aqua · aqua is an international journal which publishes original scientific articles in the fields...

aquaInternational Journal of Ichthyology

Vol. 12 (3), 10 July 2007

AquapressISSN

0945-9871

Managing Editor:

Heiko BleherVia G. Falcone 11, 27010 Miradolo Terme (PV), ItalyTel.: +39-0382-754707 Fax: +39-0382-754129E-mail: [email protected]

Scientific Editor:

Friedhelm KruppCurator of FishesSenckenberg Research Institute and Natural History MuseumSenckenberganlage 2560325 Frankfurt am Main, GermanyTel: +49-69-7542.1255Fax: +49-69-7542.1253E-mail: [email protected]

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

Wilson J. E. M. Costa Laboratório de Ictiologia Geral e Aplicada, Departamento de Zoologia, Universidade Federal do Rio de Janeiro, Brasil

Axel Meyer Lehrstuhl für Zoologie und Evolutionsbiologie, Universität Konstanz, Germany

Paolo Parenti Department of Enviromental Sciences, University of Milano-Bicocca, Milan, Italy

Mário de PinnaMuseu de Zoologia da USP, São Paulo, Brazil

John E. RandallBishop Museum, Honolulu, Hawaii, U.S.A.

Richard WinterbottomCentre of Biodiversity & Conservation Biology, Royal Ontario Museum, Toronto, Canada

Scopeaqua is an international journal which publishes originalscientific articles in the fields of systematics, taxonomy,biogeography, ethology, ecology, and general biology offishes. Papers on freshwater, brackish, and marine fisheswill be considered. aqua is fully refereed and aims at pub-lishing manuscripts within 2-4 months of acceptance. Inview of the importance of color patterns in species identi-fication and animal ethology, authors are encouraged tosubmit color illustrations in addition to descriptions ofcoloration. It is our aim to provide the international sci-entific community with an efficiently published journalmeeting high scientific and technical standards.

Call for papersThe editors welcome the submission of original manu-scripts which should be sent in digital format to the scien-tific editor. Full length research papers and short notes willbe considered for publication. There are no page chargesand color illustrations will be published free of charge.Authors will receive one free copy of the issue in whichtheir paper is published and an e-print in PDF format.

Subscription NoticeAt least one volume (4 issues) of aqua is being publishedper year, each issue comprising 48 pages (including cover). The subscription rate (for one volume = 4 issues) isfrom volume 12 on: Personal subscription: Euro 75,00(incl. priority mail); Institutional subscription: Euro 140,00(incl. priority mail). Subscription enquires should be sentto the publisher at the address given below or by e-mailto: [email protected]

aqua binderBinders for Volumes of aqua are available at cost priceEuro 12,50 (US$ 15.00) plus postage Euro 8,00 (US$10.00). Notice: aqua Volumes 1(1)-5(4) = 1st. binder; Volumes 6(1)-9(4) = 2nd. binder; Volumes 10(1)-13(4)= 3rd. binder.

Special PublicationSince 2003 Aquapress publishes a series of Special Publi-cations, which are produced at irregular intervals. All Spe-cial Publications have about 100 or more pages and areavailable separately from regular issues of aqua. Enquiriesabout subscriptions and prices should be sent to the pub-lisher at the address given here above or by e-mail to:[email protected]

ISSN 0945-9871Publisher: Aquapress, Redazione aqua, I-27010 Miradolo Terme (Pavia), ItalyPrinter: PRINTO s.r.o. - Czech RepublicCopyediting and layout: Rossella Bulla© 2007 aqua, International Journal of Ichthyology

aqua - International Journal of Ichthyology

AbstractThe following four new species of the Indo-Pacific gob-

iid fish genus Vanderhorstia are described and illustrated incolor: V. attenuata from the Solomon Islands, with 11 dor-sal and anal soft rays, 39 scales in longitudinal series onbody, 15 median predorsal scales (none in remaining threespecies), and a caudal fin with the middle five rays veryprolonged and tapering (more uniformly pointed inremaining species); V. auronotata from the MoluccaIslands, with 14 dorsal soft rays, 15 anal soft rays, andabout 55 scales on body; V. belloides from Papua NewGuinea with 16 dorsal soft rays, 17 anal soft rays, and 66-68 scales; and V. dorsomacula from Papua New Guineawith 17 dorsal soft rays, 18 anal soft rays, 78 scales, and aslender body (depth 6.6 in SL). One undescribed speciesof the genus is illustrated from a specimen that is nowmissing. The total number of species of Vanderhorstia isnow 17; at least 17 more remain to be described.

ZusammenfassungDie folgenden vier Arten der Gobiiden-Gattung Vander-

horstia werden beschrieben und in Farbfotos veranschaulichtabgebildet: V. attentuata von den Salomon-Inseln mit 11dorsalen und analen Weichstrahlen, 39 Schuppen in derLängsreihe auf dem Rumpf, 15 medianen prädorsalenSchuppen (keine bei den anderen drei Arten) und einerSchwanzflosse, bei der die mittleren fünf Strahlen sehr starkverlängert sind und allmählich spitz zulaufen (bei denanderen drei Arten einheitlich spitz); V. auronotata von denMolukken mit 14 dorsalen Weichstrahlen, 15 analenWeichstrahlen und etwa 55 Schuppen auf dem Rumpf; V.belloides von Papua-Neuguinea mit 16 dorsalen, 17 analenWeichstrahlen und 66-68 Schuppen; und schließlich V. dor-somacula von Papua-Neuguinea mit 17 dorsalen, 18 analenWeichstrahlen, 78 Schuppen und einem schlanken Körper(Körperhöhe 6,6 x in SL). Eine weitere unbeschriebene Artwird mit dem Foto eines Exemplars belegt, das nicht mehrvorhanden ist. Die Gesamtzahl der Vanderhorstia-Artenbeträgt damit 17; von mindestens 17 weiteren steht dieBeschreibung noch aus.

RésuméLes quatre nouvelles espèces, ci-dessous, du genre de Gob-

iidés Vanderhorstia de l’Indo-Pacifique sont décrites et illus-trées en couleurs : V. attenuata des îles Salomon, avec 11rayons mous à la dorsale et à l’anale, 39 écailles en rangéeslongitudinales sur le corps, 15 écailles médianes prédorsales(absentes pour les trois autres espèces), et une caudale aveccinq rayons médians très allongés et plutôt pointus (plusuniformément pointus pour les espèces restantes) ; V.auronotata des îles Moluques, avec 14 rayons mous dorsaux,15 à l’anale et environ 55 écailles sur le corps ; V. belloides dePapouasie-Nouvelle-Guinée avec 16 rayons mous à la dor-sale, 17 à l’anale, et 66-68 écailles ; et V. dorsomacula dePapouasie-Nouvelle-Guinée avec 17 rayons mous dorsaux,18 à l’anale, 78 écailles et un corps mince (6,6 de LS en hau-teur). Une espèce du genre non décrite est illustrée sur based’un spécimen égaré pour le moment. Le nombre total d’espèces de Vanderhorstia est à présent de 17 ; au moins 17autres restent à déterminer.

SommarioQuattro nuove specie del genere di gobide dell’Indo-Paci-

fico Vanderhorstia sono descritte e illustrate nella loro co-lorazione naturale: V. attenuata delle Isole Salomone, con11 raggi molli dorsali e anali, 39 scaglie in serie longitudi-nali sul corpo, 15 scaglie predorsali mediane (assenti nellealtre tre specie) e una pinna caudale con i cinque raggi cen-trali molto prolungati e affusolati (più uniformementeappuntiti nelle altre specie); V. auronotata delle IsoleMolucche, con 14 raggi molli dorsali, 15 raggi molli analiand circa 55 scaglie sul corpo; V. belloides da Papua NuovaGuinea con 16 raggi molli dorsali, 17 raggi molli anali e66-68 scaglie. Infine, V. dorsomacula anch’essa da PapuaNuova Guinea con 17 raggi molli dorsali, 18 raggi mollianali, 78 scaglie e un corpo più assottigliato (altezza delcorpo 6.6 in SL). Una specie non descritta dello stessogenere è illustrata sulla base di un esemplare andato perso.Il numero complessivo di specie del genere Vanderhorstia èora di 17, ma almeno altrettante rimangono da descrivere.

INTRODUCTIONGobiid fishes of the genus Vanderhorstia Smith,

1949 live symbiotically in a burrow in sedimentarysubstrata with snapping shrimps of the genus

aqua, International Journal of Ichthyology

Descriptions of four new shrimpgobies of the genus Vanderhorstiafrom the western Pacific

John E. Randall

Bishop Museum, 1525 Bernice St., Honolulu, HI 96817-2704, USA. E-mail: [email protected]

Received: 13 October 2006 – Accepted: 10 Janurary 2007

aqua vol. 12 no. 3 - 10 July 200789

Alpheus. The shrimps (generally a pair) constructand maintain the burrow. The goby serves as thesentinel at the entrance, quickly taking shelter inthe burrow with the approach of danger (Karplus,1987). Because of their retiring habits, these gob-ies are difficult to collect. During the period of1973 to 1987, the author collected specimens offour undescribed species of this genus from theSolomon Islands, Indonesia, and Papua NewGuinea. One is represented by two specimens, andthe other four by single specimens. No additionalspecimens have been obtained, and with no planfor more material, these species are described here.The genus Vanderhorstia was proposed by Smith

(1949) for Gobius delagoae Barnard, 1937 fromMozambique, but without a description. Smith (l.c.) stated that it is closely allied to CryptocentrusValenciennes, 1837, differentiating it only by hav-ing the gill membranes completely free from theisthmus. Smith (1959) provided a diagnosis forVanderhorstia and described a second species, V.ornatissima, from Mozambique and the Seychelles. Hoese in Smith & Heemstra (1986) provided the

following diagnosis for Vanderhorstia: gill openingfrom rear preopercular margin to below eye; cheekpapillae pattern longitudinal; head pores Type A.Scales cycloid or ctenoid. V a disc. D VI+I,12-16;A I,12-17; P 17-19; LS 47-67. Vertebrae 26. Hegave the number of species as about eight, notingthat the genus is easily confused with Amblyeleotrisand Cryptocentrus, which have a transverse cheekpapillae pattern. Shibukawa & Suzuki (2004) described Vander-

horstia papilio from the Ryukyu Islands, addingthat their assignment of their species to Vander-horstia is provisional because of the present weakdiagnosis of the genus. They recognized 11 speciesof Vanderhorstia; however, V. praealta Lachner &McKinney, 1981 and V. lanceolata Yanagisawa,1978 are now regarded as species of Tomiyam-ichthys Smith (Koichi Shibukawa, pers. comm.). Four more new species of Vanderhorstia have

recently been described: V. bella Greenfield & Lon-genecker, 2005 from Fiji; V. nannai Winterbottom,Iwata & Kozawa, 2005 from Palau and the Philip-pines; V. nobilis Allen & Randall, 2006 fromIndonesia and the Philippines; and Vanderhorstiaopercularis Randall, 2007 from the Red Sea. Morespecies remain to be named, as indicated by twobooks on Japanese gobies: Hayashi & Shiratori(2003) illustrated seven species as Vanderhorstiasp., and Senou et al. (2004) list 10 species as unde-

scribed in the genus (however, their Vanderhorstiasp. 4 is now recognized as a species of Tomiyam-ichthys).The new species described below all have cephalic

sensory pores as illustrated for Vanderhorstia aurop-unctata Tomiyama, 1955 and V. mertensi Klause-witz, 1974 by Prince Akihito in Masuda et al.(1984: figs. 117 and 119), with the same ninepores of the oculoscapular canal and the three ofthe preopercular canal (the latter species, however,is undescribed, distinct from the Red Sea V.mertensi). The pattern of sensory papillae on thehead is also very similar to that of these two species,with two near-horizontal series of close-set smallpapillae on the cheek.Figure 1 represents an undescribed species of the

genus that was collected by the author in northeastBali from a depth of 23 m in 1990. It was mea-sured as 33 mm SL and cataloged as BPBM 36706.It has since been lost (Fig. 1). All of the new species described below have 5

branchiostegal rays, 10 + 16 vertebrae, and a dorsalpterygiophore formula 3-22110 (Fig. 1A of Bird-song et al., 1988).

MATERIALS AND METHODSThe type specimens of the new species are

deposited in the Bernice P. Bishop Museum, Hon-olulu (BPBM).Lengths of specimens are given as standard

length (SL), measured from the median anteriorpoint of the upper lip to the base of the caudal fin(posterior end of the hypural plate); body depth ismeasured at both the origin of pelvic fins and theorigin of the anal fin, and body width at the ori-gin of the pectoral fins; head length is taken fromthe upper lip to the posterior end of the opercularmembrane, and head width over the posteriormargin of the preopercle; orbit diameter is thegreatest fleshy diameter, and interorbital width theleast fleshy width; snout length is measured fromthe median anterior point of the upper lip to thenearest fleshy edge of the orbit; upper-jaw lengthfrom the same anterior point to the posterior endof the maxilla; cheek depth is the least depth per-pendicular from the ventral edge of the suborbitalto the fleshy edge of the orbit; caudal-peduncledepth is the least depth, and caudal-pedunclelength the horizontal distance between verticals atthe rear base of the anal fin and the caudal-finbase; lengths of spines and rays are measured totheir extreme bases; caudal- and pectoral-fin

aqua vol. 12 no. 3 - 10 July 2007 90

Descriptions of four new shrimpgobies of the genus Vanderhorstia from the western Pacific

lengths are the length of the longest ray; pelvic-finlength is measured from the base of the pelvicspine to the tip of the longest pelvic soft ray. Mor-phometric data presented in Table I are given aspercentages of the standard length. Proportionalmeasurements in the text are rounded to the near-est 0.05. The count of scales in longitudinal series is made

from above the dorsal end of the gill opening to thebase of the caudal fin (scales small anteriorly andnot in regular rows, so counts often difficult toduplicate precisely); scales in transverse series arecounted from the origin of the anal fin anterodor-sally to the base of the first dorsal fin; gill rakers arecounted on the first gill arch, those on the upperlimb listed first; rudiments are included in thecounts.Specimens were collected with rotenone and by

multiprong Hawaiian sling spear.

Vanderhorstia attenuata n. sp.(Fig. 2, Table I)

Holotype: BPBM 38734, male, 36.3 mm, SolomonIslands, Guadalcanal, north side, 11.3 km west ofHoniara, sand and reef adjacent to wreck of Japan-ese transport ship, 38-48 m, rotenone, J. E. Randalland G. R. Allen, 13 July 1973.Diagnosis: Dorsal rays VI-I,11; anal rays I,11; pec-

toral rays 19; scales in longitudinal series 39; medianpredorsal scales 15; no scales on head except dorsallybehind eye; scales present on prepectoral area; bodynot very elongate, the depth 5.35 in SL; orbit diam-eter 3.2 in head length; third dorsal spine longest,3.95 in SL; caudal fin long, the middle five rays pro-longed into a tapering filament as long as rest of fin,the total fin length 2.15 in SL; pectoral fins reachingto above origin of anal fin, 3.6 in SL; pelvic fins justreaching anus, 4.9 in SL; color as in Fig. 2.

aqua vol. 12 no. 3 - 10 July 200791

John E. Randall

Fig. 1. Vanderhorstia sp., BPBM 36706, 33 mm SL, Bali, Indonesia (specimen lost). Photo by J. Randall.

Fig. 2. Holotype of Vanderhorstia attenuata, BPBM 38734, 36.3 mm, Guadalcanal, Solomon Islands. Photo by J. Randall.

Description: Dorsal rays VI-I,11; anal rays I,11;all dorsal and anal soft rays branched, the last tobase (each major branch of last ray divided dis-tally); pectoral rays 19, the uppermost and lowerone or two unbranched; pelvic rays I,5, all soft raysbranched, the fifth rays joined medially; segmentedcaudal rays 17, 14 branched; 6 upper and 7 lowerunsegmented caudal rays; longitudinal scale series39 (some scales missing, especially posteriorly),progressively smaller anteriorly; transverse scalerows 13; median predorsal scales 15, nearly reach-ing interorbital space; no scales on cheek or oper-cle; scales present above opercle but not extendingforward above preopercle; scales as large as bodypresent on prepectoral area; median prepelvicscales 8; circumpeduncular scales 13; gill rakers 4 +12; pseudobranch with 9 short fleshy lobes. Body not very elongate, the depth at pelvic-fin

origin 5.35 in SL; body compressed, the width atpectoral-fin origin 1.7 in body depth; head length3.7 in SL; head slightly compressed, the width 1.4in body depth; snout short, 4.75 in head length;orbit diameter 3.2 in head length; interorbitalspace narrow, the least width 15.4 in head length;caudal-peduncle depth 2.3 in head length; caudal-peduncle length 1.55 in head length.Mouth strongly oblique, forming an angle of

about 50° to horizontal axis of body, the lower jawprojecting; mouth large, the maxilla reaching tobelow center of eye, the upper-jaw length 2.2 inhead length; front of upper jaw with four smallincurved canine teeth on each side, the third pairlargest, with an irregular inner row of smallincurved teeth and a pair of very strongly incurved,inwardly depressible canines behind the secondfront teeth and larger than the largest front tooth;

aqua vol. 12 no. 3 - 10 July 2007 92


Table I. Proportional measurements of type specimens of new species of the gobiid fish genus Vanderhorstia as percentagesof the standard length.

Species attenuata auronotata belloides belloides dorsomacula

Type designation holotype holotype holotype paratype holotypeBishop Museum number 38734 32322 32550 40479 32489Standard length (mm) 36.3 28.3 40.5 47.7 39.5Sex male female female male femaleBody depth (at P2 origin) 18.7 18.0 17.3 14.7 15.2Body depth (at A origin) 15.5 13.9 15.0 12.5 13.4Body width 10.9 10.6 10.3 10.0 8.4Head length 27.0 28.3 26.5 25.6 27.0Head width 13.3 11.6 11.6 12.3 10.9Snout length 5.7 4.9 5.1 5.1 5.2Orbit diameter 8.5 7.0 7.4 6.7 7.4Interorbital width 1.4 1.8 1.7 2.1 2.5Cheek depth 1.6 1.4 2.0 2.0 1.7Upper-jaw length 12.4 13.4 11.5 10.9 11.9Caudal-peduncle depth 11.8 8.8 7.4 7.1 8.1Caudal-peduncle length 17.5 12.3 10.8 10.6 11.4Predorsal length 32.1 32.9 30.7 29.7 30.1Preanal length 55.5 53.3 52.5 51.7 53.6Prepelvic length 31.2 28.7 26.3 27.2 27.7Base of dorsal fins 52.1 54.8 32.0 25.7 58.0First dorsal spine 10.9 17.9 19.7 22.8 19.9Fourth dorsal spine 24.6 20.9 32.0 25.7 24.8Fifth dorsal spine 13.7 14.6 15.8 20.1 20.3Spine of second dorsal fin 10.6 10.3 9.8 9.7 10.3Longest dorsal ray 21.5 15.6 13.5 14.8 34.5Base of anal fin 26.4 34.4 34.5 33.8 7.6Anal spine 6.9 7.5 8.1 8.5 14.8Longest anal ray 20.8 15.7 13.8 15.3 33.4Caudal-fin length 46.3 41.3 32.2 38.2 33.4Pectoral-fin length 28.8 27.6 24.7 25.7 26.3Pelvic-spine length 8.3 7.4 7.4 7.5 7.6Pelvic-fin length 19.5 20.5 22.4 23.7 21.5

side of jaw with 13 well spaced slender incurvedteeth and a inner row of smaller teeth; anteriorthird of lower jaw with two to three irregular rowsof incurved teeth, the four front teeth on each sidelargest but smaller than largest upper teeth, fol-lowed medially by a very large recurved canine(largest in jaws) and a single row of small incurvedteeth along side of jaw; no teeth on vomer orpalatines; edge of lips smooth, the inner surfacepapillose; tongue broadly rounded, its upper sur-face with papillae; roof of mouth with well spacedpapillae; no mental flap. Gill opening broad, extending forward to within

a pupil diameter of a vertical at rear edge of orbit;gill membranes attached only anteriorly to isth-mus, with no free fold; gill rakers slender, thelongest nearly equal in length to longest gill fila-ments.Posterior nostril nearly round, with a slight rim,

directly in front of fleshy edge of orbit at level ofcenter of eye, with a large nasal pore directlyanterodorsal; anterior nostril a short membranoustube, higher posteriorly but without a flap,anteroventral to posterior nostril nearly midway toedge of snout above upper lip; cephalic sensorypores typical of the genus; a row of 12 well spaced,prominent, sensory papillae rimming lower part oforbit, commencing behind anterior nostril; a seriesof 9 close-set, small papillae paralleling upper jaw,starting below anterior edge of pupil, then anglinghorizontally across cheek (13 additional papillaecounted, but what seem to be remnants of papillaereach half way across cheek); another series of 9close-set small papillae passing horizontally poste-rior from ventral edge of orbit (remnants suggestthe series extends two-thirds distance to preopercu-lar margin).Scales on body progressively smaller anteriorly;

scales cycloid anteriorly, ctenoid posteriorly, themost anterior ctenoid scales reaching to below baseof third dorsal spine; predorsal scales reaching for-ward to within a half pupil diameter of a vertical atposterior edge of orbit; scales on side of napeextending forward above opercle but not abovepreopercle; cycloid scales on prepectoral area aslarge as body scales; no scales on fins except base ofcaudal fin where present in three rows (countedobliquely), covering a semicircular area, the small-est scales posterior.Origin of first dorsal fin above middle of pelvic-

fin base, the predorsal length 3.1 in SL; dorsalspines slender and flexible, none filamentous; first

dorsal spine 2.5 in SL; fourth dorsal spine longest,4.05 in SL; last membrane of first dorsal fin torn,but probably not reaching origin of second dorsalfin; spine of second dorsal fin 9.45 in SL; penulti-mate dorsal soft ray longest, 4.65 in SL; origin ofanal fin below a point slightly anterior to base ofsecond dorsal soft ray, the preanal length 1.8 in SL;anal spine 12.5 in SL; penultimate anal soft raylongest, 6.4 in head length; caudal fin long, themiddle five rays prolonged into a tapering filamentas long as rest of fin, the total fin length 2.15 in SL;pectoral fins pointed, the twelfth and thirteenthrays longest, reaching to above base of first anal softray, 3.5 in SL; prepelvic length 3.2 in SL; pelvicfins short, not reaching anus, 5.15 in SL; pelvicspine 2.35 in length of longest pelvic ray; pelvicfrenum torn.C o l o r o f h o l o t y p e i n a l c o h o l : pale tan

with four broad diffuse dusky bars on body, onebelow posterior half of first dorsal fin, two belowsecond dorsal fin, each with a dark-edged paleband in lower part, and one posteriorly on caudalpeduncle, the pigment concentrated to form amidlateral darker spot; a dusky brown spot ofabout pupil size above base of pectoral fin; twofaint parallel dark lines extending obliquely acrosscheek and opercle from above posterior end of jaws(the dark edges of an orange band in life); no nar-row dark streak between maxilla and premaxilla;first dorsal fin translucent with a large dusky area,mainly between fourth and fifth spines and extend-ing broadly onto adjacent membranes; second dor-sal and anal fins translucent with a middle duskystripe; caudal fin pale yellowish with a faintly pink-ish broad middle longitudinal band and broaddusky pinkish upper and lower margins; paired finspale yellowish.C o l o r w h e n f r e s h : pale grayish blue, white

ventrally on abdomen and head, with four duskybars on body, the two beneath second dorsal finwith a dark-edged orange band; a brown blotchlarger than pupil just behind upper end of gillopening; postorbital head and nape bluish brownwith oblique, dark-edged, orange bands and elon-gate spots from below and behind eye onto nape,continuing dorsally on body as two irregular rowsof small, dark-edged orange spots that are mainlyadjacent anteriorly and progressively more sepa-rated posteriorly; dorsal fins pale blue, the firstwith oblique dusky orange bands, the second withtwo broad orange stripes; caudal fin pale blue, suf-fused with orange, with a broad orange bar across

aqua vol. 12 no. 3 - 10 July 200793

John E. Randall

base and a broad submarginal basal orange band;anal fin nearly white basally, with a pale blue marginand broad submarginal orange band faintly edgedbelow with pale blue; pelvic fins bluish white. Etymology: This species is named attenuata from

the Latin meaning “drawn out” or “tapering”, inreference to the prolonged tapering filamentformed by the middle five caudal rays.Remarks: The single specimen of Vanderhorstia

attenuata was collected with rotenone at Guadal-canal from a depth of 38 to 48 m without knowl-edge of its habits, but the species probably lives sym-biotically in a burrow with a pair of alpheid shrimps. Of the known species of the genus, V. attenuata is

most similar to V. macropteryx (Franz, 1910), V.auropunctata, V. papilio, all from Japan, and V.nannai from Palau and the Philippines. All have arelatively stout body for the genus, 11 dorsal andanal soft rays, and a low longitudinal scale count(range from 27 to 45). Vanderhorstia attenuata ismost easily separated from the others by theunusual shape of the caudal fin. Vanderhorstiamacropteryx and V. auropunctata have a broadlyrounded caudal fin, while the fins of V. papilio andV. nannai have two pointed posterior projections.Vanderhorstia attenutata also differs from the foursimilar species in its short first dorsal fin and itsdistinctive color pattern. It is closest meristically toV. macropteryx in having 15 median predorsalscales (14-18 in V. macropteryx; counts from Ikedaet al., 1995, who redescribed the species). The pre-dorsal-scale counts for the other three species rangefrom 0 to 11.

Vanderhorstia auronotata n. sp.(Fig. 3, Table I)

Holotype: BPBM 32322, female, 28.3 mm,Indonesia, Molucca Islands, Ambon, Ambon Bay,about 1.5 km west of Poka, sloping silty sand bot-tom, 20-24 m, rotenone, J. E. Randall, 2 October1987. Diagnosis: Dorsal rays VI-I,14; anal rays I,15;

pectoral rays 17-18; scales in longitudinal seriesabout 55; no scales on head, nape, or prepectoralarea; body elongate, the depth 5.55 in SL; orbitdiameter 4.05 in head length; third dorsal spinelongest, 3.95 in SL; caudal fin long and verypointed, 2.5 in SL; pectoral fins reaching to aboveorigin of anal fin, 3.6 in SL; pelvic fins just reach-ing anus, 4.9 in SL; color as in Fig. 3.Description: Dorsal rays VI-I,14; anal rays I,15;

all dorsal and anal soft rays branched, the last tobase (each major branch of last ray divided dis-tally); pectoral rays 17 or 18, the upper two andlowermost unbranched; pelvic rays I,5, all soft raysbranched, the fifth rays joined medially; segmentedcaudal-fin rays 17, 14 branched; upper and lowerunsegmented caudal rays 7; longitudinal scaleseries about 55 (most scales missing), progressivelysmaller anteriorly; transverse scale rows about 15;no scales on head, nape anterior to gill opening, orprepectoral area; median prepelvic scales 10; cir-cumpeduncular scales 13; gill rakers 3 + 13;pseudobranch with 10 short fleshy lobes.Body elongate, the depth at pelvic-fin origin 5.55

in SL; body compressed, the width at pectoral-finorigin 1.4 in body depth; head length 3.55 in SL;head slightly compressed, the width 1.3 in body

aqua vol. 12 no. 3 - 10 July 2007 94


Fig. 3. Holotype of Vanderhorstia auronotata, BPBM 32322, 28.3 mm SL, Ambon, Molucca Islands. Photo by J. Randall.

depth; snout short, the length 5.8 in head length;orbit diameter 4.05 in head length; interorbitalspace narrow, the least width 15.7 in head length;caudal-peduncle depth 3.2 in head length; caudal-peduncle length 2.3 in head length.Mouth oblique, forming an angle of about 30° to

horizontal axis of body, the lower jaw projecting;mouth large, the maxilla nearly reaching a verticalat posterior edge of orbit, the upper-jaw length 2.0in head length; front of upper jaw with four smallincurved canine teeth on each side, the two lateralpairs largest, an inner row of small incurved teeth,and two inner pairs of nearly recumbent canines aslarge as largest canine of front row; side of jaw pos-terior to canines with an outer row of 16 wellspaced progressively smaller incurved teeth, and aninner row of still smaller teeth; front of lower jawwith two to three rows of incurved teeth, the outerrow largest (but smaller than teeth at front of upperjaw), ending about one-third back in jaw with arecurved canine larger than largest upper canines(two on other side, the more posterior largest); sideof jaw posterior to canine with a single row of smallincurved teeth; no teeth on vomer or palatines; roofof mouth with prominent well spaced papillae; edgeof lips smooth, the inner surface papillose; tonguetip rounded, the upper surface and roof of mouthwith scattered papillae; no distinct mental flap. Gill opening broad, extending forward to a verti-

cal at posterior edge of orbit; gill membranesattached only anteriorly to isthmus, with no freefold; gill rakers slender, the longest nearly equal inlength to longest gill filaments.Posterior nostril a moderately large oval aperture

just before fleshy edge of orbit a little above centerof eye, with a slight rim, with the nasal poredirectly dorsoanterior; anterior nostril a shortmembranous tube without a posterior flap,anteroventral to posterior nostril half distance toedge of snout above upper lip; cephalic sensorypores typical of the genus; cheek with a close-setseries of 19 small papillae commencing above mid-dle of upper jaw, paralleling jaw edge for 7 of thepapillae, then continuing slightly obliquely down-ward on cheek, ending a short distance posterior toa vertical at rear edge of orbit; a series of 11 larger,more widely spaced papillae from a pair below ante-rior margin of pupil to behind middle of eye; a rowof 10 close-set small papillae commencing betweenseventh and eighth papillae of suborbital series andpassing posteriorly, parallel to lower series. Scales on body progressively smaller anteriorly;

scales cycloid anteriorly, ctenoid posteriorly (mostscales missing, so the demarcation from ctenoid tocycloid cannot be determined, other than to notethat two scales at origin of second dorsal fin arecycloid); no scales on head, nape, or prepectoralarea; no scales on fins except base of caudal finwhere revealed by scale pockets and one largectenoid scale that lies over posterior end of hypuralplate. Origin of first dorsal fin over origin of pelvic fins,

the predorsal length 3.0 in SL; dorsal spines slen-der and flexible; first dorsal spine 5.6 in SL; thirddorsal spine longest, 4.9 in SL; last membrane offirst dorsal fin nearly reaching origin of second dor-sal fin; spine of second dorsal fin 9.7 in SL; penul-timate dorsal soft ray longest, 6.4 in SL; origin ofanal fin below base of first dorsal soft ray, the pre-anal length 1.9 in SL; anal spine 13.3 in SL; penul-timate anal soft ray longest, 6.4 in head length;caudal fin moderately long and strongly pointed,2.5 in SL; pectoral fins pointed, the eleventh raylongest, reaching to above origin of anal fin, 3.6 inSL; prepelvic length 3.55 in SL; fifth pelvic raylongest, nearly reaching anus, 4.45 in SL; pelvicspine 2.8 in length of longest pelvic ray; pelvic finsjust reaching anus, 4.9 in SL; pelvic frenum presentbut damaged.C o l o r o f h o l o t y p e i n a l c o h o l : pale tan

with four brown bars of one to two scales in widthon body, the scale centers within bars pale; first barcentered below base of fourth dorsal spine, the nextthree bars below base of second dorsal fin; a semi-circular brown bar on base of caudal fin; faint nar-rower brown bars in spaces between broader darkbars, and two before first dark bar; an oblique ovalbrown spot dorsally on opercle; a smaller darkerspot above base of pectoral fin just posterior toupper end of gill opening; a brown (not black)streak on membrane between maxilla and premax-illa; snout light brown; first dorsal fin light reddishbrown, especially posterior to third spine, the pig-ment more on membranes than rays; remainingmedian fins with pale rays and pale reddish brownmembranes; paired fins pale.C o l o r w h e n f r e s h : light bluish gray, white

on abdomen, with four narrow dark brown bars onbody (scale centers within bars paler) and a broaderfifth semicircular bar at caudal-fin base; threebrown-edged orange-yellow lines between darkbars on body, and two before first dark bar (poste-rior yellow lines broken into spots); cheek andopercle with dark-edged orange spots, becoming

aqua vol. 12 no. 3 - 10 July 200795

John E. Randall

more brown dorsally on head; an oblique darkbrown blotch on upper part of opercle, and asmaller darker brown spot above pectoral-fin base;snout dark brown, the upper lip with a broadbrownish orange band; median fins with brownrays, bluish membranes, and brown-edged orange-yellow bands and spots; pelvic fins white. The largedark area anteriorly on the first dorsal fin is notpresent on the specimen (it is the result of blackbackground showing through a clearer part of thefin). The dark distal area of the second dorsal fin isalso due to this part of the fin being clearer.Etymology: This species is named Vanderhorstia

auronotata from the Latin “aurum” for gold and“nota” for mark, in reference to the many brightorange-yellow markings.Remarks: The holotype was collected with

rotenone without knowledge of its habits, but itprobably lives in association with an alpheidshrimp, like others of the genus whose habits areknown.

Vanderhorstia belloides n. sp.(Fig. 4, Table I)

Holotype: BPBM 32550, female, 40.5 mm, PapuaNew Guinea, Madang Province, Tab Island (PigIsland), 5°10.3’S,145°50.7’E, lagoon side, siltysand, 21 m, rotenone, J. E. Randall, 8 November1987.Paratype: BPBM 40479, male, 47.7 mm, samedata as holotype.Diagnosis: Dorsal rays VI-I,16; anal rays I,17;

pectoral rays 18; scales in longitudinal series 66-68;

no scales on head, nape, or prepectoral area; bodyelongate, the depth 6.6 in SL; dorsal spines pro-gressively longer to sixth, 3.95 in SL; caudal finpointed and moderately long, 3.0 in SL; pectoralfins reaching to above anus, 3.8 in SL; pelvic finsnot reaching anus, 4.65 in SL; color as in Fig. 4. Description: Dorsal rays VI-I,16; anal rays I,17;

all dorsal and anal soft rays branched, the last tobase (each major branch of last ray divided); pec-toral rays 18, the upper and lower two unbranched;pelvic rays I,5, all soft rays branched, the fifth raysjoined medially; segmented caudal rays 17, 14branched; upper and lower unsegmented caudalrays 7; longitudinal scale series 66 (68), progres-sively smaller anteriorly; transverse scale rows 19;no scales on head, nape anterior to gill opening, orprepectoral area; median prepelvic scales 9; cir-cumpeduncular scales 13; gill rakers 5 + 15;pseudobranch with 10 short fleshy lobes. Body elongate, the depth at pelvic-fin origin 5.8

(6.8) in SL; body compressed, the width at pec-toral-fin origin 1.7 (1.45) in body depth; headlength 3.8 (3.9) in SL; head slightly compressed,the width 1.5 (1.2) in body depth; snout short, thelength 5.2 (5.0) in head length; orbit diameter 3.6(3.8) in head length; interorbital space narrow, theleast width 15.6 (12.2) in head length; caudal-peduncle depth 3.6 in head length; caudal-pedun-cle length 3.6 (3.65) in head length.Mouth oblique, forming an angle of about 35° to

horizontal axis of body, the lower jaw projecting;mouth large, the maxilla nearly or just reaching avertical at posterior edge of orbit, the upper-jawlength 2.3 (2.35) in head length; front of upperjaw with three incurved small canine teeth on each

aqua vol. 12 no. 3 - 10 July 2007 96


Fig. 4. Holotype of Vanderhorstia belloides, BPBM 32550, 40.5 mm SL, Madang Province, Papua New Guinea. Photo by J. Randall.

side, the lateral pair largest; an inner row of smallincurved teeth, and two inner pairs of nearlyrecumbent canines as large as largest canine offront row; side of jaw posterior to canines with anouter row of 15-20 progressively smaller incurvedteeth and an inner row of still smaller teeth; frontof lower jaw with three rows of incurved teeth, theouter row largest (but smaller than teeth at front ofupper jaw), ending about one-third back in jawwith two large recurved canines, these followed bya single row of small incurved teeth; no teeth onvomer or palatines; edge of lips smooth, the innersurface papillose; tongue tip rounded, the uppersurface and roof of mouth with scattered papillae;no distinct mental flap. Gill opening broad, extending forward nearly to a

vertical at posterior edge of orbit; gill membranesattached only anteriorly to isthmus, with no freefold; gill rakers slender, the longest about equal inlength to longest gill filaments of gill arch.Posterior nostril a large subtriangular aperture in

front of center of eye at fleshy edge of orbit, with aslight rim; anterior nostril a short, slightly tapering,membranous tube with no posterior flap, anteror-ventral to posterior nostril, half distance to edge ofsnout above upper lip; cephalic sensory pores typi-cal of the genus; cheek with close-set series of 24(30) small papillae, commencing above middle ofupper jaw, paralleling jaw edge for about 8 pores,then continuing nearly horizontally onto cheek(obliquely downward), ending a short distanceposterior to a vertical at rear edge of orbit; a seriesof 20 larger, more widely spaced papillae rimmingventral edge of orbit from behind anterior nostrilto behind eye; a row of 10 close-set small papillaecommencing between fifteenth and sixteenthpapillae of suborbital series and angling slightlyupward as it passes posteriorly. Scales on body progressively smaller anteriorly,

the scale rows somewhat irregular, especially ante-riorly; scales cycloid anteriorly, ctenoid posteriorly,the most anterior scales with cteni on side of bodyslightly anterior to a vertical at origin of seconddorsal fin; scales cycloid for two rows below seconddorsal fin; scales ventrally on abdomen and chestcycloid; no scales on head, nape, or prepectoralarea; no scales on fins except for three rows (count-ing obliquely) on broad central base of fin. Origin of first dorsal fin above midbase of pelvic

fins, the predorsal length 3.25 (3.35) in SL; dorsalspines slender and flexible; first dorsal spine 5.1(4.4) in SL; third and fourth dorsal spines longest

and subequal, 3.1 (3.9) in SL, narrowly joined bymembrane distally, the fourth spine extending freeof membrane as a short filament; last membrane offirst dorsal fin nearly reaching origin of second dor-sal fin; spine of second dorsal fin 10.2 (10.3) in SL;penultimate dorsal soft ray longest, 7.4 (6.75) inSL; origin of anal fin below base of first to seconddorsal soft rays, the preanal length 1.9 (1.95) in SL;anal spine 12.4 (11.8) in SL; penultimate anal softray longest, 7.25 (6.55) in head length; caudal finmoderately long and pointed, 3.1 (2.6) in SL; pec-toral fins pointed, the twelfth ray longest, reachingto above origin of anal fin, 4.05 in SL (to abovefirst anal soft ray, 3.9 in SL in paratype); prepelviclength 3.8 (3.9) in SL; fifth pelvic ray longest,nearly reaching anus, 4.45 in SL (nearly reachingorigin of anal fin in paratype, 4.2 in SL); pelvicspine one-third length of longest pelvic ray; pelvicfrenum thin, the membrane nearly reaching tip ofpelvic spines.C o l o r o f h o l o t y p e i n a l c o h o l : pale tan,

the edges of scales dorsally on body brown; a seriesof six indistinct dark blotches along body fromdarker pigment at scale edges; a longitudinal seriesof 26 very short vertical brown lines (due to darkerand broader vertical part of scale edges) from upperend of gill opening to midbase of caudal fin; ablack streak on membrane in groove between max-illa and premaxilla from below posterior nostrilnearly to a vertical through center of eye; fins paleexcept reddish brown distal ends of longest dorsalspines and adjacent membranes and distal end ofmiddle caudal rays. C o l o r o f h o l o t y p e w h e n f r e s h : body

bluish gray with small dark-edged yellow spots(faint ventrally), shading to white on abdomen; arow of larger irregular yellow spots, edged in black-ish, above pectoral fin, followed by about 15 shortbrown vertical lines to caudal-fin base; head bluishgray, becoming dusky pink over operculum, withsmall round yellow spots on cheek, operculum,and branchiostegal membranes; a blackish streakbetween maxilla and premaxilla; second dorsal finpale blue with numerous dark-edged yellow spotsand a white margin; caudal fin longitudinallybanded in pale blue and yellow; anal fin pale bluewith an apparent dark stripe and faint yellow basalmarkings; pelvic fins bluish white. The anal finappears to show a black stripe in the middle, butthere is no indication of a longitudinal zone of darkpigment on the specimen in alcohol, and no indi-cation of a more transparent zone (which, if pre-

aqua vol. 12 no. 3 - 10 July 200797

John E. Randall

sent, might explain the dark stripe as black back-ground color showing through the fin).Etymology: This species is named Vanderhorstia

belloides for its resemblance to V. bella.Remarks: The two Bishop Museum specimens of

this species from Papua New Guinea were longidentified only as Vanderhorstia sp. After thedescription of V. bella by Greenfield & Longe-necker (2005) from one specimen from Fiji, thePapua New Guinea specimens were initiallybelieved to be V. bella because of the similarity incolor pattern. A close comparison, however, hasrevealed the following meristic and morphologicaldifferences: V. bella has one more dorsal and analsoft rays (the species of Vanderhorstia rarely vary inthese two counts); more scales in longitudinalseries (about 77, compared to 66-68 for V. bel-loides); a shorter snout (3.7% SL compared to5.1% for V. belloides), with a steeper dorsal profile;and the fourth dorsal and fifth anal soft rayslongest (penultimate rays longest in V. belloides).There are also some differences in color. The sec-ond dorsal fin of V. bella is described as light yellowwith a distal lavender margin (the fin of V. belloidesis pale blue with irregular dark-edged yellow spots).It should be noted that the small spots on the headand body of V. bella are pale green in the color fig-ures in Greenfield & Longenecker’s (l. c.) paper,but described as yellow by them. The green col-oration is probably an error in color reproduction.Vanderhorstia bella is described as “overlaid bylavender in many areas.” There is no lavender onthe body of V. belloides. If there were variation inblue vs. lavender of the same species with depth,one might expect the deeper water specimen to be

lavender; V. bella, however, was collected 8.3 mand V. belloides in 21 m.

Vanderhorstia dorsomacula n. sp.(Fig. 5, Table I)

Vanderhorstia sp. Randall, 2003: 1, cover photo(Lolobau Island, north coast of New Britain).

Holotype: BPBM 32489, female, 39.5 mm, PapuaNew Guinea, Tab Island (Pig Island), 5°10.3’S145°50.7’E, lagoon side, silty sand, 25 m, spear, J.E. Randall, 2 November 1987.Diagnosis: Dorsal rays VI-I,17; anal rays I,18;

pectoral rays 18; scales in longitudinal series 78; noscales on head, nape, or prepectoral area; bodyelongate, the depth 6.6 in SL; dorsal spines pro-gressively longer to fourth, 4.05 in SL; caudal finpointed and moderately long, 3.0 in SL; pectoralfins reaching to above anus, 3.8 in SL; pelvic finsnot reaching anus, 4.65 in SL; color as in Fig. 5.Description: Dorsal rays VI-I,17; anal rays I,18;

all dorsal and anal soft rays branched, the last tobase (each major branch of last ray divided); pec-toral rays 18, the upper two and lowermostunbranched; pelvic rays I,5, all soft rays branched,the fifth rays joined medially; segmented caudalrays 17, 14 branched; upper and lower unseg-mented caudal rays 7; longitudinal scale series 78(anterior scales small); transverse scale rows 17; noscales on head, nape, or prepectoral area; medianprepelvic scales 10; circumpeduncular scales 13;gill rakers 5 + 15; pseudobranch with 12 shortfleshy lobes.

aqua vol. 12 no. 3 - 10 July 2007 98


Fig. 5. Underwater photograph of Vanderhorstia dorsomacula with Alpheus cf. rapacida, New Britain. Photo by J. Randall.

Body elongate, the depth at pelvic-fin origin 6.6in SL; body compressed, the width at pectoral-finorigin 1.8 in body depth; head length 3.7 in SL;head slightly compressed, the width 1.25 in bodydepth; snout short, the length 5.2 in head length;orbit diameter 3.65 in head length; interorbitalspace narrow, the least width 10.8 in head length;caudal-peduncle depth 3.35 in head length; cau-dal-peduncle length 2.4 in head length.Mouth oblique, forming an angle of about 40° to

horizontal axis of body, the lower jaw projecting;mouth large, the maxilla reaching a vertical at poste-rior edge of orbit, the upper-jaw length 2.0 in headlength; front of upper jaw with three incurvedcanine teeth on each side, followed on side of jawwith 24 progressively smaller teeth, with an innerrow of still smaller teeth; small incurved teeth medi-ally at front of jaw, with a large, inner, nearly recum-bent canine on each side; front of lower jaw withtwo rows of moderate-size incurved teeth, the outerrow largest; two (three on right side) well spacedrecurved canine teeth in an inner row about one-third back from front of jaw, followed by one row ofabout 20 progressively smaller teeth; no teeth onvomer or palatines; roof of mouth with prominentwell spaced papillae; edge of lips smooth; tongue tiprounded; no distinct mental flap. Gill opening broad, extending forward nearly to a

vertical at posterior edge of orbit; gill membranesattached only anteriorly to isthmus, with no freefold; gill rakers slender, the longest about equal inlength to longest gill filaments of first gill arch.Posterior nostril a large, nearly round aperture in

front of center of eye at fleshy edge of orbit, with aslight rim, except posteriorly; anterior nostril ashort membranous tube with no posterior flap,anterorventral to posterior nostril half distance toedge of snout above upper lip; cephalic sensorypores typical of the genus, with the nine pores ofthe oculoscapular canal and the three of the preop-ercular canal; a close-set series of 24 small papillaeabove and parallel to posterior third of upper jaw,continuing slightly obliquely downward as it passesonto cheek, ending a little posterior to a vertical atrear edge of orbit; a series of 23 larger, more widelyspaced papillae rimming ventral edge of orbit fromposterior to anterior nostril to behind center of eye;a short row of 6 close-set small papillae commenc-ing between nineteenth and twentieth papillae ofsuborbital series and angling slightly upward as itpasses posteriorly. Scales on body progressively smaller anteriorly,

the scale rows somewhat irregular, especially ante-riorly; scales cycloid anteriorly, ctenoid posteriorly,the most anterior scales with cteni on lower side ofbody nearly to a vertical at base of sixth dorsalspine; scales in one to two rows below second dor-sal fin cycloid; scales ventrally on abdomen andchest cycloid; no scales on head or prepectoral area;no scales on fins except for three rows on broadcentral base of fin. Origin of first dorsal fin above rear base of pelvic

fins, the predorsal length 3.3 in SL; dorsal spinesslender and flexible, none filamentous (at least infemale holotype); first dorsal spine 6.6 in SL;fourth dorsal spine longest, 4.05 in SL; last mem-brane of first dorsal fin ending at origin of seconddorsal fin; spine of second dorsal fin 9.7 in SL;penultimate dorsal soft ray longest, 6.85 in SL; ori-gin of anal fin below base of first dorsal soft ray, thepreanal length 1.85 in SL; anal spine 13.2 in SL;penultimate anal soft ray longest, 6.75 in headlength; caudal fin moderately long and pointed,3.0 in SL; pectoral fins pointed, the twelfth raylongest, reaching to above anus, 2.75 in SL; pre-pelvic length 3.25 in SL; pelvic fins short, notreaching anus, 4.65 in SL; pelvic spine one-thirdlength of longest pelvic ray; membrane of pelvicfrenum nearly reaching tip of pelvic spines.C o l o r o f h o l o t y p e i n a l c o h o l : pale yel-

lowish with series of seven indistinct brownblotches passing posteriorly from upper end of gillopening above pectoral fin, continuing as longitu-dinal row of 17 narrow short faint brown bars(from pigment on scale edges), to base of caudalfin; an oblique blackish streak on membrane ingroove between maxilla and premaxilla below frontedge and center of orbit; fins pale except for a blackspot slightly larger than pupil in middle of firstdorsal fin between fourth and fifth spines. C o l o r i n l i f e o f a n a d u l t i n d i v i d u a l

f r o m t h e M a d a n g P r o v i n c e o f Pa p u aNe w G u i n e a : body bluish gray with numerousvery small yellow spots (one per scale) and a seriesof narrow dark red bars of two to three scales inlength along side of body posterior to pectoral fin;cheek and opercle below eye blue with purple-edged yellow spots; a narrow black streak betweenpremaxilla and maxilla; dorsal fins translucent bluewith rows of pale yellow spots, the first dorsal witha black spot larger than pupil in middle of finbetween fourth and fifth spines; caudal fin andouter half of anal fin translucent pale yellow withpale blue stripes; pelvic fins translucent pale blue.

aqua vol. 12 no. 3 - 10 July 200799

John E. Randall

Remarks: The holotype and only specimen wascollected with a multiprong sling spear in 25 m atTab Island (also known as Pig Island) in the MadangProvince on the north coast of Papua New Guineain 1987. The same species was photographed under-water at Lolobau Island on the south coast of NewBritain at a depth of 20 m in 2002, but the fish wasnot collected. This goby lives in association with thesnapping shrimp Alpheus cf. rapacida (identificationprovided by Arthur Anker).A second photograph taken at the New Britain

site shows two shrimps out of the burrow in con-tact with the goby This photograph was used onthe cover of the March, 2003 issue of “I.O.P. Div-ing News”, with a short article in Japanese indicat-ing the need for additional specimens.The species of Vanderhorstia most similar in color

to V. dorsomacula is one from the CalamianesIslands, Philippines illustrated as Vanderhorstia sp.in Allen et al. (2003: 312, middle two figures).Their right-hand figure, identified as the female, ismore deep-bodied than V. dorsomacula. The figureto the left is the more slender male. Unfortunatelythey obtained no specimens. Another similarspecies, also identified only to genus, is illustratedby an underwater photograph in Hayashi & Shira-tori (2003: 159, lower figure). It differs from V.dorsomacula in the broadly rounded first dorsal finof very different color.

ACKNOWLEDGEMENTSI thank Loreen R. O’Hara for X-rays, Arnold Y.

Suzumoto for curatorial assistance, and Gerald R.Allen, David W. Greenfield, and Helen A. Randallfor review of the manuscript.

REFERENCESALLEN, G. R., & RANDALL, J. E. 2006. Vanderhorstia

nobilis, a new species of shrimpgoby from Indonesia andthe Philippines. aqua, Journal of Ichthyology and AquaticBiology. 12 (1): 39-44.

ALLEN, G. R, STEENE, R., HUMANN P. & DELOACH, N.2003. Reef Fish Identification Tropical Pacific. 457 pp.New World Publications, Jacksonville, FL.

BIRDSONG, R. S., MURDY, E. O. & PEZOLD, F. L. 1988. Astudy of the vertebral column and median fin osteology

in gobioid fishes with comments on gobioid relation-ships. Bulletin of Marine Science 42: 174-214.

GREENFIELD, D. W. & LONGENECKER, K. R. 2005. Van-derhorstia bella, a new goby from Fiji (Teleostei: Gobi-idae). Proceedings of the California Academy of Sciences 56:619-623.

HAYASHI, M. & SHIRATORI, T. 2003. Gobies of Japanesewaters. 223 pp. TBS Buritanica, Tokyo (in Japanese).

IKEDA, Y., NAKABO, T., & HIRAMATSU, W. 1995.Redescription of Vanderhorstia macropteryx (Perciformes:Gobiidae) with designation of a neotype. Japanese Jour-nal of Ichthyology 42: 303-310.

KARPLUS, I. 1987. The association between gobiid fishesand burrowing alpheid shrimps. Annual Review ofOceanography and Marine Biology 25: 507-562.

KLAUSEWITZ, W. 1974. Eilatia latruncularia n. gen. n. sp.und Vanderhorstia mertensi n. sp. vom Golf von Aqaba(Pisces: Gobiidae: Gobiinae). Senckenbergiana Biologica55: 205-212.

MASUDA, H., AMAOKA, K., ARAGA, C., UYENO, T. &YOSHINO, R. (eds.) 1984. The Fishes of the Japanese Archi-pelago. Vol. 1, text: xxii + 437 pp. Vol. 2, plates. TokaiUniversity Press, Tokyo.

RANDALL, J. E. 2003. Vanderhorstia sp. I.O.P. Diving News14: l (in Japanese).

RANDALL, J. E. 2007. Vanderhorstia opercularis, a newshrimpgoby from the northern Red Sea. Electronic Jour-nal of Ichthyology 3 (1): 18-25.

SENOU, H., SUZUKI, T., SHIBUKAWA, K. & YANO, K. 2004.A Photographic Guide to the Gobioid Fishes of Japan.Heibonsha, Ltd., Tokyo. 534 pp. (in Japanese).

SHIBUKAWA, K., & SUZUKI, T. 2004. Vanderhorstia papilio,a new shrimp-associated goby from the Ryukyu Islands,Japan (Perciformes: Gobiidae: Gobiinae), with com-ments on the limits of the genus. Ichthyological Research51: 113-119.

SMITH, M. M. & HEEMSTRA, P. C. (eds.). 1986. Smiths’ SeaFishes. xx + 1047 pp. Macmillan South Africa, Johannes-burg.

SMITH, J. L. B. 1949. Forty-two fishes new to SouthAfrica, with notes on others. Annals and Magazine of Natural History, (12) 2: 97-111.

SMITH, J. L. B. 1959. Gobioid fishes of the families Gobiidae, Periophthalmidae, Trypauchenidae, Taenioidi-dae and Kraemeriidae of the western Indian Ocean.Ichthyological Bulletin 13: 185-225.

WINTERBOTTOM, R., IWATA, A. & KOZAWA, R. 2005. Van-derhorstia nannai, a new species of burrow-associatedgoby from Palau and the Philippines (Pisces: Gobiidae).aqua, Journal of Ichthyology and Aquatic Biology 9 (3):109-114.

aqua vol. 12 no. 3 - 10 July 2007 100


AbstractHoplolatilus erdmanni is described on the basis of two

specimens, 129.8-137.2 mm standard length, collected inthe vicinity of Triton Bay, Irian Jaya Barat Province (west-ern New Guinea), Indonesia. Approximately 15 individu-als were observed on gentle rubble slopes at depths of 42to 60 m. The fish were seen solitarily, in pairs, or in triosgenerally in close proximity to large rubble mounds, whichthey construct. It is most similar to H. fronticinctus, awide-ranging Indo-west and central Pacific species. Thenew taxon differs in having fewer lateral-line scales as wellas several colour pattern features that include about 15 to17 reddish-brown bars on the side, an obliquely angled,blue-edged yellow-orange stripe behind the eye, a thin bluestripe across the operculum, a neon blue region across theentire caudal peduncle, and vivid red central caudal fin-rays.

ZusammenfassungHoplolatilus erdmanni wird auf der Grundlage anhand von

zwei Exemplaren mit 129,8-137,2 mm Standardlängebeschrieben, die in der Nähe der Triton-Bucht, Provinz IrianJaya Barat (West-Neuguinea), Indonesien, gefangen wur-den. Rund 15 Exemplare wurden über einem flachenGeröllhang in Tiefen von 42 bis 60 Metern beobachtet. DieTiere traten einzeln, paarweise oder in Dreiergruppen auf,gewöhnlich in unmittelbarer Nähe großer Steinhaufen, diesie selber errichten. Diese neue Art ähnelt sehr, in ersterLinie, H. fronticinctus, einer weit verbreiteten Art aus demIndischen Ozean und dem zentralen Pazifik. Sie unterschei-det sich aber durch eine geringere Zahl an Schuppen in derSeitenlinie und durch mehrere Farbmerkmale: 15 bis 17rötlich-braune Streifen an der Seite, ein schräger gelb-orangefarbener Streifen mit blauem Rand hinter dem Auge,ein dünner blauer Streifen über dem Kiemendeckel, einneonblauer Farbbereich über den gesamten Schwanzstielhinweg sowie lebhaft rote mittlere Schwanzflossenstrahlen.

RésuméHoplolatilus erdmanni est décrit sur base de deux spéci-

mens, 129,8-137,2 mm de LS, collectés à proximité de Tri-ton Bay, province de Barat, en Irian Jaya (ouest de la Nou-

velle-Guinée), Indonésie. Près de 15 individus ont étéobservés sur des débris en pente douce, à des profondeursde 42 à 60 m. Les poissons évoluaient en solitaire, en cou-ple ou en trios, généralement très près de grands amas dedébris qu’ils érigent. L’espèce se rapproche le plus de H.fronticinctus qui est largement distribué dans l’Indo-Paci-fique occidental et central. La nouvelle espèce se distinguepar un nombre moins élévé d’écailles sur la ligne latéraleainsi que par plusieurs détails du patron de coloration quiconprennent de 15 à 17 barres brun rougeâtre sur le flanc,une ligne à angle oblique, jaune-orange à liserés bleus, der-rière l’oeil, une fine ligne bleur sur l’opercule, une zonebleu néon sur tout le pédoncule caudal et des rayons cen-traux de la caudale d’un rouge vif.

SommarioHoplolatilus erdmanni è descritto sulla base di due esem-

plari di 129.8-137.2 mm di lunghezza standard, raccolti nel-le vicinanze di Triton Bay, Irian Jaya Barat Province (NuovaGuinea occidentale), Indonesia. Altri individui (circa 15) so-no stati osservati lungo la parete della scarpata continentalea profondità comprese tra i 42 e i 60 m. I pesci si presentava-no solitari, a coppie o in gruppi di tre generalmente pressolarghi tumuli di pietrisco, che essi stessi costruivano. La nuo-va specie è molto simile a H. fronticinctus, una forma ampia-mente diffusa nell’Indo-Pacifico occidentale e nel Pacificocentrale. Ne differisce per avere un minor numero di scagliein linea laterale ma anche per molti caratteri distintivi dellacolorazione che includono 15-17 bande bruno rossastre suifianchi, una striatura obliqua di colore giallo-arancio conmargine blu dietro l’occhio, una sottile stria blu che attra-versa l’opercolo, una regione blu neon che attraversa l’interopeduncolo caudale e, infine, raggi caudali centrali di colorrosso vivo.

INTRODUCTIONThe circumtropical family Malacanthidae, popu-

larly known as tilefishes, is divisible into two sub-families: Latilinae and Malacanthinae (Nelson2006). The latter group contains two genera,Malacanthus Cuvier, 1829 and Hoplolatilus Gün-

aqua vol. 12 no. 3 - 10 July 2007101


Hoplolatilus erdmanni, a new species of sand tilefish (Pisces: Malacanthidae)from western New Guinea

Gerald R. Allen

Department of Aquatic Zoology, Western Australian Museum, Locked Bag 49, Welshpool DC, Perth,Western Australia 6986. E-mail: [email protected]

Received: 19 February 2007 – Accepted: 12 May 2007

ther, 1887, which are common over sand and rub-ble bottoms in the vicinity of coral reefs. All of theknown species, except for M. plumieri (Bloch,1786) of the Atlantic, are distributed in the Indo-west and central Pacific region. The genus Hoplo-latilus largely escaped attention until the advent ofscientific scuba diving due to its relatively deepdwelling habits. These long, slender fishes typicallyhover above burrows, which they construct in softsubstrates. Randall & Dooley (1974) provided the first taxo-

nomic revision of Hoplolatilus, recognizing the fol-lowing five species (general distribution indicatedin parentheses): H. cuniculus Randall & Dooley,1974 (widespread Indo-west and central Pacific),H. fourmanoiri Smith, 1963 (Vietnam to SolomonIslands), H. fronticinctus Günther, 1887 (wide-spread Indo-west Pacific), H. oreni Clark & Ben-Tuvia, 1973 (Red Sea), and H. starcki Randall &Dooley, 1974 (widespread western and centralPacific). Randall (1981) published a second reviewpaper in which three additional species wereincluded: H. chlupatyi Klausewitz, McCosker, Ran-dall & Zetzsche, 1978 (Philippines), H. marcosiBurgess, 1978 (Philippines and Indonesia toSolomon Islands), and H. purpureus Burgess, 1978(Philippines and Indonesia to Solomon Islands. The most recent in-depth treatment of the genus

was published by Earle & Pyle (1997) in theirdescription of a new species, H. pohle from south-eastern Papua New Guinea. They provided a tablecomparing important diagnostic features for the 10species mentioned above as well as H. luteus Allen& Kuiter, 1989 from eastern Indonesia. They alsodiscussed an additional species, H. geo Fricke &Kacher, 1982, described from the Red Sea withouttype specimens on the basis of photographs takenfrom a submersible. Coverage of the family wasalso provided by Dooley (1999), who included asynopsis of defining characters and an illustratedkey to the species occurring in the western Pacificwith the exception of H. luteus and H. pohle. Earle & Pyle (1997) prophetically suggested that

more new species would likely be discovered in thefuture due to the inadequately collected deep sandand rubble habitat that is typical for the genus. Thepresent paper describes the twelfth known memberof the genus, which was first noticed by Mark Erd-mann during a Conservation International marinebiological survey in the vicinity of Triton Bay, IrianJaya Barat Province (southwest New Guinea) ofIndonesia in April 2006. Eight individuals were

sighted on rubble bottoms between 42-60 m. Erd-mann succeeded in spearing one specimen on thisoccasion and another was obtained during a sec-ond visit to the area in January 2007.

MATERIALS AND METHODSLengths of specimens are given as standard length

(SL) measured from the anterior end of the upperlip to the base of the caudal fin (posterior edge ofhypural plate); head length (HL) is measured fromthe same anterior point to the posterior edge of theopercular flap; head depth is measured at the levelof the posterior margin of the preopercle; cheekdepth is measured vertically from the lower rim ofthe orbit to the lower margin of the preopeculum;opercular length is measured from posterior mar-gin of the preoperculum horizontally to the tip ofthe opercular spine; suborbital depth is measuredvertically from the lower rim of the orbit to theventral edge of the head; body depth is the maxi-mum depth taken vertically between the belly andbase of the dorsal spines; body width is the maxi-mum width just posterior to the gill opening;snout length is measured from the anterior end ofthe upper lip to the anterior edge of the eye; orbitdiameter is the horizontal fleshy diameter, andinterorbital width the least fleshy width; upper jawlength is taken from the front of the upper lip tothe posterior end of the maxilla; caudal peduncledepth is the least depth, and caudal pedunclelength is the horizontal distance between verticalsat the rear base of the anal fin and the caudal finbase; caudal fin length is the horizontal lengthfrom the posterior edge of the hypural plate to avertical at the tip of the longest ray; pectoral finlength is the length of the longest ray; pelvic finlength is measured from the base of the pelvic spineto the tip of the longest soft ray. Only the poredscales are counted in the lateral-line between theupper edge of the operculum and the hypuralcrease (excludes three pored scales on caudal-finbase). Gill raker counts are presented as separatecounts for the upper and lower limbs as well as acombined count. The last fin ray element of thedorsal and anal fins is branched near the base andis counted as a single ray.Counts and proportions appearing in parentheses

apply to the paratype if different from the holo-type. If counts of bilateral characters differ on eachside the values are separated by a slash and given asleft/right. Vertebral counts were obtained fromradiographs of both specimens. Proportional mea-

aqua vol. 12 no. 3 - 10 July 2007 102

Hoplolatilus erdmanni, a new species of sand tilefish (Pisces: Malacanthidae) from western New Guinea

surements expressed as percentage of the standardlength are provided in Table I. Type specimens aredeposited at Pusat Penelitian dan PengembanganOseanologi, Jakarta, Indonesia (NCIP) and theWestern Australian Museum, Perth (WAM).

SYSTEMATICS

Hoplolatilus erdmanni n. sp. Triton Tilefish (Figs 1-2; Table I)

Holotype: NCIP 6315, 137.2 mm SL, Northwestentrance to Selat Iris, 03°53.757’S 134°06.638’E,near Triton Bay, Irian Jaya Barat Province, Indone-sia, 55 m, spear, M.V. Erdmann, 25 April 2006.Paratype: WAM P.32860-012, 129.8 mm SL,Northwest entrance to Selat Iris 03°53.757’S134°06.638’E, near Triton Bay, Irian Jaya BaratProvince, Indonesia, 52 m, spear, M.V. Erdmann,30 January 2007.Diagnosis: The new species differs from its con-

geners in having fewer lateral-line scales; about 15to 17 reddish-brown bars on the side; an obliquelyangled, blue-edged yellow-orange stripe behind theeye; a thin blue stripe across the operculum; a neonblue region across the entire caudal peduncle, andvivid red central caudal fin-rays.Description: Dorsal rays X,13 (IX,14); anal rays

II, 12; pectoral rays 16/17 (17); principal caudal-fin rays 17; gill rakers 8 + 19 = 27 (8 + 18 = 26);pored lateral-line scales 80/76 (81/84); transversecheek scale rows 11 (12); transverse opercular scalerows 10; scales above lateral line to origin of dorsal

fin 21 (22); scales below lateral line to origin ofanal fin 33 (36); vertebrae 10 +14.Body moderately elongate, compressed, its great-

est depth 3.7 (3.5) in SL; greatest body width 2.0(1.9) in greatest depth; caudal peduncle depth 2.3,caudal peduncle length 1.6 (1.7), both in HL; headblunt, its length 3.6 (3.7) in SL; head depth 1.2 inHL; snout length 3.9 (3.8) in HL; upper jawlength 2.2 in HL; cheek depth 4.3 (4.8) in HL;opercular length 3.1 (3.1) in HL; snout to verticalmargin of preopercle 1.4 (1.3) in HL; orbit diam-eter 3.8 (4.1) in HL; suborbital depth 4.0 (3.4) inHL; fleshy interorbital width 2.8 (2.9) in HL.Mouth inferior, oblique, extending ventro-poste-

riorly at approximately 28-29 degree angle belowhorizontal axis of body; maxilla reaching verticalabout midway between middle of orbit and poste-rior rim of pupil; front of upper jaw with 2-3enlarged, curved canines on each side of symph-ysis; 4-5 irregular rows of tiny villiform teeth atfront of jaws, posterior to canines, tapering to 1-2rows along side of jaw; posteriormost tooth muchenlarged and directed anteriorly; lower jaw with18-19 enlarged conical teeth in outer row, posteri-ormost tooth enlarged and directed anteriorly; 3-4irregular rows of tiny villiform teeth at front ofjaws, posterior to enlarged outer row teeth, taper-ing to 1-2 rows along side of jaw; palatine, vomer,and tongue edentate. A pronounced wart-like pro-jection of white skin on inner edge of clavicleunder operculum near pectoral-fin base. Lateral-line pores in low arching profile; pores of

cephalic system clearly visible, arranged in pattern

aqua vol. 12 no. 3 - 10 July 2007103

Gerald R. Allen

Fig. 1. Hoplolatilus erdmanni, freshly collected male holotype, 137.2 mm SL, near Triton Bay, Irian Jaya Barat Province,Indonesia. Photo by G. R. Allen.

similar to that of H. fronticinctus as illustrated byRandall & Dooley (1974), including 4 pores oneach dentary, 6 pores on preopercular margin, 8pores in area immediately above preoperculum-operculum, 9 circumorbital pores, and 4 supraor-bital-snout pores; median interorbital pore absent;preoperculum with 17/14 (19/17) serrae, includ-ing enlarged spine at preopercular angle; opercularspine flat, broad-based, slight double curvature,roughly forming equilateral triangle with concavesides and thickened centre, less than pupil diame-ter in length, not extending beyond opercularmembrane; scales extending anterior on head tolevel of posterior rim of orbit; scales generallyctenoid, except mostly cycloid and smaller on headregion; most of caudal fin scaled, remaining finsnaked except pectoral-fin base with small scales.Dorsal fin nearly uniform in height except for

lower anterior spinous portion; base of dorsal fin1.7 in SL; origin of dorsal fin over upper pectoralfin base; predorsal length 3.1 (3.3) in SL ; dorsalspines thin, short, increasing in length; first spineabout one-third length of fourth spine; last dorsalspine 3.3 (3.5) in HL; soft portion of dorsal finnearly uniform in height; first soft ray slightlyahead of level of anus; first soft dorsal rayunbranched, remaining rays becoming increasinglybranched posteriorly, last ray branched at base;penultimate soft dorsal ray notably longer thanadjacent rays, length 1.5 in HL; anal fin nearly uni-form in height, rays slightly shorter than dorsal-finrays; base of anal fin 3.5 (3.4) in SL; origin of finbelow base of first or second dorsal soft ray; two

short anal spines, the first rudimentary, its length2.5 (2.2) in length of second spine; branching offirst soft anal ray barely detectable, remaining raysbecoming increasingly branched posteriorly, lastray branched at base; penultimate soft anal raynotably longer than adjacent rays, length 2.0 inHL; pectoral fins pointed, reaching a vertical atbase of first dorsal soft ray; length of longest pec-toral ray 3.7 (3.8) in SL; all but uppermost pectoralray branched; stout upper ray about one-thirdlength of longest pectoral ray; pelvic fins more orless pointed, their origin slightly anterior to lowerpectoral-fin base, length 6.9 (6.6) in SL, reachingtwo-thirds length of pectoral fins; pelvic-fin spineabout two-thirds length of longest pelvic ray; allpelvic-fin rays branched; caudal fin emarginate, itslength 4.4 (4.9) in SL; upper and lowermost prin-cipal caudal rays unbranched, remaining principalrays branched. C o l o u r o f h o l o t y p e w h e n f r e s h (Fig.

1): head and body generally blue grey, darker dor-sally (especially on head and anterior body); about15 diffuse, yellow bars along dorsal half of body; abroad, blue-edged yellowish-white stripe fromsnout to lower edge of eye, then angling upward toupper margin of preoperculum; cheek mauve;anterior portion of lower operculum bright blue; ayellowish band extending posteriorly from upperrear part of eye to region immediately above pec-toral fin base; lips charcoal or dark grey; belly andthoracic region yellowish white; dorsal, anal, andcaudal fins yellow-orange; tip of upper caudal-finlobe whitish, separated from yellow-orange colour

aqua vol. 12 no. 3 - 10 July 2007 104


Fig. 2. Underwater photograph of Hoplolatilus erdmanni, approximately 160 mm TL, 52 m depth, near Triton Bay, Irian JayaBarat Province, Indonesia. Photo by G. R. Allen.

of fin by diagonal brownish band, which continuesalong dorsal margin of fin to base; pectoral andpelvic fins translucent whitish except uppermostpectoral ray dusky blackish.C o l o u r i n l i f e (Fig. 2): upper half of head

and adjacent antero-dorsal portion of body gener-ally grey, grading to brown along base of posteriorhalf of dorsal fin; about 17, ventrally tapering red-dish-brown bars on side, extending just below mid-lateral axis, interspersed with anteriorly-taperingbluish-white bars of about equal width, posterior-most 2-3 light and dark bars very faint and diffuse;lower half of head and side white to bluish-whiteexcept caudal peduncle which is brilliant neonblue, more intense on upper half; a blue-edged,yellow-orange stripe from snout to lower edge ofeye, then angling upward to upper margin of pre-operculum, continued posteriorly as thin bluestripe to lower base of opercular spine; blue marginon edge of operculum above opercular spine; lipsand cheek pale blue; dorsal fin red with narrowblue margin; anal fin light blue on basal portiongrading to reddish on most of fin; caudal fin redcentrally grading to pale brown or dusky yellowishon upper and lower edges; pelvic fins whitish toblue-white on anterior edge grading to translucentposteriorly; pectoral fins translucent whitish exceptuppermost pectoral ray neon blue and narrow, ven-trally tapered wedge of yellow-orange on upperbase of fin; a bluish area behind opercular spine,immediately above upper origin of pectoral fin.C o l o u r i n a l c o h o l : head and body greyish

brown, darker dorsally with very faint indication of

about 15 narrow brown bars along dorsal two-thirds of body; relatively broad whitish stripe fromsnout to lower edge of eye and continuing behindeye (although faint) to upper margin of preopercu-lum, its upper and lower margins dark grey, espe-cially pronounced on upper margin of snout stripe;fins translucent whitish except for narrow blackishband along distal margin of dorsal fin and duskygrey rays adjacent to dorsal and ventral margins ofcaudal fin; also a diagonal band of grey across dor-sal lobe of caudal fin that is confluent with duskygrey dorsal edge of fin.Remarks: Although admittedly lacking genetic

evidence, Earle & Pyle (1997) suggested thatHoplolatilus was divisible into three groups basedon selected meristic and morphological features.The first group containing H. chlupatyi, H. fron-ticinctus, H. geo, and H. pohle have relatively fewsoft dorsal and anal rays (usually 13 and 12 respec-tively), generally fewer lateral-line scales (81-97),fewer preopercular serrae (1-21) and a relativelydeeper body (3.4-4.1 in SL, except 5.1-5.6 in H.chlupatyi). Members of this group are also knownto construct impressively large mounds at theentrance to their burrows, sometimes approaching1 m in height and 2-3 m in diameter. Hoplolatilusgeo is provisionally placed in the group on the basisof photographic evidence (no specimens have beencollected). The second group contains H. fourmanoiri, H.

luteus, and H. oren. It is characterised by a highnumber of soft dorsal and anal rays (21-23 and 18-20 respectively), moderately high number of lat-

aqua vol. 12 no. 3 - 10 July 2007105

Gerald R. Allen

Fig. 3. Underwater photograph of Hoplolatilus fronticinctus, approximately 160 mm TL, 40 m depth, Triton Bay, Irian JayaBarat Province, Indonesia. Photo by G. R. Allen.

eral-lines scales (92-106), relatively few gill rakerson the first arch (16-19), and a slender body (4.9-6.3 in SL). It may also differ in usually having 11rather than 10 precaudal vertebrae, although thesingle known specimen of H. luteus has 10. The third group containing H. cuniculus, H. mar-

cosi, H. purpureus, and H. starcki has a combina-tion of features that include III to IX rather than Xdorsal spines, 16-34 soft dorsal rays, 14-20 softanal rays, 95-140 lateral-line scales, 20-27 gill rak-ers, and 27-58 preopercular serrae. The new species is clearly a member of the first

group based on its combination of 13 soft dorsalrays, 12 soft anal rays, 76-84 lateral-line scales, 26-27 gill rakers on the first arch, 14-19 preopercularserrae, and relatively deep body (3.5-3.7 in SL).The paratype, which has IX dorsal spines and 14soft rays, is probably aberrant in this respect. Acount of X,13 is also apparent on an uncollectedindividual photographed at the type locality. Mor-phologically H. erdmanni is most similar to H.fronticinctus (Fig. 3). They differ mainly in thenumber of lateral line scales (76-84 for H. erd-manni versus 85-92), although it is difficult to fullyassess this feature on the basis of only two speci-mens. However, the two species exhibit severalimportant colour pattern differences including thepresence of about 15-17 reddish-brown bars in H.erdmanni (body plain in H. fronticinctus), obliquelyangled, blue-edged yellow-orange stripe behind eye(horizontal stripe under eye in H. fronticinctus),thin blue stripe across the operculum (absent in H.fronticinctus), neon blue region across entire caudalpeduncle (forming dorsal saddle only in H. fron-ticinctus), and vivid red central caudal fin-rays(paler red in H. fronticinctus).Hoplolatilus erdmanni is known only from the

type locality and a nearby location near theentrance of Triton Bay in Irian Jaya Barat Provinceof western New Guinea. Both areas are exposed toperiodic strong currents. About 15 individualswere observed on gentle rubble slopes at depths of42-60 m. The fish were seen solitarily, in pairs, ortrios generally in close proximity to large rubblemounds. These constructions were estimated torange from 50-60 cm in height and 1.8-2 m indiameter. When approached too closely or fright-ened by spear shots, the fish quickly retreated intotheir burrow, situated at the apex of the mound.The Triton Bay area is particularly rich for sandtilefishes with regular sightings of H. chlupatyi, H.

fronticinctus, H. luteus, H. cuniculus, H. marcosi, H.pohle, and H. purpureus. Surprisingly, H. starcki,which is generally common throughout easternIndonesia, was not seen. There is a certain degreeof depth zonation for the various species with H.cuniculus occurring between about 30-40 m, H.fronticinctus in about 40-50 m, and the remainingspecies including H. erdmanni below 50 m. Thedeepest dwelling member of the group, H. pohle,was mainly seen below about 55 m. There was nosuggestion of behavioural interaction between themorphologically similar H. erdmanni and H. fron-ticinctus, which were always seen in conspecificgroups.Etymology: The species is named erdmanni in

honour of Mark V. Erdmann, who was the first toobserve this species and collector of the type speci-mens. He has also generously assisted with theauthor’s ichthyological investigations of the Bird’sHead Peninsula of western New Guinea.

ACKNOWLEDGEMENTSI am especially grateful to Conservation Interna-

tional and the Indonesian Department of NatureConservation (PHKA) for sponsoring the 2006expedition and especially to the Walton FamilyFoundation for their interest and generous supportof CI’s Bird’s Head Seascape marine conservationinitiative. I also thank Graham Abbott for assistingwith diving activities on both the 2006 and 2007visits to Triton Bay. The crews of our live-aboardboats, M.V. Citra Pelangi and M.V. Seahorse pro-vided excellent logistic support during these trips.

REFERENCESDOOLEY, J. K. 1999. Brangiostegidae (= Malacanthidae).

In: FAO species identification guide for fishery purposes. Theliving marine resources of the Western Central Pacific. (Eds.K. E. Carpenter & V. H. Niem). 4 (2): 2630-2648. FAO,Rome.

EARLE, J. L. & PYLE, R. L. 1997. Hoplolatilus pohle, a newspecies of sand tilefish (Perciformes: Malacanthidae)from the deep reefs of the D’Entrecasteaux Islands, PapuaNew Guinea. Copeia 1997(2): 382-387.

NELSON, J. S. 2006. Fishes of the World. Fourth Edition.John Wiley & Sons, Hoboken, New Jersey, 624 pp.

RANDALL, J. E. 1981. A review of the Indo-Pacific SandTilefish (Perciformes: Malacanthidae). Freshwater andMarine Aquarium 4 (12): 39-46.

RANDALL, J. E. & DOOLEY, J. K. 1974. Revision of theIndo-Pacific branchiostegid fish genus Hoplolatilus, withdescriptions of two new species. Copeia 1974 (2): 457-471.

aqua vol. 12 no. 3 - 10 July 2007 106


AbstractThe rare and interesting curimatid fish Curimatopsis

microlepis Eigenmann & Eigenmann, 1889 (Teleostei:Characiformes), the generic attribution of which is provi-sional, has been rediscovered recently in the Rio Purusregion, Brazil by Heiko Bleher. The recent specimens aredescribed; characteristics such as the squamation and thecolour, until now poorly known or not known at all, aregiven. Its type locality, Jatuarana, as well as some otherlocalities of the Thayer Expedition in Brazil in 1865 and1866, are discussed and corrected. Faunal lists of these sta-tions and a study of the faunal similarity are reported.

Résumé L’intéressant et rare poisson curimatide Curimatopsis

microlepis Eigenmann & Eigenmann, 1889 (Teleostei:Characiformes), dont l’attribution générique semble provi-soire, a été récemment redécouvert dans le Rio Purusregion, Brésil, par Heiko Bleher. Les spécimens récentssont décrits, avec mention de characters peu connus ouinconnus tels que la scalation et le patron de coloration. Salocalité typique Jatuarana, ainsi que d’autres localités del’Expédition Thayer au Brésil en 1865 et 1866, sont dis-cutées et corrigées. Des listes faunistiques de ces stations,accompagnées d’une étude de similitude faunistique, com-plätent cette note.

ZusammenfassungDer seltene und interessante Barbensalmler Curimatopsis

microlepis Eigenmann & Eigenmann, 1889 (Teleostei:Characiformes), dessen Gattungs-Zuordnung als vorläufigzu gelten hat, wurde kürzlich von Heiko Bleher im RioPurus in Brasilien wiederentdeckt. Die jetzt gefundenenExemplare werden beschrieben; bisher kaum oder garnicht bekannte Merkmale der Beschuppung und Farbge-bung werden genannt. Anschließend wird die Typus-lokalität Jatuarana sowie einige weitere Fundorte derThayer-Expedition in Brasilien in den Jahren 1865 und1866 diskutiert und korrigiert. Es folgen Faunenlisten

dieser Stationen und eine Analyse der faunistischen Sta-tions-Similarität.

SommarioIl raro e interessante curimatide Curimatopsis microlepis Ei-

genmann & Eigenmann, 1889 (Teleostei: Characiformes),la cui attribuzione al genere Curimatopsis deve considerarsiprovvisoria, è stato recentemente riscoperto da Heiko Blehernella regione del Rio Purus, in Brasile. Gli esemplari sonostati descritti; vengono fornite caratteristiche relative allescaglie e alla colorazione, tratti finora poco noti o del tuttosconosciuti. La località tipo, Jatuarana, e altre località men-zionate dalla Spedizione Thayer compiuta in Brasile tra il1865 e il 1866, sono discusse e corrette. Infine, sono ripor-tate le liste delle specie animali presenti in queste stazioniinsieme a considerazioni sulle somiglianze nella compo-sizione faunistica.

INTRODUCTIONIn his revision of the genus Curimatopsis, Vari

(1982) pointed out the peculiarities of Curimatop-sis microlepis Eigenmann & Eigenmann, 1889, acurimatid characiform fish with an isolated posi-tion within the genus by its relatively large size andits numerous scales, whereas the other species areremarkable by their nanism (small size) and regres-sion of the scales. The species is rare, in compari-son with the other pigmy Curimatopsis species,which were always well represented in every expe-dition, with hundreds of specimens collected. Vari(1982) was able to study only two specimens dat-ing from more than a century ago, but these werenot very well preserved and partly without scales:the holotype, MCZ 20344, 83.7 mm SL, from“Jatuarana”, collected by Mr. Navez for the ThayerExpedition (see note), and a specimen MNHN 09-226, 89.0 mm SL, collected by M. Jobert in the

aqua vol. 12 no. 3 - 10 July 2007107


Rediscovery of the rare curimatid fish Curimatopsis microlepisEigenmann & Eigenmann, 1889 (Teleostei: Characiformes) in the Rio Purus basin, Brazil, with a note on some Amazonian localities

of the Thayer Expedition to Brazil in 1865 and 1866

Jacques Géry*

10 Chemin du Plantier, F-24200 Sarlat, France

Received: 08 April 2006 – Accepted: 15 March 2007

*On 15 June 2007 Dr Jacques Géry passed away. An orbituary will be published in issue 12(4)

“Villa des Tonnintins sur la rive gauche du rioSolimõens” (Pellegrin 1909), both representing theonly published biological material of the species.There are unpublished records of three other spec-imens: 1) USNM 268035, collected by M. Gould-ing, 1979 in Rio Téfé; 2) USNM 268867, col-lected in a lago near Berurí, Solimões, 1985 and 3)FMNH 54925, collected by M. Ellis in HubabuCreek, Guyana. The latter is worth investigating, asit was not mentioned by Eigenmann (1912) andperhaps there could be a mistake on the label (forCurimatopsis macrolepis).Recently H. Bleher & N. Khardina collected

additional specimens of this interesting fish in thebasin of the middle and lower Rio Purus, Brazil.The rediscovery justifies the present redescriptionas well as the correction of the type locality of thespecies, along with discussion of other collectingstations of the Thayer Expedition.

MATERIALS AND METHODSMeasurements were taken with digital calipers and

expressed as fractions of standard length (SL) andhead length (HL) in the description, and in per-centages of the SL and HL in Table I for easier com-parison with modern works such as that of Vari(1982). During measurements, the head was mea-sured excluding the opercular membrane, the eyewas measured horizontally, the snout and postocu-lar length are projection lengths, the bony interor-bital was measured posterior to the supraocularbone, the number of pored scales of the lateral line(LL) are indicated in parentheses, and unbranchedsoft rays are given in small roman numerals. Theterm “redescription”, widely used nowadays in tax-

onomy, signifies a supplementary description. Spec-imens are stored in the following institutions:FMNH – Field Museum of Natural History,

Chicago, USAMCZ – Museum of Comparative Zoology, Har-

vard, USAMHNG – Musée d’Histoire naturelle de Genève,

SwitzerlandMNHN – Musée national d’Histoire naturelle,

Paris, FranceMTDF – Staatlisches Museum für Tierkunde,

Dresden, Germany MZUSP – Museu de Zoologia, Universidade de

São Paulo, Brazil USNM – United States National Museum of

Natural History, Washington, USASeveral local terms are used throughout the text

and they are defined as follows: lago (lake usuallyconnected with the main river, but not all), furo(canal usually connecting two water bodies),igarapé (small river) and igapó (inundated zone).

RESULTSRedescription of Curimatopsis microlepis Eigen-

mann & Eigenmann, 1889 (Figs 1 & 3)Material examined: MZUSP 820021, 66.2 mm

SL, Lago Solitario near Baturité, SW of Tapauá onthe middle Rio Purus, Brazil, H. Bleher & N.Khardina, August 2002; MTDF 274931, 1, 64.7mm SL, same data as previous sample; MHNG,58.1 mm SL, same date as previous sample; coll.pers. Géry 1059.1.2003, 63.6 mm SL; coll. pers.Géry N° 1059.1.2006, 72.7 mm SL, Lago Aiapuánear Communidade de Oliveira, lower Rio Purus,Brazil, H. Bleher, 10 September 2006.

aqua vol. 12 no. 3 - 10 July 2007 108

Rediscovery of Curimatopsis microlepis Eigenmann & Eigenmann, 1889, with a note on some Amazonian localities of the Thayer Expedition 1865-1866

Fig. 1. Curimatopsis microlepis, from Lago Solitario near Baturité, SW of Tapauá, after preservation in formalin. Photo by N. Khardina.

Diagnosis: Body shape typical of a curimatid fish,without striking features except the large head witha pronounced postorbital part, the broad, flat pre-orbital and interorbital with the mouth and thenostrils superior, the gape almost as broad as theinterorbital, the relatively long snout, compressedhorizontally, the relatively large dorsal and caudalfins and finally the absence of any black mark suchas humeral or caudal spot, lateral band etc. that arecommon in the group. The dorsal part of the bodyis speckled with tiny chromatophores and the over-all colour is light brownish, with some reflective sil-ver on live fish. The species is immediately identi-fiable by the very small but free maxilla, the rela-tively large number of scales (up to 64 in a longi-

tudinal series and 31 in a transverse series), theincomplete lateral line with no more than 5-13pored scales, associated with a body three timeslarger than the other Curimatopsis species.Description (Table I): Largest SL 72.7 mm; body

relatively deep, 2.95-3.41 in the SL; dorsal fin ante-rior to mid-body, the predorsal distance 1.09-1.26in the postdorsal distance, and fairly elevated, thelargest specimens with the second simple and firstbranched rays elongate, forming a short filamentabout 10-20% longer than the other rays but notextending to the adipose where the dorsal fin isdepressed, ca 30% of the SL; anal fin short, falci-form; pectoral fin very short, not longer than thedistance of its tip to the origin of the pelvic fin;

aqua vol. 12 no. 3 - 10 July 2007109

Jacques Géry

SL 72.65 66.15 64.65 63.55 58.10Largest depth 29.3 33.8 31.8 32.9 32.7 Predorsal 49.0 50.6 49.7 ?45.4 48.8Prepelvic 54.5 54.7 53.3 52.2 52.9Peduncle depth 11.8 11.2 11.9 11.4 11.4Head length 32.5 33.7 33.4 32.5 34.3Eye diameter 25.0 26.0 24.7 25.1 26.8Interorbital 33.0 33.2 32.7 32.2 32.7Maxilla 10.4 11.1 11.1 10.9 11.3Snout 24.5 21.9 20.8 24.8 23.7Postorbital 42.2 47.2 44.4 48.3 48.1

Table I. Proportions of the five specimens of Curimatopsis microlepis from the Rio Purus region (in percentages of the SL forthe depth, predorsal, prepelvic, peduncle depth and HL, and in the HL for the eye, interorbital, maxilla, snout and postor-bital part).

Fig. 2. Biotope of Curimatopsis microlepis, Lago Solitario on the middle Rio Purus near Tapauá. Photo by H. Bleher.

pelvic fin behind level of dorsal-fin origin, the pre-pelvic distance 0.95-0.99 in the postpelvic distance,its tip not extending to first anal ray, with the begin-ning of a filament; peduncle relatively deep andshort, its depth 1.09-1.29 in its length, 11.8% of theSL in the largest specimen; no apparent sexualdimorphism contrary to the other species of thegenus (except perhaps the higher dorsal fin); uppercaudal lobe probably somewhat longer than thelower one (but most of the lower rays partly bittenoff and regenerated), with beginning of elongated

upper outer rays; predorsal region rounded with asmall median crest; postdorsal region rounded, witha small median depression; prepelvic region flat witha fairly broad median depression, the anglesrounded; postpelvic region forming a short, bluntkeel, triangular in section; dorsal rather high andpointed, II,8I or II,9; anal III,7 (first unbranchedray rudimentary, last branched one double); caudallobes well developed, pointed; longitudinal scales (5-7)62-64 in total, including about 3 on caudal base;transverse scales (26?)28-31 from dorsal to pelvic

aqua vol. 12 no. 3 - 10 July 2007 110


Fig. 3a-b.Two specimens of Curimatopsis microlepis alive from Lago Aiapuá, photographed before preservation: a) semi-adult;b) adult with its details of the anterior part. Photos by N. Khardina.

a

b

specimens, to find a complete row of each “category”to be quite certain of this important character(except the predorsal series where the count is some-what subjective).Head large, 2.90-3.10 in the SL; eye quite large,

putatively without a negative allometry, with a nar-row anterior adipose lid, its diameter 3.75-4.10 inHL, situated almost entirely in the anterior half ofthe head; postorbital part of head long, 1.8-1.9times the eye’s diameter and more than twice thesnout (in projection); bony interorbital (behind

base (13-14 above the row corresponding to the lat-eral line and 14-15 from that row to the pelvic); 26-30 scales irregularly set along the predorsal line witha bare zone after the occiput corresponding to 5-6scales, and another one just before the dorsal fin cor-responding to 2-3 scales; about 7 scales transversallyacross the breast, the scales irregular without a dis-tinct series along the median line; 26-30 aroundcaudal peduncle; the scales seem to be fragile(explaining the difficulties of counting them on theold specimens) but it was possible, at least on some

aqua vol. 12 no. 3 - 10 July 2007111

Jacques Géry

1 2 3 4 5 6 7 8

CHARACIDAETetragonopterusT. argenteus + + + +T. chalceus + +

MoenkhausiaM. grandisquamis + +M. chrysargyrea + +M. comma TM. oligolepis + +M. ceros TM. gracilima T M. lepidura + + + M. colletti +Bario steindachneri +Gymnocorymbus thayeri + + + + Ctenobrycon hauxwellianus + + +

AstyanaxA. paucidens +A. multidens T

HyphessobryconH. serpae TH. melazonatus + THemigrammusHe. ocellifer +He. microstomus + He. hyanuary T He. cupreus T +Stichonodon insignis + + Stethaprion erythrops +Iguanodectes spilurus +Phenacogaster pectinatus +Leptobrycon jatuaranae TAphyocharax agassizi

ERYTHRINIDAE Hoplia malabaricus + + + + + +

1 2 3 4 5 6 7 8

Erythrinus erythrinus + +Hoplerythrinus unitaeniatus + + + +

LEBIASINIDAEPyrrhulina brevis + +P. semifasciata + + + +Copella nattereri + + +C. guttata

ANOSTOMIDAE - PROCHILODIDAE Leporinus moralesi +L. fasciatus + +L. f. altipinnis TL. friderici + + +L. pachyurus +L. varia + +L. taeniata + + +L. garmani T + Schizodon fasciatus + + + + +Sch. vittatus + Prochilodus taeniurus +

CURIMATIDAECurimatopsis macrolepis + +C. microlepis TC. spiluropsis + + + + + +C. leucostictus T C. plumbeus TC. serpae TC. alburna +C. a. lineata T C. vittata + C. knerii + C. bimac. trachyst. +Psectrogaster amazonica T +Ps. rutiloides + Potamorhina latior + + + +P. laticeps +P. pristigaster +

Table II (1). Table of presence of characiform species collected by the Thayer Expedition in eight stations. 1 = Lago do Aleixo;2 = Lago José Fernandez; 3 = Paraná do Janauari; 4 = Serpa; 5 = Silves; 6 = Jutahy; 7 = Cudajás; 8 = Jatuarana (for compu-tation of similarities between collecting stations, see text; + = presence; T = presence of type specimens).

the supraorbital) broad, 3.0-3.1 in HL (2.56 mea-sured on the largest specimen, including the supra-orbital); maxilla very short but free, about 9 timesin HL; snout 4.0-4.8 in HL, mouth superior(somewhat like that of a cyprinodontid), the gape,only visible from above, almost as broad as theinterorbital, which, together with the relativelylong snout, compressed horizontally, gives theappearance of a duck’s beak, though much shorter;nostrils without tube, close together, very anterior,completely on the top of the head near the outerend of premaxilla instead of being lateral; anteriornostril round, posterior one bean-shaped, sepa-rated by a narrow space equal to that separating theanterior nostril from the upper lip margin; poste-rior nostril separated from the anterior border ofthe supraorbital by a space twice as large; anteriorpart of palate with one median and two lateral clus-ters of tubercles barely developed and almost indis-tinct in some specimens, situated near the border ofthe upper jaw; gill-membrane united to the isthmusfar forward; gill rakers absent or reduced to veryminute mamilla-like tubercles numbering 45 or 50on the lower limb of the first arch. The colour pattern is quite uniform: some metal-

lic reflections on the fishes in vivo and yellowishdorsal, adipose and caudal fins, absent in formalin;overall colour light brownish on the dorsal partand yellowish on the ventral one (in adults), with-out black marks except a dark dorsal band fromocciput to upper accessory caudal rays, circling thedorsal, about 2 scales broad; and the black longestunbranched dorsal rays visible in vivo. Through

the lens, the following, almost microscopic, pat-tern of freshly preserved fishes can be seen: entirebody and head speckled with chromatophores,rare on abdomen and dense on the dorsal region,where they are concentrated mainly on the caudalhalf of each scale; chromatophores very small anddense on the top of the head, the lips, the maxil-lary border and the antorbital, and less dense butmuch enlarged and conspicuous on the superiorpart of the opercle, which appears very dark,leaden coloured, in contrast to the golden irides-cence of the subopercle and the upper part of eye;numerous chromatophores inside the mouthalong the margin of the lower jaw, representingperhaps the “U-shaped dark bar in base of mouth”described by Eigenmann & Eigenmann (1889a);base of anal fin with a series of parallel lines eachcorresponding to a ray; unbranched rays of dorsaland pectoral fins with dark border anteriorly; darkthin edging on the tips of the dorsal and caudalrays; first anal rays lighter than the rest of the fin.According to the collectors, the fish in vivo hadnothing distinguishing it from the other plaincurimatids. The photograph of the smaller fish,shot in situ after capture in the Lago Aiapuá (Fig.3a) shows a quite ordinary silvery specimen withpale brownish-yellowish paired dorsal, adiposeand caudal, while the close-up of the anterior partof the larger specimen (adult?) shows light browncolour of the flanks and white pectoral fins.Discussion: In the context of a revision that can

be now regarded as a standard in Characiformesichthyology (Vari’s 1982 supplementary descrip-tion), these specimens are smaller than the old ones(largest 89 mm SL), with a smaller snout andinterorbital (probably owing to different methodsof taking measurements) and fewer scales in thelateral line (7 vs 12-13). Without doubt, theybelong to a population of C. microlepis quite dis-tant from the type locality (500 km by air, see notehere below), and they have not evolved much. Thequestion of its attribution to the genus Curimatop-sis is not the purpose of the present paper. Sufficeit to say that the absence of shared derived charac-ters (such as sexual dimorphism), as well as itshabitus and head structures, do not support thistaxonomic position.Note on the type locality “Jatuarana” of Curimatopsis microlepis and on other collecting stations of the Thayer expeditionThe collecting stations of the old Brazilian expe-

ditions are difficult to trace, owing sometimes to

aqua vol. 12 no. 3 - 10 July 2007 112


Stations 2 3 4 5 6 7 8

1 .20 .31 .14 .18 .45 .27 .31

2 .23 .27 .31 .32 .17 .20

3 .18 .32 .27 .33 .19

4 .44 .18 .12 .07

5 .27 .34 .16

6 .23 .19

7 .13

Table II (2). Coefficients of similarity (%) calculated fromthe data of Table II (1):

the loss of labels, but chiefly to the inconsistencies ofmaps before the middle of the twentieth century: forexample the map in Eigenmann (1917) has, amongother discrepancies, the Lago Tefé represented onthe wrong side of the Solimões (Fig. 4). Moreover,the habit of giving the same name to several locali-ties and of changing the names of the localities andeven of the rivers causes problems, as pointed out byDick (1977) with respect to the great Thayer Expe-dition. To cite a few examples: the Solimõesbecomes the Amazon downstream of the mouth ofthe Rio Negro; Manaus, formerly Manaos, was oncetermed Barra, and Forte do Rio Negro; in 30 years,the place name Tapauá has moved about 200 kmeast on the Rio Purus, passing from the mouth ofthe Rio Tapauá to the mouth of the Rio Itaparaná,but persisting west, resulting in two villages with thesame name on the same river (see Annex I). Fortu-nately a number of dubious stations were updatedby Higuchi (1992) and only a few, concerning theAmazon basin, need to be revised here.The lack of information or, worse, wrong informa-

tion, is critical as it affects the scientifically impor-tant type locality (terra typica or locus typicus).

One of the most critical localities is “Jatuarana”where C. microlepis was found. Jatuarana is the com-mon name of a Brycon (Characidae), a food fishcommon in the Amazon; local fishermen had a ten-dency to name a sector of a water body by the nameof the most abundant or interesting fish there.At the end of 1865, Mr. Navez (spelled Naves by

Higuchi 1992), possibly from Manaus, took partin the greatest collecting expedition ever made inSouth America, that of the Nathaniel Thayer Expe-dition, directed by Professor Louis Agassiz andcomprising more than 30 members, includingBrazilian volunteers. Alone, or more probably withLouis Agassiz (according to Böhlke 1954), he col-lected a number of characiform species in a placetermed “Jatuarana”, spelled Jutuarana by Fowler(1948) as well as in a “lake” termed “Hyanuary”. In“Jatuarana”, Navez collected at least 14 species orsubspecies, including five new ones: Hemigrammuscupreus Durbin, 1918; Aphyocharax agassizi (Stein-dachner 1883); Leptobrycon jatuaranae Eigen-mann, 1915; Leporinus fasciatus altipinnis Borodin,1929 and Curimatopsis microlepis Eigenmann &Eigenmann, 1889. In “Lago Hyanuary”, he

aqua vol. 12 no. 3 - 10 July 2007113

Jacques Géry

Fig. 4. The city Teffe (=Tefé) on Eigenmann’s (1917) map with the correct location, but L. Teffe (= Lago Tefé) on the wrongside of the Rio Solimões. The route of the Thayer Expedition is in solid line (red in the original).

secured two new species, Moenkhausia ceros Eigen-mann, 1908 and Hemigrammus hyanuary Durbinin Eigenmann (1918), some topotypes ofCyphocharax plumbeus (Eigenmann & Eigenmann1889), and a number of other species, difficult toseparate from those collected in the same lake byD. Bourget and Major Coutinho, also active mem-bers of the expedition. For example, the tinyCarnegiella schereri Fernández-Yépez, 1950 wasdescribed from “Lake Hyanuary” without record ofthe collector, with the only information: “collectedin Lake Hyanuary (Januari) by Thayer Expeditionto Brazil in January 1866”, implying perhaps anexplanation of the origin of the name of the col-lecting station given by the collectors. This puta-tive explanation is not plausible because a LakeJanuari or Janauari existed before the expeditionand still exists.Despite his important record, Mr. Navez is not

cited among the collaborators of the expedition,neither by Eigenmann (1917) nor Dick (1977);moreover the localities where he collected are not

correctly recorded. His name appears only at theoccasion of the description of new species collectedby the Thayer expedition1. Borodin (1929) mentioned only “Jatuarana,

Brazil; Steindachner (1883) added to the locality“Ein Geschenk des Herrn Prof. L. Agassiz”. OnlyEigenmann & Eigenmann (1889a, b), Eigenmann(1908, 1915, 1917) and Durbin in Eigenmann(1918) added the name of the collector: “Navez”.Dick (1977, map fig. 9) was apparently the first

to pinpoint Jatuarana and to give a sketch locatingthe site on the lower course of the Rio Roosevelt, atributary of the Rio Aripuana (Rio Madeira basin)at about 7°30’ S. This location is accepted withoutcomment by Vari (1982, map fig. 14).A study of the route of the Thayer Expedition

checked against maps shows that the location of

aqua vol. 12 no. 3 - 10 July 2007 114


Fig. 5. Map of Central Amazon region as known during the Thayer Expedition (after H. Kiepert, 1867 – 1:2000000) with

1 Of which only a part has been completely studied yet: Curimatidae,Erythrinidae, Lebiasinidae, Anostomidae and Characidae partim etc.For some reason, Eigenmann could not continue his studies of thecollections of the Museum of Comparative Anatomy at Harvard, andlater sent Haseman to South America to complete the record.

4°

3° 63

63 62

62

Jatuarana given by Dick (1977) is misplaced byabout five degrees of latitude. There is (or was)indeed a village or some houses termed Jatuarana onthe bank of Rio Roosevelt (bottom of OperationalNavigation Chart M-26, U.S. Air Force millionthmap), an almost unexplored region in 1865. Themembers of the Thayer expedition never went thatfar, as confirmed by the narrations of the expeditionand the map in Eigenmann (1917). They concen-trated on the Amazon itself, usually anchoring theirboat and starting from a base camp (chiefly Manaus,but also Tefé-Sitio and some other places) andexploring mainly the lakes and the lower course ofthe tributaries of the central basin, the richest in theworld for freshwater species. Individual expeditionmembers went far upstream in only a few instances:Tocantins, Tapajós, Xingú up to the first cataractsand Negro up to Rio Branco. The Rio Roosevelt isthus to be deleted from the list of candidates, as ifthey had gone that far south, there would have beenmore than a single, remote, collecting station, as inthe other rivers.

Another Jatuarana, the Igarapé Jatuarana nearSamuel on the Rio Jamari (basin of the RioMadeira), is likewise improbable, for the same rea-sons. Fatuaranua, near Mauhes (visited by mem-bers of the expedition, but not by Navez) is alsounlikely, owing to its different spelling, and thesame is true for Igarapé Jutuarana, a small tributaryof the lower Rio Uatumã. Lastly, the “Jatuarana”that appeared recently on maps within the IlhaXiborena (separating the Negro from the Solimões)is modern, being 100 years younger. Only one Igarapé Jatuarana seems to be the

appropriate candidate: according to the two mil-lionth map of the Amazon established by Kiepert(1867) and from the names in use in the secondhalf of the nineteenth century, often of indigenousorigin, this igarapé is easily reachable by boat fromManaus, the principal base-camp: it is situatedabout 40 km downstream, on the left bank of theRio Amazonas at the middle of the big Ilha doCareiro, at mid-distance between the “encontro daságuas” (mouth of Rio Negro) and the next island,the Ilha das Onças, and opposite the Lago do Rei(Fig. 5), often visited by the local fishermen.Nowadays, the names of the igarapés of the left

bank downstream from Manaus (Paraná CoaraMirim and Coara Açú, Igarapé Mocambo) havebeen omitted from maps, whereas two well-knownmodern names were not (or not at their presentplace) on the ancient maps: Rio Puraquequara, thenext rio downstream from Manaus, and Rio Pretoda Eva. This perhaps explains the error in Dick(1977). The name reappears only in the detailedmodern IBGE map of 1998 (after Dick’s research)(Fig. 6). Checking the archives of the MCZ wouldhave shown anyhow that the co-ordinates of Jatu-arana, as given by Agassiz, correspond almost per-fectly to the above described locality (except thatthe latitude is four or five minutes further south),corresponding to the northern bank of the Ilhaopposite the igarapé, which empties into the south-ern bank of the northern arm of the Amazon. Theco-ordinates of the modern map are (at the mouthof the igarapé) 59°40’W and 03°04’S. Higachi’s(1992) putative location of an igarapé or lago Jatu-arana near Barreirinha SW of Parintins does notcorrespond to these co-ordinates and is thus notconsidered as probable.Other discrepancies between ancient maps and

Dick’s (1977) locations could be found (somealready succinctly updated by Higuchi 1992). Itappears that some of the maps produced by Dick

aqua vol. 12 no. 3 - 10 July 2007115

Jacques Géry

F. January, the Ig. Jatuarana L. Arary (marked with red spots).

61

61

aqua vol. 12 no. 3 - 10 July 2007 116


Fig. 6a-b. Maps which show the important localities discussed: a) Município de Manaus today with the lagos Iranduba andJanauari (and other localities discussed extisting today); b)Município de Manaus in 1965: from left to right, Janauari, Aleixoand Jatuarana. The solid circles represent Manacapuru and Manaus 1965. IBGE map 1998 (a) and a drawing by J. Géry (b).

a

b

(1977) are unreliable. The corrections of certainlocalities, when possible, may thus be useful forworkers on Amazonian fauna. They concern espe-cially “Arary”, “Lago Hyanuary”, “Lago Aleixo”(Lago do, or Lago Alexo), “Lago José Fernandez”and “Curupira”, said to be a lake by Dick (1977).“Arary” is mentioned has having been explored by

Stephen Van R. Thayer. Dick (1977) cited onlyone Arary (Arari), a little town located, accordingto his map (fig. 11 in Dick 1977), on Rio Itapi-curu in the State of Maranhão. Indeed, there is anArari between Rio Mearim (into Rio Pindaré) andRio Itapicuru. However, the examination of thefauna cited in the literature (Thayer foundMoenkhausia lepidura but not a single Hemigram-mus marginatus, for example) shows that we aredealing with another, Amazonian, Arary: Higuchi(1992) mentioned a Lago Arari, situated in the Ilhade Marajó (Pará). However, Kiepert’s (1867) map,referred to above, gives a Lago Arary on the rightside of the Rio Amazonas just downstream of themouth of the Rio Madeira (Fig. 5), i.e. on the wayto Serpa (now Itacoatiara, see glossary in Annex 1)where Thayer collected at the same time. Owing tothe similarities of fauna, it is obvious that the“Arary” of Dick (1977) and of Higuchi (1992) are

15 and 10 degrees (of longitude this time) respec-tively away from the putative real one. This error isnot scientifically serious, owing to the absence oftype specimens in the collection. The positions of the other “lagos” Hyanuary, José

Fernandez and Aleixo, are perhaps inaccurate byless than a degree, but the scientific consequencesof an eventual error should not be underestimated,because the faunas are quite different in the acidbrown waters of the Rio Negro and in the less acid,yellow turbid waters of the Solimões upstreamfrom the mouth of the Negro. For example the firstlake, now termed Janauari, depends on the Soli-mões and not on the Negro (with nevertheless apossible connection during floods).Dick (1977) mentioned the Rio Hyanuary “on

the western side of the Rio Negro”; but in his map(fig. 10), he locates the Lago Hyanuary or Januaryon the left bank (eastern side) of the Rio Negroupstream from Manaus and from the other lakes:Lago Irandubo (not in the index, a mistake forIranduba [spelled Ueranduba in Eigenmann’s time]on the Solimões, where Coutinho collected at leastone characid, Moenkhausia lepidura), Lago Aleixo(not in the index) and Lago José Fernandez (alsospelled José-Fernandez). On modern maps, there

aqua vol. 12 no. 3 - 10 July 2007117

Jacques Géry

Fig. 7. Dendogram of the coefficients of similarity between eight Amazonian stations of the Thayer Expedition (see text).

8 Jatuarana

1 Aleixo (Alexo)

6 Jutahy

3 Janauary

7 Cudajas

4 Serpa

5 Silves

2 José Fernandes

20%25%

30%45% 100%

are no such names along the lower Rio Négro.Instead, from east to west we have the IgarapésTarumã Grande and Tarumazinho, Lago Tupé andRio Cuieiras. Lake Hyanuary (also termed Paraná do Januari or

F. Janauari) is cited as being on the left bank of theRio Solimões, connecting it with the Rio Négro,in the descriptions of species collected. According to the nineteenth century map, Janu-

ary is indeed a furo on the left bank of the RioSolimões in the great peninsula do Caldeirão(including Ilha Xiborena), nearly separating theNegro and the Solimões, roughly from Manaca-puru to Ilha de Marchantaria. This area, as in mostparts of the Central Amazon basin, is an intricacyof lagos, furos, igarapés and igapós. Paraná doJanauari is situated east of Ilha Paciencia, aboutopposite the small Ilha J. Venâncio. It empties intothe Solimões and also into the Rio Negro via twofuros. This has some biogeographical and faunalimportance, as a filtration effect of the fauna hasbeen shown in the Caldeirão region through theFuro Paracuuba or other connections (Lamarque,pers. comm.). The Paraná do Janauari exists onmodern maps a little further eastward, owing pos-sibly to the modifications of the Amazon’s coursein almost 140 years (but more probably owing tothe imprecision of old maps). Its co-ordinates werenoted during a stage in Manaus in 1986: 60°05’Wand 03°13’S (HIGUCHI 1996 gives the same longi-tude but 03°12’ in latitude). Janauari and Irandubaare thus close together, if not belonging to the sameexpanse of water. The Lakes Castanha andJanauaca, well known by the fishermen, are some-what west of the same longitude, but on the otherbank of the Solimões. The three “lakes”, submittedto a fairly high fishing pressure from both the fish-ermen and the scientists, are easily reachable byboat from Manaus, where some members of theexpedition had their headquarters.Lago Aleixo is situated in Dick’s (1977) map (fig.

10) on the Rio Negro, but is not mentioned onancient maps. According to the accounts of theirtravels, Thayer explored the “Ilha Cudajaz” (nowIlha Codajás, more than 100 km upstream fromthe mouth of Rio Purus) with Bourget, and“Jutahy” (between Tonantins and Fonte Boa) withJames and Talisman. Alone, he visited Serpa (stillexisting on the other side opposite Itacotiara, map9 in Dick, 1977), Silva (now Silves) (Lake Saraca,map 9 in Dick, 1977 = Furos de Saraca on oldmaps, about 50 km downstream of Itacoatiara),

Lake Aleixo (erroneously located in Dick’s 1977map 9) and Arary (map 11 in Dick, 1977, on RioItapicuru). Concerning the lake Aleixo, Dick(1977) states: “Thayer and Bourget were at LagoCudajas for ten days, 27 November to 6 December(Thayer also collected at Lago Aleixo), and Jameswent to Manacapuru for ten days.”Actually Jutahy, as well as Cudajas, are on the

Solimões well upstream from Manaus, whereasArari, Serpa and Silves are on the Amazon welldownstream from Manaus, downstream from themouth of the Rio Madeira: Thayer alone couldthen have collected only near these localities(including Manaus). It was nevertheless impossibleto find a Lago Aleixo or Alexo, certainly of someimportance, on the old maps. We have thus to rely upon the original label

(archives of the MCZ) which states: station Thayer#073, “Rio Negro near Lago Alexo”, E of ManausLat. 03°05’S, Long. 59°53’ W, (corroborated byHiguchi 1992). Actually the co-ordinates “3°05’/59°53’” are those from a spot just east of the RioNegro (remember that Manaus is exactly on the60th meridian), i.e. where the Amazon starts: themention “Rio Negro near Lago Alexo” would thensignify a beach opposite the mouth of the Solimõesat the meeting of the waters, now at the limit of theextension of Manaus (now extremely populated,about four million inhabitants) towards the east.Alexo should then be only a few kilometres westfrom the type locality of Curimatopsis, the IgarapéJatuarana (Fig. 6b). However, the geographic distribution of the few

species collected in Lago Aleixo (see Annex II, 4),which range from the upper Rio Negro inVenezuela, to Rio Madeira and Rio Xingú, includ-ing the Lago Janauca close to Manaus, is a bitproblematic. For instance Moenkhausia grandis-quamis and Tetragonopterus argenteus have neverbeen found to date in the lower waters of the RioNegro (but have in its upper course), although theyare often sympatric in the Amazon Basin(Tetragonopterus chalceus being vicariant inGuyana). This distribution, together with the sim-ilarities of the fauna with that of the Lago do Cas-tanha, Janauary, José Fernandez etc., indicate thatthe water of Lake Aleixo is more like that of theAmazon than that of the lower Rio Negro. Curupira could be located. According to the map

(fig. 9 in Dick 1977), Curupira is a lake on theParaná Urariá that empties into the Amazon(under the name Paraná do Ramos) at Parintins

aqua vol. 12 no. 3 - 10 July 2007 118


(delimiting the south of the Ilha Tupinambarana),and is also connected with the Rio Madeira. This isone of the most complex regions of the centralbasin, where place names have constantly changed.The lake appears on modern detailed maps at58°30’W and 03°45’S, whereas Higuchi (1992)gives 58°33’W and 03°46’S (not very different).Altogether very few specimens were gathered there.“Lago José Fernandez” cannot be precisely

located. It was probably named after an inhabitantof its border, and the name is now lost. Thearchives of the MCZ contain no localities nor co-ordinates, and mention only: “presumably nearManaus”, which is accepted by Higuchi (1992).In an attempt to solve these hypotheses and other

dilemma, a numerical method was used: twenty-eight coefficients of similarity (formula = 2xC[common species]/A+B [sum of the species of thetwo samples]) were calculated. They concern theeight most critical stations and only the few speciesrecorded in the literature (Annex II). As several col-lectors spent only a brief period of time at each sta-tion, the method has drawbacks: the similarity canreflect that of the collectors rather than that of thefauna, and the incomplete sampling may introducea bias: for example E. Honda spent a year samplingthe Lago do Castanha when he collected about 70characiform species (unpublished), whereas in asimilar biotope, but in only 10 days, members ofthe Thayer Expedition collected less than half ofthis number (taking into account the specimensnot yet completely studied). Despite this bias,which renders the attempted dendogram (Fig. 7)provisional, some inferences could be made. Thecoefficients range from 0.07 (most differentcharaciform fauna) to 0.45 (closest fauna), with anaverage of 0.24. Two clusters can be found, eachbased on a pair of localities with similar fauna:Aleixo and Jutahy, with Janauary, Cudajás andputatively Jatuarana; Serpa and Silves, with JoséFernandez. The most divergent faunas are that ofAleixo cf. Silves and Jatuarana cf. Serpa.Concerning those stations that were wrongly

located or have still not been located, the calcu-lated dendogram is not of much use. Jatuarana’sfauna is rather distinct, being only quite similar tothat of Aleixo (0.31) and very different from thatof Serpa (0.07), Cudajás (0.13) and Silves (0.16).Despite the putative geographical proximity, itbelongs hypothetically to a different faunal group,possibly for ecological reasons in this area of con-fluence of two types of waters.

Despite some inconsistencies, the localities of theThayer expedition are now somewhat betterknown, with the exception of those of the materialtransmitted by Major Coutinho and said to havebeen collected by Dr. Justa, which are withoutlabels. They might have been collected near Man-aus (as suggested by Eigenmann (1917), eventhough Dr. Justa did collect in Rio Paraiba doNorte.To sum up, the type localities of at least seven

characiform species that were incorrectly locatedare corrected as follows:Jatuarana = Igarapé Jatuarana on the left bank of

Rio Amazonas, about 40 km by air east-north-eastof the port of Manaus, Amazonas, Brazil,59°40’W and 03°04’S (the latitude given in thearchives of the MCZ indicate that the bank of theIlha do Careiro faces the igarapé). The ecology ofC. microlepis agrees with the collection details, inthat the igarapé, very close to its mouth into thelago, represents the usual biotope of the species. Lago Hyanuary = Parana do Janauari, on the left

bank of Rio Solimões, about 10 km by air south ofthe port of Manaus, Amazonas, Brazil, 60°15’Wand 03°13’S.“Lago Alexo” (now spelled Aleixo) = small igarapé

into the Amazon near the mouth of the Rio NegroE of Manaus, 59°53’ W and 03°05’S. According tothe co-ordinates given in the archives of the MCZ,it lies about 25 km by air west of Jatuarana, justdownstream from the Encontro Das Águas anddefinitely belonging to the basin of the Amazon(even if the label of the MCZ says “Rio Negro”). Itis now within the limits of the city of Manaus.It is likely that the unknown locality, the “Lake”

José Fernandez, is located in the same regionaround Manaus.Tonnantins (now Tonantins, a village on the

mouth of the Rio Tonantins), where Dr. Jobert col-lected more than 50 characiformes species includ-ing the second known specimen of Curimatopsismicrolepis, is about 50 km downstream from themouth of the Rio Içá and not about 100 kmupstream as shown on the map (fig. 14) in Vari(1982).

ACKNOWLEDGEMENTSHeiko Bleher and Natasha Khardina provided the

biological material and prepared the illustrations,Axel Zarske procured the rare Kiepert map andprepared the electronic version of the Higuchi(1992) document, Heraldo Britski, Claude Weber

aqua vol. 12 no. 3 - 10 July 2007119

Jacques Géry

and Flávio C. T. Lima corrected some geographicaldata. They also improved the manuscript greatlyand are warmly thanked for their help.

REFERENCESBÖHKE, J. 1954. Studies on Fishes of the Family Characi-

dae. N° 6. A Synopsis of the Iguanodectinae. Annals andMagazine of Natural History, London, Ser. 12, 7: 97-104.

BORODIN, N. A., 1929. Notes on some Species and Sub-species of the genus Leporinus Spix. Memoirs of theMuseum of Comparative Zoology at Harvard College 50(3): 269-290, pls.1-17.

DICK, M. M. 1977. Stations of the Thayer Expedition toBrazil.1865-1866. Brevoria, Museum of ComparativeZoology, Cambridge, Mass. 444: 1-37.

EIGENMANN, C. H. 1908. Preliminary descriptions of newgenera and species of Tetragonopterid Characins. Bulletinof the Museum of Comparative Zoology at Harvard College52 (6): 91-106.

EIGENMANN, C. H. 1912. The fresh-water fishes of BritishGuiana, including a study of the ecological grouping ofspecies and the relation of the fauna of the plateau to thatof the lowlands. Memoirs of the Carnegie Museum, Pitts-burgh 5 (67): 1-578, pls.1-103.

EIGENMANN, C. H. 1915. The Cheirodontinae, a subfamilyof minute Characid Fishes of South America. Memoirs ofthe Carnegie Museum 7 (1): 1-99, pls. I-XVII. Addendaand corrigenda Cheirodontinae, id. 7: 1919-1920, xix-xii.

EIGENMANN, C. H. 1917. The American Characidae.Memoirs of the Museum of Comparative Zoology at Har-vard College 43 (1): 1-102.

EIGENMANN, C. H. 1918. The American Characidae.Memoirs of the Museum of Comparative Zoology at Har-vard College 43 (2): 103-208.

EIGENMANN, C. H. & EIGENMANN, R. S. 1889a. Prelimi-nary descriptions of Characinidae. Preliminary descrip-tions of new species and genera of Characinidae. The WestAmerican Scientist 6 (7-8): 7-8.

EIGENMANN, C. H. & EIGENMANN, R. S. 1889b. A Revi-sion of the Edentulous Genera of Curimatinae. Annals ofthe New York Academy of Sciences 4: 409-440.

FOWLER, H. W. 1948. Os peixes de agua doce do Brasil, 1aentraga. Arquivos de Zoologia do Estado de Sao Paulo 6:1-204.

HIGUCHI, H. 1992. An updated list of ichthyological collecting stations of the Thayer Expedition to Brazil.http://www.mcz.harvard.edu/Deparments/Fish/thayer.htm

HIGUCHI, H., 1996. An updated list of ichthyological col-lecting stations of the Thayer Expedition to Brazil. Elec-tronic version (1996)http://www.oeb.harvard.edu//thayer.htm etc.

KIEPERT, H. 1867. Zur Karte des Amazonenstromes.Zeitschrift der Gesellschaft für Erdkunde zu Berlin, 2, 453-454, Tafel V.

MENEZES, N. A. 1976. On the Cynopotaminae, a new

subfamily of Characidae (Osteichthyes, Ostariophysi,Characoidei). Arquivos de Zoologia, S. Paulo 28 (2): 1-91.

PELLEGRIN, J. 1909. Characinidés du Brésil rapportés parM. Jobert. Bulletin du Muséum National d’HistoireNaturelle, 2e série, Paris 15: 147-152.

STEINDACHNER, F. 1883. Beiträge zur Kenntnis derFlussfische Südamerika’s. (IV). Denkschriften der kaiser-lichen Akademie der Wissenschaften in Wien. Mathema-tisch-Naturwissenschaftliche Klasse 46: 1-43, pls. i-vii(Char.: 11-16, 25-26, 32-41).

VARI, R. P. 1982. Systematics of the Neotropical Chara-coid Genus Curimatopsis (Pisces: Characoidei). Smithson-ian Contributions to Zoology 373: 28 pp.

Annex I Ancient toponyms (place names) with their modernaccepted versionsBARRA do RIO NEGRO or FORTE do RIO NEGRO =

Manaus (Manaos)BASTOS = a farm on the Rio Alegré, tributary of alto Rio

GuaporéBRAGANCA (Haseman) = Bragança near the Atlantic

coast east of the State of Pará CAICARA = upstream from TéféCARERO, also spelled CAREIRO = Ilha do CareiroCIPO (RIO): tributary of the Itapicuru, Eastern Brazil,

possibly connected with Rio S. Francisco (see Howes,Bull. BM(NH), Zool. ser. 43(1): 39, 1982, concerningthe Coll. Ch. Cumberland). F. Lima (pers. comm.09.2005) corrected the name of Itapicuru into Rio dasVelhas (according to the locality mentioned by Låtken),which surely belongs to the São Francisco drainage.

COARY = Coari (between Codajás and Téfé)CODAJÁS or CUDAJAS or ILHA CODAJAZ, Solimões

about 100 km upstream from Rio Purus (= Casa de Silvade Solimões ? between Manacapuru and Coari)

CUPAI (RIO) = not on recent and ancient maps. Accord-ing to Böhlke (1954), probably from the label of Piabucaspilura Günther (coll. Stevens): “branch of the Amazon,800 mi. from the sea, Pará, Brazil”. By air, 800 miles(1287 km) correspond to the distance between the seaand Manaus (State of Amazonas); thus, there is only onealternative: 1) the 800 miles were calculated by the routeof the boat, and concern the smallest mile, the statuteone (1.609 km). By the northern course, the state of Paráis passed at Rio Nhamundá (about 610 statute miles,about 980 km). Only the southern route (Belém-Breves-Pauxís-Gurupá etc.) is credible. The only toponym atthe far west of State of Pará that can be confused with“Cupai” is Lago Grande do Curuai, about 700 milesfrom the sea on the right bank, at the same longitude asÓbidos; 2) The other alternative, which is favoured here,is a labelling error, which might have been “80 mi. fromthe sea (at Belém)” and would explain the distance inmiles. Stevens collected mainly around Belém.

CURUPIRA (visited by Coutinho, Thayer Expedition,before December 6, 1865) = not on ancient maps, shownon the USAF Operational Aeronautical Chart but not

aqua vol. 12 no. 3 - 10 July 2007 120


named, existing in the detailed modern maps as a smalllake on the southern part of the Ilha Tupinambarana, emp-tying into the Paraná Urariá through the Rio Curupira,and a small Lago Maria Cupira a few km south, on theother side of the Paraná. There is also a Lago Curipira a fewkilometres from Oriximiná on Rio Trombetas on the leftside, State of Pará, near Óbidos, which was visited bymembers of the Thayer Expedition, but in January, 1866.

EGA (EGAS or EDA) = TeféFONTEBOA (Thayer) = Fonte Boa, upstream from

mouth of rio JuruáFORTE DO RIO NEGRO = see Barra do Rio NegroFORTE PAUXIS = see Pauxís = ÓbidosFORTE PRINCIPE = Forte Principe da Beira, on the Rio

GuaporéGURUPA (Thayer) = CurupáIÇÁ = Santo Antônio do Içá at the mouth of the Rio IçáIRISANGA = Oriçanga, near Rio Mogi-Guaçu (MENEZES,

1976)ITACOTIARA = see Serpa; probably new name for São

Antonio de MaraJAMUNDA = Rio Nhamundá, between states of Pará and

AmazonasJANUACA, JANUARY = see lagoJATUARANA = Igarapé Jatuarana downstream from mouth

of Rio Negro, left bank of Solimões, opposite Ilha doCareiro

JUTAHY = Foz de Jutaí, Solimões between Tonantins andFonte Boa

JUTICA = downstream from TéféLAGO ALEXO or ALEIXO (Thayer) = unknown (not on

maps); probably not far from Manaus (Fig. 6b)LAGO HYANUARY = lake and furo S of Manaus, emp-

tying into SolimõesLAGO JANAUCA (and LAGO CASTANHA) = lagos

emptying into the Solimões on the right (southern) sideabout at mid-distance between the longitudes of Man-acapuru and of Manaus

LAGO JOSE ASSU = Lago José Uaçu near the mouth ofRio Tupinambaranas, just east of Parintins, co-ordinates2°37 S and 56°38 W according to Böhlke (1954)

LAGO JOSÉ FERNANDEZ or JOSÉ-FERNANDEZ(Thayer) = not on maps; collected by Major J. M. S.Coutinho Nov.-Dec. 1865, “presumably near Manaus”(Archives of the MCZH)

LAGO MANACAPURU = name unchanged, left side ofAmazon between Vila Rica and mouth of Rio Negro

LAGO MAXIMO = east of Villa Bella (Parintins); accord-ing to Böhlke (1954), connected with the Furo de Ramosnear Parintins; near the junction with Amazon – revisitedrecently by Heiko Bleher (2004)

MACACOS (RIO): 1- near Rio de Janeiro (not connectingthe Rio Pará with the Amazon, as stated by Eigenmann)

MACACOS (Rio dos): 2- also a tributary of Rio das Vel-has in Minas Gerais

MACIEL = a farm on the Rio Barbados, tributary of altoRio Guaporé

MAI JOANA (IGARAPÉ do) = close to the intersection of

the Rio Negro with the Rio Amazonas proper, emptyinginto the Rio Negro according to J. Böhlke (1954); notnear Manacapuru

MANAOS = Manaus, formerly Forte do Rio Negro(1788) or Barra do Rio Negro, or Barra

MATO GROSSO = Vila Bela, Villa Bella de MattoGrosso, today Vila Bela da Santissima Trinidade

MATURA (1788) = Porto do Moz, or Foz do Moz(Thayer), near the mouth of Rio Xingú = Porto de Moz

MAUHES = MauésMONTE ALEGRE (Thayer) = downstream from San-

tarém, on the left Amazon River bankOBYDOS or OBIDOS = see Forte PauxísPARA = SANTA MARIA de BELEM = BelémPARU (1788) = AlmeirimPARINTINS = see Villa Bella da ImperatrixPAUXIS (FORTE PAUXIS 1788) = Óbidos on the lower

Amazon on Eigenmann’s map, actually just downstreamof the mouth of the Rio Trombetas

PEDREIRA = downstream from mouth of Rio BrancoPORTO do MOZ = Porto de Moz, near the mouth of Rio

XingúRAMOS, rio (Thayer) = Boa Vista do Ramos, on the Pa.

do RamosSAO PAULO (Thayer) = São Paulo de Olivença, SolimõesSAO ANTONIO de MARA: see SerpaSERPA = Itacoatiara, downstream from mouth of Rio

Madeira, on the left side of the Amazon riverSILVA = Silves, NE Itacoatiara SANTAREM = see Tapajós (1788)TAJAPURU or TAJUPURU (Thayer) = Caruma ? near

Gurupá = CurupáTAPAJOS (1788) = SantarémTEFFE = Tefé TONNANTINS (also TONNINTINS) = Tonantins,

between Foz de Jutaí and Santo Antônio do IçáVILLA BELLA da IMPERATRIX (Amazon) = on a 1867

map, situated downstream from the Ilhas Pacoval and dasOnças, upstream (and on the opposite side) from mouthof Rio Nhamundá; now Parintins

Annex II (1)Characiformes species sympatric with C. microlepis inthe locus typicus, Jatuarana (in the literature):Moenkhausia lepidura (Kner, 1859)Moenkhausia collettii (Steindachner, 1883)Tetragonopterus argenteus Cuvier, 1816Hemigrammus cupreus Durbin, 1918Leptobrycon jatuaranae Eigenmann, 1915Aphyocharax agassizi (Steindachner, 1883)Iguanodectes spilurus (Günther, 1864)Hoplias malabaricus (Bloch, 1794)Leporinus fasciatus altipinnis Borodin, 1929Semaprochilodus taeniurus (Humboldt & Valenciennes, 1833)Cyphocharax spilurus (Günther, 1864)Potamorhina latior (Spix, 1829)Psectrogaster rutiloides (Kner, 1859)

aqua vol. 12 no. 3 - 10 July 2007121

Jacques Géry

Annex II (2) Characiform species sympatric with C. microlepis in Ton-nantins (now Tonantins, between Foz de Jutaí and SantoAntônio do Içá), after Pellegrin 1909 (not checked):Tetragonopterinae spilurus Valenciennes, 1849Tetragonopterinus gibbicervix Pellegrin, 1909 (or Ctenobrycon spilurus (Valenciennes, 1849))Tetragonopterinus chalceus (Agassiz, 1829)Bryconops melanurus (Bloch, 1794)Astyanax abramis (Jenyns, 1812)Charax gibbosus (Linnaeus, 1758)Moenkhausia grandisquamis (Müller & Troschel, 1844)Moenkhausia collettii (Steindachner, 1883)Aphyocharax alburnus (Günther, 1869)Aphyocharax agassizi (Steindachner, 1883)Lutkenia insignis Steindachner, 1876Roeboides affinis Günther, 1868Cynopotamus limaesquamis (Cope, 1878) (= amazonus (Günther, 1868) ?)Cynopotamus kneri (Steindachner,1878)Acestrorhynchus falcirostris (Cuvier, 1819)Boulengerella maculata (Valenciennes, 1849)Cynodon gibbus Spix, 1829Myletes macropomus Cuvier, 1818 (= Myletes nigripinnisCope, 1878, a synonym)Pygocentrus nattereri (Kner, 1860)Brycon stubelii Steindachner, 1883Iguanodectes spilurus (Günther, 1864)Hoplias malabaricus (Bloch, 1794)Pyrrhulina semifasciata Steindachner, 1875Copeina guttata (Steindachner, 1875)Nannostomus unifasciatus Steindachner, 1876Rhytiodus microlepis Kner, 1859Leporinus friderici (Bloch, 1794)Leporinus bahiensis Steindachner, 1875 (?)Leporinus hypselonotus Günther, 1868Anostomus vittatus Valenciennes, 1849Anostomus trimaculatus Kner, 1859Anostomus fasciatus Agassiz, 1829Anostomus taeniatus Kner, 1859Memiodus gracilis Günther, 1864Hemiodus semitaeniatus Kner, 1859Hemiodus longiceps Kner, 1859Anodus elongatus Spix, 1829Caenotropus labyrinthicus (Kner, 1859)Chilodus punctatus Müller & Troschel, 1844Psectrogaster ciliata (Müller & Troschel, 1844)Curimatella dorsalis (Eigenmann & Eigenmann, 1889)

(under the name Curimatella alburna var. caudimaculataPellegrin, 1909)Curimatus bimaculatus Steindachner, 1876Curimatus plumbeus Eigenmann & Eigenmann, 1889 Curimatus leuciscus Günther, 1868Curimatus vittatus Kner, 1859Curimatus cyprinoides (Linnaeus, 1766)Curimatus latior Spix, 1829Prochilodus ortonianus Cope, 1878Prochilodus amazonensis Fowler, 1906Prochilodus taeniurus (Humboldt & Valenciennes, 1833)

Annex II (3)Characiform species sympatric with C. microlepis in theLago Solitario and the Lago Aiapuá, Rio Purus region:Triportheus angulatus (Spix, 1829)Chalceus erythrurus (Cope, 1870) auct.Ctenobrycon hauxwellianus (Cope, 1870)Hyphessobrycon species, close to minimus Durbin, 1909Moenkhausia chrysargyrea (Günther, 1864)Hemigrammus bellottii (Steindachner, 1882)Hemigrammus levis Eigenmann, 1908Poptella compressa (Günther, 1864)Acestrorhynchus minimus Menezes, 1969Mylossoma duriventre (Cuvier, 1818)Metynnis (Myleocollops) hypsauchen (Müllet & Troschel,1844) Serrasalmus rhombeus (Linnaeus, 1766))Crenuchus spilurus Günther, 1863Potamorhina pristigaster (Steindachner, 1876)

Annex II (4) Characiform species from the “Lago do Aleixo”Tetragonopterus argenteus Cuvier, 1816Moenkhausia grandisquamis (Müller & Troschel, 1844)Astyanax paucidens (Ulrey, 1894)Hoplias malabaricus (Bloch, 1794)Erythrinus erythrinus (Bloch & Schneider, 1801)Pyrrhulina brevis Steindachner, 1875 Psectrogaster amazonica Eigenmann & Eigenmann, 1889(lectotype)Cyphocharax spiluropsis (Eigenmann & Eigenmann, 1889)(cited as spilurus)Cyphocharax leucostictus (Eigenmann & Eigenmann, 1889)(paralectotype)Psectrogaster rutiloides (Kner, 1859) (cotype of isognatus)Potamorhina latior (Spix, 1829).

aqua vol. 12 no. 3 - 10 July 2007 122


AbstractRivulus lazzarotoi n. sp., a member of the R. santensis

species group from the Jurumirim River basin, southeasternBrazil, is described. It is hypothesized to be closely related toR. simplicis as both uniquely have numerous neuromasts inthe infraorbital series. Rivulus lazzarotoi differs from R. sim-plicis in having its dorsal-fin origin posteriorly inserted anda longer pectoral fin in males. It is distinguished from allspecies of the R. santensis group in possessing minute pelvicfins, a unique colour pattern on the anal fin in males and alarge and round caudal spot in females. Monophyly of theR. santensis species group is further supported by a derivedarrangement of infraorbital neuromasts.

ResumoRivulus lazzarotoi, espécie nova, um membro do grupo de

espécies R. santensis da bacia do rio Jurumirim, sudeste doBrasil, é descrita. Ela é hipotetizada ser estreitamenteaparentada a R. simplicis pelas duas exclusivamente pos-suírem numerosos neuromastos na série infra-orbital. Rivu-lus lazzarotoi difere de R. simplicis por possuir a origem danadadeira dorsal inserida posteriormente e nadadeirapeitoral mais longa em machos, e se distingue de todas asespécies do grupo R. santensis por possuir nadadeiras pélvi-cas diminutas, um padrão de colorido exclusivo denadadeira anal e mancha caudal grande e redonda emfêmeas. O monofiletismo do grupo de espécies R. santensis éadicionalmente sustentado por uma distribuição derivada deneuromastos infra-orbitais.

ZusammenfassungRivulus lazzarotoi n. sp. wird beschrieben, der zur R. san-

tensis-Artengruppe aus dem Becken des Jurumirim Flusses,SO-Brasilien, zugeordnet wird. Die Hypothese einer engenVerwandtschaft mit R. simplicis wird aufgestellt, da beideArten zahlreiche infraorbitale Neuromasten aufweisen. Dieneue Art unterscheidet sich aber von R. simplicis durch dieweiter hinten ansetzende Rückenflosse und eine längereBrustflosse bei den Männchen. Von allen anderen Artender R.-santensis-Gruppe unterscheidet sich die neue Artdurch besonders kleine Bauchflossen, ein unverwechsel-

bares Farbmuster an der Analflosse bei Männchen undeinen großen, runden Schwanzfleck bei Weibchen. DieMonophylie der R. santensis-Artengruppe wird weiterhindurch eine abgeleitete Anordnung der infraorbitalen Neu-romasten gestützt.

RésuméRivulus lazzarotoi n. sp., un représentémet du groupe d’e-

spèces R. santensis du bassin de la rivière Jurumirim, ausud-est du Brésil, est décrit. On écrit l’hypothèse de prox-imité avec R. simplicis puisque les deux espèces sont seulesà avoir de nombreux neuromastes dans les séries infra-orbitaires. Rivulus lazzarotoi se distingue de R. simplicispar une insertion postérieure de l’origine de l’anale et deplus longues pectorales chez les mâles. Il se distingue detoutes le espèces du groupe R. sanensis par de minusculesnageoires pelviennes, un patron de coloration unique del’anale des mâles et une grande tache arrondie sur la cau-dale des femelles. La monophylie du groupe R. santensis estrenforcées par un arrangement dérivé de neuromastesinfra-orbitaires.

SommarioSi descrive Rivulus lazzarotoi n. sp., un membro del

gruppo di specie R. santensis raccolto nel bacino del fiumeJurumirim, Brasile sudorientale. Si ipotizza che sia stretta-mente imparentato con R. simplicis, poiché entrambi con-dividono la caratteristica unica di avere numerosi neuro-masti nella serie infraorbitale. Rivulus lazzarotoi differisceda R. simplicis per aver l’inserzione della pinna dorsale piùarretrata e una pinna pettorale più lunga nei maschi. Si di-stingue da tutte le specie del gruppo R. santensis per averepinne pelviche minute, una colorazione particolare dellapinna anale nei maschi e una grande macchia caudale ton-deggiante nelle femmine. L’origine monofiletica delgruppo di specie R. santensis è ulteriormente confortatadalla disposizione dei neuromasi infraorbitali.

INTRODUCTIONThe New World killifish genus Rivulus Poey,

aqua vol. 12 no. 3 - 10 July 2007123


Rivulus lazzarotoi n. sp. (Teleostei: Cyprinodontiformes: Rivulidae), a new killifish from the Jurumirim River basin, southeastern Brazil

Wilson J. E. M. Costa

Laboratório de Ictiologia Geral e Aplicada, Departamento de Zoologia, Universidade Federal do Rio deJaneiro, Caixa Postal 68049, CEP 21944-970, Rio de Janeiro, RJ, Brasil. E-mail: [email protected]

Received: 14 November 2006 – Accepted: 06 June 2007

1860 is represented by the assemblage known as theRivulus santensis species group in the river basins ofthe eastern coast of Brazil (Costa 1998, 2003, 2006).Species of this assemblage are found in shallowstreams and pools within or close to rainforests. Itcomprises seven species: Rivulus depressus Costa,1991, from the João de Tiba River basin, Bahia state;R. nudiventris Costa & Brasil, 1991, from theItapemirim River basin, Espírito Santo state; R.janeiroensis Costa, 1991, from an area between theMacabu and Guandu river basins, Rio de Janeirostate; R. simplicis Costa, 2004, from the coastal plainsaround Paraty, Rio de Janeiro state; R. santensis Köh-ler, 1906, from the area between Ubatuba, São Paulostate, and Guaratuba basin, Paraná state; R. luelingiSeegers, 1984, from the area between Guaratubabasin, Paraná state, and the coastal basins of northernSanta Catarina state; R. haraldsiolii Berkenkamp,1984, from the coastal river basins of Santa Catarinastate (Köhler 1906; Berkenkamp 1984; Seegers1984; Costa 1991, 2004; Costa & Brasil 1991). Anew species with a distinctive colour pattern andminute pelvic fins, collected in the Jurumirim Riverbasin, Rio de Janeiro state, is described here.

MATERIALS AND METHODSMeasurements and counts follow Costa (1995).

Measurements are presented as percentages of stan-dard length (SL), except for those related to headmorphology, which are expressed as percentages ofhead length. Fin-ray counts include all elements.Number of vertebrae, gill rakers and pectoral,pelvic and caudal-fin rays were recorded only fromcleared and stained specimens. The compoundcaudal centrum was counted as a single element.Meristic characters for the holotype are given inparentheses. Osteological preparations were made

according to Taylor & Van Dyke (1985). Termi-nology follows Hoedeman (1958) for frontal squa-mation and Costa (2001) for cephalic neuromastseries. The abbreviation c&s means specimenscleared and stained for bone and cartilage. Mater-ial is deposited in UFRJ, Departamento de Zoolo-gia da Universidade Federal do Rio de Janeiro, Riode Janeiro, Brazil.

Rivulus lazzarotoi n. sp.(Figs 1-2)

Holotype: UFRJ 6435, 7, 29.7 mm SL; Brazil:Estado do Rio de Janeiro: Município de Angra dosReis, stream close to the road RJ-155, rio Juru-mirim basin, 22°55’24.7”S 44°18’51.4”W, altitude9 m; W. J. E. M. Costa, C. P. Bove, B. B. Costa &J. P. Moreira, 11 March 2006.Paratypes: UFRJ 6436, 4 7, 21.7-27.8 mm SL, 26, 22.0-23.8 mm SL; collected with holotype.UFRJ 7211, 1 7, 31.4 mm SL, 8 6, 21.0-25.7mm SL; same locality; H. Lazzaroto, D. Eberienos,I. Montenegro & J. Prata, 22 August 2005. UFRJ7213, 2 7, 17.3-26.5 mm SL, 2 6, 19.0-27.1 mmSL (c&s); same locality; H. Lazzaroto, M. Rhein-gantz & R. Curcio, 1 October 2005.Diagnosis: Rivulus lazzarotoi is distinguished

from all other congeners in having the combina-tion of the following features: infraorbital neuro-masts 1 + 26-27 + 1, arranged in a zigzag row;pelvic fin minute, tip not reaching anus, pelvic-finlength 4.4-6.7% of SL; pectoral-fin length 19.0-21.2% of SL; anal fin orange-yellow to reddishorange on distal half, with distal narrow black mar-gin in males; large, round black spot on dorsal por-tion of caudal-fin base in females.

aqua vol. 12 no. 3 - 10 July 2007 124


Fig. 1. Rivulus lazzarotoi, UFRJ 6435, male, holotype, 29.7 mm SL (some hours after collection); Brazil: Estado do Rio deJaneiro: Angra dos Reis. Photo by W. J. E. M. Costa.

Description: Morphometric characters appear inTable I. Largest male examined 31.4 mm SL;largest female examined 25.7 mm SL; dorsal pro-file slightly convex, almost straight, from snout toend of dorsal-fin base, approximately straight oncaudal peduncle; ventral profile gently convex fromlower jaw to anal-fin origin, nearly straight alongcaudal peduncle; body slender, subcylindrical ante-riorly, slightly wider than deep, to compressed pos-teriorly; greatest body depth at level of pelvic-finbase; jaws short, snout blunt to slightly pointed.Dorsal and anal fins rounded in both sexes,

slightly longer in males; caudal fin elliptical; pec-toral fin rounded, its posterior margin reachingvertical anterior to pelvic-fin base, about 60-70%distance between bases of pectoral and pelvic fins;pelvic fin minute, tip not reaching anus; pelvic-fin

bases medially separated by short interspace, aboutequal to pelvic-fin base width; dorsal-fin origin invertical through base of eleventh or twelth anal-finrays, between neural spines of vertebrae 21 and 23;anal-fin origin between pleural ribs of vertebrae 14and 16; dorsal-fin rays 8-9 (9); anal-fin rays 14-15(15); caudal-fin rays 32-34; pectoral-fin rays 14;pelvic-fin rays 5-6.Scales large, cycloid; body and head entirely scaled,

except on anteroventral surface of head; no scales ondorsal and anal-fin bases; scales extending on ante-rior 40% of caudal fin; frontal squamation E-pat-terned, frontal scales circularly arranged around A-scale without free margins; E-scales not overlapping;no row of scales anterior to H-scale; supraorbitalscales 2; longitudinal series of scales 33-34 (34);transverse series of scales 7; scale rows around caudalpeduncle 16; minute contact organ on posteriormargin of some scales of midventral portion of flankin males, often absent; no contact organs on fin rays.Cephalic neuromasts: supraorbital 3 + 3, parietal 2,

anterior rostral 1, posterior rostral 1, infraorbital 1 +26-27 (27) + 1, preorbital 2, otic 1, post-otic 2,supratemporal 1, median opercular 1, ventral oper-cular 2, preopercular 2 + 4-5 (4), mandibular 4 + 2,lateral mandibular 3, paramandibular 1; one neuro-mast per scale of lateral line, sometimes absent infew scales; two neuromasts on caudal-fin base.

aqua vol. 12 no. 3 - 10 July 2007125


Holotype Males Females(n = 5) (n = 4)

Standard length (mm) 29.7 21.7-31.4 22.0-25.7

Percentages of standard lengthBody depth 19.1 18.8-20.6 17.9-21.2Caudal peduncle depth 12.8 12.4-13.6 11.7-14.1Predorsal length 76.8 76.7-78.7 77.2-78.8Prepelvic length 55.1 53.8-57.6 54.0-57.3Length of dorsal-fin base 9.9 8.9- 9.9 8.5- 9.9Length of anal-fin base 21.2 20.4-23.6 19.0-21.0Caudal-fin length 38.0 36.4-41.4 36.1-40.4Pectoral-fin length 19.5 19.4-21.7 19.0-21.2Pelvic-fin length 6.3 5.2- 6.7 4.4- 6.6Head length 25.4 25.4-27.6 24.6-27.9

Percentages of head lengthHead depth 65.4 60.4-65.5 59.5-65.4Head width 77.0 73.8-77.0 74.0-78.1Snout length 13.1 12.5-13.4 11.6-13.0Lower jaw length 19.5 18.5-22.6 18.0-20.2Eye diameter 33.0 31.3-37.3 33.7-35.7

Table I. Morphometric data of Rivulus lazzarotoi.

Fig. 2. Rivulus lazzarotoi, UFRJ 6436, female, paratype,23.8 mm SL (some hours after collection); Brazil: Estado doRio de Janeiro: Angra dos Reis. Photo by W. J. E. M. Costa.

Basihyal subtriangular, greatest width about 50%of length; basihyal cartilage about 25% of totallength of basihyal; six branchiostegal rays; secondpharyngobranchial teeth 1-2; gill rakers on firstbranchial arch 1 + 9; vomerine teeth 3; dermo-sphenotic present; ventral process of posttemporalabsent; total vertebrae 33-34.C o l o r a t i o n . Males: Side of body pale blue,

red around some scales to form irregularly shapedoblique marks; dorsum light orange, with longitu-dinal row of transversely elongated brown spots onlateral portion of dorsum; venter reddish orange;side of head orange, opercle pale golden; jaws red-dish orange; iris light blue; dorsal fin hyaline, dis-tal portion pale blue with narrow dark grey mar-gin; anal fin orange-yellow to reddish orange ondistal half, with marginal narrow black line on dis-tal portion, light blue with red dots on basal por-tion; caudal fin pale yellow with light blue dots,dorsal portion pale blue with narrow dark greymargin, ventral portion reddish orange with mar-ginal narrow black line; pectoral fin yellowish hya-

line; pelvic fin reddish orange with narrow ante-rior black margin.Females: Side of body light brown with pale blue

dots; dorsum light yellowish brown, with longitu-dinal row of transversely elongated brown spots onlateral portion of dorsum; venter pale yellow; sideof head light brown, opercle pale golden; jaws paleorange; iris light yellow; dorsal fin hyaline, withthree transverse brown stripes; anal fin light bluewith brown dots on basal half, orange-pink on dis-tal half; caudal fin hyaline with 5-6 transverse rowsof small brown spots; large, round black spot ondorsal portion of caudal-fin base, reaching dorsalmargin of fin, preceded by small light blue spot;pectoral fin hyaline; pelvic fin orange-hyaline.Distribution and habitat: Known only from the

type locality, a stream within dense secondary for-est vegetation; water clear, pH 6.1; during collec-tion of the holotype, the water temperature was25.5°C at 10:40 a.m. Etymology: The new species is named after Hen-

rique Lazzaroto, its first collector.

aqua vol. 12 no. 3 - 10 July 2007 126


Fig. 3. Diagrammatic representation of neuromasts of the head, lateral view, of Rivulus lazzarotoi, UFRJ 6435, male, holo-type, 29.7 mm SL. Abbreviations: fio, first infraorbital neuromast; lio, last infraorbital neuromast; lma, last mandibular neu-romast; lm, lateral mandibular series; mo, median opercular neuromast; ot, otic neuromast; pa, parietal neuromast; pm, para-mandibular neuromast; po1-6, preopercular neuromast 1-6; pob, preorbital series; pot, post-otic series; pso, posterior supra-orbital series; st, supratemporal neuromast; vio, ventral infraorbital series; vo, ventral opercular series.

Discussion: Monophyly of the genus Rivulus isweakly supported by morphological characters (e.g. Costa 2006), but it is not corroborated by mol-ecular data (Hrbek & Larson 1999, Murphy et al.1999). However, some assemblages of Rivulus arewell corroborated both by morphology and mole-cules (Costa 2006). This is the case of the Rivulussantensis species group, which has been diagnosedby an apomorphic curved shape of the ventralprocess of the angulo-articular bone (Costa 1998,2006). Another feature not found elsewhereamong aplocheiloids except in the R. santensisspecies group is the peculiar arrangement of theneuromasts of the infraorbital series (Fig. 3).Uniquely in species of this group, the neuromastsare arranged in a zigzag pattern, instead of aligned

around the orbit as in other aplocheiloid fishes.Besides this uncommon arrangement, infraorbitalneuromasts are more numerous in the R. santensisspecies group than in other congeners (1 + 19-28 +1, vs. 1 + 9-16 + 1). Among species of the R. santensis group, higher

counts of infraorbital neuromasts are found in R.lazzarotoi (1 + 26-27 + 1) and R. simplicis (1 + 24-28 + 1, vs. 1 + 19-24 + 1 in other species of the R.santensis group), suggesting that they are moreclosely related species. Both species are knownfrom isolated plains in the Ilha Grande Bay (Fig.4), which is enclosed by hills, some reaching about1000 m height. Rivulus simplicis is endemic to asmall plain area around the city of Paraty, whereasR. lazzarotoi occurs in another small plain area, in

aqua vol. 12 no. 3 - 10 July 2007127


Fig. 4. Geographic distribution of Rivulus lazzarotoi (diamond – green colour) and R. simplicis (dot – red colour).

the Jurumirim River basin, which is about 90 kmfrom Paraty along the coast.Rivulus lazzarotoi is distinguished from R. simpli-

cis in having its dorsal-fin origin in a verticalthrough the base of the eleventh or twelth anal-finrays (vs. through the base of ninth or tenth anal-finrays) and a longer pectoral fin in males (pectoral-fin length 19.4-21.7% SL vs. 17.5-19.2% SL).Rivulus lazzarotoi differs from all other species ofthe R. santensis group, including R. simplicis, byhaving a minute pelvic fin, its tip not reaching theanus, pelvic-fin length 4.4-6.7 % SL (vs. reachingurogenital papilla, pelvic-fin length 6.7-11.3% SLin other species of the R. santensis group, except R.nudiventris, in which the pelvic fin is absent), analfin orange-yellow to reddish orange on distal half,with distal narrow black margin in males (vs.greenish yellow with a narrow red margin in R.simplicis, or greenish yellow with a dark grey toblack margin in the remaining species of thegroup) and a large, round black spot on the dorsalportion of the caudal-fin base in females, extendingto fin margin (vs. small, usually triangular spot, notreaching caudal-fin margin).

ACKNOWLEDGEMENTSThanks are due to Claudia P. Bove, Bruno B. Costa

and João Pedro P. Moreira for help during collec-tions, and to Henrique Lazzaroto for collecting andmaking available material from the new species. Thisstudy was supported by CNPq (Conselho Nacionalde Desenvolvimento Científico e Tecnológico –Ministério de Ciência e Tecnologia) and ProgramaCientistas do Nosso Estado by FAPERJ (Fundaçãode Amparo à Pesquisa do Estado do Rio de Janeiro).

REFERENCESBERKENKAMP, H. O. 1984. Eine neue Bachlingsart aus

dem Staat Sta. Catarina/Brasilien, Rivulus haraldsioliispec. nov. (Pisces – Rivulidae). Amazoniana 8: 429-439.

COSTA, W. J. E. M. 1991. Description of two new speciesof the genus Rivulus (Cyprinodontiformes: Rivulidae)from eastern South American coastal plains. Revue Suissede Zoologie 98: 581-587.

COSTA, W. J. E. M. 1995. Pearl killifishes - the Cynolebiatinae:

systematics and biogeography of the neotropical annual fishsubfamily. Tropical Fish Hobbyist, Neptune City, 128 pp.

COSTA, W. J. E. M. 1998. Phylogeny and classification ofRivulidae revisited: origin and evolution of annualismand miniaturization in rivulid fishes. Journal of Compar-ative Biology 3: 33–92.

COSTA, W. J. E. M. 2001. The neotropical annual fish genusCynolebias (Cyprinodontiformes: Rivulidae): phylogeneticrelationships, taxonomic revision and biogeography.Ichthyological Exploration of Freshwaters 12: 333-383.

COSTA, W. J. E. M. 2003. Family Rivulidae (South Ameri-can annual fishes). In: Check list of the freshwater fishes ofSouth and Central America (Eds. R. E. Reis, S. O. Kullan-der & C. J. Ferraris Jr.): 526-548. Edipucrs, Porto Alegre.

COSTA, W. J. E. M. 2004. Rivulus simplicis n. sp. (Cyprin-odontiformes: Rivulidae): a new killifish from the coastalplains of south-eastern Brazil. aqua, Journal of Ichthyologyand Aquatic Biology 8: 103-108.

COSTA, W. J. E. M. 2006. Relationships and taxonomy ofthe killifish genus Rivulus (Cyprinodontiformes:Aplocheiloidei: Rivulidae) from the Brazilian Amazonasriver basin, with notes on historical ecology. aqua, Jour-nal of Ichthyology and Aquatic Biology 11: 133-175.

COSTA, W. J. E. M. & BRASIL, G. C. 1991. Description ofa new species of Rivulus (Cyprinodontiformes: Rivulidae)from the coastal plains of eastern Brazil. IchthyologicalExploration of Freshwaters 1: 379-383.

HOEDEMAN, J. J. 1958. The frontal scalation pattern insome groups of toothcarps (Pisces, Cyprinodontiformes).Bulletin of Aquatic Biology 1: 23-28.

HRBEK, T. & LARSON, A. 1999. The diapause in the killi-fish family Rivulidae (Atherinomorpha, Cyprinodontif-ormes): a molecular phylogenetic and biogeographic per-spective. Evolution 53: 1200-1216.

KÖHLER, W. 1906. Diesjährige Neuheiten in Wort undBild, 3: Neu importierte Rivulus-Arten. Blätter fürAquarien- und Terrarienkunde 17: 404-408.

MURPHY, W. J., THOMERSON, J. E. & COLLIER, G. E.1999. Phylogeny of the neotropical killifish family Rivul-idae (Cyprinodontiformes, Aplocheiloidei) inferred frommitochondrial DNA sequences. Molecular and Phyloge-netic Evolution 13: 289-301.

SEEGERS, L. 1984. Zur Revision der Rivulus-Arten Südost-Brasiliens, mit einer Neubeschreibung von Rivulusluelingi n. sp. und Rivulus caudomarginatus n. sp. (Pisces:Cyprinodontidae: Rivulinae). Zoologische Beiträge 28:271-320.

TAYLOR, W. R. & VAN DYKE, G. C. (1985) Revised proce-dures for staining and clearing small fishes and other ver-tebrates for bone and cartilage study. Cybium 9: 107-109.

aqua vol. 12 no. 3 - 10 July 2007 128


AbstractCynolebias paraguassuensis n. sp., a median sized species

from the Paraguaçu River basin, northeastern Brazil, isdescribed, constituting the first record of a rivulid fish forthat basin. It is similar to C. itapicuruensis and C. vazabar-risensis in having dentary teeth strongly bowed. It is distin-guished from C. itapicuruensis and C. vazabarrisensis inpossessing a longer pectoral fin in males and light blue dotsover the entire anal fin in males.

ResumoCynolebias paraguassuensis Costa, Suzart & Nielsen, n.

sp., uma espécie de tamanho médio da bacia do rioParaguaçu, nordeste do Brasil, é descrita, constituindo oprimeiro registro de um peixe rivulídeo para aquela bacia.Ela se assemelha a C. itapicuruensis e C. vazabarrisensis porpossuir os dentes do dentário fortemente arqueados. Ela sedistingue de C. itapicuruensis e C. vazabarrisensis por pos-suir nadadeira peitoral mais comprida em machos e pontosazul claros sobre toda a nadadeira anal em machos.

ZusammenfassungCynolebias paraguassuensis n. sp., eine Art mittlerer

Größe, wird aus dem Paraguaçu-Becken, NO-Brasilien,beschrieben. Es ist der erste Nachweis eines Rivuliden ausdiesem Becken. Die neue Art ähnelt C. itapicuruensis undC. vazabarrisensis in dem Merkmal der stark gebogenenKieferzähne. Sie unterscheidet sich von C. itapicuruensisund C. vazabarrisensis durch längere Brustflossen bei denMännchen und hellblaue Punkte auf der gesamten After-flosse beim Männchen.

RésuméOn décrit ici Cynolebias paraguassuensis n. sp., une espeè

de taille moyenne du bassin de la rivière Paraguaçu, aunord-est du Brésil, et il s’agit du premir relevé d’unRivulidé pour ce bassin. Elle est analogue à C. itapicuruen-sis et à C. vazabarrisensis par les dents de denture fortementarquées. Elle se distingue de C. itapicuruensis et de C. vaza-barrisensis par de plus longues pectorales chez les mâles etdes taches bleu clair sur toute l’anale des mâles.

SommarioSi descrive Cynolebias paraguassuensis n. sp., una specie di

medie dimensioni del bacino del fiume Paraguaçu, Brasilenordorientale, che rappresenta la prima segnalazione dirivulide in questo bacino. È simile a C. itapicuruensis e C. vazabarrisensis per avere denti molto ricurvi. Si di-stingue da C. itapicuruensis e C. vazabarrisensis poiché imaschi posseggono una pinna pettorale più lunga e pun-tini blu chiaro lungo tutta la pinna anale.

INTRODUCTIONCynolebias Steindachner, 1876 is a genus of killi-

fishes inhabiting temporary pools of central andnortheastern Brazil. Twelve species were recognizedas valid in the recent taxonomic revision of thegenus (Costa, 2001). One species, C. griseus Costa,Lacerda & Brasil, 1990 is found in the upper sec-tion of the rio Tocantins basin, central Brazil, inthe savanna-like region of Cerrado (Costa et al.1990, Costa, 2001). All other species of Cynolebiasare endemic to northeastern Brazil, usually occur-ring in the vast semi-arid area known as Caatinga,or sometimes in more humid areas on the coastalplains: C. porosus Steindachner, 1876 from thecoastal plains of Pernambuco state; C. microphthal-mus Costa & Brasil, 1995 from the Jaguaribe Riverbasin and adjacent coastal areas; C. itapicuruensisCosta, 2001 from the Itapicuru River basin; C.vazabarrisensis Costa, 2001 from the VazabarrisRiver basin; and C. albipunctatus Costa & Brazil,1991 C. altus Costa, 2001 C. attenuatus Costa,2001 C. gibbus Costa, 2001 C. gilbertoi Costa,1998 C. leptocephalus Costa & Brasil, 1993 and C.perforatus Costa & Brasil, 1991 from the São Fran-cisco River basin (Steindachner 1876; Costa &Brasil 1991, 1993; Costa, 1995, 1998, 2001). Thenew species herein described was collected in theParaguaçu River basin, which occupies a large

aqua vol. 12 no. 3 - 10 July 2007129


Cynolebias paraguassuensis n. sp. (Teleostei: Cyprinodontiformes: Rivulidae), a new seasonal killifish from the Brazilian Caatinga, Paraguaçu River basin

Wilson J. E. M. Costa1, Rogério Suzart2 and Dalton T. B. Nielsen2

1) Laboratório de Ictiologia Geral e Aplicada, Departamento de Zoologia, Universidade Federal do Rio deJaneiro, Caixa Postal 68049, CEP 21944-970, Rio de Janeiro, RJ, Brasil. E-mail: [email protected]) Laboratório de Zoologia, Departamento de Biologia, Universidade de Taubaté, Praça Marcelino

Monteiro 63, CEP 12030-010, Taubaté, São Paulo, SP, Brasil.

Received: 12 Decmber 2006 – Accepted: 10 May 2007

drainage area of northeastern Brazil and the mainriver extends for 496 km. This constitutes the firstrecord of a rivulid for this basin.

MATERIALS AND METHODSMeasurements and counts follow Costa (1995).

Measurements are presented as percentages of stan-dard length (SL), except for those related to headmorphology, which are expressed as percentages ofhead length. Fin-ray counts include all elements.Numbers of vertebrae, gill rakers and pectoral, pelvicand caudal-fin rays were recorded only from clearedand stained specimens; the compound caudal cen-trum was counted as a single element. Osteologicalpreparations were made according to Taylor & VanDyke (1985). Terminology follows Hoedeman(1958) for frontal squamation and Costa (2001) forcephalic neuromast series. The abbreviation c&smeans specimens cleared and stained for bone andcartilage. Material is deposited in UFRJ, Departa-mento de Zoologia da Universidade Federal do Riode Janeiro, Rio de Janeiro, Brazil.

Cynolebias paraguassuensis n. sp.(Figs 1-2)

Holotype: UFRJ 6454, 7, 58.6 mm SL; Brazil:Estado da Bahia: Município de Itaberaba, tempo-rary pool close to the road BA-488, near riachoNovo, a tributary of rio Paraguaçu, about 12°20’S40o00’W, R. Suzart, May 2006.Paratypes: UFRJ 6455, 5 6, 42.6-43.9 mm SL;collected with holotype. UFRJ 6456, 1 7, 43.8

mm SL; UFRJ 6457, 1 7, 40.5 mm SL, 2 6,42.0-42.3 mm SL (c&s); same locality; R. Suzart& D. T. B. Nielsen, February 2006.Diagnosis: Distinguished from all congeners by

the following combination of features: dentaryteeth strongly bowed, posterior margin reachingvertical through base of fifth or sixth anal-fin ray inmales and light blue dots over entire anal fin inmales.Description: Morphometric data appear in Table

I. Largest male examined 58.6 mm SL; largestfemale examined 43.9 mm SL; dorsal profile con-vex from snout to end of dorsal-fin base, approxi-mately straight on caudal peduncle; ventral profileconvex from lower jaw to end of anal-fin base,nearly straight along caudal peduncle; body mod-erately deep, compressed; greatest body depth atlevel of pectoral-fin base; jaws short, snout blunt.Dorsal and anal fins pointed in males, rounded in

females; anal fin tip with short filamentous rays inmales; caudal fin rounded; pectoral fin elliptical, itsposterior margin reaching vertical through base offifth or sixth anal-fin ray in males and verticalbetween pelvic-fin base and anus in females; pelvicfin pointed, tip reaching between base of third andfifth anal-fin rays in males and between base of sec-ond and third anal-fin rays in females; pelvic-finbases medially separated by short interspace, aboutequal pelvic-fin base width; dorsal-fin origin in ver-tical through base of sixth anal-fin ray in males,between base of third and fifth anal-fin rays infemales; dorsal-fin origin between neural spines ofvertebrae 14 and 15; anal-fin origin betweenpleural ribs of vertebrae 11 and 12; dorsal-fin rays15-17; anal-fin rays 19-21 in males, 18-19 infemales; caudal-fin rays 31; pectoral-fin rays 15;pelvic-fin rays 6.Scales small, cycloid; body and head entirely

scaled, except on anteroventral surface of head; noscales on dorsal fin base; scales slightly extendingover middle of anal-fin base; scales extending onanterior 25% of caudal fin; frontal region withsmall scales irregularly arranged; scale with all mar-gins free on central portion of frontal region;supraorbital scales 9-10; longitudinal series ofscales 36-39; transverse series of scales 15; scalerows around caudal peduncle 20; contact organsabsent on flanks; minute papillate contact organson inner surface of five dorsalmost pectoral-fin raysin males.Cephalic neuromasts: supraorbital 27-28, parietal

3-5, anterior rostral 3, posterior rostral 3, infraor-

aqua vol. 12 no. 3 - 10 July 2007 130

Cynolebias paraguassuensis n. sp. (Teleostei: Cyprinodontiformes: Rivulidae): a new seasonal killifish from the Brazilian Caatinga, Paraguaçu River basin

Fig. 1. Cynolebias paraguassuensis, UFRJ 6454, male holo-type, 58.6 mm SL, above, UFRJ 6455, female paratype,42.6 mm SL, below; Brazil: Estado da Bahia: Itaberaba.Photo by W. J. E. M. Costa.

bital 37, preorbital 4, otic 7-8, post-otic 6-7,supratemporal 2, median opercular 1, ventral oper-cular 3, preopercular plus mandibular 51-55, lat-eral mandibular 6 + 9, paramandibular 1; one neu-romast per scale of lateral line; two neuromasts oncaudal-fin base.Teeth conical, dentary teeth strongly bowed; basi-

hyal subtriangular, greatest width about 65% oflength; basihyal cartilage about 25% of total lengthof basihyal; six branchiostegal rays; second pharyn-gobranchial absent; gill-rakers on first branchialarch 3 + 11; vomerine teeth absent; dermosphe-notic absent; ventral process of posttemporalabsent; total vertebrae 34.

C o l o r a t i o n . Males: Side of body dark brown,with narrow pale golden bars and vertical rows ofpale blue dots; dorsum brown; venter pale orange;side of head brown, opercle pale golden; smallblack spots on dorsoposterior portion of head sideand anterodorsal portion of trunk side; jawsbrown; iris dark brown; dorsal and caudal fins darkbrown with bluish white dots; anal fin reddishbrown with light blue dots; pectoral fin hyaline;pelvic fin reddish brown.Females: Side of body light brown with pale dark

grey bars; one or two black spots on anterocentralpart of flank; dorsum light brown; venter lightgrey; side of head light brown, opercle pale golden;

aqua vol. 12 no. 3 - 10 July 2007131

Wilson J. E. M. Costa, Rogério Suzart and Dalton T. B. Nielsen

Holotype Males Females(n = 3) (n = 5)

Standard length (mm) 58.6 40.5-58.6 42.6-43.9Percentages of standard lengthBody depth 33.6 29.6-33.6 30.0-31.5Caudal peduncle depth 15.9 13.8-15.9 13.2-13.6Predorsal length 63.9 63.8-66.1 63.9-67.3Prepelvic length 51.0 51.0-55.8 54.1-57.0Length of dorsal-fin base 28.8 25.9-28.8 24.4-25.7Length of anal-fin base 31.5 30.8-32.3 25.2-27.5Caudal-fin length 34.4 34.4-36.9 30.7-35.7Pectoral-fin length 27.2 27.2-28.3 23.8-26.0Pelvic-fin length 9.1 9.1-10.8 9.9-11.5Head length 31.3 31.3-34.8 31.6-32.8Percentages of head lengthHead depth 96.7 88.8-96.7 89.2-94.1Head width 72.8 72.0-75.2 74.7-79.1Snout length 12.4 12.3-13.1 12.8-13.5Lower jaw length 23.1 22.0-25.6 19.5-23.2Eye diameter 22.9 22.7-26.7 26.6-28.2

Table I. Morphometric data of Cynolebias paraguassuensis.

Fig. 2. Cynolebias paraguassuensis, UFRJ 6456, young male paratype, 43.8 mm SL (one week after collection); Brazil: Estadoda Bahia: Itaberaba. Photo by W. J. E. M. Costa.

jaws light brown; iris pale yellow; dorsal fin hya-line, with faint grey spots on basal portion; caudal,pectoral, pelvic and anal fins hyaline.Distribution and habitat: The new species is

known only from the type locality, a temporarypool in the Caatinga, near Riacho Novo,Paraguaçu River basin, Bahia, northeastern Brazil.Etymology: The name paraguassuensis is an allu-

sion to the occurrence of the new species in theParaguaçu River basin.Discussion: Cynolebias paraguassuensis is a mem-

ber of a clade containing C. itapicuruensis and C.vazabarrisensis diagnosed by the dentary teethbeing strongly bowed, a condition not found else-where among cynolebiatines (“Clade d” in Costa2001). It is distinguished from C. itapicuruensisand C. vazabarrisensis by having a longer pectoralfin in males (posterior margin reaching verticalthrough base of fifth or sixth anal-fin ray vsbetween base of second and fourth rays) and lightblue dots over the entire anal fin in males (vs lightblue dots on basal portion of fin, light blue stripeson distal portion). Species of Cynolebias usually are large voracious

killifishes, supposedly preying on small species ofthe genus Simpsonichthys Carvalho, 1959 (Costa,2001). However, C. paraguassuensis is a mediansized killifish species, smaller than its closelyrelated species. The largest male collected in Feb-ruary was 43.8 mm SL and at the end of the rainyseason three months later, the largest male was 58.6mm SL. Other species of Cynolebias studied in thefield during recent years, including the most basallineages of the genus, reach between about 80 and135 mm SL (Costa, 2001). The only exception isC. gilbertoi, a species endemic to the middle SãoFrancisco River basin, which reaches 57.9 mm SLas maximum adult size, a condition considered tobe apomorphic by Costa (2001), and similar to themaximum size recorded for C. paraguassuensis.However, the examination of osteological featuresclearly indicates that C. gilbertoi and C. paraguas-suensis are not closely related species, and conse-quently the smaller size was independentlyacquired in both species. Cynolebias gilbertoi is amember of a clade, also including C. perforatus, C.altus, C. leptocephalus, C. attenuatus, and C. gibbus(“Clade e” in Costa, 2001), diagnosed by derivedfeatures such as anteroventral portion of autopala-tine vestigial or absent and vomerine teeth present,

conditions not occurring in C. paraguassuensis (e. g.anteroventral portion of autopalatine long, vomer-ine teeth absent). The dentary teeth of C. gilbertoiare not bowed as in C. paraguassuensis, C. itapicu-ruensis and C. vazabarrisensis. Cynolebias paraguas-suensis also differs from C. gilbertoi by possessing31 caudal-fin rays (vs. 26-29), 15 pectoral-fin rays(vs. 12-13), 3 ventral opercular neuromasts (vs. 5-6), and 36-39 scales on the longitudinal series (vs.31-35).

ACKNOWLEDGEMENTSThis study was partly funded by CNPq (Con-

selho Nacional de Desenvolvimento Científico eTecnológico – Ministério de Ciência e Tecnologia)and FAPERJ (Fundação de Amparo à Pesquisa doEstado do Rio de Janeiro).

REFERENCESCOSTA, W. J. E. M. 1995. Pearl killifishes - the Cynolebi-

atinae: systematics and biogeography of the neotropicalannual fish subfamily. Tropical Fish Hobbyist. NeptuneCity, 128 pp.

COSTA, W. J. E. M. 1998. Cynolebias gilbertoi, a newspecies of annual fish (Cyprinodontiformes: Rivulidae)from the rio São Francisco basin, northeastern Brazil.Cybium 22: 237-243.

COSTA, W. J. E. M. 2001. The neotropical annual fish genusCynolebias (Cyprinodontiformes: Rivulidae): phylogeneticrelationships, taxonomic revision and biogeography.Ichthyological Exploration of Freshwaters 12: 333-383.

COSTA, W. J. E. M. & BRASIL, G. C. 1991. Three newspecies of Cynolebias (Cyprinodontiformes: Rivulidae)from the São Francisco basin, Brazil. Ichthyological Explo-ration of Freshwaters 2: 55-62.

COSTA, W. J. E. M. & BRASIL, G. C. 1993. Two newspecies of Cynolebias (Cyprinodontiformes: Rivulidae)from the São Francisco basin, Brazil, with notes on phy-logeny and biogeography of annual fishes. IchthyologicalExploration of Freshwaters 4: 193-200.

COSTA, W. J. E. M., LACERDA, M. T., C. & BRASIL, G. C.1990. Description de deux nouvelles espèces du genreCynolebias du bassin du Rio Tocantins (Cyprinodontif-ormes, Rivulidae). Revue Française d’Aquariologie 17: 9-14.

HOEDEMAN, J. J. 1958. The frontal scalation pattern insome groups of toothcarps (Pisces, Cyprinodontiformes).Bulletin of Aquatic Biology 1: 23-28.

STEINDACHNER, F. 1876. Ichthyologische Beiträge (V).Sitzungsberichte der Kaiserlichen Akademie der Wis-senschaften in Wien (Abteilung 1) 74: 49-240, pls. 1-15.

TAYLOR, W. R. & VAN DYKE, G. C. 1985. Revised proce-dures for staining and clearing small fishes and other ver-tebrates for bone and cartilage study. Cybium 9: 107-109.

aqua vol. 12 no. 3 - 10 July 2007 132

Cynolebias paraguassuensis n. sp. (Teleostei: Cyprinodontiformes: Rivulidae): a new seasonal killifish from the Brazilian Caatinga, Paraguaçu River basin

BLABER, S. J. M. 1980. Fish of the Trnity inlet systemof North Queensland, with notes on the ecologyof fish faunas of tropical Indo-Pacific estuaries.Australian Journal of Marine and FreshwaterResearch 31:137-46.

DAY, J. H., BLABER, S. J. M., & WALLACE, J. H. 1981.Estuarine fishes. In: Estuarine Ecology with Parti-cular Reference to Southern Africa. (Ed. J.H. Day.):197-221. A. A. Balkema, Rotterdam.

DIMMICH, W. W. 1988. Ultrastructure of North Ame-rican cyprinid maxillary barbels. Copeia 1988 (1):72-79.

TREWAVAS, E. 1983. Tilapiine Fishes of the GeneraSarotherodon, Oreochromis and Danakilia.British Museum (Natural History), London, 583 pp.

7. Submission of manuscript and illustrations: Themanuscript and illustrations must be submitted digi-tally to the Scientific Editor:

Dr. Friedhelm KruppSenckenberg Research Institute

Senckenberganlage 2560325 Frankfurt am Main, Germany

E-mail: [email protected]: +49-69-7542.1255, Fax: +49-69-7542.1253

to whom all subsequent correspondence shall beaddressed. Texts, tables, and graphs should be inMicrosoft-compatible electronic form.

After the paper has been accepted for publication, illus-trations as high-resolution TIF files or original pho-tographs (ideally transparencies; otherwise glossyprints, preferably in the size in which they will appear -the type area of aqua is 158 x 224 mm, one column is76 mm wide) must be sent to:

Aquapress, The Managing Editor Via G. Falcone 11,

27010 Miradolo Terme (Pavia), ItalyE-mail: [email protected]

Authors should retain copies of all materials for refer-ence.

Proofs of accepted papers will be sent as PDF files by e-mail attachment to the corresponding author.

8. Evaluation of manuscripts: manuscripts will be eval-uated by the editors and referees. Papers are accepted onthe understanding that they have not and will not bepublished elsewhere.

9. Reprints: Authors will receive one free copy of theissue in which their paper appears and an e-print inPDF format. Additional copies may be ordered fromAquapress.

1. Manuscript preparation: manuscripts must be sub-mitted in English. In exceptional cases aqua may pro-vide translations of manuscripts written in French, Ger-man, Italian, or Spanish.

Manuscripts must be word-processed in MicrosoftWORD and submitted in an electronic form. Generic,specific, and sub-specific names must be italicised. Allpapers must conform to the International Code of Zoo-logical Nomenclature. Authors are strongly advised tofollow the format set out in previous publications ofaqua.

2. Title: the title must be short and should preciselyidentify the main topic of the article. Names of generaor species are followed by the systematic group to whichthey belong. Author name(s) are given in full beneaththe title, followed by the complete mailing and e-mailaddress(es).

3. Abstract: the abstract should not exceed 250 wordsand draw attention to the principal conclusions. Itshould not contain any uncommon abbreviations or lit-erature citations. The inclusion of abstracts in other lan-guages is optional.

5. Subject matter: the text of the manuscript is usuallyarranged in four main sections: Introduction, Materialsand methods (including a key to abbreviations),Results, and Discussion. Other subdivisions may bechosen depending on the material presented. Acknowl-edgements should be placed between the text and refer-ences.

Scientific names of genera, species, and subspeciesshould be followed by the name(s) of author(s) and theyear of publication on first mention. A description of anew taxon must contain the following sections: Mater-ial, Diagnosis, Description, and Affinities. Holotypeand paratypes must be clearly identified, the institutionin which they have been deposited named, and the cat-alogue numbers given. Private collections are notacceptable as repositories for holotypes.

Synonyms must be arranged in chronological order.Identification keys must be dichotomous.

The metric system and SI units must be used. Temperatures are given in °C. Fractions should not beused.

6. References to literature: the name-year system mustbe used. The list of references should be placed at theend of the paper, alphabetically arranged according toauthor name. Only those publications cited in the papermay be included. Titles of journals must be given infull.

Examples of correct reference formats:

Guidelines for Authors

aquaInternational Journal of Ichthyology

Vol. 12 (3), 10 July 2007

Contents:

John E. Randall: Descriptions of four new shrimpgobies of the genus Vanderhorstiafrom the western Pacific ........................................................................................................... 89-100

Gerald R. Allen: Hoplolatilus erdmanni, a new species of sand tilefish (Pisces: Malacanthidae) from western New Guinea ............................................................................. 101-106

Jacques Géry: Rediscovery of the rare curimatid fish Curimatopsis microlepisEigenmann & Eigenmann, 1889 (Teleostei: Characiformes) in the Rio Purus basin, Brazil, with a note on some Amazonian localities of the Thayer Expedition to Brazil in 1865 and 1866 .................................................................................................................................. 107-122

Wilson J. E. M. Costa: Rivulus lazzarotoi n. sp. (Teleostei: Cyprinodontiformes: Rivulidae), a new killifish from the Jurumirim River basin, southeastern Brazil ........................ 123-128

Wilson J. E. M. Costa, Rogério Suzart and Dalton T. B. Nielsen:Cynolebias paraguassuensis n. sp. (Teleostei: Cyprinodontiformes: Rivulidae), a new seasonal killifish from the Brazilian Caatinga, Paraguaçu River basin ............................. 129-132

Papers appearing in this journal are indexed in: Zoological Record; Biolis – Biologische Literatur Information Senckenberg;

www.aquapress-bleher.com; www.aquapress-bleher.com; www.Joachim-Frische.com; www.aquageo.com

Cover photo: Curimatopsis microlepis from the Lago Aiapuá, lower Purus basin, Amazonas, Brazil. Photo by H. Bleher.

Collecting Curimatopsis microlepis in the Lago Aiapuá region, lower Purus basin, Amazonas, Brazil. Photo by N. Khardina.

aqua · aqua is an international journal which publishes original scientific articles in the fields...

Documents

Transcript of aqua · aqua is an international journal which publishes original scientific articles in the fields...