They may be brilliantly colored, and a few reach a maximum recorded size of some 60 cm SL, although most are small, inconspicuous, benthic tropical coastal marine or coral reef fishes. Well over 2,000 or an estimated one in every 10 to 20 species of bony fishes is a goby (suborder Gobioidei or order Gobiiformes, used interchangeably in this book), yet much knowledge about gobies is broadly dispersed and as cryptic as the fishes themselves. The editors make a bold claim for gobies as (p. vi) unparalleled among vertebrates in their ability to diversify and adapt. They are right. The vertebrate with the shortest lifespan is a goby: the diminutive Eviota sigillata, which lives in tropical inshore and coral reef habitats throughout the Indo-West and Central Pacific, completes its entire life cycle in about two months (Depczynski and Bellwood, 2005). One or another species of the genus Trimmatom is a perpetual entrant in the contest for the title of world s smallest vertebrate. The mudskippers (Periophthalmus and relatives) are the most terrestrial of bony fishes; embryos are tended in mudflat burrows that males excavate by sucking up mud and regurgitating it as pellets at the surface. Males ventilate the burrows by filling their buccopharyngeal cavity with air and expelling the air during forays into the burrows (Ishimatsu et al., 1998). All gobies are oviparous; adults may be of fixed sex, change sex (from male to female, or from female to male), or be hermaphrodites (Cole, 1990). The Biology of Gobies, remarkably, is the first attempt to consolidate our knowledge of goby biology gobiology to the cognoscenti, systematics, ecology, and biogeography. The effort is international and collaborative: 33 authors from 15 countries wrote 25 chapters in four sections on goby systematics (10 chapters), zoogeography (six chapters), general biology (four chapters) and ecology (five chapters). The editors are optimists who claim that the book contains (p. v) the background information for others to join in on the study of this diverse group. Here, they are just partly right. This book has been written by many of the world s leading goby researchers and it is rich with their original observations and mature insights, yet its production is uneven, the text is repetitive, and reproduction of illustrations ranges from good to poor. Many of the black and white figures are faint or blurry and the color plates, collected together in the back of the book, are mostly fuzzy and their text illegible. The 25-page subject, select author, and taxonomic Index was likely produced with off-the-shelf book indexing software. Index entries such as Tiny size 362 or Walking 352 are amusing, but not useful. Despite these drawbacks, The Biology of Gobies is the best and most complete goby compendium today and the only one that we will likely have for some time. This book is marketed and distributed by CRC Press which sells electronic access to chapters on its website: Gobioid monophyly has been well-supported by both morphology (e.g., Miller 1973, 1992; Springer 1983) and molecules (Thacker and Hardman, 2005) and is not questioned here. It is perhaps the firmest ground on which goby systematics stands. Following a brief Preface, the book opens with two introductory chapters, one on gobioid morphological classification and the other on gobioid molecular systematics. From both the morphological and molecular perspectives, the gobioid sistergroup is unknown. The nine gobioid families, following the conventional classification of Nelson (2006), are reviewed in eight chapters, the first of which, Systematics of the Rhyacichthyidae, by Helen K. Larson, sets the style for the family studies that follow: it is a short and informative literature review that includes some personal insights into gobiology and ends with a list of references. Unfortunately, the systematics chapters may contradict each other. For example, Larson classifies two genera, 239 aqua vol. 18 no October 2012

62 Rhyacichthys and Protogobius, in the family Rhyacichthyidae, and Akihisa Iwata classifies six genera, including Terateleotris, in the family Odontobutidae. In contrast, Christine Thacker uniquely classifies Terateleotris in the Rhyacichthyidae. Ideally, these systematics chapters would have been collected together into one, multiauthored, comprehensive disquisition to include a review of the classification of each family, a consensus classification of gobioids, at least to genus, presented in a table, an illustrated discussion of the characters that diagnose major clades, and the collected references. The zoogeography chapters are more descriptive than analytical or comparative, yet present a wealth of data that may be used to interpret the distributional history of gobies worldwide. Not surprisingly, despite a marine larval period of weeks to months for many species, gobies exhibit marked endemism. The 21 species of neon goby, genus Elacatinus, and their color morphs, have (p. 159) very restricted, non-overlapping, geographic distributions in the western Atlantic and eastern Pacific. Many mudskippers (p. 101) are endemic to relatively restricted geographic areas. The majority of marine, euryhaline and amphidromous gobies, 51 of 87 species (p. 221) that live in western, central and southern Africa, are endemics. Their broad distributions, high species diversity, and discrete areas of endemism make gobies ideal for elucidating global biogeographic patterns. This high degree of endemism potentially will translate into highly informative biogeographic patterns when relationships among endemic areas are proposed and tested (Parenti and Ebach, 2009). The sheer abundance of gobies, in numbers of species and individuals, and their small size means that they play a critical role in the maintenance of healthy ecosystems. The nine chapters in the sections on General Biology and Ecology review an array of topics for which gobies are already well-known, such as the mutualistic association between gobies and alpheid shrimps, the cleaning behavior of Elacatinus species, and the excavation of burrows by mudskippers, and those for which they are less so, such as their life as predator and prey. One chapter by C. Dieter Zander is an eclectic historical review of life history observations that span nearly eight decades. Here is some data from Zander on the importance of goby fry in fisheries (p. 328): Fishery on goby fry is concentrated in several localities of the tropic region, especially in Philippines, Hawaii, Caribbean and Reunion Mostly post-larvae of the genera [Sicyopterus] and Sicydium are caught during the emigration to the sea from freshwater where the adults spawn Yields vary between 60,000 t/year in Luzon, 2 t in Hawaii, 0.9 t in [Dominica] (Caribbean) or a [mean of] 40 t in Reunion. At nearly 700 pages, the book bursts with such observations, some original, others compiled from the literature, on the biology of gobies. Gobiologists, new and old, will find it useful, even stimulating. It s a good start. LITERATURE CITED COLE, K. S Patterns of gonad structure in hermaphroditic gobies (Teleostei: Gobiidae). Environmental Biology of Fishes 28: DEPCZYNSKI, M. & BELLWOOD, D Shortest recorded vertebrate lifespan found in a coral reef fish. Current Biology 15(8): ISHIMATSU, A., HISHIDA, Y., TAKITA, T., KANDA, T., OIKAWA, S., TAKEDA, T. & KHOO, K. H Mudskippers store air in their burrows. Nature 391: MILLER, P.J The osteology and adaptive features of Rhyacichthys aspro (Teleostei: Gobioidei) and the higher classification of gobioid fishes. Journal of Zoology, London 171: MILLER, P. J The sperm duct gland: a visceral synapomorphy for gobioid fishes. Copeia 1992 (1): NELSON, J. S Fishes of the World, Fourth Edition. John Wiley & Sons, Hoboken, NJ. PARENTI, L. R. & EBACH, M. C Comparative Biogeography: Discovering and Classifying Biogeographical Patterns of a Dynamic Earth. University of California Press, Berkeley, CA. SPRINGER, V. G Tyson belos, new genus and species of western Pacific fish (Gobiidae, Xenisthminae), with discussions of gobioid osteology and classification. Smithsonian Contributions to Zoology 390: THACKER, C.E. & HARDMAN, M. A Molecular phylogenetics of basal gobioid fishes: Rhyacichthyidae, Odontobutidae, Xenisthmidae, Eleotridae (Teleostei: Perciformes: Gobioidei). Molecular Phylogeny and Evolution 37(3): Lynne R. Parenti Division of Fishes National Museum of Natural History Smithsonian Institution PO Box 37012, NHB MRC 159 Washington, DC , USA aqua vol. 18 no October

63 Guidelines for Authors 1. Manuscript preparation: manuscripts must be submitted in English. In exceptional cases aqua may provide translations of manuscripts written in French, German, Italian, or Spanish. Manuscripts must be word-processed in Microsoft WORD and submitted in an electronic form. Generic, specific, and sub-specific names must be italicised. All papers must conform to the International Code of Zoological Nomenclature. Authors are strongly advised to follow the format set out in previous publications of aqua. 2. Title: the title must be short and should precisely identify the main topic of the article. Names of genera or species are followed by the systematic group to which they belong. Author name(s) are given in full beneath the title, followed by the complete mailing and address(es). 3. Abstract: the abstract should not exceed 250 words and draw attention to the principal conclusions. It should not contain any uncommon abbreviations or literature citations. The inclusion of abstracts in other languages is optional. 4. Subject matter: the text of the manuscript is usually arranged in four main sections: Introduction, Materials and methods (including a key to abbreviations), Results, and Discussion. Other subdivisions may be chosen depending on the material presented. Acknowledgements should be placed between the text and references. Scientific names of genera, species, and subspecies should be followed by the name(s) of author(s) and the year of publication on first mention. A description of a new taxon must contain the following sections: Material, Diagnosis, Description, and Affinities. Synonyms must be arranged in chronological order. Identification keys must be dichotomous. Holotype and paratypes must be clearly identified, the institution in which they have been deposited named, and the catalogue numbers given. Private collections are not acceptable as repositories for holotypes. DNA sequences must be archived in GenBank, DNA Databank of Japan or European Molecular Biology Laboratory. Voucher specimens associated with DNA sequences must be deposited in a recognized research collection accessible to the professional community. Accession numbers for sequences and catalog number for vouchers must be included as a table in an appendix to the final manuscript. The metric system and SI units must be used. Temperatures are given in C. Fractions should not be used. 5. References to literature: the name-year system must be used. The list of references should be placed at the end of the paper, alphabetically arranged according to author name. Only those publications cited in the paper may be included. Titles of journals must be given in full. Examples of correct reference formats: BLABER, S. J. M Fish of the Trinity inlet system of North Queensland, with notes on the ecology of fish faunas of tropical Indo-Pacific estuaries. Australian Journal of Marine and Freshwater Research 31: DAY, J. H., BLABER, S. J. M., & WALLACE, J. H Estuarine fishes. In: Estuarine Ecology with Particular Reference to Southern Africa. (Ed. J.H. Day.): A. A. Balkema, Rotterdam. DIMMICH, W. W Ultrastructure of North American cyprinid maxillary barbels. Copeia 1988 (1): TREWAVAS, E Tilapiine Fishes of the Genera Sarotherodon, Oreochromis and Danakilia. British Museum (Natural History), London, 583 pp. 6. Submission of manuscript and illustrations: The manuscript and illustrations must be submitted digitally to the Scientific Editor: Dr Frank L. Pezold College of Science & Engineering Texas A&M University Corpus Christi 6300 Ocean Drive Corpus Christi, TX frank.pezold@tamucc.edu to whom all subsequent correspondence shall be addressed. Texts, tables, and graphs should be in Microsoft-compatible electronic form. After the paper has been accepted for publication, illustrations as high-resolution TIF files or original photographs (ideally transparencies; otherwise glossy prints, preferably in the size in which they will appear - the type area of aqua is 158 x 224 mm, one column is 76 mm wide) must be sent to: Aquapress, The Managing Editor Via G. Falcone 11, Miradolo Terme (Pavia), Italy aquapress@aquapress-bleher.it Authors should retain copies of all materials for reference. Proofs of accepted papers will be sent as PDF files by e- mail attachment to the corresponding author. 7. Evaluation of manuscripts: manuscripts will be evaluated by the editors and referees. Papers are accepted on the understanding that they have not and will not be published elsewhere. 8. Reprints: Authors will receive one free copy of the issue in which their paper appears and an e-print in PDF format. Additional copies may be ordered from Aquapress.

64 aqua International Journal of Ichthyology Vol. 18 (4), 15 October 2012 Contents: F. B. M. Vermeulen, W. H. Suijker and G. E. Collier: Laimosemion paryagi (Cyprinodontiformes: Aplocheiloidei: Rivulidae), a new species from the upper Mazaruni river drainage of Western Guyana Stefano Valdesalici: Nothobranchius kardashevi and Nothobranchius ivanovae (Cyprinodontiformes: Nothobranchiidae): two new annual killifishes from the Katuma River drainage, western Tanzania David Bierbach, Madlen Ziege, Claudia Zimmer, Rüdiger Riesch, Ingo Schlupp, Lenin Arias-Rodriguez and Martin Plath: Male Grijalva mosquitofish (Heterophallus milleri Radda, 1987) increase individual mating preferences in front of an audience John E. Randall and Rachel J. Arnold: Uranoscopus rosette, a new species of stargazer (Uranoscopidae: Trachinoidei) from the Red Sea Michelle R. Gaither and John E. Randall: On the validity of the cirrhitid fish genus Itycirrhitus Gerald R. Allen and Peter J. Unmack: A new species of Rainbowfish (Chilatherina: Melanotaeniidae), from the Sepik River System of Papua New Guinea Book review: Shore Fishes of Easter Island. John E. Randall and Alfred Cea Book review: The Biology of Gobies. R. A. Patzner, J. L Van Tassell, M. Kovačié and B. G. Kappor (editors) Papers appearing in this journal are indexed in: Zoological Record; BioLIS Biologische Literatur Information Senckenberg; Cover photo: Nothobranchius kardashevi, adult male, about 35 mm SL, not preserved, Karira stream, western Tanzania. Photo by I. Ivanova. Type locality of Nothobranchius kardashevi and N. ivanovae. Residual pool in Karira stream, western Tanzania. See pages