Comparing the Tadpoles of Hyla geographica and Hyla semilineata

Abstract

-External morphology, internal oral anatomy, and chondrocranial anatomy were examined for tadpoles of Hyla geographica from the Amazon rainforest, Brazil, and Hyla semilineata from the Atlantic rainforest, Brazil. Here, we provide morphological larval data to help diagnose these closely related species. Scanning electron microscopy analysis of buccal morphology showed the most distinctive features between these species: the distance between the lingual papillae in the buccal floor of H. geographica is three times greater than that distance in H. semilineata, and the relative size of the lingual papillae in H. geographica is less than half their size in H. semilineata. Although the chondrocranium of both species is identical, and the external morphology of the larvae of both taxa is very similar, they differ greatly in size at most developmental stages. A multivariate analysis of covariance, corrected for stage and size, also showed a significant difference between morphometric measurements of the two species. These support the existence of two separate taxa. The identity of Hyla geographica and H. semilineata has not been satisfactorily resolved. Hyla geographica was described in 1824 by Spix, who also described a variety identified as H. geographica var. sive semilineata. These taxa were synonimized by Bokermann (1966). Subsequently, Lutz (1973) stated that there are no clear-cut differences between the two taxa, and Duellman (1973) also placed both taxa under the name H. geographica. More recently, Silveira and Caramaschi (1989) compared adult morphology and argued for the specific recognition of H. semilineata for the taxa inhabiting the coastal region of Brazil, from the State of Alagoas to the State of Santa Catarina. These authors differentiated H. semilineata from H. geographica by its narrower head, more rounded snout, smaller tympanum, larger nuptial pad, and color pattern. General descriptions of tadpoles and their oral discs have been used as taxonomic characters sufficient for species differentiation (Bokermann, 1963; Altig, 1970). However, the descriptions of the larvae of Hyla geographica and Hyla semilineata in the literature are very similar, suggesting that a more detailed study of the oral disc and of other morphological characters may be important for distinguishing these species. The tadpoles of H. geographica and H. semilineata are classified as Orton type IV larvae (Orton, 1953), because they possess a horny beak furnished with denticles and a single, sinister spiracle. These tadpoles, like most hylids, are adapted to lentic environments. Characteristically, they posses an ovalshaped body, with a Present Address: Departamento de Zoologia, Universidade Estadual Paulista, C.P. 199, 13506-900 Rio Claro, Sao Paulo, Brasil. Corresponding author. tail about twice the size of the body, and an antero-ventral mouth. (Orton, 1953; Bokermann, 1963; Kenny, 1969; Duellman and Lescure, 1973; Duellman and Trueb, 1994). Bokermann (1963) described the tadpoles from the Atlantic rainforest (Piassaguera, State of Sao Paulo, Brasil) as being blue/black with gray limbs and having a dental formula of 2(2)/4. Kenny (1969) described the tadpoles from Trinidad in a similar manner as did Duellman and Lescure (1973) for the species from French Guiana (identified incorrectly as Osteocephalus taurinus; see Caldwell, 1989), although the latter observed a 3(1)/5 dental formula at Stage 37. We examined larvae of Hyla geographica and Hyla semilineata from allopatric populations and here we describe their morphometric differences, external and internal oral structures, and chondrocranial anatomy, presenting features that help to distinguish the two species. MATERIALS AND METHODS Tadpoles of Hyla geographica were collected at the Reserva Florestal Adolfo Ducke, Manaus, State of Amazonas, Brasil in January, 1996. Tadpoles of Hyla semilineata were collected at the following localities: Poco das Antas, Mongagua, State of Sao Paulo, Brasil in April, 1994 and October 1995; Picinguaba Beach, Ubatuba, State of Sao Paulo, Brasil in January 1992; Massaguacu Beach, Caraguatatuba, State of Sao Paulo, Brasil, in September, 1993; and Recanto Silvestre, Blumenau, State of Santa Catarina, Brasil in January, 1996. Specimens are deposited in the CFBH collection of the Departamento de Zoologia, Instituto de Biociencias, UNESP, Rio Claro, Sao Paulo, Brazil. Tadpoles were preserved in 10% buffered formaldehyde solution. Most of the tadpoles This content downloaded from 207.46.13.193 on Thu, 08 Sep 2016 04:24:24 UTC All use subject to http://about.jstor.org/terms A. D'HEURSEL AND R. O. DE SA were preserved shortly after collection, but some were reared in the laboratory to verify identification. Tadpoles were staged according to Gosner's table of normal development (Gosner, 1960). Measurements were taken with calipers to the nearest 0.1 mm. Morphometric measurements of 20 specimens of each species from Stage 31 to 39 follow Lavilla and Scrocchi (1986). Morphometric measurements were: total length (TL), body length, from snout to vent (BL), tail length, from vent to tip of tail (TaL), maximum body width (BW), body width at eye level (BWE), body width at nostril level (BWN), maximum body height (BH), maximum tail height (TaH), tail muscle height (TMH), rostro-spiracular distance (RSD), frontonasal distance (FN), naso-ocular distance (NO), extraorbital distance (EO), interorbital distance (IO), extranarial distance (EN), internarial distance (IN), eye diameter (E), nostril diameter (N), and rostral gap size (RG). Morphometric variation between the 40 larval specimens of Hyla geographica and H. semilineata was assessed by a principal component analysis (PCA) of the variance-covariance matrix of logtransformed values (Neff and Marcus, 1980). A PCA was also produced after Burnaby (1966) correction for size. To test for significance of shape between the two species, a multivariate analysis of covariance (MANCOVA) was undertaken controlling size and developmental Stage. The STATISTICA for Windows (StatSoft, Inc., 1996) computer program was used for the statistical analysis. The external structure of the oral disc was examined in tadpoles between Stage 25 to 42. Seven specimens of each taxa were dissected for observation of the internal oral anatomy under scanning electron microscopy (SEM) and prepared as follows: The dissections were submitted to ultrasonic cleaning for 20 min. They were then fixed in 4% glutaraldehyde for 2 h at room temperature, followed by three 15 min washes with 0.1 M phosphate buffer and post fixed in 1% osmium tetroxide for 2 h room temperature. Three 15-min washes with 0.1 M phosphate buffer followed. Specimens were next dehydrated using 15 min changes of an ethanol series (35%, 50%, 70%, 80%, 95%, 3 x 100%). Specimens were critical point dried in CO2, mounted on aluminum stubs and sputter coated with gold/palladium to 30 nm thickness, using a Hummer VII sputtering system. Internal oral anatomy was examined under a Hitachi s-2300 scanning electron microscope at 15kV and photographed using Polaroid 55 positive/negative film. Terminology of internal surface features follows Wassersug (1976). Tadpoles were cleared and double-stained with Alcian Blue and Alizarin Red following Dingerkus and Uhler (1977). Chondrocranium was observed under a stereomicroscope (LeicaWild M3C) and drawings were made with the aid of a camera lucida. The description of the chondrocranium was based on larvae at Stage 30. Terminology follows de Sa (1988) and Haas (1995).