156. Jahrgang Stuttgart 2000 - Dendrobates.org

1 downloads 0 Views 3MB Size Report
Toes without webbing. Adpressed ... webbing and tadpoles with pointed tail and LTRF 2 (2) / 3 (1-2) (e.g. SLLVERSTO E ... es webbed feet (e.g. COLOMA] 995).




d e r G e s e ll s c h a f t fur Naturkunde in W u r t t e m b e r g •

156. Ja h rg an g S tu tt g ar t 20 00





Jh Gas f Jalurkde. Wurllemberg

156. Jahrgang

Stullgarl. 15. Delemb.u 2000

Jh. Ges. Naturkde. WOrttemberg

156. Jahrgang

Stuttgart, 15. Dezember 2000

A New Genus of Aposematic Poison Frog (Amphibia: Anura: Dendrobatidae) from the Upper Amazon Basin, with Notes on Its Reproductive Behaviour and Tadpole Morphology Eine neue aposematische Giftfrosch-Gattung (Amphibia: Anura: Dendrobatidae) aus dem oberen Amazonas-Becken mit Anmerkungen zu deren Reproduktionsverhalten und Kaulquappenmorphologie By/von STEFAN LOTTERS, Bonn, KARL-HEINZ JUNGFER, Gaildorf, ALEX WIDMER, Ziirich •

With 4 figures Mit 4 Abbildungen

Introduction The neotropical poison frogs (Dendrobatidae) are a conspicuous group of almost 200 small, often colourful and poisonous species. Dendrobatids are also remarkable for their complex territorial, mating and parental care behaviour. Numerous data suggest the family to comprise a monophyletic unit of New World anurans (e.g. VENCES et al. 2000). Several genus names have been erected (FROST 1985). For many years three of them were widely accepted, Colostethus COPE, 1866, Dendrobates WAGLER, 1830, and Phyllobates DUMERlL & BmRoN, 1841 (e.g., SAVAGE 1968; StLVERSTONE 1975, 1976). According to different authors (e.g. VENCES et al. 2000), Colostethus is a paraphyletic group including the cryptically coloured, usually non-toxic dendrobatids with more basal behaviour. Aromobates MYER & PAOULLa & DALY, 1991, Mannophryne LA MARCA, 1992, and Nephelobates LA MARCA, 1994 ,,1991" are related to Colostethus and also primitive. The names Dendrobates and Phyllobates as understood by SAVAGE (1968) and SILVERSTONE (1975, 1976) were applied to the aposematic taxa. Specie of these genera are also characterised by more or less toxic skin secretions as well as derived behavioural traits. MYERS et al. (1978) restricted Phyllobates only to some trans-Andean species allied to Phyllobates bie%,. DUMtRIL & B,BRoN, 1841. The remaining taxa formerly referred to this genus were transferred to Dendrobates. Subsequently, the latter genus wa splitted when Minyobates MYERS, 1987 [for several small species formerly allied to the D. minutus group of Sn.VI',RSTONF. (1975) and additional ont'sl .md Epipedobates M YI',RS, 1987 Ifor species referred to Phyllo/1atl's b II \'I'RJh.

(iel.

Narurkde. WUlrr. 156 (20011)



23 4

ST EF AN LO nE RS , KA RL -H EI NZ JU N ,FE R, AL EX WI DM ER

STONE (1 97 6) an d ad di tio na l on es ] w er e er ec te d. Epipedobates w as fu rth er di vi de d w he n Allobates ZlMMERMANN & ZIMMERMANN, 19 88 (fo r Prostherapis femoralis BOULENGER, 18 84 ) an d Phobobates ZIMMERMANN & ZLVlMERMANN, 19 88 [fo r Dendrobates silverstonei MYERS & DALY, 19 79 an d th e Phyllobates trivittatus gr ou p of SILVERSTONE (1 97 6) ] w er e de sc rib ed . A pa rt fr om m or ph ol og y, de nd ro ba tid ge ne ric ch ar ac te ris tic s incl ud e sk in to xi n pr of ile s, an d be ha vi ou r (e.g., MYERS 19 87 ; ZI MMERMANN & ZIMMERMANN 19 88 ). So m e of th e re ce nt ly de sc rib ed ge ne ra ha ve be en di sc us se d co nt ro ve rs ia lly (e.g., MYERS et al. 19 91 ) re ga rd le ss of th e la ck of co m pr eh en si ve ph yl og en et ic stu di es . Re ce nt ly , in fr a- fa m ili al re la tio ns hi ps of de nd ro ba tid fro gs ha ve be en acce ss ed on th e ba sis of m ito ch on dr ia l D N A se qu en ce s (e.g. CLOU GH & SUMMERS 20 00 ; VENCES et al. 20 00 ). In de pe nd en tly , di ffe re nt au th or s ro ug hl y ca m e to th e fo llo w in g co nc lu si on s: 1. Dendrobates an d Phyllobates ar e th e m os t de riv ed po is on fr og ge ne ra fo rm in g a cl ad e; 2. Epipedobates is th ei r m or e pr im iti ve sis te r gr ou p; 3. Allobates is a va lid ge nu s cl us te rin g w ith th e m or e ba sa l de nd ro ba tid s (Colostethus se ns u la to ); 4. of do ub tfu l ge ne ric po si tio n ar e Minyobates an d Phobobat es, w hi ch sh ou ld pr ob ab ly be st be re ga rd ed as ju ni or sy no ny m s of Dendr obates an d Epipedobates, re sp ec tiv el y (d . VENCES et al. 20 00 ). As a re su lt, ev ol ut io na ry re la tio ns hi ps am on g th e de riv ed ap os em at ic de ndr ob at id ge ne ra (a nd th e va lid ity of th ei r pr op os ed na m es ) ar e ge ne ra lly we ll un de rs to od . Th es e ar e al so su pp or te d by ot he r th an on ly ge ne tic ch ar ac te rs (e.g. ZIMMERMANN & ZIMMERMANN 19 88 ; MYERS et al . 19 96 ). W he n in cl ud ed in D N A se qu en ce st ud ie s, a sp ec ie s or ig in al ly de sc rib ed as Phyllobates azureiventris KNELLER & HENLE, 19 85 (Figs. 1, 2) al w ay s ap pe ar ed ap ar t fr om th e ot he r ge ne ra in a po si tio n of its ow n [v al id fo r all ge ne s se qu en ce d, i.e. 12 S, 16 S an d cy to ch ro m e b (CLOUGH & SUNLVlERS 20 00 ; VENCES et al . 20 00 ; WIDMER et al. in pr es s) ]. Th e ge ne ric pl ac em en t of th is sm al l ap os em at ic de nd ro ba tid fr om th e ea st er n ve rs an t of th e A nd es of Pe ru ha s al w ay s be en di sc us se d co nt ro ve rs ia lly . In th e or ig in al de sc rip tio n, a cl os e re la tio ns hi p w ith th e tra ns -A nd ea n ge nu s Phyl lobates se ns u st ric to w as pr op os ed (KNELLER & HENLE 19 85 ). In co nt ra st , M YER & BURROWES (1 98 7) su gg es te d a po ss ib le re la tio ns hi p w ith th e ci s- A nd ea n pictlls gr ou p of Epipedobates. A s a re su lt, MYERS (1 98 7) in cl ud ed az ureiventris in Epipedobates. M or ph ol og ic al ly , ad ul t Epipedobates an d Phyl/obate s ar e m or e or less id en tic al an d th er ef or e ex te rn al ch ar ac te rs ar e pr ob ab ly n t us ef ul to al lo ca te azureiventris. A cc or di ng to HENLE (1 99 2) , th e ge ne ric pl ac em en t re m ai ns to be re so lv ed be ca us e th e re pr od uc tiv e be ha vi ou r (a de sc rib ed by KNELLER 19 87 ) sh ow s ch ar ac te ris tic s of bo th Ph yl/obates an d Epipedobates. LOTTERS & KNELLER (2 00 0) de sc rib ed th e ad ve rti em nt ca ll of azureiventris an d co nc lu de d th at it w as ve ry si m ila r to tri ll ca lls kn ow n in th e sp ec ie s of Phyllobates se ns u st ric to an d so m e Ep ip ed ob at e. en et il' da ta su gg es t azureiventris to be a ba sa l pa rt of a si te r 1 d to DemJrolhltt's (in cl ud in g Minyobates) + Phyllobates, th at al 0 co nt in Epip edobat . (inJh.

• N tu rk d.

Urn. I

\ 000)

A New Genus of Aposematic Poi on Frog

3A

235

3B

Fig. 1. CryptophyLLobates azureiventris, subadult specimen of 18 mm SVL. Note colour pattern with light labial line, light lateral line and light dorsolateral line ending on posterior back. Abb.1 CryptophyLLobates azureiventris, subadultes Tier von 18 mm Kopf-RumpfLange. Beachte das Zeichnungsmuster mit leuchtender Labiallinie, leuchtendem Lateralstreifen und leuchtendem Dorsolateralstreifen, der auf dem Riicken ended. Fig. 2. Captive male parent of Cryptophyllobates azureiventris carrying aU tadpoles of one clutch together. Abb.2. Mannliches Elterntier von Cryptophyllobates azureiventris in Gefangenschaft, das aUe Larven eines Geleges gemeinsam transporriert. Fig. 3. Tadpole of CryptophyLLobates azureiventris (SMNS 9282: 2), in stage 39 in life in dorsal (A) lateral view (B). Abb.3. Lebende Larve von CryptophyLLobates azureiventris (SMNS 9282: 2) im Entwicklungsstadium 39 in Dorsal- (A) und Lateralansicht (B).

cluding Phobobates) and, in one case, a species of currently classified a a Colostethus. But azureiventris was always placed separately from Epiped bates (CLOUGH & SUMMERS 2000; VENCES et al. 2000). CLOUGH & U ,lJ',IER. (2000) also suggested that azureiventris may be basal not only t th Epipedobates lineages but to the entire Epipedobates + (Phy/lobates + Dendrobates) clade. At least, recently WIDMER et al. (in press), equ n ing parr of the mitochondrial cytochrome b gene of all known pe ie of Ph)'/I bites and azureiventris, demonstrated that the latter clear! i n t a Ph) /lOb,ltt'S. In summary, from a genetic point of view, it i apparent that a~lIreil·£'llt,.is is more closely related to Epipedobates than to Ph)11l butes. studied the reproductive behaviour of captive azureiventris and larval morphol g '. Our Jh, (,el, Nalurkde. WOrn, J ~6 (2000)

236

STEFAN U:l1TERS, KARL-HEINZ ]UNGFER, ALEX WIDMER

data demonstrate that the species is not a member of Epipedobates, from which it differs by at least one apomorphy. It shares more - bur not all characters with Phyllobates from which it appears genetically well separated. In our opinion, azureiventris forms a genus of its own. Because no name is available, we describe it as follows.

Systematics

Cryptophyllobates gen. nov. Type species: Phyllobates azureiventris KNELLER & HENLE, 1985. Etymology: From the Greek kryptos (hidden) + phyllos (leaf) + bates (a walker), in reference to (a) its secretive way of life in the forest leaf litter (d. KNELLER & HENLE 1985) and (b) its relative resemblance to the genus Phyllobates. Gender masculine. Diagnosis: Medium-sized dendrobatid genus with adult SVL ca. 25-28 rom (females larger than males). Maxillary teeth present. Medial lingual process absent. Surface of dorsum slightly granular. Toes and fingers not broadened; tips slightly broadened, dorsally with paired raised areas (scutes). Toes without webbing. Adpressed first finger in length equal to second. Dermal flap above cloaca absent. Ground colour black with aposematic pattern: complete light labial, dorsolateral and lateral lines present with dorsolateral line ending on posterior back (Fig. 1); ventral side marbled. Light proxirnoventral calf spot lacking. Cephalic amplexus absent in mating. Advertisement caLI a trill. All larvae from one clutch carried together by the male parent (Fig. 2). Larval development in pools with more or less running water. Tadpoles with dextral vent tube and rounded tail fin, tail in length> 70 % of total length (Fig. 3 A, B). Labial tooth row formula (LTRF) 2 (2) / 3 (Fig. 4). Allobates differs from Cryptophyllobates by having basal foot webbing and tadpoles with pointed tail and LTRF 2 (2) / 3 (1-2) (e.g. SLLVERSTO E 1976; ZIMMERMANN & ZIMMERMANN 1988). Epipedobates (including Phobobates) shows cephalic amplexus in mating, lacks a complete light lateral line, dorsolateral line never ends on posterior back and several species develop a light proximoventral calf spot (e.g. MYERS 1987). From Dendrobates (including Dendromedusa GISTEL, 1848, Hylaplesia SCHLEGEL in BOlE, 1827 and Minyobates) the new genus can be distinguished by hand morphology [in Dendrobates the first finger is shorter than the second when adpressed and finger discs are expanded (e.g. SILVERSTONE 1975)], the pre ence of light dorsolateral and lateral lines as well as maxillary teeth (absent in Dendrobates). In the latter genus the tadpoles are carried singly or in mall groups (usually not the whole clutch) either by a male or female parent. Larval development occurs in phytotelmata. Phyllobates alway lack a lateral line and, if present, the dorsolateral line ends in the groin, not on th poterior back (e.g. MYERS et al. 1978). Cryptophyl/obates differs from Colostethus (including Pr th COPf., 1868, Hyloxales JIM~NI!Z Ill! 1 A ESPADA, 1871 nd Ph II Jh. Ge . Nalurkd.

lint. I

(

A New Genu of Aposematic Poi on Frog

237

Fig. 4. Oral disc of the tadpole of Cryptophyllobates azureiventris in stage 39 (SMNS 9282: 2). Scale equals 1 mm. Abb.4. Mundfeld der Larve von Cryptophy/lobates azureiventris Enrwicklungsstadium 39 (SMNS 9282: 2). Die Linie entspricht 1 mm.

JIMENEZ DE LA ESPADA, 1871,) by having aposematic coloration with a dorsolateralline ending on the posterior back (usually absent but present in C. nexipus FROST, 1986), lacking webbing [present in many Colostethus (e.g. SILVERSTONE 1975; COLOMA 1995)], lacking a medial lingual process [as present in some Colostethus (GRANT et al. 1997)] and absence of cephalic amplexus [usually present in Colostethus (e.g. ZIMMERMANN & ZIMMERMANN 1988)]. From other cryptically coloured dendrobatids, Cryptophyllobates differs also by the following characters: from Aromobates by smaller size, from Mannophryne by absence of a collar pattern on the throat, from Nephelobates by lacking a dermal flap above the cloaca (MYERS et al. 1991; LA MARCA 1992, 1994). Species included: Phyllobates azureiventris l 70 % ve rs us ca . 56 -6 4 % in la rv ae of Phyllobates sp p. be tw ee n st ag es 25 an d 41 (b as ed on da ta pr ov id ed by DONNELLY et al . 19 90 ; se e also MYERS et al. 1978)], a m or e po st er io r po si tio n of th e sp ira cl e op en in g [9 m m be hi nd sn ou t in st ag e 37 ve rs us ca . 6.9 m m in a P. lugubris ta dp ol e of th e sa m e st ag e (D on ne lly et al. 1990)] an d a la rg er or al di sc w id th [2.58 in st ag e 37 ve rs us ca . 2. 25 m m in a P. lugubris ta dp ol e of th e sa m e sta ge (D on ne lly et al. 1990)]. O th er m or ph om et ric di ffe re nc es w e as su m e to lie w ith in th e va ria tio n of th e di ff er en t species. Sy st em at ic D is cu ss io n Cu rr en tly , th er e ar e ei gh t va lid de nd ro ba tid fr og ge ne ra . U nf or tu na te ly , th es e ar e no t un ifo rm ly de fin ed an d pl es irn or ph ic ch ar ac te rs ar e pr ob ab ly to o he av ily in vo lv ed (c om p. MYERS et al. 19 91 ; MYERS 19 87 ). Cryptophyllobates m os t cl os el y re se m bl es Epipedobates an d Phyllobates. Fr om th e fo rm er it di ffe rs by at le as t on e ap om or ph y, i.e. th e la ck of ce ph al ic am pl ex us du rin g m at in g. Th e la tte r ch ar ac te r is al so ab se nt in th e de riv ed Phyllobates + Dendrobates cl ad e an d th e m or e pr im iti ve Allobates. Th is su gg es ts th at ce ph al ic am pl ex us ha s be en lo st at le as t tw ic e in de nd ro ba tid ev ol ut io n (e.g. VENCES et al. 20 00 ). If Cryptophyllobates w as ba sa l to Epip edobates (d . CLOUGH & SUMMERS 20 00 ; VENCES et al . 20 00 ), ce ph al ic am pl ex us w ou ld ha ve in de pe nd en tly di sa pp ea re d on e m or e tim e. If Cryptophy llobates w a ba sa l to th e en tir e Epipedobates + (Phyllobates + Dendrobates ) cl ad e a alte rn at iv el y pr op os ed by CLOUGH & SUMMERS (2 00 0) , it w ou ld pr ob ab ly be re la te d to Allobates. In co nt ra st to th e "h ig he r" de nd ro ba tid s, bo th th es e ge ne ra ex hi bi t a lig ht la te ra l lin e (w hi ch m ay be in te rp re te d as a pl es io m or ph ic ch ar ac te r) . H ow ev er , th e ta dp ol e of Cryptophyllobates re se m bl e larva e of Epipedobates (a lth ou gh no t all) an d es pe ci al ly Phylloba tes m or e (alth ou gh pr ob ab ly ch ar ac te ris ed by lim ite d yn ap om or ph ie s on ly ); fu rth er , Cryptophyllobates la ck s fo ot w eb bi ng (a de riv ed ch ar ac te r la ck in g in Allobates). Allobates sp p. ar e of te n sa id to be sli gh tly ap os em at ic (e.g. E CE et al . 20 00 ) w hi ch is no t th e ca se in Cryptophyllobates [b ut ne ith er in A. ;:,1para (SILVERSTONE, 1976)). M or eo ve r, du e to its br ig ht co lo ur pa tte rn , we su sp ec t th at th e la tte r pr od uc es re la tiv el y po te nt sk in to xi ns (in All b