11 August 2003 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON 116(2):317-329. 2003. Lßptodactylus caatingae, a new species of frog from eastern Brazil (Ampliibia: Anura: Leptodactyjidae)
W. Ronald Heyer an4 Flora A. Junc4 (WRH) Amphibians and Reptiles, MRC 162, RQ. Box 37012, Smithsonian Institution, Washington, D.C. 20Q13-7012, U.SA,, e-mail: [email protected]; (FAJ) Departamento de Ciencias Biológicas, Laboratorio de Animais Peçonhentos e Herpetologia, Universidade Estadual de Feira de Santana, Campus Universitario•:km 03 (BR 116), 44031-460 Feira de Santana, Bahia, Brazil, e-mail: [email protected]
Abstract.•^A iiew species of frog of the genus Leptodactylus is described from eastern Brazil. The new species differs little morphologically from L. latinasus, but the advertisement calls are very different. The geographic dis- tribution of L. latinasus an4 the new species, together with the respective levels of differentiation of morphology and call, is mirrored in the species pair L. bufonius and L. troglodytes. Presumably the same historical event or events lead to differentiation of these species pairs.
In a revision of the Leptodactylus fuscus Estadual de Feira de Santana) were bor- species group (Heyer 1978), Leptodactylus rowed and data were taken from them. Data latinasus was characterized as having a ma- for Leptodactylus latinasus are those used jor distribution pattern in Argentina, Uru- in Heyer (1978) with some additional data guay, and the State of Rio Grande do Sul, for eye-nostril distance, tympanum diame- Brazil and a few disjunct localities in east- ter, and belly patterns (data for these char- em Brazil. The late Dr. Adao J. Cardoso acters were not evaluated in the study pub- recorded the advertisement call of a male L. Ushed as Heyer 1978). latinasus from the State of Bahia in Brazil Measurement data include snout•vent and brought the recording to the attention length (SVL), head length, head width, eye- of WRH, pointing out that its call was very nostril distance, thigh length, shank length, different from the other known recordings and foot length, following Hey er et al. of L. latinasus from Argentina and Rio (1990) except for the eye-nostril distance Grande do Sul, Brazil. More recently, FAJ being measured by calipers as the distance collected a series of specimens of this east- from the anterior comer of the eye to mid- em Brazil form from the fossil sand dune nostril and the tympanum diameter being region of Bahia. The purpose of this paper measured by calipers as the maximum di- is to describe the eastern Brazilian popula- ameter of the tympanum including the an- tion currently identified as L. latinasus as a nulus. new species and provide new field obser- Measurement data were analyzed using vations for the taxon. the software program SYSTAT 10 for prin- cipal component analysis (Stenson and Wil- Materials and Methods kinson 2000) and discriminant function All specimens of the new species in the analysis (Engelman 2000). MZUSP, UEFS, and ZUEC collections The advertisemeijt call analyzed of the (museum abbreviations follow Levitón et new species is USNM recording 234, cut 1, al. 1985 with UEFS being the Universidade from Joazeiro, Fazenda Mary, Bahia, Bra- 318 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON zil, 7 March 1990, air temperature 26°C, Advertisement Calls 1910 h, voucher ZUEC 8833, by Adao J. The call of the new species is given at Cardoso. All calls were recorded from the an average rate of 160/min. The call dura- single voucher specimen. Call rate was tion ranges from 64 to 77 ms. Each call has based on eight samplings of the recording 7-8 distinct pulses (Fig. 1), given at a rate for a total of 396 calls. Other call parame- between 102 to 109 pulses/sec. The domi- ters are evaluated from analysis of 10 calls. nant frequency for the entire call has two For comparative purposes, a recording of peaks (Fig. 2), the slightly less loud peak Leptodactylus latinasus, USNM recording ranges from 1072 to 1121 Hz, the loudest 19, cut 1, from Embarcación, Salta, Argen- peak ranges from 1543 to 1648 Hz. The call tina, 22 December 1971, air temperature is frequency modulated as rising from the 22°C, time not noted, unvouchered, by W. beginning of the call to a maximum fre- Ronald Heyer, was analyzed. Call rate was quency at about % the call duration, then based on three samplings of the recording slightly falling to the end of the call; the for a total of 57 calls. Other call parameters lowest dominant frequency ranges from 934 are evaluated from analysis of seven calls to 963 Hz, the highest dominant frequency (non-overlapping calls in a chorus) with fil- ranges from 1543 to 1648 Hz. There are no ter bandpass settings of 600 and 11,205 Hz. indications of harmonic structure in the call. Advertisement calls were analyzed using (Fig. 3). Canary 1.2 software (Charif et al. 1995). The call of the new species is dramati- The calls were digitized at a sample rate of cally different from the call of Leptodac- 22,050 Hz, sample size of 16 bits. Call tylus latinasus in terms of amplitude struc- component terminology follows Duellman ture (Fig. 1, distinct pulses present in the and Trueb (1986) and Heyer et al. (1990). new species, absent in L. latinasus) and fre- Call duration was measured from the wave- quencies (Fig. 3, Table 1). The call of the form. Dominant frequency was determined new species from Ibiraba, Brazil sounds using the spectrum analysis in Canary with very similar to Cardoso's recording from settings of analysis resolution filter band- Joazeiro, Brazil, and very different from the width 349.70 Hz, frame length 256 points, calls of Argentinian recordings of L. lati- grid resolution time 128 points, overlap nasus and L. fuscus, another species in the 50%, frequency 43.07 Hz, FFT size 512 same species group that occurs in some of points, window function hamming, ampli- the same localities as the new species (FAJ, tude logarithmic, clipping level •80 dB. pers. obs.) The differences between the Minimum and maximum call frequencies calls of the new species and L. latinasus were measured from the audiospectrogram convincingly demonstrate that they repre- display (spectrogram in Canary terminolo- sent two distinct species. gy) using settings of analysis filter band- width 174.85 Hz, frame length 512 points, Morphology grid resolution 256 points, overlap 50%, The following variation occurs in the frequency 43.07 Hz, FFT size 512 points, new species. window function hamming, amplitude log- Dorsal patterns include well-defined to arithmic, clipping level •80 dB, display weakly defined single or double mid-dorsal style smooth. Call amplitude modulation dark chevrons (as in Heyer 1978, Fig. lA, was evaluated from visual inspection of ex- C, p. 3), chaotically placed small to medi- panded waveform displays. Harmonics um sized dark markings, or almost uniform. were determined using both expanded wave A very interrupted light or dark mid-dorsal form displays and power spectrum analysis pin stripe is present or absent. displays. Lip stripes range from a light stripe well- VOLUME 116, NUMBER 2 319
b
••jW Fig. 1. Wave form of advertisement call of Leptodactylus latinasus (a) and the new species from eastern Brazil (b); scale bars = 0.02 sec. defined both above and below from the tip The specimens from Ibiraba, Bahia have of the snout passing under the eye and con- a greater development of belly pattern than tinuing under the tympanum onto the com- observed from other localities. The Ibiraba missural gland to indistinct or indiscernible. specimens range from having an almost All of these conditions occur in specimens uniform distribution of densely packed me- from Ibiraba, Bahia; all but one of nine lanophores (pale to the eye, but melano- specimens from Itiúba, Bahia have well-de- phores distinct under magnification), a var- fined stripes. iegated pattern of melanophores, to groups All but one specimen were scored as hav- of melanophores scattered over the belly. ing a distinct light stripe on the lower face Specimens from the other localities either of each posterior thigh surrounded by irreg- lack a belly pattern, have a weakly varie- ularly shaped broad dark outlines with the gated pattern of melanophores, or have only rest of posterior thigh having a relatively a few scattered melanophores lateralmost uniform scattering of melanophores to a on the belly. mottled pattern. One specimen from Ibira- Many individuals lack any indication of ba, Bahia has a series of light spots, rather dorsolateral folds. Several individuals have than a continuous light stripe. very interrupted folds from the eye to the The upper shanks have irregularly shaped sacrum. One individual has a very inter- dark transverse bands. rupted fold from the eye to the groin. Most 320 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON •50 dû -60 CO G (D -70- -80- -T- 0 12 3 4 Frequency in kHz Fig. 2. Power spectram of advertisement call of new species of Leptodactylus from eastem Brazil. a b 5- N 0 O c 0) 2- er 1 - M M N ,^..^. ^-.-> -U.^^..-.,:-.--..:^-.-^«..*^ i. ¿iii^i^-^v-4i.-s-^ ^i^lavA»,-n^ 1-a ífrái .-,..-¡i.t^-fel 2 0 Time in seconds Fig. 3. Audiospectrograms of advertisement calls of Leptodactylus latinasus (a) and the new species from eastern Brazil (b). VOLUME 116, NUMBER 2 321 individuals have complete to broken lateral folds. The upper tibia has scattered to many distinct white tubercles. Twenty of 21 spec- fQ 'ils-* S imens have many distinct white tubercles on the outer tarsal surface and sole of foot. One individual has scattered distinct white tubercles on the outer tarsal surface. ü Males range from 32.1 to 36.9 mm SVL '•S tí tí 3 (3 y o O ^ (n = 8), females 36.2 and 39.1 mm (n = o o r- o >n OÍ UM en CO r) 2). The dorsal pattern, lip stripes, posterior thigh pattern, upper shank pattern, mid-dor- I sal pin stripes, upper shank texture, outer tí c J3 tarsal texture, and sole of foot texture traits c are indistinguishable between the new spe- cies and L. latinasus. .Sc| o g.« Belly pattern data were not taken for the & Heyer (1978) study. WRH examined bellies of all USNM specimens of L. latinasus from Argentina, Uruguay, and Rio^ Grande do Sul, Brazil. Forty-nine individuals lack any belly pattern. One specimen from Ar- gentina (USNM 319587) and two speci- ooo o ot~- odd tí mens from Uruguay (USNM 535968, 539970) have a few melanophores on the lateralmost venter and/or in the arm inser- •T3 o Ö tion portion only of the venter. One indi- o o.u vidual from Uruguay (USNM 535969) has ot/i ••B a pair of small blotches just behind the arm on the sides of the chest. None of the Ar- gentina, Uruguay, or Rio Grande do Sul, Brazil specimens examined have as exten- sive venter patterns as found in some indi- ON ^ d viduals of the new species. Body fold development was not scored in the same way for the Heyer (1978) study o o q \o d oi and this study. A series of 20 well-pre- i3 served adult Leptodactylus latinasus from o. Salta Province, Argentina examined for this study shows the same expressions of lateral folds as in the new species, but with dif- 2 " ferent frequencies of occurrences of indi- •p M I DO vidual states. Only one of the 20 L. lati- a> taSá +ä2 '= nasus has a very broken dorsolateral fold tí .C 73 &I m u M from the eye to % distance to the sacrum; tí tí the other 19 show no indications of dorso- •s -S tí tí lateral folds. Thus, the new species appar- bj) W) tJO td Ü I -5 < -^ ffl ently has a greater frequency of any dor- 322 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASfflNGTON solateral fold expression than occurs in L. latinasus. There are two sets of measurement data. A more extensive data set (in terms of num- U o ber of individuals) is available for all adults ^ ^ w (M §8 examined for the Heyer (1978) study for é •a SVL, head length, head width, thigh length, 00 N ^ shank length, and foot length. Additional \o t^ ^O data for eye-nostril distance and tympanum diameter are available for a smaller number of adults for both the new species and Lep- todactylus latinasus. Due to the small number of adults avail- able for the new species, principal compo- nent analyses are the primary statistical tool used to explore measurement variation in adults of the new species and Leptodactylus •* 00 es latinasus (Table 2). The principal compo- •S M . d Ö Ö nent analyses for males show a greater dis- It tinction between the two species using the data for all variables versus the data set Ö lacking eye-nostril distance and tympanum 11 n 1 1 1 r 1 r -2-10 1 2 -20246 Principal Component 1 Principal Component 1 Fig. 4. Male measurement data principal component analyses. A. Reduced variable data set. B. Variable set including eye-nostril distance and tympanum diameter data. Circles = Leptodactylus latinasus, filled triangles = new species. "1 1 1 1 r r -2-10 1 2 2-10 1 Principal Component 1 Principal Component 1 Fig. 5. Female measurement data principal component analyses. A. Reduced variable data set. B. Variable set including eye-nostril distance and tympanum diameter data. Circles = Leptodactylus latinasus, triangles = new species. 324 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON tylus latinasus are very similar morpholog- indistinct dorsolateral folds or lacks them. ically. There are no characters that com- Leptodactylus caatingae, fragilis, and lati- pletely diagnose the two species. There are nasus have considerable morphological and a few character states that demonstrate color pattern overlap and cannot be consis- some differentiation between the species. If tently diagnosed from each other with these the only data available were the morpho- characters. The advertisement call of L. la- logical data, we would interpret the varia- tinasus is not pulsed and has a high broad- tion to be within-species and would recog- cast frequency (3000-3780 Hz); the call of nize only one species for the two taxa, as L. caatingae is pulsed and has a lower was done in the Hey er (1978) study. broadcast frequency (940-1620 Hz). The Given that the advertisement call data in- advertisement call of L. fragilis is longer dicate that the specimens from eastern Bra- (0.19 sec) than the call of L. caatingae zil represent a species distinct from Lepto- (0.07 sec). Leptodactylus caatingae, fragil- dactylus latinasus and there is no name is, and latinasus have allopatric distribu- available for it, we hereby describe the tax- tions: L. fragilis from southernmost Texas, on lacking a name as: United States to north coastal Venezuela; L. caatingae in eastern Brazil; and L. latinasus Leptodactylus caatingae, new species in southern South America (Argentina, Bo- livia, Rio Grande do Sul, Brazil, Uruguay). Figs. 6-7 Description of holotype.•Snout rounded Holotype.•ZUEC 8833, an adult male from above, acutely rounded in profile; can- from Brazil, Bahia, Joazeiro, 09°25'S, thus rostralis indistinct; lores obtusely con- 40°30'W. Collected by Adâo J. Cardoso, 7 vex in cross section; tympanum well de- March 1990. fined, moderate size, horizontal diameter Referred specimens.•BRAZIL; BAHIA; about % eye diameter; vocal slits elongate, Bom Jesus da Lapa, 13°15'S, 43°25'W, parallel to lower jaw, starting about mid- UMMZ 109991(2); Ibiraba, 10°48'S, tongue level; vocal sac single with distinct 42°50'W, UEFS 686, 688-695; Itiuba, lateral expansions indicated by pronounced 10°42'S, 39°51'W, MZUSP 38556-38559, folds under lower jaws; vomerine teeth in 38561-38563, USNM 547844-547845; almost straight line patches separated from Sao José do Rio Grande, 11°49'S, 44°44'W, each other by about V2 length of a single UMMZ 109992; ESPÍRITO SANTO; Sao tooth row, well behind and between almost Mateus, 18°44'S, 39°51'W, MCZ A-92142 round choanae; finger lengths II ~ IV llillilil Uli Uli llillilil • • jai' •^¥-' r '-*. ,1. -•..•;«in Fig. 6. Holotype of Leptodactylus caatingae. 326 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON Fig. 7. Leptodactylus caatingae, UEFS 695, photo by F. A. Juncá. toe surfaces lacking ridges or fringes, no posterior face of thigh with a very distinct indication of basal webbing between toes; light ventral stripe, bordered by dark subarticular tubercles distinct, pungent; no brown, rest of posterior face of thigh mot- metatarsal fold; tarsal fold low, broad, glan- tled tan and brown; lateral chin area mottled m dular, extending % distance of tarsus, either with brown, extending onto vocal folds, rest abutting inner metatarsal tubercle or with a of venter immaculate. slight gap between fold and tubercle; upper Etymology.•The name, Latinized from shank surface with scattered white tuber- the Portuguese word caatinga, refers to the cles; outer tarsal surface with many white characteristic distribution of this species tubercles; sole of foot with many white tu- within the Caatinga Morphoclimatic Do- bercles. main (Ab'Sáber 1977). SVL 35.3 mm, head length 14.2 mm, Larval characteristics.•Unknown. head width 12.7 mm, eye-nostril distance Advertisement call.•See previous de- 3.3 mm, tympanum diameter 2.9 mm, fe- scription. mur length 13.4 mm, shank length 14.1 Karyotype. •Unknown. mm, foot length 14.9 mm. Distribution.•Currently known from a Dorsum blotched tans and browns; a few localities in eastern Brazil (Fig. 8). AU very interrupted series of dark dots from but one of the localities occur in the Caa- between the nostrils to just past the sacrum; tinga Morphoclimatic Domain (Ab'Sáber upper limbs with weakly and irregularly de- 1977). The southernmost locality, in Espir- fined darker transverse bands; face with ito Santo, occurs in the Atlantic Forest dark canthal stripe only well defined just in Morphoclimatic Domain (Ab'Sáber 1977). front of eye; moderately well defined light The locality of Rio Grande at Sao José, Mi- stripe from tip of snout passing under eye nas Gerais, Brazil for UMMZ 109992 given •-i^^-^.- ^^^tSÜ and tympanum and extending to end of in Hey er (1978) was incorrect and is Sao commissural gland; ñanks with indistinctly José do Rio Grande, Bahia. Natural history.•FAJ made the follow- ••t. f • darker brown blotches on a tan background; r V Ai •< •-•• ^--^^ •lit: f. , .-*^"^'' ' \ ^J. . . - V - vii-; v> ^ '•A, /ñ^^.... .i'^s!^- >\'.-.¿tí' •Aaf' í!^.----. •':: VOLUME 116, NUMBER 2 327 Fig. 8. Distribution map of Leptodactylus caatingae and L. latinasus. Type locality of L. caatingae indicated by dot; other localities by circles. Predicted distribution range of L. latinasus indicated by shaded area. 328 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON ing observations on specimens from Ibira- represent a single species. There is variation ba, Bahia. On March 2000, in the rainy sea- in belly pattern development among speci- son, about 10-20 males were calling within mens of L. caatingae from different local- grass bordering a flooded forest area that ities. Advertisement calls are very impor- was totally dry the day before. One of these tant in evaluating species limits in the frogs males was calling in an exposed position that have been known as L. latinasus. Un- and FAJ could observe and collect it. The fortunately, we have only a recording from air temperature was 25°C. Other specimens a single frog of L. caatingae. Recordings were collected in pitfall traps, which were are needed from the other population sam- put in on one edge of the Icatú River. In ples we have designated as referred speci- that area the vegetation is a typical gallery mens of L. caatingae in order to determine forest with trees up to 5-7 m tall. In the whether there are additional species con- same area Leptodactylus fiiscus, ocellatus, tained in the specimens we have examined. and podicipinus were calling too. In August Because of the observed belly pattern var- 2000 the same area was dry and no indi- iation and lack of additional recordings, we viduals of L. caatingae were calling, al- think it is inappropriate to designate para- though some L. podicipinus males were typic specimens for L. caatingae. heard. The currently known distribution of L. caatingae is confounding. The Atlantic For- Discussion est Morphoclimatic Domain locality of Sao Mateus, Espirito Santo is peculiar. There The call rates of Leptodactylus latinasus are very few species of frogs that occur in analyzed for this paper are very different both caatinga and Atlantic Forest areas. from those analyzed by Barrio (1965, see Copeland Hartt collected MCZ A-92142 in Table 1) and may impact the call rate de- 1865/1866. The available evidence is quite scribed for L. latinasus herein. The differ- convincing that the specimen was collected ences are real and not due to different from the Atlantic Forest Morphoclimatic means of analysis. The major difference be- Domain and most probably Sao Mateus it- tween the recordings involved is that Bar- self. Hartt collected amphibians and reptiles rio's recordings appear to be of a single from five other localities in addition to Sao calling individual, whereas the recording of Mateus in the States of Bahia and Espirito L. latinasus analyzed herein was a chorus Santo. All of these five localities are also recording, although there seemed to be one from the Atiantic Forest Morphoclimatic primary individual that was louder (i.e., Domain. There are no habitat data for MCZ closer to the microphone), but it is possible A-92142. There are at least three possibil- that two or more individuals were calling ities to explain a presumably caatinga equally loudly from the microphone in a adapted frog being collected at Sao Mateus: manner that their calls did not overlap with (1) the Sao Mateus population inhabits each other. Thus, it is possible that the dif- open areas within the Atlantic Forest Mor- ference in calling rates of L. latinasus as phoclimatic Domain, such as sandy beaches presented in Table 1 could be due either to along the ocean coast; (2) the Sao Mateus more than one individual actually being an- population actually represents a new cryptic alyzed in the Salta recording, or L. latinasus species of this complex; and (3) this distri- increases calling rates when in a chorus of bution represents another example of how conspecifics. The recorded call of L. caa- little we really understand about neotropical tingae has every indication that it is from a frog distributions. Collecting new material single calling male, however. from Sao Mateus of L. caatingae, including We are not certain that all of the speci- recording its advertisement call, could re- mens we have examined of L. caatingae solve the current distributional anomaly. VOLUME 116, NUMBER 2 329 The pattern of differentiation and distri- by Copeland Hartt. Greg Schneider provid- bution described herein for L. caatingae ed J. Bailey's collection records to resolve and L. latinasus is very similar to that the locality for UMMZ 109991. P E. Van- found between two other members of the zolini and George R. Zug reviewed the same species group (the L. fuscus group): manuscript. Research for this paper was L. bufonius and troglodytes. These two lat- supported in part by the Conselho de De- ter species are very similar morphologically senvolvimento Científico (CNPq) and and differ markedly in their advertisement UEFS (FAJ), the Neotropical Lowlands Re- calls (Heyer 1978). Leptodactylus bufonius search Program, Smithsonian Institution has a distribution in the open formations of (WRH, Richard P. Vari, Principal Investi- Argentina, Bolivia, Mato Grosso do Sul, gator) and the National Science Foundation Brazil, and Paraguay. There is an extensive (award 9815787 to Rafael de Sa and distributional hiatus between L. bufonius WRH). We thank these individuals and in- and L. troglodytes, with L. troglodytes oc- stitutions for their support. curring in open formations of northeast Brazil (compare maps in Figs. 34 and 68 in Literature Cited Heyer [1978]). This similar pattern suggests Ab'Sáber, A. N. 1977. Os dominios morfoclimáticos that the two species pairs shared the same na América do Sul. Primeira aproximaçâo.• differentiation history, probably the ances- Geomorfologia (52): 1-22 + map. tor of each having a general distribution Barrio, A. 1965. Afinidades del canto nupcial de las throughout the diagonal of open formations especies cavícolas del genero Leptodactylus (Anura-Leptodactylidae).•Physis 25:401-410. in South America. Whatever event isolated Charif, R. A., S. Mitchell, & C. W. Clark. 1995. Ca- the populations from northeast Brazil from nary 1.2 user's manual. Cornell Laboratory of the populations in and adjacent to Argen- Ornithology, Ithaca, New York, 229 pp. tina was probably the same event that led Duellman, W. E., & L. Trueb. 1986. Biology of am- phibians. McGraw-Hill Book Company, New to speciation of both L. bufonius-troglo- York, St. Louis, San Francisco, 670 pp. dytes and L. caatingae•latinasus. This hy- Engelmann, L. 2000. Discriminant analysis. Pp. 1-275- pothesis would be corroborated if the levels 1-326 in SYSTAT® 10. Statistics I, SPSS Inc., of molecular differentiation between the Chicago. 663 pp. pairs of species were identical. If that turns Hayek, L-A. C, W. R. Heyer, & C. Gascon. 2001. Frog morphometrics: a cautionary tale.•Alytes 18: out to be true, the molecular data might pro- 153-177. vide a suggestion of when the allopatry oc- Heyer, W. R. 1978. Systematics oí Xhe fuscus group of curred, which, in turn, might agree with the the frog genus Leptodactylus (Amphibia, Lep- hypothesis proposed by Vanzolini (1997: todactylidae).•Natural History Museum of Los Angeles County, Science Bulletin (29):l-85. 80-84) of separation of dry formation pop- Heyer, W. R., A. S. Rand, C. A. G. Cruz, O. L. Peixoto, ulations by fragmentation due to Pleisto- & C. E. Nelson. 1990. Frogs of Boracéia.^Ar- cene wet forest expansions during more quivos de Zoología 31:231-410. mesic periods than present. James, E C, & C. E. McCulloch. 1990. Multivariate analysis in ecology and systematics: panacea or Pandora's Box?•Annual Review of Ecology Acknowledgments and Systematics 21:129-166. Levitón, A. E., R. H. Gibbs, Jr., E. Heal, & C. E. Addison Wynn produced Fig. 8. P. E. Dawson. 1985. Standards in herpetology and Vanzolini loaned materials from the Museu ichthyology: Part I. Standard symbolic codes de Zoología da Universidade de Sao Paulo for institutional resource collections in herpe- tology and ichthyology.•Copela 1985:802- and provided advice on the formation of the 832. name for the species described as new in Stenson, H., & L. Wilkinson. 2000. Factor analysis. this paper. Adriana A. da Silva, Flávia M. Pp. I-327-I-363 in SYSTAT 10. Statistics I. P. Lima, Silvana A. da Silva, and Fabio R. SPSS Inc., Chicago, 663 pp. Vanzolini, P. E. 1997. The silvestrii species group of C. Castelucci made the difficult fieldwork Amphisbaena, with the description of two new at Ibiraba successful. José Rosado assisted Brasilian species (Reptilia: Amphisbaenia).• with information about materials collected Papéis Avulsos de Zoología 40:65-85.