aqua International Journal of Ichthyology

Vol. 23 (4), 15 March 2018

Aquapress ISSN 0945-9871 aqua - International Journal of Ichthyology

Managing Editor: Scope aqua is an international journal which publishes original sci- Heiko Bleher entific articles in the fields of systematics, , bio - Via G. Falcone 11, geography, ethology, ecology, and general biology of . 27010 Miradolo Terme (PV), Italy Papers on freshwater, brackish, and marine fishes will be Tel. & Fax: +39-0382-754129 considered. aqua is fully refereed and aims at publishing E-mail: [email protected] manuscripts within 2-4 months of acceptance. In view of www.aqua-aquapress.com the importance of color patterns in identification and ethology, authors are encouraged to submit col- or illustrations in addition to descriptions of coloration. It is our aim to provide the international scientific community Scientific Editor: with an efficiently published journal meeting high scientific Flávio C. T. Lima and technical standards. Museu de Zoologia da Universidade Estadual de Campinas “Adão José Cardoso” Call for papers Rua Albert Einstein s/n, Caixa Postal 6109, The editors welcome the submission of original manu- 13083-863 Campinas, São Paulo, Brazil scripts which should be sent in digital format to the scien- tific editor. Full length research papers and short notes will Tel. +55 019-3521-6385 be considered for publication. There are no page charges E-mail: [email protected] and color illustrations will be published free of charge. Au- thors will receive one free copy of the issue in which their paper is published and an e-print in PDF format. Editorial Board: Subscription Notice Gerald R. Allen At least one volume (4 issues) of aqua is being published Department of Aquatic Zoology, per year, each issue comprising 36-64 pages (incl. cover). Western Australian Museum, Perth, Australia The subscription rate (for one volume = 4 issues): Personal subscription: Euro 75,00 (online edition) – Euro Nina G. Bogutskaya 100,00 (online edition + print edition incl. priority mail); Zoological Institute of the Russian Academy of Institutional subscription: Euro 200,00 (print edition) – Sciences, St. Petersburg, Russia Euro 800,00 (online edition) – Euro 1000,00 (online edi- tion + print edition incl. priority mail). Subscription en- Frank Pezold quires should be sent to the publisher at the address given College of Science & Engineering below or by e-mail to: [email protected] Texas A&M University – Corpus Christi, TX Special Publication Friedhelm Krupp Since 2003 Aquapress publishes a series of Special Publica- Curator of Fishes, Senckenberg Research Institute tions, which are produced at irregular intervals. All Special and Natural History Museum, Frankfurt am Main, Publications have about 100 or more pages and are available Germany separately from regular issues of aqua. Enquiries about sub- scriptions and prices should be sent to the publisher at the Axel Meyer address given here above or by e-mail to: Lehrstuhl für Zoologie und Evolutions biologie, [email protected] Universität Konstanz, Germany Claim policy Paolo Parenti Orders are regarded as firm and are not refundable. Taking Department of Enviromental Sciences, out an annual subscription commits you for a minimum of University of Milano-Bicocca, Milan, Italy one year. Subscription will be renewed automatically. If you have received a renewal invoice and do not wish to continue Mário de Pinna with the subscription, please submit the cancellation in writ- Museu de Zoologia da USP, São Paulo, Brazil ing within one month to: [email protected] or Fax: +39 0382754129 John E. Randall Bishop Museum, Honolulu, Hawaii, U.S.A. ISSN 0945-9871 Publisher: Aquapress, Redazione aqua, I-27010 Miradolo Terme (Pavia), Italy Helen K. Larson www.aqua-aquapress.com Curator Emeritus, Fishes - Museum and Printer: C.N.S. Srl – Bergamo – Italy Art Gallery of the Northern Territory. Darwin, Copyediting and layout: Rossella Bulla Australia © 2018 aqua, International Journal of Ichthyology aqua, International Journal of Ichthyology

Cynolebias akroa, a new species of annual (: ) from the rio Preto, São Francisco basin, northeastern Brazil

Dalton Tavares Bressane Nielsen1, Mayler Martins2, Luciano Medeiros de Araujo3, Fabio Origuela de Lira4 and Amer Faour5

1) Laboratório de Zoologia, departamento de Biologia, Universidade de Taubaté, Av. Tiradentes180, CEP: 12030-180, Taubaté, SP, Brazil. E-mail: [email protected] 2)Instituto Federal Minas Gerais- Campus Bambuí- Fazenda Varginha-Estrada Bambuí-Medeiros Km 5, CEP: 38900-000, MG, Brazil. E-mail: [email protected] 3) Fundação Zoobotânica de Belo Horizonte – Aquário da Bacia do São Francisco Avenida Otacílio Negrão de Lima, 8000; Belo Horizonte, MG Brasil. E-mail: [email protected] 4) Meandros Consultoria Ambiental Ltda., Rua Maria Cândida Cerqueira, 75, Santa Amélia, Belo Horizonte/MG, Brasil. Email: [email protected] 5) C/ Las Fuentes 4 bajo C - 37439 - San Cristóbal de la Cuesta - Salamanca. E-mail: [email protected]

Received: 20 October 2017 – Accepted: 25 November 2017

Keywords als mit den anderen Cynolebias-Arten vom mittleren Rio- Caatinga domain, Hypsolebias, Cynolebias parnaibensis, São-Francisco-Becken. Cynolebias akroa sp. n. unterschei- introduced species. det sich von den anderen Arten der Gattung durch das Farbmuster der Männchen, die Stellung der Rückenflosse Abstract relativ zu der Afterflosse, die Zahl der Bauchflossenstrah- A new species of Cynolebias is described from temporary len, die Form der Urogenitalpapille und das Muster der pools from the Rio Preto drainage, São Francisco basin, Neuromasten im Kopfbereich. Bahia, Brazil, within the Caatinga domain. The type-local- ity is an altered aquatic environment, with the presence of Resumo exotic fish species and in contact with a permanent water- Uma nova espécie de Cynolebias é descrita a de uma poça course. Cynolebias akroa sp. n., along with Hypsolebias temporária da drenagem do Rio Preto, bacia do São Fran- faouri, that inhabits the same annual pool, are in great dan- cisco, Bahia Brasil, no bioma da Caatinga, em um ambien- ger of extinction. Cynolebias akroa sp. n., appears to be te aquático alterado, com introdução de espécies exóticas e more closely related to Cynolebias parnaibensis, with which em contato com um curso de água permanente. Cynolebias it shares several features, than to the remaining species of akroa sp. n., juntamente com Hypsolebias faouri, que habita Cynolebias from the middle Rio São Francisco basin. a mesma poça anual, estão em grande perigo de extinção. Cynolebias akroa sp. n. differs from other species of Cynolebias akroa sp. n., parece estar mais relacionada com Cynolebias by the male color pattern, relative position of Cynolebias parnaibensis, por uma série de características co- dorsal fin to the anal fin, pelvic-fin rays counts, shape of muns, do que com as demais espécies de Cynolebias do mé- the urogenital papilla, and cephalic neuro mast pattern. dio rio São Francisco. Cynolebias akroa sp. n. difere de ou- tras espécies do gênero Cynolebias pelo padrão de colorido Zusammenfassung dos machos, posição relativa da nadadeira dorsal em rela- Beschrieben wird eine neue Cynolebias-Art aus temporä- ção a nadadeira anal, número de raios na nadadeira pélvica, ren Tümpeln im Einzugsgebiet des Rio Preto, São-Francis- forma da papila urogenital e padrão dos neuromastos cefá- co-Becken, Bahia, Brasilien, im Bereich der Caatinga. Die licos. Typus-Lokalität ist eine veränderte aquatische Umwelt, wo exotische Fischarten vorkommen und Kontakt zu einem Résumé dauerhaften Wasserlauf besteht. Cynolebias akroa sp. n., Une nouvelle espèce de Cynolebias est décrit à partir de ebenso Hypsolebias faouri, der denselben zeitweiligen Tüm- mares temporaires du Rio Preto, drainage du bassin du São pel bewohnt, sind stark von Ausrottung bedroht. Cynole- Francisco, Bahia, Brésil, dans le domaine de caatinga. La bias akroa sp. n. ist offenbar näher mit Cynolebias parnai- localité type est une altération de l'environnement aquati- bensis verwandt – mehrere Merkmale stimmen überein – que, avec la présence d'espèces de poissons exotiques et en

113 aqua vol. 23 no. 4 - 15 March 2018 Cynolebias akroa, a new species of annual fish (Cyprinodontiformes: Rivulidae) from the rio Preto, São Francisco basin, northeastern Brazil contact avec un cours d'eau permanent. Cynolebias akroa Fifteen Cynolebias species occurs at the middle São n. sp., avec Hypsolebias faouri, qui habite le même taux an- Francisco basin, while Cynolebias microphthalmus nuel d'extérieure, sont en grand danger d'extinction. Cyno- Costa & Brasil (2001) occurs in the Rio Jaguaribe lebias akroa n. sp., semble être plus étroitement liée à Cy- basin, Cynolebias parnaibensis Costa et al. (2010) nolebias parnaibensis, avec laquelle elle partage plusieurs ca- ractéristiques, que pour les autres espèces de Cynolebias du occurs in the Rio Parnaíba basin, Cynolebias itapi- milieu du bassin du Rio São Francisco. Cynolebias akroa n. curuensis Costa (2001) occurs in the Rio Itapicuru sp. diffère des autres espèces de Cynolebias par l'homme, basin, Cynolebias paraguassuensis Costa et al. position relative de nageoire dorsale à la nageoire anale, (2007), occurs in the Rio Paraguaçu basin, rayons pelviens-chefs, la forme de la papille urogénitale, Cynolebias vazabarriensis Costa (2001) occurs in neuro-mast céphalique et motif. the Rio Vaza-Barris basin, and Cynolebias griseus Costa et al. (1990) occurs in the Rio Tocantins Sommario basin. Una nuova specie di Cynolebias è descritta da pozze tem- All species of Cynolebias have an annual re- poranee del sotto bacino del Rio Preto, bacino idrografico del Rio São Francisco, Bahia, Brasile, all'interno del domi- production cycle, with an elaborated reproductive nio Caatinga. La località tipo è un ambiente acquatico al- behavior (Belote & Costa 2003), including the terato, con la presenza di specie ittiche esotiche e in con- production of sounds by males (Costa 2010). In tatto con un corso d'acqua permanente. Cynolebias akroa contrast to most other Rivulidae, the genus sp. n., insieme a Hypsolebias faouri, che abita nella stessa Cynolebias contains medium to large sized species, pozza annuale, sono in grande pericolo di estinzione. Cy- which can surpass 80 mm SL; the largest known nolebias akroa sp. n., sembra essere più strettamente impa- species, C. altus, presents 133.3 mm SL (Costa rentato con Cynolebias parnaibensis, con cui condivide di- 2001). Despite their size, they are not used for hu- verse caratteristiche, rispetto alle restanti specie di Cynole- bias dal bacino centrale del Rio São Francisco. Cynolebias man comsuption because they are said to have a akroa sp. n. differisce dalle altre specie di Cynolebias per la very bad taste. In some areas species of the genus colorazione del maschio, per la posizione della pinna dor- are called “peixe-sabão” (“soap fish”), but more of- sale rispetto alla pinna anale, per il numero dei raggi della ten they are popularly known as “peixe-das-nu- pinna pelvica, per la forma della papilla urogenitale e per il vens” (“cloud s fish”). pattern neurommastico cefalico. In most of the known localities of species of Cynolebias, they are sympatric with species of the INTRODUCTION annual fish genus Hypsolebias. It has been suggested The first annual fish species described from in that species of Cynolebias prey on species of the South America was Cynolebias porosus Steindachner, genus Hypsolebias (Wildekamp 1995). 1876. The stated type-locality was “Pernambuco” The new species described herein was found in a (now Recife, the capital city of the state of Pernam- tributary of the left bank of the rio São Francisco, buco in northeastern Brazil), but as discussed by situated farther away from its main channel than Costa (2001), it is difficult to ascertain the precise any other species of the genus. The aim of the pre- origin of this species, which was never collected sent contribution is to describe formally this unde- again in nature. scribed species. For a long time, the genus Cynolebias was a catch- all taxon, harboring species which are currently as- MATERIALS AND METHODS signed to several other genera of Rivulidae. In fact, a Measurements were taken point-to-point under a great proportion of the South American annual fish stereomicroscope with a digital caliper to the near- species that were described before the 1980 s were est 0.01 mm, on the left side of the specimens, then assigned to the genus Cynolebias. With the in- whenever possible, following Costa (1995, 2007). crease of studies on the family, along with fieldwork Measurements are expressed as percentages of stan- and discovery of new species, new genera were cre- dard length (SL), except subunits of the head, ated for species previously assigned to Cynolebias, which are recorded as percentages of head length such as Austrolebias, Nematolebias, , (HL). Hypsolebias, Ophthalmolebias, Xenurolebias, Plesi- In the description, counts of vertebrae and pleural olebias, Cynopoecilus, Leptolebias, Terranatos, and ribs were taken from one male and one female Maratecoara. cleared and stained (c&s) paratypes, which were The majority (15 out of 21) of the species of prepared according to Taylor & Van Dyke (1985). Cynolebias were described during the last 15 years. Terminology for frontal squamation follows aqua vol. 23 no. 4 - 15 marzo 2018 114 Dalton Tavares Bressane Nielsen, Mayler Martins, Luciano Medeiros de Araujo, Fabio Origuela and Amer Faour Hoedeman (1958) and Costa (2006). For vertebral Cynolebias akroa, n. sp. counts, the caudal compound centrum was count- (Figs 1-2; Table I) ed as a single element. Osteological features includ- ed in the description are those considered phyloge- Holotype: ZUEC 14768, male, 75.0 mm SL: Bra- netically informative by recent studies on zil, Bahia, Prazeres, temporary pool at road BA- Cynolebias (Costa 2001, 2014, 2017; Costa et al. 225, rio Preto basin (Rio São Francisco basin), 2010). Institutional abbreviations follow Sabaj 11°23’24.8”S 44°54’11.2”W, altitude 522 m, 25 (2016), with addition of UNITAU (Universidade January 2014; Mayler Martins, Fabio Origuela de de Taubaté). Lira & Luciano Medeiros de Araujo. Paratypes: ZUEC 14769, 5 males 35.1-62.5 mm

Fig.1. Cynolebias akroa, ZUEC 14768, male, holotype, 75.0 mm SL: Brazil, Bahia, Santa Rita de Cássia. Photo by D. Nielsen.

Fig. 2. Cynolebias akroa, ZUEC 14769, female, paratype, 59.2 mm SL: Brazil, Bahia, Santa Rita de Cássia. Photo by D. Nielsen.

115 aqua vol. 23 no. 4 - 15 March 2018 Cynolebias akroa, a new species of annual fish (Cyprinodontiformes: Rivulidae) from the rio Preto, São Francisco basin, northeastern Brazil

Table I. Morphometric and meristic data for the holotype cephalus, C. microphthalmus, C. oticus, C. parietalis, (H) and paratypes of Cynolebias akroa. C. parnaibensis, C. paraguassuensis, C. rectiventer, C. vazabarriensis, C. gorutuba, and 5th-7th in C. ob- H Paratypes scurus, and C. roseus), number of anterior rostral Male Males n=5 Females n=6 cephalic neuromasts 6 (vs. 2-3 in C. obscurus, C. Standard length (mm) 75.0 35.1-62.5 44.5-61.3 ochraceus, C. oticus, C. paraguassuensis, C. parietalis, Percents of standard length C. parnaibensis, C. elegans, and C. gorutuba, and 3- Body depth 28.8 26.4-29.8 29.0-30.9 4 in C. rectiventer), and urogenital papilla in males Caudal peduncle depth 15.1 11.1-14.4 13.6-15.6 cylindrical and elongated (vs. cylindrical and Pre-dorsal length 62.5 61.3-68.4 64.0-70.0 short). Pre-pelvic length 52.0 47.5-52.9 50.3-56.6 Females of Cynolebias akroa differs from females Length of dorsal-fin base 24.0 22.1-25.6 19.7-23.8 of the remaining species of Cynolebias by overall color pattern light gray, with 10-12 pale, large Length of anal-fin base 27.2 27.4-30.6 17.4-28.2 brown bars, and black spot in the center of body Caudal-fin length 26.5 27.3-30.7 20.0-28.2 sometimes absent (vs. overall color pattern light Pectoral-fin length 25.0 21.4-27.0 19.3-23.6 gray without bars and black spot always present in Pelvic-fin length 7.5 4.2-7.3 5.7-7.4 C. microphthalmus, C. porosus, C. vazabarriensis, C. Head length 35.5 32.5-33.8 31.9-34.6 perforatus, C. altus, C. leptocephalus, C. attenuatus, Percents of head length C. gibbus, and C. rectiventer, light gray body with Head depth 82.6 76.8-80.5 70.0-82.4 faint gray vertical bars, and one or two dark gray to Head width 64.5 60.3-63.4 27.5-33.6 black spots in C. itapicuruensis, C. paraguassuensis, Snout length 13,9 13.8-14.4 11.1-11.6 and C. parnaibensis, pale brownish grey, lacking Lower jaw length 25.2 24.2-27.3 22.1-24.9 spot on anteriormedial portion of flank, and 1-3 Eye diameter 24.8 24.1-28.3 20.1-23.3 dark brown blotches on humeral region in C. ob- scurus, light gray body with faint gray vertical bars, SL, 6 females 44.5-61.3 mm SL, 2 c&s, 54.7-77.9 one or two dark gray to black blotches in C. roseus), mm SL, same date as holotype. lower caudal fin length 20.0-28.2 % SL (vs. 34.1- Diagnosis: Cynolebias akroa differs from the re- 38.0 % SL in C. ochraceus or 28.7-32.4 % SL in C. maining species of Cynolebias by presenting overall parietalis). color pattern metallic golden-green, with 11-13 Additionally, Cynolebias akro can be distinguished large, light black bars (vs. 16-18 narrow, dark from the remaining species of the Cynolebias, ex- brown bars in C. obscurus, 17-19 yellow brown cept C. gorutuba, by a higher number of post-otic bars separated by narrow dark golden interspace in cephalic neuromasts (11, vs. 5-8 in C. obscurus, C. C. oticus, faint gray vertical bars (sometimes ab- ochraceus, C. oticus, C. paraguassuensis, C. parietalis, sent) in C. microphthalmus, C. vazabarriensis, C. C. parnaibensis, and C. rectiventer). itapicuruensis, C. perforatus, C. porosus, C. altus, C. Description: Morphometric data are presented in leptocephalus, C. attenuatus, C. gibbus, narrow pale Table I. Largest male specimen examined 75.0 mm golden bar in C. parnaibensis), color pattern of cau- SL, largest female 77.9 mm SL. Dorsal and ventral dal fin hyaline, without dots (vs. light gray in C. profiles convex between snout and anterior part of gorutuba, dark brown in C. paraguassuensis, dark caudal peduncle, nearly straight on caudal peduncle. grey with white dots in C. microphthalmus, C. poro- Body moderately deep, slightly compressed, greatest sus, and C. vazabarriensis, grey with bluish white body depth at level of vertical just anterior to pelvic- dots in C. itapicuruensis, C. rectiventer, C. gilbertoi, fin base. Snout very short, blunt. Dorsal and anal C. obscurus, and C. parietalis, yellowish gray with fins slightly pointed in males, without filaments, white dots in C. perforatus, C. altus, C. lepto- rounded in females. Pectoral-fin long, rounded, cephalus, C. attenuatus, C. gibbus, and C. par- reaching vertical between base of 2nd and 3rd anal- naibensis, dark ochre-yellow with bluish white dots fin rays in males, and between pelvic-fin base and in C. roseus, greenish yellow in C. oticus), seven anus in females. Tip of pelvic fin reaching between pelvic fin rays (vs. 5 in C. griseus and 6 in the re- base of 1st and 3rd anal-fin rays in males, 1st anal-fin maining species), dorsal-fin origin through base of rays in females. Pelvic-fin bases medially separated 8th or 9th anal-fin ray (vs. 3rd-6th in C. altus, C. at- by interspace. tenuatus, C. gibbus, C. itapicuruensis, C. lepto- Dorsal-fin origin through base of 8th or 9th anal- aqua vol. 23 no. 4 - 15 marzo 2018 116 Dalton Tavares Bressane Nielsen, Mayler Martins, Luciano Medeiros de Araujo, Fabio Origuela and Amer Faour fin ray in males, and above base of 4th or 5th anal- Second pharyngobranchial teeth absent. Gill-rak- fin ray in females; dorsal-fin origin between neural ers on first branchial arch 10, 8 in the second arch. spines of vertebrae 15-17 in both sexes. Anal-fin ori- Vomerine teeth absent. Urogenital papilla cylindri- gin between pleural ribs of vertebrae 13 in males, cal, large and elongated. Total vertebrae 32-33. and pleural ribs of vertebrae 12 in females. Dorsal- Coloration: Males. Side of body metallic fin rays 17 in males, 15-16 in females; anal-fin rays golden-green, with 11-13 large light black bars, 19 in males, 18-20 in females; caudal-fin rays 28-30; fading ventrally before reaching the belly and the pectoral-fin rays 13-14; pelvic-fin rays 7. anal-fin base. Opercular region metallic golden- Frontal scales irregularly arranged; 28 small green. Venter whitish. Head metallic golden, with supraorbital scales. Longitudinal series of scales 37- black lines along laterosensory series of neuromasts 39; transverse series of scales 16-18; scale rows around orbit and on postorbital region. Iris white around caudal peduncle 22. No contact organ on with black vertical bar through eye. Dorsal-fin hya- flank. Minute papillate contact organs on inner line with gray pigmentation on interradial mem- surface pectoral fin-rays in males. No contact or- branes of fin basis, and thin brown lines on inter- gans on pelvic and unpaired fins. radial membranes at distal portion of fin. Caudal- Cephalic neuromasts: supraorbital 29-39, parietal fin hyaline. Anal-fin hyaline with brown pigmenta- 4, anterior rostral 6, posterior rostral 5, infraorbital tion in the last rays. Dorsal and pelvis fins hyaline. 4+29-30, preorbital 3, otic 9, post-otic 11, Females. Sides of body light gray, with 10-12 supratemporal 5, median opercular 3, ventral oper- pale, large brown bars; single rounded black spot in cular 3-5, preopercular + mandibular 47-57, lateral the center of body, sometimes absent. Venter mandibular 8-9, paramandibular 1. whitish. Head sides yellowish gray, pale golden on One neuromast on each scale of lateral line. Two opercular region; small dark reddish brown spots neuromasts on caudal-fin base. along laterosensory series of neuromasts around or-

Fig. 3. Map from northeastern Brazil (inset) showing known localities for Cynolebias from the rio São Francisco basin.

117 aqua vol. 23 no. 4 - 15 March 2018 Cynolebias akroa, a new species of annual fish (Cyprinodontiformes: Rivulidae) from the rio Preto, São Francisco basin, northeastern Brazil bit and on postorbital region. Iris orangish yellow solved calcium (Ca) 60 mg/l, dissolved copper (Cu) with dark reddish brown vertical bar through cen- 0 mg/l. The temperature on the water surface was ter of eye. Dorsal and anal fins hyaline, with small approximately 28°C and, in the deepest portion and blue metallic spots between last rays. Caudal-fin at the margins, approximately 25°C. The region’s hyaline with small metallic blue spots. Paired fins annual average temperature is 28°C, with maxi- hyaline. mum of 34°C and minimum of 20°C. The rainy Distribution (Fig. 3): Known only from the type season ranges from December to March. The period locality, a temporary pool beside the highway BA- of collection coincided with a “Veranico” (Indian 225, at Santa Rita de Cássia, rio Preto, rio São Summer), a phenomenon which consists of a dry Francisco basin, Bahia state, Brazil. period with high temperatures within the rainy sea- Habitat (Fig. 4): The type-locality is a typical an- sons (Nielsen 2008). Together with Cynolebias akroa nual pool of the Caatinga, located about 96 km were also found Hypsolebias faouri, Astronotus ocella- north of Barreiras. The substrate is composed of clay tus and Oreochromis niloticus (Figs 5-6). and sand, with a slightly dark water. There is a dense The pool is close to a road near a small village. A cover of aquatic vegetation, mostly composed by family living in this village introduced Oreochromis Echinodorus sp. and Nymphaea sp. The average niloticus and Astronotus ocellatus on the pool, as a depth of the pool is 100 cm, with deepest portions means of having an alternative source of protein. about 130 cm, pH 6.5 and a low electric conductiv- At one end the pool was dug so the water would re- ity (146µs). Other physico-chemical parameters main throughout the year, so part of the pool does were: total hardness (GH) 2°dGH, carbonate hard- not dry, which explains the survival of the intro- ness (KH) 4°dKH, dissolved iron (Fe) 1.0 mg/l, dis- duced species.

Fig. 4. Type locality of Cynolebias akroa, Brazil, Bahia, temporary pool at road BA-225. Photo by M. Martins. aqua vol. 23 no. 4 - 15 marzo 2018 118 Dalton Tavares Bressane Nielsen, Mayler Martins, Luciano Medeiros de Araujo, Fabio Origuela and Amer Faour

Etymology: In honor Akroá ethnic, an ethnic The area of occurrence of the new species is locat- group belonging to Macro-Jê linguistic trunk, that ed approximately 175 km in a straight line from lived in the region until the century XIX, when the Rio São Francisco, but the real distance from they were decimated. its main channel, considering the actual rivers courses is approximately 570 km. The middle Rio DISCUSSION São Francisco basin lies within the Caatinga do- Cynolebias akroa is easily distinguished from its main, an area with a relatively high rate of biodi- congeners by the presence of light dark vertical bars versity, with high level of speciation for rivulids, in the flank. This color pattern is unique within but with a consequently limited distribution of the genus Cynolebias, other species of the genus each species. The accelerated pace of environmen- may present vertical bars, but never as wide and tal degradation, due to the disorderly occupation conspicuous as the ones present in the new species. of the area by human activities, poses a great risk of extinction for rivulid species occurring at the Caatinga domain (Costa 2017). Cynolebias species are commonly found in annual pools sintopically with two other annual fish species of the genus Hypsolebias, typically one species be- longing to the Hypsolebias magnificus species group and other species belonging to the Hypsolebias flav- icaudatus species group. Rarely, Cynolebias species are found alone in an annual pool. Cynolebias akroa was found syntopically with Hypsolebias faouri. Hypsolebias faouri is closely related to H. igneus (Britzke et al. 2016), a species that occurs in annual pools at the banks of the Rio São Francisco, near the Rio Preto mouth with the Rio São Francisco. Cynolebias akroa could be putatively closely related to Cynolebias roseus, considering the geographic Fig. 5. Astronotus ocellatus, a specimen of an exotic invasive proximity between the two species and a similar dis- species, collected at the type-locality of Cynolebia akroa. Photo by M. Martins. tribution of H. faouri and H. igneus. However, Cynolebias akroa appears to be more closely related to C. parnaibensis, since both species share the fol- lowing characteristics: color pattern with vertical bars, absence of contact organs in the body sides of males vomerine teeth absent, second pharyngob- ranchial teeth absent, a similar count of dorsal-fin rays (17 in males vs. 16-18 in other species), a com- mon count of dorsal-fin rays in females (15-16), a common count of anal-fin rays (19 in males and 18-20 in females), a common count of caudal-fin rays (28-30), a common count of pectoral-fin rays (13-14), frontal scales irregularly arranged, dorsal fin origin vertebrae 15-17, and the presence of minute papillate contact organs on inner surface pectoral fin-rays in males. Interestingly, the clade formed by H. faouri and H. igneus is closely related to the clade formed by H. coamazonicus and H. martinsi (Britzke et al. 2016), which is similar to the putative biogeographic relation between C. akroa Fig. 6. Oreochromis niloticus, a specimen of an exotic inva- from the middle Rio São Francisco basin with C. sive species, collected at the type-locality of Cynolebia parnaibensis from the Rio Parnaíba basin. akroa. Photo by M. Martins. In spite of the modification of its habitat, with

119 aqua vol. 23 no. 4 - 15 March 2018 Cynolebias akroa, a new species of annual fish (Cyprinodontiformes: Rivulidae) from the rio Preto, São Francisco basin, northeastern Brazil the establishment of a connection to a permanent COSTA, W. J. E. M. 2009. Morphology of the teleost pha- water body and with the invasion of exotic fish that ryngeal jaw apparatus in the Neotropical annual killifish potentially can act as predators as Astronotus ocella- genus Cynolebias (Cyp rinodontiformes: Aplocheiloidei: tus and Oreochromis niloticus, Cynolebias akroa still Rivulidae). Cybium 33: 145-150. COSTA, W. J. E. M. 2010. Historical biogeography of survives in its habitat. Cynolebias akroa is not the cynolebiasine annual killifishes inferred from dispersal-vic- most abundant species of annual fish at its habitat, ariance analysis. Journal of Biogeography 37: 1995-2004. being outnumbered by Hypsolebias faouri, with a COSTA, W. J. E. M. 2011. Parallel evolution in population ratio of 5: 1 in relation to H. faouri, ichthyophagous annual killifishes of South America and which may indicate that C. akroa preys on H. Africa. Cybium 35: 39-46. faouri, as reported for other syntopic associations COSTA, W. J. E. M. 2014. Six new species of seasonal kil- among annual rivulids (e.g., Wildekamp 1995). lifishes of the genus Cynolebias from the São Francisco The morphological characteristics of the teeth, typ- river basin, Brazilian Caatinga, with notes of C. porosus (Cyprinodontiformes: Rivulidae). Ichthyological Explo- ical of piscivorous fishes, common to all Cynolebias ration of Freshwaters 25: 79-96. species and shared with other annual fishes from COSTA, W. J. E. M. 2017. Description of two endangered both South America and Africa, as Austrolebias new seasonal killifish species of the genus Cynolebias from elongatus and Nothobranchius ocellatus, which are the São Francisco River basin, Brazilian Caatinga reported to prey on smaller sympatric congeners (Cyprinodontiformes, Aplocheilidae). Zoosystematics and (Costa 2006, 2009, 2011, 2017), also points to Evolution 93 (2), 333-341. this possible relation of predator-prey between COSTA, W. J. E. M. & BRASIL, G. C. 2001. The neotropi- Cynolebias akroa and Hypsolebias faouri. cal annual fish genus Cynolebias (Cyprinodontiformes: Rivulidae): phylogenetic relationships, taxonomic revi- sion and biogeography. Ichthyological Exploration of Fresh- ACKNOWLEDGEMENTS waters 12: 333-383. Thanks are due to Itamar Alves Martins from COSTA, W. J. E. M., LACERDA, M.T.C & BRASIL, G. C. Universidade de Taubaté (UNITAU) for laboratory 1990. Description de deux nouvelles espèces du genre support and Flávio C. T. Lima (ZUEC), who read Cynolebias du bassin du Rio Tocantins (Cyprinodontifor- the manuscript and offered useful suggestions. mes, Rivulidae). Revue Française d’Aquariologie et Herpe- tologie 17: 9-14. REFERENCES COSTA, W. J. E. M, SUZART, R. & NIELSEN, D. T. B. 2007. Cynolebias paraguassuensis n. sp. (Teleostei: Cyprinodon- BELOTE, D. F. & COSTA, W. J. E. M. 2003. Reproductive behavior of the Brazilian annual fish Cynolebias albipunc- tiformes: Rivulidae), a new sazonal killifish from Brazi- tatus Costa & Brasil, 1991 (Teleostei, Cyprinodontif- lian Caatinga, Paraguaçu River basin. aqua, International ormes, Rivulidae): a new report of sound pro duction in Journal of Ichthyology 12 (3): 129-133. fishes. Arquivos do Museu Nacional 61: 241-244. COSTA, W. J. E. M., RAMOS, T. P. A., ALEXANDRE, L. C., RAMOS, R. T. C. 2010. Cy nolebias parnaibensis, a new BRITZKE, R., NIELSEN, D. T. B. & OLIVEIRA, C. 2016. Description of two new species of annual fishes of the seasonal killifish from the Caatinga, Parnaíba River basin, Hypsolebias antenori species group (Cyprinodontiformes: northeastern Brazil, with notes on sound pro ducing Rivulidae), from Northeast Brazil. Zootaxa 4114 (2): courtship behavior (Cyprinodontiformes: Rivulidae). 123-138. Neo tropical Ichthyology 8: 283-288. HOEDEMAN, J. J. 1958. The frontal scalation pattern in COSTA, W. J. E. M. 1995. Pearl killifishes: The Cynolebiati- nae. Systematics and Biogeography of a Neotropical annual some groups of tooth carps. Bulletin of Aquatic Biology 1: fish subfamily (Cyprinodontiformes: Rivulidae). TFH Pub- 23-28. lications, Neptune City, 128pp. NIELSEN, D. T. B. 2008. Simpsonichthys e Nematolebias. Cabral Editora e Livraria Universitária, Taubaté, 235pp. COSTA, W. J. E. M. 2001. The neotropical annual fish genus Cynolebias (Cyprinodontiformes: Rivulidae): phylogenetic SABAJ, M. H. 2016. Standard symbolic codes for institu- relationships, taxonomic revision and biogeography. tional resource collections in herpetology and ichthyolo- Ichthyological Exploration of Freshwaters 12: 333-383. gy: an Online Reference. Version 6.5 (16 Aug 2016). Electronically accessible at http://www.asih.org/, Ameri- COSTA, W. J. E. M. 2006. Descriptive morphology and phylogenetic relationship among species of the Neotrop- can Society of Ichthyologists and Herpetologists, Wash- ical annual killifish genera Nematolebias and Simp- ington, DC. sonichthys (Cyprinodontiformes: Aplocheiloidei: Rivuli- TAYLOR W. R. & VAN DYKE G. C. 1985. Revised pro- dae). Neotropical Ichthyology 4: 1-26. cedures for staining and clearing small fishes and other ver- tebrates for bone and cartilage study. Cybium 9: 107-109. COSTA, W. J. E. M. 2007. Taxonomic revision of the sea- sonal South American killifish genus Simpsonichthys WILDEKAMP, R. H. 1995. A world of killies: atlas of the (Teleostei: Cyprinodontiformes: Aplocheiloidei: Rivuli- oviparous cyprinodontiform fishes of the world, volume 2. dae). Zootaxa 1669: 1-134. American Killifish Association, Mishawaka, 384 pp. aqua vol. 23 no. 4 - 15 marzo 2018 120 aqua, International Journal of Ichthyology

Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”?

Mandy T. Etpison1 and Patrick L. Colin2

1) Etpison Museum, PO Box 7049, Koror, Palau 96940. Email: [email protected] 2) Coral Reef Research Foundation, PO Box 1765, Koror, Palau 96940. Email: [email protected]

Received: 13 December 2017 – Accepted: 05 March 2018

Keywords am Morgen zu den Laichplätzen, schlossen sich zu Gruppen Predation, aggregation, feeding frenzy, gray reef shark, zusammen und bewegten sich über der Rifffläche auf und lunar periodicity. ab und zogen dabei die Aufmerksamkeit von Beutegreifern auf sich. Um die Mittagszeit steigen sie vom Riff auf und Abstract begeben sich ins freie Wasser jenseits vom Riff. Graue Spawning aggregations of the moorish idol (MI) and or- Riffhaie folgen ihnen, greifen sie an der Oberfläche an und angespine surgeonfish (OSS) were found on the western verzehren viele von ihnen in einem Fressrausch. Ein hoher barrier reef of Palau. MI aggregated around the first quar- Prozentsatz der aufsteigenden erwachsenen HF wird von ter moon from Dec. to Mar., with largest groups in Jan. den Haien gefressen, nur wenige können in die sichere Zone and Feb. Fish arrived near the sites in the morning, des Riffs zurückkehren. KD versammeln sich in denselben grouped together and moved up and down the reef face up Monaten, aber in der Zeit des letzten Mondviertels – wobei in late morning attracting the attention of predators. At es hierüber weniger Berichte gibt. Die Beobachtungen bei mid-day they ascend from the reef out into open water beiden Fischarten, dass sie weit nach oben steigen und sich away from the reef. Gray reef sharks follow them and at- zum Ablaichen vom Riff entfernen, gelten als ungewöhn- tack at the surface in a feeding frenzy. A high percentage of lich; das Verhalten wird als „Blauwasserlaichablage“ bezeich- the ascending adults are eaten and few return safely to the net, und es gibt wenige weitere Beispiele. Bisher hatte man reef. OSS aggregated in the same months, but on the last die Bedeutung von Riffhaien als Einfluss auf das Laichver- quarter moon with fewer observations being made. The halten von Rifffischen als bedeutungslos bis „mäßig“ observation of both fishes ascending high above and mov- eingestuft (nur wenige laichende Fische würden von Haien ing away from the reef to spawn is unusual and is termed gefressen, hieß es bisher). Das Beispiel der HF und KD, bei “blue water spawning” with only a few similar examples denen sehr viele Individuen von Beutegreifern gefressen known. Previously the importance of reef sharks in influ- werden, ist als Extrem zu werten, das man bisher nur von encing reef fish spawning behavior has been reported as Zackenbarsch-Gruppen kennt. Das Auftreten von Haien an non-existent to “moderate” (a few spawning fish taken by den Laichplätzen während der Gruppenbildung und Laich- sharks). This example of many individuals being taken by ablage lässt auf eine enge Beziehung zu den Rifffischen predators represents an extreme only reported previously schließen. Die offensichtlich hohe Erfolgsquote beim Er- for a grouper aggregation. The occurrence of sharks at the beuten von HF und KD mag eine Besonderheit dieser er- site during aggregation and spawning is indicative of a forschten Laichplätze sein, und es ist wahrscheinlich, dass close relationship with reef fishes. The apparent high rate die einzelnen Fische sich grundsätzlich iteropar verhalten, of predation on spawning MI and OSS may be specific to also Verluste durch erneutes Ablaichen ausgleichen können. these study sites and it is likely individual fishes are gener- ally iteropareous. Résumé Les concentrations de frai de l'idole maure (MI) et orange- Zusammenfassung spine poisson chirurgien (OSS) ont été trouvés sur la bar- Am westlichen Barriereriff Palaus konnte an Halfterfischen rière de corail de l'ouest de Palau. Regroupés autour de la MI (HF) Zanclus cornutus und an Kuhkopf-Doktorfischen Premier quartier de lune à partir de Décembre à Mars, avec (KD) Naso lituratus gruppenweises Ablaichen beobachtet des groupes plus importants en Janvier et Février. Les pois- werden. HF versammelten sich etwa zur Zeit des ersten sons sont arrivée en proximité des sites le matin, regroupés Mondviertels von Dezember bis März, mit den größten et déplacé vers le haut et vers le bas le récif de face dans la fin Gruppierungen im Januar und Februar. Die Fische kamen de matinée attire l'attention des prédateurs. À la mi-journée

121 aqua vol. 23 no. 4 - 15 March 2018 Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”? ils montent du récif out simplement dans l'eau loin du récif. few decades (Sadovy de Mitcheson and Colin Les requins gris de récif de les suivre et d'attaque à la surface 2012) but basic information is still lacking for dans une frénésie d'alimentation. Un pourcentage élevé de many species that aggregate or are likely to do so, l'ordre croissant adultes sont mangés et peu de retourner en limiting the understanding of relationships of toute sécurité dans le récif. OSS regroupés dans le même mois, mais le dernier quartier de lune avec moins d'observa- spawning behavior, interspecific relationships and tions effectuées. L'observation des les deux poissons ascen- environmental conditions. In many areas aggrega- dant au-dessus et en s'écartant de la reef pour frayer est in- tions have been heavily fished, often to ecological habituelle et est appelé "l'eau bleue" de frai avec seulement extinction (they no longer form). Palau in the west- quelques exemples similaires connues. Auparavant, l'impor- ern tropical Pacific is an exception, retaining nu- tance des requins de récif à influencer le comportement de merous spawning aggregations, making it an ideal frai des poissons de récif a été rapporté comme inexistant à location for their study. It also has fishermen and "modérée" (un peu de la fraie des poissons pris par les re- tourist dive guides who are observant and often quins). Cet exemple d'un grand nombre de personnes prises par les prédateurs représente un extrême seulement rap- discover new knowledge about fish reproductive portés précédemment pour une agrégation de mérous. L'ap- behavior adding to traditional knowledge known parition de requins sur le site au cours de l'agrégation et la for decades, if not centuries (Johannes 1981, reproduction est l'indice d'une relation étroite avec poissons Sadovy 2007). Many such reports are now validat- de récif. L'apparente taux élevé de prédation sur les frayères ed scientifically by in situ observations with infor- MI et de l'OSS peuvent être spécifiques à ces sites d'étude et mation on aggregating fish species also included in il est probable que les poissons sont en général iteropareous. a number of popular books and publications, such as Colin (2009) and Etpison (2009, 2014). Sommario Presso la barriera corallina occidentale di Palau sono state Most reef fishes with planktonic eggs, both male- osservate aggregazioni di idoli moreschi (MI) e di pesci female pairs and group spawners with 3 or more unicorno arancione (OSS) legate alla fase della deposizione fish, remain within sight of shelter when spawning delle uova. Gli MI si sono riuniti intorno al primo quarto and release their gametes within 5-10 m of struc- di luna da dicembre a marzo, con gruppi più numerosi a tures to which they can quickly retreat for shelter gennaio e febbraio. I pesci sono arrivati vicino ai siti al from predators. They engage in a rapid vertical or mattino, si sono raggruppati e mossi su e giù per la barriera angled movement away from the bottom, the a tarda mattinata attirando l'attenzione dei predatori. A “spawning rush”, with eggs and sperm released at metà giornata sono risaliti dalla scogliera e si sono portati in acque aperte lontano dalla barriera corallina. Gli squali the peak and fishes quickly returning to near the grigi della scogliera li hanno seguiti e attaccati in preda a bottom. This behavior is generally believed to re- una frenesia alimentare. Un'alta percentuale di adulti è sta- duce egg predation from benthic-based plankti- ta mangiata e pochi sono ritornati sani e salvi nella barriera vores (eggs released high above the reef) (Nemeth corallina. Gli OSS si sono ammassati negli stessi mesi, ma 2012) and predation on the spawners by piscivores nell'ultimo quarto di luna si sono fatte meno osservazioni. (Molloy et al. 2012). L'osservazione di entrambi i pesci che risalgono la colonna The aggregation and spawning of two reef fishes, d’acqua e si allontanano dalla scogliera per deporre le uova moorish idol, Zanclus cornutus (Linnaeus, 1758) è insolita e viene definita "deposizione in acque blu" con solo pochi esempi simili noti. Precedentemente l'impor- (hereafter referred to as “MI”) and the orangespine tanza degli squali di barriera nell'influenzare il comporta- surgeonfish, Naso lituratus (Bloch & Schneider, mento di deposizione dei pesci nella barriera corallina è 1801) (hereafter referred to as “OSS”), violate this stata segnalata come inesistente o "moderata" (pochi pesci generality by swimming high above and far away riproduttori catturati dagli squali). Questo esempio di from the reef to release their gametes, even with a molti individui presi dai predatori rappresenta un caso es- high predation risk from sharks. Our information tremo segnalato in precedenza solo per un'aggregazione di predominantly concerns MI, but includes some cernie. La presenza di squali nel sito durante l'aggregazione comparative data on OSS, which use the same e la deposizione delle uova è indicativo di una stretta re- lazione con i pesci della barriera corallina. L'apparente alto spawning sites on a different lunar schedule. We tasso di predazione su MI e OSS durante la deposizione term this reproductive strategy “blue water spawn- delle uova può essere specifico per questi siti di studio ed è ing” in which the spawning fishes range 50-100 probabile che questi pesci siano specie iteropare. meters away from the shelter of the reef, often to the surface, to spawn. While doing this, the adults INTRODUCTION may suffer high predation rates from sharks. The The reproduction of reef fishes forming spawning large numbers of sharks which gather, then focus aggregations has become better known in the last their attention on the spawning fish, begs the ques- aqua vol. 23 no. 4 - 15 marzo 2018 122 Mandy T. Etpison and Patrick L. Colin tion of whether sharks gather to specifically target Indo-Pacific tropics. Colin et al. (2012: 513, 515) the spawning aggregations and also why these fish reported a few instances of pair and aggregation put themselves at such risk of predation?. spawning in Palau from December through March The MI is an iconic fish found throughout the with peaks in January and February. The schooling

Figs 1a-b. (a) Landsat 7 images of the main Palau Island group with area shown in right panel indicated. (b) Detailed area of western barrier reef with locations mentioned in text indicated.

123 aqua vol. 23 no. 4 - 15 March 2018 Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”? of adults preparatory to spawning on outer reef Feb 2010, 3) Dec 2011-Jan 2012, 4) Jan 2015, 5) slopes has been known since the 1990s (Etpison Jan 2016 and 6) Dec 2016-Jan 2017. Additional ob- 1994, 2009) while in 2009 the ascent of large num- servations made on an irregular basis at a series of bers from the reef in mid-day was first seen (Etpison other outer reef areas; Peleliu, New Drop-off, Shark 2014). Elsewhere a few pair spawns were reported at City, Sandy Paradise and Elas (Fig. 1b). Vertical and Johnston (Sancho et al. 2000) and Enewetak Atolls oblique aerial photographs of the sites were used to (Colin & Bell 1991) but none for group spawning. plot the areas of fish presence and their movements The maximum life span is not known but they are (Fig. 2). Underwater photographs from these sites reported to live up to several years in aquaria. There were used to identify and map underwater locations is no external way to distinguish sexes. relative to the aerial images. GPS surveys of areas es- The OSS is common and distinctive on outer reef tablished latitude/longitude positions of features on slopes often seen some distance above the bottom. the reef and helped to quantify the geography of the Johannes (1981) indicated Palau fishermen report- fish movements. ed aggregation, but provided no specific informa- Observations were made while SCUBA diving or tion. OSS occur in large schools, recognized as pos- snorkeling. Initially observation dives were made at sible spawning aggregations, off the Palau barrier all seasons, lunar phases and time of day by many ob- reef from December to March, often in the compa- servers regarding presence/absence of fishes. Over ny of other acanthurids (Etpison 2004, 2009). several years the times during which groups of a hun- These schools were subject to attack by groups of dred to several thousand fish (termed a “running gray reef sharks, Carcharhinus amblyrhynchos, school”) move together one direction along the reef hereafter referred to as “GRS” (Colin 2012). Taylor face, then turn in near unison at the ends of their et al. (2014) reported a life span maximum of 14 swim pattern and move in the opposite direction, oc- years in Guam while sexual maturation occurred at curred were narrowed down and associated with par- 15 and 18 cm fork length for male/female OSS. ticular seasons and lunar phases. Behavior was docu- mented using diver operated digital still and video MATERIALS AND METHODS cameras with the numbers of fishes participating in Most observations were made at two locations on the aggregation and ascending to spawn counted the outer slope of the western barrier reef; “Blue from photographs. After groups of fishes left the reef Corner” (hereafter referred to as “BC”, 7°08.08’N; to spawn due to the speed of their swimming, divers 134°13.25’E) and “Siaes Corner” (hereafter referred could no longer follow, and we used boats to track to as “SC”, 7°18.85’N; 134°13.22’E) (Fig. 1); sites the fishes from the surface. At those times documen- that are visited daily by many divers. Our observa- tation was limited to GoPro cameras held over the tions on occurrence and behavior were centered on side of the boat on poles to record the activity. six aggregation periods: 1) Jan 2009, 2) Dec 2009-

Fig. 2a-b. (a) Physiography of the aggregation/spawning at Siaes Corner (SC) reef. (b) Physiography of aggregation/ spawn- ing site to Blue Corner (BC) reef. aqua vol. 23 no. 4 - 15 marzo 2018 124 Mandy T. Etpison and Patrick L. Colin

RESULTS aggregation occurs during decreasing temperature A. Geography and currents (or tides) of the ag- from late December (about 29°C) to March gregation sites. Palau has a semi-diurnal tide with (around the annual low of 28°C). spring amplitudes up to 2.3 m on full and new The western barrier reef of Palau is typically 1.0- moons around sunrise and sunset. The first and 1.5 km wide between lagoon and ocean (Fig. 1) with last quarter (“half”) moons have neap tides with alternating cross reef (lagoon-ocean) currents driven high tides occurring at mid-day with amplitude by the tides. Our primary study sites, while part of around 0.7-1.2 m. Water temperatures (Fig. 3) the barrier reef, are different from typical areas. BC range annually between about 28° and 30°C while has a distinct promontory and a much wider reef flat

Fig. 3. Annual daily mean temperature profile for Ulong Rock 11 m depth, near study sites, averaged for the years 2009- 2016 and the first half of 2017. Green arrows indicate full seasonal aggregation and spawning period while red arrow indi- cates the peak period.

125 aqua vol. 23 no. 4 - 15 March 2018 Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”? with islands on it somewhat restricting the lagoon- area, angle away from the reef and move towards ocean flow. Currents in such areas tend to be domi- the surface. They are subsequently driven further nated by along reef oceanic flows which often di- offshore by the sharks which follow their ascent. verge at the promontories. SC has a projection of At SC the MI were observed to migrate to the site reef, less prominent than BC, a similarly wide reef from both the north and south. This area has two flat (but without islands), and along reef currents distinct “resting areas” while the location where the which diverge (Fig. 1b). There are still some tidally- fish ascend to spawn lies between them, 100 and influenced currents at both sites, the patterns of 450 m away. The reef corner where ascent occurs which have been characterized from diver observa- has a distinct transition with its southern wall, ver- tions over many years (Etpison 2009, 2016) as “in- tical from depths of 20 m to 50 m, transitioning to coming” (rising) and “outgoing” (ebbing). The cur- a slope leading downward to depths of about 60 m, rents at these sites need further quantification. then becoming near vertical. In roughly the hour The two sites on the western barrier reef are 32 km prior to spawning sharks gather in large numbers apart along the reef edge (20 km straight line dis- along the vertical wall at the ascent area. tance) (Fig. 1). During winter, when the aggrega- B. Moorish idols. Numbers and behavior in ag- tions occur, Palau normally has northeast trade gregations. From the late 1990s to 2005, the MI ag- winds of 10-15 knots and the sites are protected gregations observed every year by dive guides at BC from NE winds and waves in the lee of the reef. were estimated to be 5,000-10,000 fish. During During the summer, the non-spawning season for 2009-2012 the numbers aggregating were estimated these species, the western reefs often have strong visually at BC appeared to decrease while the num- south to southwesterly monsoon winds, which pro- bers at SC increased. During 2015-2017 the largest duce rough conditions with high surf on the reef. groups, estimated at 1,000-3,000 individuals, were The distinct promontory at BC gradually slopes seen at SC while only a few hundred fish were seen at its outer end becoming steep to vertical faces aggregating at BC each season. These numbers must around its perimeter starting at 5 to 30 m depth. A be considered only as rough estimates obtained from discrete area (termed the “resting area”) about 300 observations without any means to accurately quan- m north of the promontory end and only about tify the numbers of individuals. 60-70 m in length is utilized by MIs during rising Stages of aggregation and spawning. There are tides as a gathering place prior to moving several sequential stages and events involved with up and away to spawn (Fig. 2). Fish seen migrating the aggregation and spawning. These will be de- to this site from the south come around the tailed in sequence. promontory of BC from distances of at least 1 km. Schooling and aggregation along the reef – Oc- It is uncertain whether fish also migrate to the site currence of aggregations, aggregation size and from the north so overall the catchment area for numbers. Starting in 2009 aggregations were seen this aggregation is uncertain. When ready to at BC and SC for roughly 6 days around the first spawn, the fish will rise from the middle of this quarter moon; a period of neap tides with high wa-

Figs 4a-b. (a) Aggregation of MIs “running” along reef at BC, 7 Jan 2017, 11:18 AM one day after first quarter moon. The photo has an approximately 668 fish based counts of individuals visible in photo. (b) A portion of an aggregation is seen “running” along the reef (with diver for scale) on 15 Jan 2016, 12:40 PM, two days before first quarter moon. Photos copy- right M. T. Etpison. aqua vol. 23 no. 4 - 15 marzo 2018 126 Mandy T. Etpison and Patrick L. Colin ter occurring at mid-day. Spawning ascents with aside to get closer. The white tip reef sharks and gi- the most fish were generally seen from two days be- ant trevally jam themselves under the rocks trying fore to the day after the first quarter moon. The ag- to get to the fish, often sustaining visible scraps and gregation appears at the site during the morning on scratches as a result. Humphead wrasses, groupers the rising tide, moving up and down the reef (“run- and twin-spot snappers, relying more on their eye- ning”) as a tight school at 20 to 35 m depth, fish sight and speed, wait for others to flush out the turning in synchrony with some distance between hiding fish and then pursue it. The moray eels are individuals. Groups varied in size with a minimum often able to maneuver into locations where the of 100-200 individuals up to multiple hundreds. fish is sheltering and, if not quickly eaten by the For example, one photos (Fig. 4) has approximate- eel, the MI will subsequently flee the shelter. Once ly 668 individuals visible in it. again in the open, the waiting predators then at- The school moving up and down the reef was of- tack and often succeed in capturing it. ten trailed by large numbers of grey reef sharks, The ascent of MI towards the surface and away Carcharhinus amblyrhynchus (subsequently referred from the reef. While swimming and then milling to as GRS). Other predators such as small white tip along the reef no behavior by MI interpreted as reef sharks (Triaenodon obsesus), giant trevally courtship was seen. However, near the time of high (Caranx ignobilis), brown-marbled grouper (Epi- tide the movement of the group up and down the nephelus fuscoguttatus), twin-spot snapper (Lutjanus reef ceases, the fish moving towards the ascent lo- bohar), humphead wrasse (Cheilinus undulatus) cation after the tide switches to outgoing and as- and moray eels (Gymnothorax javanicus) wait on cending somewhat as a group (Fig. 6). The GRS re- the reef and work as multi-species hunting packs. main close by, often to the side and slightly above As the tide rose towards high water, these predators of the group. About one half to one hour after high attempted to separate individuals or small groups tide the MI start the ascent and spawning soon fol- of MIs from the school (Fig. 5a) by rushing the lows. They start by initiating partial ascents in group from a distance. The individual MIs are fair- which the fish rise as a group for several meters, but ly safe when in the larger school on the reef, as the then turn back towards the reef. We interpret this school itself is not attacked, but if an individual is as a “false start” to the ascent; a type of behavior separated, several predators immediately dive to- seen in many other reef fishes prior to the actual wards it and, if not eaten immediately, force it to spawning ascent and gamete release (Sadovy de take shelter in crevices or beneath rocks or coral Mitcheson and Colin 2012). Numerous GRS con- heads (Fig. 5b). Once a fish is cornered under shel- tinue to shadow their movements and then return ter, the predators chaotically try to push each other to near the bottom when the MI do.

Figs 5a-b. (a) Three small white tip reef sharks, a giant trevally, a twin spot snapper and a large male humphead wrasse at- tempting to prey upon a MI which has taken shelter under a small coral head. Photo taken January 2016 (b) Predators going after MI under coral head. Two moray eels are beneath the coral head with only their sides visible through small openings in the reef. Photo taken December 2014. Photos copyright M. T. Etpison.

127 aqua vol. 23 no. 4 - 15 March 2018 Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”?

Groups of fish ascending, as counted from photos Presence/Predation. The actual spawning of as- with clearly distinguishable individuals, numbered cending groups was seldom observed, although 290, 311, 360 and 544 (Figs 7 and 8), while others Colin et al. (2012) did see one instance of mass where individuals could not be readily counted had spawning at the conclusion of the ascent. Given apparently similar numbers. Groups approaching the behavior of the fish swimming upward away 1,000 fish rise towards the surface for spawning. from the reef, we make the assumption, in those The aggregation often splits up when preparing to instances seen, that their intention was to spawn in ascend and not all individuals ascend at the same very shallow water, as seen previously. The presence time. In early years we believed the grey reef sharks of sharks following the ascending groups compli- were actively herding groups forcing them to move cated observations. In January 2015 we followed away from the reef, however, since 2009 we have several groups of fish up from the reef out into repeatedly observed MI schools move off the reef open water, but lost sight of them and the sharks. and ascend by themselves, at times with no sharks At short time later, while we were hovering about 6 nearby. The column of fish leaving the reef remains m below the surface in mid-water with no fish cohesive with the individual fish staying a few body around, a small school of MIs came to us out of lengths apart and turning in synchrony. When open water, then milled about apparently trying to splitting into several smaller schools during ascent, hide amongst the diving observers. No sharks were the subgroups often move in different directions. present at that time and a few minutes later the The GRSs do not make obvious moves to force the school suddenly darted away towards the surface fish to ascend, but rather seem to anticipate when and out of sight. Spawning by them was not ob- and where this will take place. Once clear of the served. Due to the potential danger to diving ob- reef the ascending column of fish forms an elon- servers from sharks when the MIs reach the surface gate mass or ball (Fig. 7). The GRS take up posi- and presumably spawn, divers would return to the tion at the lower end of the group (Fig. 8), not im- boat and then motor outward to areas where the peding their rise, but remaining close beneath the thrashing of the sharks attacking the MIs was read- MI potentially discouraging their return to the ily apparent on the surface. reef. In essence, the ascent now is irreversible and In January 2015, 2016 and 2017 we observed and the fish are committed to continuing the ascent to took photos/video of the spawning movements of its conclusion. As the group of MI and sharks as- the MI at the two sites. At SC in January 2016 over cend from the reef, they immediately get caught in 100 grey reef sharks (numbers determined from whatever currents are moving along the reef face. photographs) gathered in the spawning area. In These, combined with the now falling tide bring- January 2017 the GRS numbered over 200 at their ing water from the lagoon across the reef to the peak. During morning incoming tides as the aggre- ocean, tend to push the groups even further away gation formed on the fore reef, fifty or more GRS from the reef. were seen trailing the MI aggregation; swimming Spawning Behavior in Blue Water and Shark in an unhurried manner, but occasionally diving at and attacking the schools. When the MI school gathered to ascend from the reef, the sharks be- came extremely agitated and aggressive, following the MI school in a tight cluster. We had seen close to 100 GRS daily along the reef for several days, but when the MI ascended off the reef, they were joined by an additional 100 GRS (numbers deter- mined from photographs), which had evidently been present in open water off the reef just beyond the limits of visibility (about 30 m). The MIs split into several groups once they had risen away from the reef. One group was seen to re- turn directly back to the reef while all others swam Fig. 6. Gray reef sharks holding station above an aggrega- directly away from the reef and up towards the sur- tion of MIs shortly before they ascended off the reef to face (Fig. 8). Unable to keep up with the idols and spawn. 15 Jan 2016. Photo copyright M. T. Etpison. sharks on scuba, observers returned to the boat. A aqua vol. 23 no. 4 - 15 marzo 2018 128 Mandy T. Etpison and Patrick L. Colin short time later the boat arrived at an offshore lo- surface, ending up, based on GPS positions, three cation where five schools of MI being attacked by km from the nearest reef (Fig. 10). Numbers of packs of GRS were observed on the surface a hun- GRS were shadowing small groups of actively dred meters or more away from each other (Fig. 9). swimming MI at the surface but once the MI num- Several small silvertip sharks (Carcharhinus albi- bers were reduced and dispersed by the predation, marginatus) were also filmed trailing the grey reef those individual MI left appeared to be more effec- sharks. Our boat had to move at 10 knots speed tive in eluding the sharks. However, the sharks just to keep up with movements of the frenzied were seen to slowly pick off the exhausted MI, fishes on the surface. GoPro cameras on poles, held which had nowhere to hide or shelter, one by one. over the side of the boat while it motored along, MI also tried to station themselves under our boat, were immediately and repeatedly attacked by GRS; which was moving along at several knots. The pres- one camera housing was crushed by a shark bite ence of the boat trailing the MI and sharks may while another was badly scratched after several well have affected their interactions, but by the sharks bit it. At this time the sharks were in a full time our boat-based observations were broken off, feeding frenzy and would certainly have attacked there were only a few MI left on the surface. Our anything in the water or moving on the surface. observers returned to the reef and dove again, and After several minutes the chaotic feeding frenzy, during subsequent observations, no MI were seen during which many of the MIs were eaten, it returning directly to the reef. ceased, the remaining MI formed schooling groups Post spawning behavior. On some occasions when observed to have up to 20 individuals which swim- observers remained on the reef after the spawning ming together at the surface in open water. The ascent started and MI disappeared offshore, rem- sharks continued to chase the remaining fish at the nants of the schools of MIs were observed returning

Figs 7a-g. Variation in size and shape of ascending groups of MI at BC and SC. (a-d) General groups ascending from the reef, numbers not specified. Ascending groups of (e) approximately 290 fish, (f) 311 fish, (g) approximately 392 fish. A num- bers based on counts of individuals from photos. Photos copyright M. T. Etpison.

129 aqua vol. 23 no. 4 - 15 March 2018 Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”?

Fig. 8. Ascending aggregation of MIs (approximately 544 fish visible in photo) being shadowed by gray reef sharks, as they move towards the surface, Jan 2015. Photo copyright M. T. Etpison. aqua vol. 23 no. 4 - 15 marzo 2018 130 Mandy T. Etpison and Patrick L. Colin to the reef from open water after their ascent. It is same as MI. They have a different lunar timing, ag- unknown if those returning fish had successfully gregating and spawning on the last quarter moon spawned or aborted their run out into open water (approaching the new moon) instead of the first prior to spawning. Almost immediately they dis- quarter moon. OSS have been seen to use an addi- persed and started feeding on the shallow reef slopes tional location, called “New Drop Off” (approxi- during the remainder of the outgoing tide, mixing mately 7°6.15’N; 134°14.29’E), not known as a with other individuals which may not have been spawning site for MI. They are more challenging ready to spawn. The sharks and other predators ap- than MI for divers to observe as they move more peared uninterested in pursuing them. rapidly along the reef, their groups are less compact The morning after MI spawning, if within the and higher up in the water column (Fig. 11), mak- multi-day spawning window, the process usually ing it difficult to follow them for any period of repeated, but over days of a given lunar month time. Randall (2001) confirmed that those N. litu- spawning period, the schools and numbers of fish ratus with caudal fin extensions represent mature visibly diminished each day. What had been several males, while mature females lack these. Examining thousand fish seen during the first month/days of several photographs of orange spine surgeonfish aggregation diminished to a few hundred. Similar- “running” along the reef during times of aggrega- ly, in the days following MI spawning, most sharks tion indicates, based on caudal fin extensions, the dispersed and were not seen at the sites in such presence of considerably more females than males numbers, suggesting the GRSs were gathered D. Additional Species of Surgeonfishes using specifically to target the spawning. While the attri- sites. The bignose unicornfish, Naso vlamengii, has tion levels of the aggregating/spawning popula- been seen to aggregate at BC and SC on days from tions are difficult to quantify, we have observed first quarter to full moon in October and November, when spawning and shark predation are high dur- spawning high in the water column and are also pur- ing December/January, the fish often do not aggre- sued by GRS. Their aggregations are difficult for gate and spawn in February/March, perhaps a crit- divers to observe as they usually swim quickly and at ical mass needed to spawn is no longer present. a distance away from the reef in strong currents. At Other years if aggregation/spawning does not start BC the blackstreak surgeonfish, Acanthurus nigri- until January, it will continue into March. cauda, (Etpison 2004, 2009) has also been observed C. Orange spine surgeonfish: Aggregation and to join together in large mixed aggregating schools spawning. Although we have much less informa- to spawn in the summer months, pursued by GRS tion, OSS use the same aggregation sites (BC, SC) and other reef predators. We are still gathering more with the same seasonal pattern, December to information on the exact dates and timing on these March, with a peak in January or February; the unusual mixed spawning aggregations.

Figs 9a-d. Attacks by gray reef sharks on MIs commence once the small fish are at or near the surface, and the sharks drive the now vulnerable MIs out to sea away from the reef. Photos copyright M. T. Etpison.

131 aqua vol. 23 no. 4 - 15 March 2018 Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”?

DISCUSSION ocean water level differences. Where the shallowest Only a few sites on Palau’s western barrier reef are portion of barrier reef is narrow (about 200 m known to have MI and OSS spawning aggregations. width), tidal currents cross the reef cleanly, moving These are tourist dive sites, visited nearly every day, between lagoon and ocean or vice versa (Colin and if large numbers of sharks and spawning fishes 2009, Fig. 2.17). The present aggregation areas are are present this would be noted. This offers some not as simple, with a much broader shelf between confidence that the presently known aggregations, lagoon and ocean, as well as projections producing seen during specific limited periods, are not occur- eddies as along reef currents pass by them. There is ring at other times. We have many fewer observa- need for a detailed study of the currents associated tions of OSS than MI, due to the difficulties men- with these sites to better understand the relation- tioned in working with the former species. We aim ship of currents to spawning. to rectify this discrepancy in the future and also in- In their spawning behavior MI and OSS ascend clude more species occurring in these areas. and move away from the reef sufficiently far before MI and OSS are common on outer reefs to 60 m releasing eggs and sperm at locations where, even depth but few reliable quantitative data on fish den- with 30+ m water visibility, the reef is no longer vis- sity are available. Qualitatively the numbers occur- ible. Only a few fishes with similar strategies are ring within spawning schools could likely be drawn known in Palau. The large bumphead parrotfish, from a relatively small catchment area with migra- Bolbometopon muricatum, aggregates on the new tion distances less than a few km. It is reasonable to moon and rises above and moves far away from the assume with over 300 km of barrier and outer fring- reef to spawn at morning high tide (based on 45 ing reefs many more aggregation/spawning sites for days pers. obs.) and no predation attempts by sharks these fishes exist in Palau. Many areas are seldom vis- (and seldom the presence of any sharks) were seen. ited by divers, but using known times of aggrega- This was also noted by Roff et al. (2017) based on tion, additional areas can be examined to discover one day’s observations. The large twin-spot snapper, additional aggregations, expanding knowledge of re- Lutjanus bohar, aggregates in the thousands every production for management purposes. month before the full moon at some reef promonto- The use of same sites and seasons by MI and OSS ries (Sakaue et al. 2016, pers. obs.) and moves high for spawning, but on opposite phases of the quarter in the water and far off the reef to spawn early in the moon (MI first quarter, OSS last quarter), means morning. Bull sharks (Carcharhinus leucas) and they do not directly compete for reef space for ei- oceanic blacktip sharks (Carcharhinus limbatus) oc- ther spawning preliminaries or ascent/spawning. cur with them and normally a few attacks during Their larvae would also not be entrained into the spawning, often successful, occur every day. same water masses, reducing competition for early It is very surprising that a relatively small fish like life history food resources. Both have their spawn- the MI would ascend so far above and away from ing during lunar periods of neap tides with mid- the reef to spawn. Among comparably-sized deep day high tides. Neap periods have smaller tidal cur- bodied reef fishes, such as butterflyfishes and an- rents on and off the reef, generated by lagoon- gelfishes, for which there are spawning observa-

Figs 10a-b. (a) After the initial feeding frenzy, small groups of MI form at the surface trying to avoid shark predation. (b) Remnant group of MI at surface in blue water with GRS in background, photo taken from boat using GoPro camera on pole. Photos copyright M. T. Etpison. aqua vol. 23 no. 4 - 15 marzo 2018 132 Mandy T. Etpison and Patrick L. Colin tions, they are not known to do so. It is possible could also be attributed the normally tapering num- some of the species in these families which normal- bers of fish within most aggregations after a peak day ly station themselves high above the reef (such as (Sadovy de Mitcheson and Colin 2012). members of the chaetodonid genus Hemitau- The success rate for spawning (releasing all or a richthys) might have behavior similar to MI, but major proportion of gametes) prior to predation is their spawning is unknown. unknown. The undisturbed spawning seen in prior It is clear that at our study sites many MI ascending years (Colin et al. 2012) indicates the gamete release from the reef to spawn are eaten by sharks during is similar to that known for other reef fishes. The the process. There are two stages to the predation process of releasing gametes took a short time (min- event, an initial feeding frenzy lasting several min- utes?) and was not a single massive release. When a utes followed by a period of slow attrition of the re- spawning group is attacked, it is likely only a portion maining MI found schooling at the surface. During of the fish will be ingested prior to releasing gametes, this time, currents carry the groups along or away and it is also likely that as some are eaten, a portion from the reef. Qualitatively only a small number of of their gonadal products may be dispersed into the fish are seen returning to the reef after ascending, water as the body of the fish is ripped apart. perhaps only 10% of individuals return to the same Given our present state of knowledge, we surmise reef area. The remainder are presumed eaten by that spawning by MI at the sites examined has a sharks or when far off the reef in open water drifting high likelihood of death for the spawning adults. to end up far away from the reef area where they as- In that sense, this activity might be considered cended, their possible return to the reef not noted “suicidal” for the individual fish, but spawning is a and fate unknown. Subsequent days during the behavior for which there is a high incentive to un- spawning period had decreasing numbers of MI dertake. Swimming en masse up and away from each day, potentially due to predation, but this the shelter of the reef to spawn is unusual, particu-

Fig. 11. Spawning aggregation of OSS, Naso lituratus, at BC with gray reef sharks visible on the reef and within the overall aggregation, 18 Jan 2009, day of last quarter. There are at least 377 OSS (313 females, 64 males as identified by the males having visible caudal fin extensions). Photo copyright M. T. Etpison.

133 aqua vol. 23 no. 4 - 15 March 2018 Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”? larly for a small species vulnerable to predation. capture them except in a chaotic feeding frenzy Prior to the spawning ascent, a few individuals may type situation. Potentially the MI might follow be taken by predators, but only when individuals sound back to the nearest reef, as they are far be- are separated from the larger group and attacked yond visual range. from multiple directions by several predators. Palauan fishermen had a traditional term, Plutek, While ascending MI are not typically attacked and for rare occasions when they have observed packs of only when they are near the surface do attacks extremely aggressive sharks swimming fast in tight commence, initially as a very chaotic feeding frenzy formation outside the reef (Johannes 1981: 142) in which the sharks rush the schools of MI from all and their response would be to immediately get out sides, contorting themselves and biting wildly. De- of the water. In January 2015 we surfaced in open spite this activity, video footage indicates the sharks water about 100 m out from the reef after losing have a hard time ingesting the small fish, since sight of the MI schools and the pursuing sharks. their mouths are on their ventral surface, the prey While waiting on the surface for our boat to pick us is small and they are not biting chunks out of a up, packs of grey reef sharks (Fig. 12) came under us large prey. Even amid the thrashing and confusion out of the blue, swimming spread out like a carpet of the frenzy, individual MI can be seen successful- in formation back to the reef after finishing their ly evading predation for some time. Over minutes, open water predation; a perfect example of plutek. however, individuals make mistakes and are taken, Why these fish do not spawn closer to the sub- normal predation avoidance tactics are no longer strate, like many other reef fishes, is unknown. The effective, and the numbers of prey fish decrease. “running” schools are followed and attacked by Each remaining fish receives increasing attention predators prior to their ascent, it is possible fish from multiple sharks and the MI become exhaust- would face the same risk of predation even if ed by the sheer numbers of sharks chasing them. spawning occurred closer to the reef. While rising At some point, after the initial frenzy has sub- further off the bottom may enhance the chances of sided, the remaining MI form small groups which their eggs surviving, this question needs further swim together aimlessly in open ocean some dis- consideration as we cannot yet identify any com- tance away from the nearest reef and, based on pelling reason why the fishes ascend so high in the video filmed from boats, can swim without being water column to spawn. constantly attacked by sharks. GRS continue to Is the high predation rate on spawning adults part trail them. staging infrequent attacks, but the small of the life history trajectory for these fishes? While numbers of fish left appeared to enjoy a modest de- there are certainly instances of semelparous (one gree of protection from further attacks. It might be time spawning followed by death) life histories argued this was due to the inability of the sharks to among fishes (salmon being a prime example), it

Fig. 12. Probable Plutek group of sharks returning to the reef after venturing offshore in pursuit of MI. Photo copyright M. T. Etpison. aqua vol. 23 no. 4 - 15 marzo 2018 134 Mandy T. Etpison and Patrick L. Colin seems likely MI and OSS (like nearly all other reef at the time, Colin (2012) suggested that the risks fishes) rely on an iteroparous (spawning multiple of shark predation were relatively minor in aggre- times) life history and the high predation risk at our gation spawning events and did not remove large sites is a product of the unusually large numbers of numbers nor influence behavior of spawning fish- predators often present at the spawning areas. es. That generality has been upset by the present The large numbers of sharks present at the sites be- observations and those of Mourier et al. (2016) re- fore spawning, as well as the movement of the fishes quiring that this assumption of limited influence away from the protection of the reef, is remarkable. by predators on spawning behavior be reexamined. There are examples, such as Mourier et al. (2016) in Our study sites (BC and SC) may be extremes in which spawning is anticipated by predators to prey the continuum of predation risk and unusual in on adult fishes, and many others where egg preda- the context of the present day due to the high tors are stationed at locations to quickly feed on re- numbers of sharks normally present, but they also leased eggs (Sadovy de Mitcheson and Colin 2012). may represent more the natural populations of The attacks on the fishes, particularly MI, at the sur- sharks and spawning of fishes without fishing pres- face is a true feeding frenzy, something seen in Palau sure by humans having removed large predators. only in association with pelagic “bait balls”, not with Given that the large numbers of sharks seen dur- reef fishes (Etpison 2016). Diving observers have ing the spawning periods are not always present, wisely chosen to exit the water and observe the be- what is the benefit to the GRS from the effort in- havior from the safety of a boat. volved in targeting of the aggregations? Individual The numbers of GRS seen at BC and SC grow sig- MI are not a large food items for a shark. Sixteen nificantly during the late winter/early spring and are individuals 95-125 mm standard length weighed roughly correlated with the presence of the aggregat- 47 to 100 g, averaging 65 g (Colin et al. 2012). ing fishes, as well as many other species reproducing The biomass of 500 ascending fish would conse- around the same time. Tourism dive operators con- quently be only 30-35 kg, yet these fish might be stantly assess their abundance, so although the num- targeted by 100 or more GRS. Predation on OSS bers are qualitative, we believe they are realistic. and other surgeonfishes would produce a food con- Vianna et al. (2012) pointed out the depths inhab- tent on the order of 200-400 g per fish, so there ited by GRS varied with water temperatures, seasons would be a higher return per fish, but perhaps still and time of day, adding some complication to visual not a particularly large return on effort. During assessment of shark abundance. Based on present mid-day, though, when MI and OSS spawning oc- knowledge, they may have also underestimated the curs there may be few other opportunities for pre- maximum numbers of GRS populations based on dation (based on the rarity of divers seeing preda- acoustic tagging and visual observations, reporting tion events at those times) and the MI and OSS there were about 100 GRS at the five sites they in- may represent the easiest species to target as a vestigated, present efforts have documented via pho- small, but reliable, food source. tographs up to 200 GRS at just SC when the MI The spawning by MI in the face of a high preda- were preparing to spawn. tion risk might suggest the spawning fish are in a Other outer reef areas in Palau, based on numbers “spawning stupor” as proposed by Johannes (1981) of GRS seen via time-lapse cameras and Baited Re- in which fish in their intent to spawn become so mote Underwater Video (S. Lindfield, pers. oblivious to predation dangers that they continue comm.), have much smaller shark populations. to follow through with their spawning no matter Why do BC and SC have so many GRS? The two the predation around them. There is no evidence sites are focal points for spawning and the lunar se- for these two species exhibiting a spawning stupor quencing of aggregations (first and last quarter) for as the individuals are fully aware and MI exhibit MI and OSS, as well as full and new moon for oth- predator avoidance by rapid, agile movements er species, extends the periods of high fish abun- when attacked making them surprisingly difficult dance over the entire lunar month, increasing the for predators to capture. Colin (2012) argued the potential benefits to GRS from remaining in the evidence for the existence of a “stupor” was based areas during winter months. on misinterpreted observations and that although The level of shark predation on aggregating fishes the predation risks may be high, spawning fishes seems to vary across locations and species. Based are extremely aware of their surroundings and the on collective observations and information known presence of predators.

135 aqua vol. 23 no. 4 - 15 March 2018 Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”?

Each year over the successive monthly spawning ETPISON, M. T. 1994. Palau Portrait of Paradise. Neco Ma- cycles, the schools of MI, initially numbering up to rine, Koror, Palau, 251 pp. thousands of individuals, are seen to diminish to ETPISON, M. T. 2009. Celebrating Palau. Mutual Publish- near zero. The little information we have on ing, Honolulu, HI, 361 pp. ETPISON, M. T. 2014. Thirty Years Palau. Etpison Muse- growth rates and life spans of MI comes from fish um, Koror, Palau, 224 pp. maintained in aquaria, but does indicate a multi- JOHANNES, R. E. 1981. Words of the lagoon. Fishing and year life span with moderate growth rate, so fish marine lore in the Palau district of Micronesia. University observed spawning are at least a few years old. An of California Press, Berkeley, 245 pp. important focus of future research will be to docu- MOLLY, P. P., COTE, I. M. & REYNOLDS, J. D. 2012. Why ment the population sizes more accurately before spawn in aggregations? In: Reef Fish Spawning Aggrega- and after the spawning season, as well as the attri- tions: Biology, Research and Management. (Ed. Y. Sadovy tion over that time. Why there are such large num- de Mitcheson & P. L. Colin): 57-83. Fish and Fisheries Series Vol. 35, Case Study 12.21. Springer Science & bers of sharks found at BC and SC is not under- Business Media. stood. Whether the promontory areas attract and MOURIER, J., MAYNARD, J., PARRAVICINI, V., BALLESTA, L., retain large numbers of fishes is uncertain, but that CLUA, E., DOMEIER, M. L. & PLANES, S. 2016. Extreme type of geomorphology is often associated with inverted trophic pyramid of reef sharks supported by high fish populations. Throughout most the spawning groupers. Current Biology 26 (15): 2011-2016. world’s reefs numbers of both predators and their NEMETH, R. M. 2012. Ecosystem aspects of species that ag- prey fishes have been greatly reduced through over- gregate to spawn. In: Reef Fish Spawning Aggregations: Bi- fishing. The remarkable behavior documented here ology, Research and Management. (Ed. Y. Sadovy de Mitch- eson & P. L. Colin): 21-55. Fish and Fisheries Series Vol. illustrates how little is known about the life histo- 35, Case Study 12.21. Springer Science & Business Media. ries of many common reef fishes. RANDALL, J. E. 2001. Surgeonfishes of the World. Mutual Publishing Co., Honolulu. ACKNOWLEDGEMENTS ROFF, G., DOROPOULOS, C., MEREB, G. & MUMBY, P. J. Mandy Etpison would like to thank NECO Ma- 2017. Mass spawning aggregation of the giant bumphead rine, Shallum, Tkel and Iked Etpison, and Edwin parrotfish Bolbometopon muricatum. Journal of Fish Biol- Maidesil for assistance in the field. Patrick Colin ogy 91: 354-361. would like to thank Lori J. B. Colin, Matt SADOVY, Y. 2007. Report on current status and exploita- tion history of reef fish spawning aggregations in Palau. Mesubed, Steve Lindfield, Emilio Basilius and Paul Western Pacific Fishery Survey Series: Society for the Conser- Collins for field assistance. This work was made vation of Reef Fish Aggregations, Volume 3. SCRFA and the possible by support from NECO Marine and Palau Conservation Society, 40 pp. Coral Reef Research Foundation. SADOVY DE MITCHESON, Y. & COLIN, P. L. (eds). 2012. Reef Fish Spawning Aggregations: Biology, Research and REFERENCES Management. Fish and Fisheries Series Vol. 35, Case Study COLIN, P. L. 2009. Marine Environments of Palau. Indo- 12.21. Springer Science & Business Media, 621 pp. Pacific Press, San Diego, 414 pp. SAKAUE, J., AKINO, H., ENDO, M., IDA, H. & ASAHIDA, T. COLIN, P. L. 2012. Striped Bristletooth – Ctenochaetus stria- 2016. Temporal and spatial site sharing during spawning tus and Brown Surgeonfish Acanthurus nigrofuscus with in snappers Symphorichthys spilurus and Lutjanus bohar Notes on Other Indo-West Pacific Surgeonfishes (Acan- (Pisces: Perciformes: Lutjanidae) in waters around Peleliu thuridae). In: Reef Fish Spawning Aggregations: Biology, Re- Island, Palau. Zoological Studies 55: 1-15. search and Management. (Ed. Y. Sadovy de Mitcheson & P. SANCHO, G., SOLOW, A. R. & LOBEL, P. S. 2000. Environ- L. Colin): 526-535. Fish and Fisheries Series Vol. 35, Case mental influences on the diel timing of spawning in coral Study 12.21. Springer Science & Business Media. reef fishes. Marine Ecology Progress Series 206: 193-212. COLIN, P. L. & BELL, L. J. 1991. Aspects of the spawning TAYLOR, B. M., RHODES, K. L., MARSHELL, A. & MCIL- activity of labrid and scarid fishes (Pisces: Labroidei) at WAN, J. L. 2014. Age based demographic and reproduc- Enewetak Atoll, Marshall Islands, with notes on other tive assessment of orangespine Naso lituratus and blue- families. Environmental Biology of Fishes 32: 229-260. spine Naso unicornis unicornfishes. Journal of Fish Biology COLIN, P. L., ETPISON, M. T. & COLLINS, P. 2012. The 85(3): 901-916. moorish idol, Zanclus cornutus. In: Reef Fish Spawning VIANNA, G. M. S., MEEKAN, M. G., PANNELL, D. J., Aggregations: Biology, Research and Management. (Ed. Y. MARSH, S. P. & MEEUWIG, J. J. 2012. Socio-economic Sadovy de Mitcheson & P. L. Colin): 513-517. Fish and value and community benefits from shark-diving tourism Fisheries Series Vol. 35, Case Study 12.21. Springer Sci- in Palau: a sustainable use of reef shark populations. Bio- ence & Business Media. logical Conservation 145 (1): 267-277.

aqua vol. 23 no. 4 - 15 marzo 2018 136 aqua, International Journal of Ichthyology

Fooled by a fish: seed camouflage by an Amazonian banjo , Bunocephalus verrucosus (Siluriformes: )

Flávio C. T. Lima1 and Gilberto N. Salvador2

1) Museu de Zoologia, Universidade Estadual de Campinas, Caixa Postal 6109, 13083-863, Campinas, SP, Brazil. E-mail: [email protected] 2) Laboratório de Ecologia e Csonservação, Universidade Federal do Pará, Rua Augusto Côrrea, 1, Belém, PA, Brazil. E-mail: [email protected]

Accepted: 05 February 2018

Keywords mentada baseada em observações de dois espécimes que hypsiura, Couepia, Chrysobalanaceae, flooded dobraram seu pedúnculo caudal contra o lado do corpo forest, Chromolucuma rubriflora. após serem capturados, na região de Tefé, Amazônia cen- tral, Brasil. Este é o segundo registro de camuflagem de se- Abstract mentes em bagres aspredinideos, o primeiro sendo para Seed camouflage (masquerade) by Bunocephalus verruco- Amaralia hypsiura. Tipos de camuflagem entre bagres as- sus, an aspredinid catfish, is herein reported based on two predinideos são aqui reportados como variando entre observations of specimens that folded their caudal pedun- “masquerade”, imitação do substrato, e colorido disruptivo. cle against the side of the body after being captured, in the Muitas sementes (tipicamente drupas) da foresta inundada region of Tefé, Central Amazon, Brazil. This is the second podem servir como modelos potenciais para Bunocephalus report of seed camouflage for aspredinid , the first verrucosus, através de sua área de distribuição, em destaque being reported for Amaralia hypsiura. Types of camouflage sementes de Chrysobalanaceae. among aspredinid catfishes are herein reported to range from masquerade, background matching, and disruptive Résumé coloration. Many seeds (mostly drupes) of the flooded for- Camouflage de semences (masquerade) par un aspredinid est may serve as potential models for Bunocephalus verruco- Bunocephalus verrucosus, poisson-chat, est rapporté dans le sus across its range, particularly seeds of Chrysobalanaceae. présent document en se fondant sur deux observations de spécimens que leur pédoncule caudal pliée contre le côté Zusammenfassung du corps après sa capture, dans la région de Tefé, Central Berichtet wird hier von einer Samenmimese (Maskierung) Amazon, au Brésil. C'est le deuxième rapport du camou- bei Bunocephalus verrucosus, einem Wels der Aspredinidae; flage des semences pour aspredinid poissons-chats, le pre- der Bericht gründet sich auf Beobachtungen in der Gegend mier étant déclaré pour Amaralia hypsiura. Types d'aspre- von Tefé, Zentralamazonien, Brasilien, an zwei Exemplaren, dinid camouflage chez les poissons-chats sont présentes de die nach dem Fang ihren Schwanzstiel an die Körperseite mascarade, d'arrière-plan, correspondance et perturbateurs legten. Es ist der zweite Bericht von Samenmimese bei de la couleur. Beaucoup de graines (surtout les drupes) de Welsen der Aspredinidae; der erste bezog sich auf Amaralia la forêt inondée peuvent servir de modèles potentiels pour hypsiura. In diesem Bericht wird ein Überblick über die Tar- Bunocephalus verrucosus dans toute son aire de répartition, nungstypen bei Welsen der Aspredinidae gegeben: Sie re- notamment les semences de Chrysobalanaceae. ichen von Mimese (Maskierung) und Somatolyse (Ver- schmelzung mit dem Hintergrund) bis zur auffälligen Fär- Sommario bung (Schrecktracht). Viele Samen (meistens Steinfrüchte) Il criptismo (mascheramento) a forma di seme di Buno- in einem überfluteten Waldstück eignen sich als Modell für cephalus verrucosus, un pesce gatto aspredinide, è qui ripor- die Tarnung von Bunocephalus verrucosus innerhalb seiner tato in base a due osservazioni di esemplari che hanno ri - mimetischen Möglichkeiten, insbesondere Samen der piegato il loro peduncolo caudale contro il lato del corpo Chrysobalanaceae (Goldpflaumengewächse). dopo essere stati catturati, nella regione di Tefé, nell'Amaz- zonia centrale, in Brasile. Questo è la seconda segnalazione Resumo di criptismo a seme per i pesci gatto aspredinidi, dopo Camuflagem de sementes (“masquerade”) por Buno- quello riportato per Amaralia hypsiura. I tipi di mimetiz- cephalus verrucosus, um bagre aspredinideo, é aqui docu- zazione qui riportati tra i pesci gatto aspredinidi compren-

137 aqua vol. 23 no. 4 - 15 March 2018 Fooled by a fish: seed camouflage by an Amazonian banjo catfish, Bunocephalus verrucosus (Siluriformes: Aspredinidae) dono il mascheramento, il camuffamento con il fondale e November 2017. Data for the map of distribution la colorazione distruttiva. Molti semi della foresta allu- of Bunocephalus verrucosus was obtained from the vionale (principalmente drupe, in particolare di literature (Mees 1988; Mol 2012; Le Bail et al. Chrysobalanaceae) possono servire come potenziali model- 2012) complemented with the records available li per Bunocephalus verrucosus. through two online databases, Species Link (splink.cria.org.br) and Global Biodiversity Infor- INTRODUCTION mation Facilit (GBIF) (www.gbif.org). Specimens Visual camouflage (of which crypsis is the best deposited in two collections (MZUSP and ZUEC) known example: Stevens & Merilaita 2009) is a were checked to confirm their identity (see Appen- widespread, although relatively poorly document- dix). Bunocephalus verrucosus is a very distinctive ed, evolutive strategy present in several freshwater aspredinid and misidentifications in collections are and marine fishes (see, e.g., Randall 2005, Roberts very likely non-existent or negligible in number. 2015, Sazima 2017). Several fishes from the Institutional acronyms follow Sabaj (2016). Distri- Greater Amazonia region (used herein to include, bution data of Couepia paraensis was obtained besides the Amazon basin, the Orinoco basin and through the above-mentioned databases. A sample the several relatively small guyanese river systems; of the Sapotaceae Chromolucuma rubriflora was de- cf. Van der Sleen & Albert 2017) are known to use posited at the UEC (Universidade Estadual de camouflage with its surroundings or inanimate ob- Campinas) herbarium. jects, typically to avoid predation (e.g., Zuanon et al. 2006; Sazima et al. 2006). Camouflaged color RESULTS patterns of fishes from the Pan-Amazon region typ- The first Bunocephalus verrucosus specimen ob- ically imitates either fallen, dead, soaken leaves or served imitating a large seed was collected with a leaves remnants (e.g., juvenile pacus; Zamprogno small handnet at a tributary of the Lago Amanã, at & Andrade 1986; the catfishes Helogenes marmora- the shoreline of a flooded igapó forest with a deep tus and Tetranematichthys quadrifilis, and the knife- layer of dead leaves and other vegetal debris. This fish Steatogenys duidae; Sazima et al. 2006), dead large (c. 10 cm SL) specimen was taken to be at twigs (e.g., Farlowella suckermouth catfishes; Ret- first a large seed by the first author, and only a zer & Page 1996: 34), or living aquatic vegetation closer look revealed it to be a fish. This observa- (e.g., the crenuchid Ammocryptocharax elegans; tion was previously reported by Roberts (2015: Zuanon et al. 2006; the suckermouth catfishes 127). The second Bunocephalus verrucosus ob- Acestridium spp.; Retzer et al. 1999). The most fa- served (ZUEC 15094, 72.1 mm SL) was collected mous example of cryptic camouflage among fishes by the first author with a handnet at the mouth of from the Greater Amazonia region, and in fact one a small igarapé emptying into the Lago Tefé, a of the most famous among all animals, is of course large ria-lake of the middle Amazon basin, in No- the leaffish, Monocirrhus polyacanthus (Britz & vember 2017. Although the specimen was identi- Kullander 2003; Catarino & Zuanon 2010; Skel- fied immediately as a Bunocephalus verrucosus, it horn et al. 2010). Recently, Roberts (2015) argued was initially mistaken as a seed during its subse- that the aspredinid catfish Amaralia hypsiura imi- quent handling by the two authors, a few hours af- tate large seeds. The present contribution reports a ter being collected. The specimen kept its caudal second case of putative seed camouflage, this time peduncle bent against the side of the body (Figs by another aspredinid catfish, Bunocephalus verru- 1a-b), in the same way as described by Roberts cosus, furthering the case that some aspredinid cat- (2015: 122) for a living Amaralia hypsiura. In fact, fishes do indeed imitate large seeds. the bent caudal peduncle, with the erect caudal fin, gave the impression of a sprout coming out of MATERIALS AND METHODS the seed, and contributed for the overall similarity Ad libitum observations were made from two alive of the fish to a seed. The specimen only straight- Bunocephalus verrucosus specimens, one collected at ened after being anesthesized and euthanized prior a tributary of the Lago Amanã (which is connected to fixation (Fig. 2). to the lower rio Japurá, Amazonas state, Brazil; IN- PA uncat.) in March 2003 and another at the DISCUSSION mouth of a small igarapé tributary to the Lago Tefé, As noticed in the introduction, Roberts (2015) Alvarães, Amazonas state, Brazil; ZUEC 15094) in was the first to call attention to the possibility that aqua vol. 23 no. 4 - 15 marzo 2018 138 Flavio C. T. Lima and Gilberto N. Salvador

Figs 1a-f. a-b) Bunocephalus verrucosus, ZUEC 15094, 72.1 mm SL, living specimen, in the “seed-camouflage mode”, with caudal peduncle bent over the side of the body. c) Seed of Couepia sp. from rio Cristalino, Alta Floresta, Mato Grosso, Brazil; d) Seeds of Virola surinamensis from an unspecified locality in the Brazilian Amazon. e-f) Fruits of maiá, Chromolucuma rubri- flora from Igarapé Uacatuna, São Gabriel da Cachoeira, Amazonas, Brazil, February 2018, showing different degrees of shrink- age (e) and different sizes (f). Photos a-b) by G. N. Salvador, photo c) by D. Sasaki, photo d) by P. S. Sena (available at: www.commons.wikimedia.org/wiki/File%3AUcuuba_casca.JPG), photo e) by F. C. T. Lima, and photo f) by G. C. Bortolo.

139 aqua vol. 23 no. 4 - 15 March 2018 Fooled by a fish: seed camouflage by an Amazonian banjo catfish, Bunocephalus verrucosus (Siluriformes: Aspredinidae)

Fig. 2. Bunocephalus verrucosus, ZUEC 15094, 72.1 mm SL, preserved specimen, in dorsal, lateral, and ventral views. Photo by E. G. Baena. aqua vol. 23 no. 4 - 15 marzo 2018 140 Flavio C. T. Lima and Gilberto N. Salvador aspredinid catfishes – specifically Amaralia hypsiu- over the side of the body (Roberts 2015; present ra – imitate seeds. Although Roberts (2015) called study). In addition, Bunocephalus verrucosus is a rel- the similarity between Amaralia hypsiura and a seed atively robust, high-bodied aspredinid, contrasting as constituting mimicry, we herein use the term with the typically more depressed body shape pre- “camouflage”, to be more consistent with the cur- sent in most of the remaining aspredinids, includ- rent terminology on the subject, where most au- ing the remaining Bunocephalus species (cf., e.g., thors consider them to be distinct (e.g. Stevens & Mees 1989; Carvalho et al. 2015), which reinforces Merilaita 2009), even though somewhat related a seed-shape overall appearance for the species. (e.g., Endler 1981), phenomena. More specifically, Other Bunocephalus species present a cryptic color seed imitation by Amaralia hypsiura and Buno- pattern, overall appearance, and behavior; B. aleu- cephalus verrucosus constitute a type of visual cam- ropsis and B. coracoideus are typically found near ouflage called “masquerade”, where a specific ob- vegetal debris and when captured remain immo- ject, such as a leaf, a twig, a stone, or a bird-drop- bile, imitating the background where they are ping, is imitated, instead of a general background found (F. C. T. Lima, pers. obs.). Some other (Stevens & Merilaita 2009: 424). Bunocephalus species, as the recently described B. It might be questioned whether this putative case hartti and B. minerim, as well as some hoplomy- of camouflage is indeed real. One of the arguments zontini aspredinids as Dupouyichthys sapito, and the pointed against the identification of examples of species of Ernestichthys (for pictures-drawings of camouflage (or else mimicry) is that we, humans, these aspredinids, see respectively Carvalho et al. are not the “targeted receivers” of the visual infor- 2015, and Stewart 1985), present a saddled color mation and as such might not be in a good posi- pattern, a disruptive color signal occurring in ben- tion to assess them in the first place (Stevens & thic fishes living in flowing environments with un- Merilaita 2009). This argument is, however, self- even substrates (Armbruster & Page 1996). Clear- defeating, because there is no way other than to re- ly, different aspredinind taxa rely in distinct modes ly on our own visual senses to identify putative ex- of camouflage: ‘masquerade’ for Amaralia species amples of camouflage or mimicry. We do not con- and Bunocephalus verrucosus, ‘background match- sider a coincidence that on two separate occasions, ing’ for B. aleuropsis and B. coracoideus, and ‘dis- two ichthyologists could be fooled into thinking ruptive coloration’ for B. hartti, B. minerim, that a fish was a seed, to the point of feeling Dupouyichthys sapito, and Ernestichthys species (ter- awestruck when finding out the truth (undoubted- minology follows Stevens & Merilaita 2009). ly the same feeling experienced by Roberts 2015). The flooded forests of the Amazon basin provide We have little doubt that this similarity is no mere a great number of possible seed models for Buno- coincidence, but instead a type of camouflage ac- cephalus verrucosus. The species occur across the quired during the course of evolution. Four main lowlands of the Amazon basin, as well as in rivers questions then arise: 1) how widespread might be from Guyana, Suriname, and French Guyana this type of camouflage among aspredinids; 2) (Mees 1988; Mol 2012; Le Bail et al. 2012) (Fig. what types of seeds are potentially serving as mod- 3), and most of its known localities lie within els; 3) what predators are potentially being fooled black- or clear-water rivers, typically in flooded by this camouflage; and 4) if this camouflage may forests. Drupe pits and arillated seeds are the types be serving to a purpose other than to disguise the of seeds that most likely serve as models for Buno- fish against predators. We will address each of these cephalus verrucosus, because of their size, hardness, questions in that order. ornamented surfaces, and imperishability in the Among aspredinids, seed camouflage is presum- submerged leaf litter. Examples of trees occurring ably only present in Amaralia (a previously mono- in flooded forests of the Amazon basin that pro- typic genus, now containing two species, A. hypsiu- duces seeds (mostly drupes) that are potentially im- ra and A. oviraptor; Friel & Carvalho 2016) and itated are the seeds of Virola spp. (Myristicaceae), Bunocephalus verrucosus. Both taxa share the pres- several Chrysobalanaceae, as species of the genera ence of conspicuous bony knobs over the skull, Couepia, Licania, and Parinari, Caryocar micro- supracleithrum, and middle nuchal plate (Roberts carpum (Caryocaraceae), Andira inermis (Faba - 2015; Friel & Carvalho 2016), which imparts ceae), Buchenavia spp. (Combretaceae), among some similarity to the irregular surface of many other examples. The ucúuba, Virola surinamensis, is seeds, and the ability of flexing the caudal peduncle a widespread and common tree occurring in flood-

141 aqua vol. 23 no. 4 - 15 March 2018 Fooled by a fish: seed camouflage by an Amazonian banjo catfish, Bunocephalus verrucosus (Siluriformes: Aspredinidae) ed forests, especially white-water flooded forests ring in flooded forests, as V. elongata, which prefer (Wittmann et al. 2010), and its seeds, although rel- black-water flooded forests (Prance 1979; Worbes atively small (the whole fruits reaches up to 3 cm 1997). Even more similar are the drupes of several in diameter), may serve as a possible model for Chrysobalanaceae, which are larger and present small-sized Bunocephalus verrucosus specimens (Fig. conspicouous striae over their surfaces (e.g., 1d), as well as the seeds of other Virola trees occur- Couepia sp.; Fig. 1c). Chrysobalanaceae trees are

Fig. 3. Map of northern South America, showing the distribution of Bunocephalus verrucosus (above) and Couepia paraensis (below), exemplifying the co-occurrence of the fish with a putative seed model. aqua vol. 23 no. 4 - 15 marzo 2018 142 Flavio C. T. Lima and Gilberto N. Salvador among the dominant elements from black- and cently fallen, floating fruits with fleshy mesocarps clear-water flooded forests (Ayres 1993; Wittmann are taken, so its resemblance to drupes most likely et al. 2011). Particularly, Couepia paraensis, a does not put in jeopardy the seed-imitating Buno- species widespread in central and western Amazon cephalus verrucosus of being accidentaly eaten by (Prance 1972), is one of the most common trees these and other frugivorous fishes. occurring in igapó (black-water flooded forest) (Prance 1979; Kubitzki 1989), and remarkably, ACKNOWLEDGEMENTS with a distribution largerly overlapping with B. The first observation of a Bunocephalus verrucosus verrucosus (Fig. 3). In addition, fruits of maiá, imitating a seed was made by the first author dur- Chromolucuma rubriflora (Sapotaceae), a tree oc- ing a field collection trip at the invitation of curring in the forests of the Rio Negro basin as well Michel F. Catarino, then at the Instituto de Desen- as in the lower Amazon basin (Pennington 1990; volvimento Sustentável Mamirauá, Tefé, Brazil. We Pennington & Edwards 2005) might serve as mod- are grateful to him for the opportunity, as well as to els as well. The relatively large (90-150 mm) fruits Jansen S. Zuanon, Leandro M. Sousa, and Jonas A. have a mesocarp that shrink after desiccation or a Oliveira, for help and comradeship in the field. long immersion in water (Fig. 1e-f), when it sinks The second observation was made during a trip of (fresh fruits float in the water; F.C.T. Lima, pers. the project “Systematics of the tetras (genera Hemi - obs.). However, in the same way as Roberts (2015: grammus, Hyphessobrycon, Thayeria, Parapristella e 126-127) remarked for Amaralia hypsiura, we do Bryconella), with emphasis on the species from cis- not consider likely that Bunocephalus verrucosus andean South America” (FAPESP grants # evolved a resemblance to a specific seed, but actu- 2011/51532-7 and 2013/20936-0 to the first au- ally present a overall similarity to seeds, allowing it thor). We are grateful to Nelson Flausino Junior, to imitate available seeds present at the submerged and again to Jonas A. Oliveira, as well as to leaf litter of a given locality. Danielle Pedrociane Cavalcante (Instituto de De- Potential predators that may be fooled by the sim- senvolvimento Sustentável Mamirauá) for the help ilarity of Bunocephalus verrucosus to a seed are evi- at Tefé. Observations of the fruits of Chromolu- dently diurnal predators present in flooded forests, cuma rubriflora at São Gabriel da Cachoeira in relying on visual cues to identify their prey. These February 2018 were funded by the same grants. include ichthyophagous birds that hunt in shallow Denise Sasaki (Programa Flora Cristalino) kindly water, as herons, finfoot, sunbittern, and kingfish- provided Fig. 1c. Eduardo G. Baena prepared Fig. ers, and some middle to large-sized ichthyo pha - 2. We thank Alessio Datovo and Osvaldo T. Oy- gous fishes, as the cichlids Cichla spp. and the akawa (MZUSP) for allowing access to specimens large-sized pike cichlids of the Crenicichla lugubris under their care. The second author was funded by group. However, we are not aware of records of Coordenação de Aperfeiçoamento Pessoal de Nível predation on aspredinids by any of these potential Superior (CAPES). We dedicate this paper to Ivan predators. Sazima, for inspiring us into looking to nature with Roberts (2015: 125) remarked that the similarity unprejudiced eyes, and seeing interesting facts to a seed may have a dual function in Amaralia, where many would see nothing. since these catfishes are oophagous, eating large eggs, apparently of loricariid catfishes (Friel & Car- REFERENCES valho 2016: 541), and its similarity to seeds may ARMBRUSTER, J. W. & PAGE, L. M. 1996. Convergence of help them to approach egg masses without being a cryptic saddle pattern in benthic freshwater fishes. En- driven away (loricariid catfishes present parental care vironmental Biology of Fishes 45: 249-257 of eggs). The diet of Bunocephalus verrucosus is cur- AYRES, J. M. 1993. As matas de várzea do Mamirauá: médio rently unknown, but more likely consists in aquatic rio Solimões. Sociedade Civil Mamirauá, Belém, 123 pp. invertebrates, which is apparently the typical diet of BRITZ, R. & KULLANDER, S. O. 2003. Family Polycentri- aspredinid catfishes, so it is uncertain whether its dae (leaffishes). In: Check List of the Freshwater Fishes of similarity to seeds may help them to acquire food South and Central America (CLOFFSCA) (Ed. R. E. Reis, S. O. Kullander & C. J. Ferraris, Jr.): 603-604. Edipucrs, itens. Although Chrysonabalanaceae fruits may be Porto Alegre. ingested by fruit-eating fishes as Lithodoras dorsalis CARVALHO, T. P., CARDOSO, A. R., FRIEL, J. P. & REIS, R. (Doradidae) and Colossoma macropomum (Ser- E. 2015. Two new species of the banjo catfish Buno- rasalmidae) (Goulding 1980: 202-205), only re- cephalus Kner (Siluriformes: Aspredinidae) from the up-

143 aqua vol. 23 no. 4 - 15 March 2018 Fooled by a fish: seed camouflage by an Amazonian banjo catfish, Bunocephalus verrucosus (Siluriformes: Aspredinidae)

per and middle rio São Francisco basins, Brazil. Neotrop- ROBERTS, T. R. 2015. Mimicry of a bean seed by the Ama- ical Ichthyology 13: 499-512. zonian aspredinid catfish Amaralia hypsiura (Kner 1855), CATARINO, M. F. & ZUANON, J. 2010. Feeding ecology of with notes on vegetative camouflage by fishes. aqua, In- the leaf fish Monocirrhus polyacanthus (Perciformes: Poly- ternational Journal of Ichthyology 21: 120-127. centridae) in a terra firme stream in the Brazilian Ama- SABAJ, M. H. 2016. Standard symbolic codes for institutional zon. Neotropical Ichthyology 8: 183-186. resource collections in herpetology and ichthyology: an online ENDLER, J. A. 1981. An overview of the relationships be- reference. Version 6.5 (16 August 2016). Electronically tween mimicry and crypsis. Biological Journal of the Lin- accessible at http://www.asih.org/, American Society of nean Society 16: 25-31. Ichthyologists and Herpetologists, Washington, D.C. FRIEL, J. P. & CARVALHO, T. P. 2016. A new species of SAZIMA, I. 2017. Deception, protection, and aggression in Amaralia Fowler (Siluriformes: Aspredinidae) from the the mangrove: three juvenile fishes and floating leaves in Paraná-Paraguay river basin. Zootaxa 4088 (4): 531-546 Southeast Brazil. aqua, International Journal of Ichthyolo- GOULDING, M. 1980. The fishes and the forest: explorations gy 23: 41-46. in Amazonian natural history. University of California SAZIMA, I., CARVALHO, L. N., MENDOÇA, F. P. & ZUANON, Press, Berkeley, xii+280 pp. J. 2006. Fallen leaves on the water-bed: diurnal camou- KUBITZKI, K. 1989. The ecogeographical differentiation of flage of three night active fish species in an Amazonian Amazonian inundation forests. Plant Systematics and Evo- streamlet. Neotropical Ichthyology 4: 119-122. lution 162: 285-304. SKELHORN, J., ROWLAND, H. M., SPEED, M. P. & RUX- LE BAIL, P.-Y., COVAIN, R., JÈGU, M., FISCH-MULLER, S., TON, G. D. 2010. Masquerade: camouflage without VIGOUROUX, R. & KEITH, P. 2012. Updated checklist of crypsis. Science 327: 51. the freshwater and estuarine fishes of French Guiana. Cy- STEVENS, M. & MERILAITA, S. 2009. Animal camouflage: bium 36: 293-319. current issues and new perspectives. Philosophical Trans- MEES, G. F. 1988. The genera of the subfamily Buno- actions of the Royal Society B 364 (1516): 423-427. cephalinae (Pisces, Nematognathi, Aspredinidae). Pro- STEWART, D. J. 1985. A review of the South American cat- ceedings of the Koninklijke Nederlandse Akademie van fish tribe Hoplomyzontini (Pisces, Aspredinidae), with Wetenschappen (series C, Biological and Medical Sciences) descriptions of new species from Ecuador. Fieldiana Zo- 91: 85-102. ology (new series) 25(1360): i-iii + 19 pp. MEES, G. F. 1989. Notes on the genus Dysichthys, subfam- VAN DER SLEEN, P. & ALBERT, J. S. 2017. General intro- ily Bunocephalinae, family Aspredinidae (Pisces, Ne- duction. In: Field guide to the fishes of the Amazon, matognathi). Proceedings of the Koninklijke Nederlandse Orinoco & Guianas (Ed. P. van deer Sleen & J. S. Albert): Akademie van Wetenschappen (series C, Biological and 13-22. Princeton University Press, Princeton. Medical Sciences) 92: 189-250. WITTMANN, F., SCHONGART, J., DE BRITO, J. M., MOL, J. H. 2012. The freshwater fishes of Suriname. Konin- WITTMANN, A. O., PIEDADE, M. T. F., PAROLIN, P., JUNK, klijke Brill NV, Leiden, 889 pp. W. J. & GUILLAUMET, J. L. 2010. Manual of trees from PENNINGTON, T. D. 1990. Sapotaceae. Flora Neotropica Central Amazonian várzea floodplains: taxonomy, ecology, 52: 1-770. and use. Editora INPA, Manaus, 298 pp. PENNINGTON, T. D. & EDWARDS, K. S. 2005. Sapotaceae. WITTMANN, F., SCHONGART, J. & JUNK, W. J. 2011. Phy- In: Flora of the Venezuelan Guyana. Volume 9. Rutaceae- togeography, species diversity, community structure and Zygophyllaceae (Ed. P.E. Berry, K. Yatskievych & B.K. dynamics of Central Amazonian floodplain forests. In: Holst): 89-137. Missouri Botanical Garden Press, St. Amazonian floodplain forests: ecophysyology, biodiversity Louis. and sustainable management (Ed. W. J. Junk, M. T. F. PRANCE, G. T. 1972. Chrysobalanaceae. Flora Neotropica Piedade, F. Wittmann, J. Schongart & P. Parolin): 61- 9: 1-409. 102. Ecological Studies 210, Springer, Berlin, xviii+618 PRANCE, G. T. 1979. Notes on the vegetation of Amazonia pp. III. The terminology of amazonian forest types subject to WORBES, M. 1997. The forest ecosystems of the flood- inundation. Brittonia 31: 26-38. plains. In: The central Amazon floodplain: ecology of a puls- RANDALL, J. E. 2005. A review of mimicry in marine fish- ing system (Ed. W. J. Junk): 223-265. Ecological Studies es. Zoological Studies 44: 299-328. 126, Springer, Berlin, xxiv+ 525 pp. RETZER, M. E., NICO, L. G. & PROVENZANO, F. 1999. ZAMPROGNO, C. & ANDRADE, G. V. 1986. Camuflagem Two new species of Acestridium (Siluriformes, Loricari- em jovens de pacu, Myleus sp. (Characiformes, idae) from southern Venezuela, with observations on Myleinae). Revista Brasileira de Biologia 46: 415-418. camouflage and color change. Ichthyological Exploration ZUANON, J., CARVALHO, L. N. & SAZIMA, I. 2006. A of Freshwaters 10: 313-326. chamaleon characin: the plant-clinging and colour- RETZER, M. E. & PAGE, L. M. 1996. Systematics of the changing Ammocryptocharax elegans (Characidiinae: stick catfishes, Farlowella Eigenmann & Eigenmann Crenuchidae). Ichthyological Exploration of Freshwaters (Pisces, Loricariidae). Proceedings of the Academy of Nat- 17: 225-232. ural Sciences of Philadelphia 147: 33-88.

aqua vol. 23 no. 4 - 15 marzo 2018 144 Flavio C. T. Lima and Gilberto N. Salvador

APPENDIX Caracaraí, igarapé do Bota Panela, cachoeira do Material examined. Bunocephalus verrucosus: Bem-Querer, c. 1˚55’N, 61˚1’W; M. Goulding, 9 Brazil, Amazonas: MZUSP 7362, 15; MZUSP Jan 1984. Pará: MZUSP 23678, 1, Lago Jacaré, rio 7363, 4; Maués, igarapé Limãozinho, 3˚24’S, Trombetas; EPA, 7-11 Oct 1969. Peru, Loreto: 57˚42’W; EPA, 4 Dec 1967. MZUSP 6208, 1, MZUSP 15296, 1, Río Nanay, Iquitos, 3˚44’S, Manaus, igarapé Jaraqui, trib. rio Negro, above 73˚18’W; M. Villacorta, 26 Sept 1979. Manaus; EPA, 22-24 Apr 1967. MZUSP 42809, 2, Manaus, igarapé Tarumãzinho; EPA, 17-18 Nov 1967. MZUSP 58966, 2, rio Negro, Anavilhanas, igapó; M. Goulding, March 1982. MZUSP 54514, 2; MZUSP 31254, 1; MZUSP 64941, 1, Barcelos, lake on island; M. Goulding, 29 Feb 1980. MZUSP 31253, 1, Rio Negro, central lake at ilha de Buiu-Açu, above rio Urubaxi; M. Gould- ing, 6 Feb 1980. MZUSP 63310, 1, Rio Negro, downstream mouth of rio Daraá, central lake; M. Goulding, 17 Feb 1980. MZUSP 59173, 1, Can- tagalo, rio Negro; EPA, 28 Jan 1972. MZUSP 61915, 3; MZUSP 84754, 1; Santa Isabel do Rio Negro, rio Aiuanã, 0˚24’S, 65˚ 2’W; EPA, 22 Oct 1972. MZUSP 62052, 6, Santa Isabel do Rio Ne- gro, rio Negro, Tapera, 0˚12’S, 64˚4’W; EPA, 2 Nov 1972. MZUSP 55117, 7, Santa Isabel do Rio Negro, igarapé em São João, 0˚24’S, 65˚2’W; EPA, 27 Oct 1972. MZUSP 109555, 1, Santa Isabel do Rio Negro, igarapé do Bitiana, próximo a boca do rio Neuixi, 0˚34’3’’S, 65˚4’5’’W; M. T. Piza et al., 8 Feb 2011. MZUSP 30699, 1, Rio Tefé, Jauar- iatuba; M. Goulding, 7 Ag 1979. ZUEC 15094, 1, Alvarães, small igarapé into Lago Tefé, 3˚ 20’00’’S, 64˚49’8’’W; F. C. T. Lima, G. N. Salvador, N. Flausino Jr. & J. A. Oliveira, 28 Nov 2017. MZUSP 6350, 2; MZUSP 6351, 2; lago Castro, mouth of rio Purus; EPA, 7-8 Nov 1967. MZUSP 23349, 7, Rio Solimões, in front of Jacaré, near- Fonte Boa; EPA, 7 Oct 1968. Acre: MZUSP 30701, 9; MZUSP 30968, 3; Tarauacá, rio Ta- rauacá, c. 8˚11’S, 70˚46’W; M; Goulding, Jul 1984. ZUEC 13501, 7, Cruzeiro do Sul, Lago Ver- melho (rio Moa), 7˚36’48’’S, 72˚48’10’’W; T. Jacó et al., 15 Jul 2015. Roraima: MZUSP 23569, 2, Caracaraí, igarapé 1 km. N of Caracaraí, c. 1˚51’N, 61˚7’W; T. R. Roberts, 5 Feb 1969. MZUSP 112954, 1; MZUSP 112777, 1; Caracaraí, igarapé Pretinho, below mouth of igara- pé Branquinho, 0˚56’32’’S, 62˚6’19’’W; O. T. Oy- akawa et al., 9 Sept 2011. MZUSP 112911, 1, Caracaraí, praia do Paricá, rio Jufari, 1˚8’41’’S, 61˚59’57’’W; O. T. Oyakawa et al., 10 Sept 2011. MZUSP 113561, 1, Caracaraí, igarapé Caicubi, trib. rio Jufari, 1˚0’54’’S, 62˚6’30’’W; O. T. Oy- akawa et al., 30 Aug 2011. MZUSP 30700, 3,

145 aqua vol. 23 no. 4 - 15 March 2018 NEWS of some Sicydiinae species from Babelthuap Island on the Palau archipelago With the help of Tom Bowling, from the BIOTA, Marine Life Nursery, a breeding establishment of 37 dif- ferent marine species, in June 2017, I was able to find many amazing Sicydiinae species, some of which are shown below, including unidentified and possibly new species on the Palau islands, to be verified by ichthy- ologists and taxonomist, as some (or most) of the species have never been recorded from Palau before.

Sicyopterus lagocephalus in nature on Palau’s Babelthaup island, June 2017. Photo by T. Bowling.

Sicyopus zosterophorus from Palau’s Babelthaup island, June 2017. Photo by H. Bleher in a aquarium.

Sicyopus zosterophorus Palau, mating pair June 2017. Photo Sicyopus aff. auxilimentus or Sicyopus aff. discordipinnis, by H. Bleher in a aquarium. very similar also to Lentipes concolor. Photo by T. Bowling.

Stiphodon pelewenis, from Palau’s Babelthaup island, June 2017. Photo by H. Bleher in a aquarium. aqua vol. 23 no. 4 - 15 marzo 2018 146 Stiphodon cf. pelewenis, from Palau’s Babelthaup island, very black. Photo by H. Bleher in a aquarium. FO01 on 30-05-2017.

Stiphodon pelewenis pair. from Palau Photo by H. Bleher. Sicyopterus surrufus. Babelthaup island, Photo by T. Bowling.

Sicyopterus surrufus, male from Palau’s Babelthaup island, June 2017. Photo by T. Bowling in a aquarium.

Stiphodon cf. percnopterygionus from Palau’s Babelthaup Stiphodon cf. percnopterygionus male and female, from island, June 2017. Photo by H. Bleher in a aquarium. Palau’s Babelthaup island, June 2017. Photo by H. Bleher.

Stiphodon cf. percnopterygionus, male, Palau. Note the amazing colour difference in this species. Photo by H. Bleher.

147 aqua vol. 23 no. 4 - 15 March 2018 Smilosicyopus leprurus, male from Palau’s Babelthaup island. Photo by H. Bleher in a aquarium.

Smilosicyopus leprurus, female from Palau’s Babelthaup is- Smilosicyopus leprurus female ventral view, from Palau’s land. Photo by H. Bleher in a aquarium. Babelthaup island. Photo by H. Bleher in a aquarium.

Smilosicyopus cf. chloe, male, from Palau’s Babelthaup island. Photo by H. Bleher in a aquarium.

Smilosicyopus cf. chloe, female from Palau’s Babelthaup island. Photo by H. Bleher in a aquarium.

Smilosicyopus aff. chloe, male, Palau, shows much similar Smilosicyopus aff. chloe, female, Palau, shows much similar with Smilosicyopus pentecost, male. Photo by H. Bleher. with Smilosicyopus pentecost, female. Photo by H. Bleher. aqua vol. 23 no. 4 - 15 marzo 2018 148 Guidelines for Authors

1. Manuscript preparation: manuscripts must be submitted BLABER, S. J. M. 1980. Fish of the Trnity inlet system of in English. In exceptional cases aqua may provide transla- North Queensland, with notes on the ecology of fish faunas tions of manuscripts written in French, German, Italian, or of tropical Indo-Pacific estuaries. Australian Journal of Ma- Spanish. rine and Freshwater Research 31:137-46. Manuscripts must be word-processed in Microsoft WORD DAY, J. H., BLABER, S. J. M., & WALLACE, J. H. 1981. Es- and submitted in an electronic form. Generic, specific, and tuarine fishes. In: Estuarine Ecology with Particular Reference sub-specific names must be italicised. All papers must con- to Southern Africa. (Ed. J.H. Day.): 197-221. A. A. Balke- form to the International Code of Zoological Nomencla- ma, Rotterdam. ture. Authors are strongly advised to follow the format set DIMMICH, W. W. 1988. Ultrastructure of North American out in previous publications of aqua. cyprinid maxillary barbels. Copeia 1988 (1): 72-79. 2. Title: the title must be short and should precisely identify TREWAVAS, E. 1983. Tilapiine Fishes of the Genera the main topic of the article. Names of genera or species are Sarotherodon, Oreochromis and Danakilia. British Muse- followed by the systematic group to which they belong. Au- um (Natural History), London, 583 pp. thor name(s) are given in full beneath the title, followed by the complete mailing and e-mail address(es). 7. Submission of manuscript and illustrations: The manu- script and illustrations must be submitted digitally to the 3. Keywords: five key words not found in the title. Scientific Editor: 4. Abstract: the abstract should not exceed 250 words and Flávio C. T. Lima draw attention to the principal conclusions. It should not Museu de Zoologia da Universidade Estadual contain any uncommon abbreviations or literature citations. de Campinas “Adão José Cardoso” The inclusion of abstracts in other languages is optional. Rua Albert Einstein s/n, Caixa Postal 6109, 5. Subject matter: the text of the manuscript is usually 13083-863 Campinas, São Paulo, Brazil arranged in four main sections: Introduction, Materials and E-mail: [email protected] methods (including a key to abbreviations), Results, and to whom all subsequent correspondence shall be addressed. Discussion. Other subdivisions may be chosen depending Texts, tables, and graphs should be in Microsoft-compatible on the material presented. Acknowledgements should be electronic form. placed between the text and references. After the paper has been accepted for publication, illustra- Scientific names of genera, species, and subspecies should be tions as high-resolution TIF files or original photographs followed by the name(s) of author(s) and the year of publica- (ideally transparencies; otherwise glossy prints, preferably in tion on first mention. A description of a new taxon must con- the size in which they will appear – the type area of aqua is tain the following sections: Material, Diagnosis, Description, 158 x 224 mm, one column is 76 mm wide) must be sent to: and Affinities. Synonyms must be arranged in chronological order. Identification keys must be dichotomous. Aquapress, The Managing Editor Via G. Falcone 11, Holotype and paratypes must be clearly identified, the insti- 27010 Miradolo Terme (Pavia), Italy tution in which they have been deposited named, and the E-mail: [email protected] catalogue numbers given. Private collections are not accept- able as repositories for holotypes. Authors should retain copies of all materials for reference. DNA sequences must be archived in GenBank, DNA Data- Proofs of accepted papers will be sent as PDF files by e-mail bank of Japan or European Molecular Biology Laboratory. attachment to the corresponding author. Voucher specimens associated with DNA sequences must be deposited in a recognized research collection accessible to 8. Evaluation of manuscripts: manuscripts will be evaluated the professional community. Accession numbers for se- by the editors and referees. Papers are accepted on the un- quences and catalog number for vouchers must be included derstanding that they have not and will not be published as a table in an appendix to the final manuscript. elsewhere. The metric system and SI units must be used. Temperatures 9. Reprints: Authors will receive one free hard copy of the are given in °C. Fractions should not be used. issue in which their paper appears and an e-print in PDF format for private/institutional use only. Additional hard 6. References to literature: the name-year system must be copies may be ordered from Aquapress. used. The list of references should be placed at the end of the paper, alphabetically arranged according to author 10. Copyright: All Scientific papers published in aqua are name. Only those publications cited in the paper may be in- considered to be the copyright of Aquapress. The authors are cluded. Titles of journals must be given in full. welcome to publish elsewhere on-line the Abstract of their paper ONLY. The complete paper is available on-line on Examples of correct reference formats: www.aqua-aquapress.com for on-line view and download. aqua International Journal of Ichthyology Vol. 23 (4), 15 March 2018

Contents:

Dalton Tavares Bressane Nielsen, Mayler Martins, Luciano Medeiros de Araujo, Fabio Origuela de Lira and Amer Faour: Cynolebias akroa, a new species of annual fish (Cyprinodontiformes: Rivulidae) from the rio Preto, São Francisco basin, northeastern Brazil ...... 113-121 Mandy T. Etpison and Patrick L. Colin: Blue Water Spawning by Moorish Idols and Orangespine Surgeonfish in Palau: Is it a “Suicide Mission”? ...... 121-136 Flávio C. T. Lima and Gilberto N. Salvador: Fooled by a fish: seed camouflage by an Amazonian banjo catfish, Bunocephalus verrucosus (Siluriformes: Aspredinidae) ...... 137-145 News of some Sicydiinae species from Babelthuap Island on the Palau archipelago ...... 146-148

Papers appearing in this journal are indexed in: Zoological Record; BioLIS – Biologische Literatur Information Senckenberg; www.aqua-aquapress.com; www.aquapress-bleher.com; http://scholar.google.it/scholar?q=aqua+international+journal+of+ichthyology&btnG=&hl=it&as_sdt=0%2C5

Cover photo: Bunocephalus verrucosus, ZUEC 15094, 72.1 mm SL, preserved specimen, in dorsal, lateral, and ventral views. Photo by E. G. Baena.

BIOTA is the company of Tom Bowling on the Palau archipelago doing sustainable aquaculture for over 4 years now breeding up to 37 marine fish species and many soft corals. Heiko Bleher was invited during June 2017 to do research on the freshwater species on the Babelthaup island. Some of it can be seen on pages 146-148. Photo by H. Bleher.