tadpole types of chinese megophryid frogs (anura: mego- phryidae

Current Zoology 57 (1): 93−100, 2011

Received June 13; 2010; accepted Sept. 10, 2010.

∗ These authors contributed equally to this work. ∗∗ Corresponding author. E-mail: [email protected]

© 2011 Current Zoology

Tadpole types of Chinese megophryid frogs (Anura: Mego-phryidae) and implications for larval evolution

Cheng LI1, 2*, Xian-Guang GUO1*, Yue-Zhao WANG1** 1 Chengdu Institute of Biology, Chinese Academy of Sciences, Chengdu, Sichuan 610041, China 2 Graduate University of the Chinese Academy of Sciences, Beijing 100039, China

Abstract The larval body shapes and oral discs of 30 frog species from the family Megophryidae from China were examined. Using a phylogenetic framework derived from a Bayesian analysis of published mitochondrial cytochrome b and 16S rRNA gene sequences, we deduced a pattern of historical change among megophryid larval forms. These larvae were categorized into four types according to their body shapes and oral discs: (A) Leptobrachiini type, (B) Lalax type, (C) Brachytarsophrys type, and (D) Megophryini type, of which B and C are novel types. Type A is characterized by a typical oral disc with multiple rows of teeth, representing the tadpole type of the most recent common ancestor of the family Megophryidae. Type B has a typical oral disc with reduced tooth rows, an elongated labium, and integumentary glands. Type C has no labial teeth and a smaller umbelliform oral disc. Type D is characterized by a lack of labial teeth, an enlarged umbelliform oral disc, representing the tadpole of the most recent common ancestor of the subfamily Megophryinae. Our analysis supports the hypothesis that the umbelliform oral disc is apomorphic and also reveals the close association between morphology and microhabitat [Current Zoology 57 (1): 93–100, 2011].

Key words Megophryidae, Tadpoles, Lalax type, Brachytarsophrys type, Novel tadpole types

The family Megophryidae is the largest and most di-verse family in the Archebatrachia, and species occur throughout southern and southeastern Asia (Dubois and Ohler, 1998; Frost, 2010). Currently, the family includes 149 species and 10 genera, which have been grouped into two subfamilies, Megophryinae and Leptobrachii-nae (Frost, 2010). The subfamily Megophryinae has tadpoles with umbelliform oral discs and reduced and/or weak keratinized beaks (Dubois and Ohler, 1998; Frost et al., 2006). The subfamily Leptobrachiinae has tad-poles with typical oral discs of an anteroventral orienta-tion and robust keratinized horny beaks (Dubois and Ohler, 1998; Frost et al., 2006). Liu and Hu (1961) first suggested that umbelliform oral discs originated from typical oral discs. Dubois and Ohler (1998) considered umbelliform oral discs (Megophryinae type) as apo-morphic and unmodified oral discs (Leptobrachiinae type) as plesiomorphic without providing any cladistic analysis.

The two types of tadpoles and their mouthparts relate well to different water layers, diet and feeding mecha-nisms. Huang et al. (1991) examined the tadpoles of 17 Chinese megophryids and found that the buccopharyn-geal structures of the tadpoles were correlated with the

shapes of oral discs. Furthermore, Huang et al. (1991) noted that the tadpoles of Megophryinae occur near the water surface of small streams and are filter feeders that consume plankton and organic debris floating near the water surface. In contrast, Leptobrachiinae tadpoles inhabit the bottoms of streams and graze on epiphyton or major detritus and organic matter on the substrate. This implies that ecological divergence might have played an important role in the diversification and ra-diation of megophryid tadpoles.

Phylogenetic analysis provides a solid foundation for inferring character evolution (Maddison, 1994). For instance, historical changes in sexual dimorphism and reproductive behavior of Vibrissaphora have been ex-amined within a phylogenetic framework (Rao and Wil-kinson, 2008), and multiple origins of reversed sexual size dimorphism were revealed based on a phylogeny of high-elevation Tibetan megophryid frogs (Fu et al., 2007). In this study, we examine the evolutionary his-tory of tadpole morphology of the family Megophryidae within a phylogenetic framework. Our objectives were (1) to describe the divergence of larval body shapes and oral discs in the family Megophryidae, (2) to test whether the umbelliform oral disc is apomorphic, and (3)

94 Current Zoology Vol. 57 No. 1

to determine the evolutionary trends of larval structure in this family. We compare larval morphology of 30 Chinese megophryid species and construct a phylogeny of the species with DNA sequence data using a parti-tioned Bayesian analysis. Furthermore, hypothetical changes of larval morphology are inferred using the parsimony principle.

1 Materials and Methods 1.1 Morphological data

Tadpoles of 30 Chinese megophryid frog species from nine genera were examined. Ten specimens of each species were studied, except Xenophrys boettgeri, for which only two specimens were available. All mate-

rial was collected from 1957 to 2004 and preserved in the Museum of Herpetology, Chengdu Institute of Bi-ology, Chinese Academy of Sciences, Chengdu, China. Specimens were fixed and preserved in 10% formalin (Table 1). Most tadpoles examined for morphology were at developmental stage 26 – 41 (Gosner, 1960). Seven qualitative larval morphological characters were re-corded: oral disc orientation, oral disc shape, oral disc papillae, labial tooth row formula, jaw sheaths, eye po-sition and body shape. Eight quantitative characters were measured with a digital caliper: total length (TOT), tip of snout to tip of tail; head-body length (HBL), tip of snout to posterior-most extent of body; head-body width (HBW), maximum width of head-body; head-body

Table 1 Tadpole information

Species Locality Field No. Total length (mm) Stages

Atympanophrys shapingensis Yuexi, Sichuan 65I0196, 65I0388 39.8±1.7 (n=10) 36–41

Brachytarsophrys platyparietus Jiulong, Sichuan 80I0329 49.9±1.5 (n =10) 38–39

Leptobrachium chapaense Jingdong, Yunnan 581550, 583636 74.9±5.1 (n =10) 36–40

Leptobrachium hainanense Wuzhi Shan, Hainan 64III0415 67.3±5.0 (n =10) 33–39

Leptolalax oshanensis Wushan, Chongqing 571847 53.7±3.9 (n =10) 26–38

Leptolalax pelodytoides Wuyishan, Fujian 64I1083 50.8±3.9 (n =10) 33–36

Ophryophryne microstoma Longzhou, Guangxi 600984, 603099 34.8±1.3 (n =10) 35–38

Oreolalax lichuanensis Lichuan, Hubei 74I0714 66.4±3.1 (n =10) 31–35

Oreolalax major Wenchuan, Sichuan 2L0009 66.7±4.9 (n =10) 36–39

Oreolalax pingii Zhaojue, Sichuan 65I0301 64.2±2.8 (n =10) 36–38

Oreolalax popei Baoxing, Sichuan 74I002 79.5±3.6 (n =10) 37–38

Oreolalax puxiongensis Yuexi, Sichuan 65II0189 46.2±3.2 (n =10) 30–32

Oreolalax rhodostigmatus Lichuan, Hubei 79I0075–2 106.3±11.5 (n =10) 26–37

Oreolalax rugosus Lijiang, Yunnan 3LW0012 53.2±9.6 (n =10) 29–39

Oreolalax schmidti Emei Shan, Sichuan 561526 48.7±2. 4 (n =10) 35–37

Scutiger boulengeri Markam, Sichuan 73I1389 60.8±4.0 (n =10) 37–39

Scutiger glandulatus Kangding, Sichuan 73I1149 62.0±2.9 (n =10) 37–39

Scutiger mammatus Zhongdian, Yunnan 3LW0046, 47, 48 46.5±7.2 (n =10) 28–39

Scutiger nyingchiensis Nyingchi, Tibet 73I0511 61.8±2.5 (n =10) 35–37

Scutiger sikimmensis Yadong, Tibet 73I0249 45.0±3.8 (n =10) 32–38

Vibrissaphora boringii Emei Shan, Sichuan 575303 83.9±3.2 (n =10) 36–38

Vibrissaphora leishanensis Leishan, Guizhou 63II0202 96.1±6.3 (n =10) 37–40

Vibrissaphora liui Wuyishan, Fujian 64I3870 88.2±8.0 (n =10) 34–38

Vibrissaphora yaoshanensis Dayao Shan, Guangxi 602228 89.1±5.9 (n =10) 27–39

Xenophrys boettgeri Wuyishan, Fujian 64I3795 42.3±6.4 (n =2) 26, 31

Xenophrys kuatunensis Wuyishan, Fujian 64I1085 41.3±1.1 (n =10) 37–39

Xenophrys jingdongensis Jingdong, Yunnan 98II2010 33.2±2.4 (n =10) 32–37

Xenophrys minor Emei Shan, Sichuan 595260 40.8±2.0 (n =10) 26

Xenophrys nankiangensis Nanjiang, Sichuan 610489B 36.5±2.0 (n =10) 28–30

Xenophrys omeimontis Emei Shan, Sichuan 570287 44.3±1.0 (n =8) 27–30

LI C et al.: Tadpole types of Chinese megophryid 95

height (HBH), maximum height of head-body; tail length (TL), juncture of midline of caudal muscle with body to tip of tail; tail height (TH), maximum height of tail; diameter of tail muscle (DTM), maximum width of tail muscle; and oral disc width (ODW). Definition of measurements followed Liu and Hu (1961) and Zhao et al. (1994). Larval morphologies and ecomorphological guilds followed Altig and Johnston (1989). 1.2 DNA sequence data

Sequence data of two mtDNA gene fragments, cyto-chrome b and 16S rRNA genes, for 23 species examined in this study were retrieved from GenBank (Table 2).

Sequences of two additional species, Vibrissaphora yaoshanensis and Xenophrys jingdongensis, were pro-vided by J. Jiang (unpublished data). Pelodytes cau-casicus was selected as the outgroup based on a recent study (García-París et al., 2003). The alignment of cy-tochrome b data was straightforward because of the conserved amino acid transcription frames. Initial alignment of the 16S rRNA gene sequences was per-formed with Clustal X1.83 (Thompson et al., 1997) and default parameters. The alignment was then refined based on published secondary structural information (Gutell and Fox, 1998).

Table 2 DNA data GenBank accession numbers

Species cyt b 16S rRNA References

Atympanophrys shapingensis AY561310 AY526203 Zheng et al. (2004a, b)

Brachytarsophrys platyparietus AY561311 AY526206 Zheng et al. (2004a, b)

Leptobrachium chapaense N/A AF285188 Zheng et al. (2004a)

Leptobrachium hainanense EU180927 EU180885 Rao and Wilkinson (2008)

Leptolalax oshanensis AY561313 AY561306 Zheng et al. (2004a, b)

Leptolalax pelodytoides AY236764 AY236797 Zheng et al. (2004a, b)

Ophryophryne microstoma AY561318 AY561309 Zheng et al. (2004 a, b)

Oreolalax lichuanensis AY561321 AY526214 Zheng et al. (2004a, b)

Oreolalax major N/A EF397252 Fu et al. (2007)

Oreolalax pingii N/A EF397259 Fu et al. (2007)

Oreolalax popei AY561319 AY526210 Zheng et al. (2004a, b)

Oreolalax rhodostigmatus AY561320 AY526212 Zheng et al. (2004a, b)

Oreolalax rugosus N/A EF397254 Fu et al. (2007)

Oreolalax schmidti N/A EF397257 Fu et al. (2007)

Scutiger boulengeri AY561322 AY526201 Zheng et al. (2004a, b)

Scutiger glandulatus N/A EF397275 Fu et al. (2007)

Scutiger mammatus N/A EF397279 Fu et al. (2007)

Vibrissaphora boringii AY561324 AY526207 Zheng et al. (2004a, b)

Vibrissaphora leishanensis N/A EF397247 Fu et al. (2007)

Vibrissaphora liui AY561325 AY526208 Zheng et al. (2004a, b)

Vibrissaphora yaoshanensis N/A Jiang (unpubl.data)

Xenophrys jingdongensis N/A Jiang ( unpubl.data)

Xenophrys minor N/A AY561308 Zheng et al. (2004a)

Xenophrys nankiangensis AY561316 AY526200 Zheng et al. (2004a, b)

Xenophrys omeimontis AY561314 AY561307 Zheng et al. (2004a, b)

Pelobates syriacus AY236773 AY236807 García-París et al. (2003)

Pelobates varaldii AY236776 AY236810 García-París et al. (2003)

Pelodytes cacasicus AY236777 AY236811 García-París et al. (2003)


1.3 Bayesian phylogenetic analysis As noted by Nylander et al. (2004) and Brandley et al.

(2005), the use of partitioned Bayesian analyses has facilitated the exploration of partition-specific evolu-tionary models, and should reduce systematic errors and result in more accurate posterior probability estimates. Thus, we reconstructed the phylogeny of 25 megophryid frogs by a partitioned Bayesian analysis implemented with MrBayes version 3.1 (Ronquist and Hulelsenbeck, 2003). MrBayes treats gaps and missing characters as missing data. Thus, gaps and missing characters will not contribute any phylogenetic information. For the Bayesian analysis, the dataset was partitioned into two sets according to the two genes for separate evaluations in the program MrModeltest 2.2 (Nylander1, 2004). The GTR+I+G model of base substitution was selected for both partitions using Akaike information criterion (AIC; Akaike, 1974). Four Monte Carlo Markov Chains were run for 10 000 000 generations and trees were sampled every 500 generations. After two million generations, the average split frequencies started to approach zero, indicative of convergence. We therefore designated the first 4 000 sample trees as “burn-in” and used the last 32 000 trees for constructing the consensus tree and esti-mating Bayesian posterior probabilities (BPP). 1.4 Ancestral state reconstruction

We examined historical changes in larval morphol-ogy by mapping tadpole types onto the resulting phy-logeny using PAUP* 4.0b10 (Swofford, 2002). The evolutionary scenario of the tadpole was reconstructed using the ACCTRAN algorithm under the reversible parsimony principle (Ronquist, 1994).

2 Results A total of 292 tadpoles were examined. These tad-

poles were classified into four types, according to their oral and buccal structures, body forms and habitats. 2.1 Type A: Leptobrachiini type

This type corresponds to the previously described bottom dweller type. Species from the genera Lepto-brachium, Oreolalax, Scutiger and Vibrissaphora share this type of tadpole, and V. boringii is shown in Fig. 1. The oral disc is large, anteroventral in general, and the ODW/HBW ratio ranges from 0.440 to 0.591 (median = 0.490) (Table 3). Marginal papillae are complete, small, and numerous except for a narrow median gap on the upper lip. Submarginal papillae are always with numer-ous labial teeth. Tooth rows are multiple and long, with

a labial tooth row formula of > 4/4. Distal anterior labial ridge and tooth row on each lip are undivided and much shorter than adjacent divided rows. Jaw sheaths are ro-bust with hypertrophied serrations. The TOT is the larg-est among the tadpole types and ranges from 46.2 to 106.3 mm (median = 68.6 mm). The eyes are positioned dorsally. The snout is round in lateral view. The belly is flattened slightly. The HBH/HBW ratio ranges from 0.742 to 0.905 (median = 0.826), and the HBL/TL ratio ranges from 0.511 to 0.733 (median = 0.587). The spir-cular tube is long with a free distal end. The tail muscle is the most massive among the tadpole types, and the DTM/TL ratio ranges from 0.125 to 0.207 (median = 0.146). The fin may be low or high; its forepart or mid-part is the highest. The tip of the tail may be pointed or round (Fig. 1). 2.2 Type B: Lalax type

This is a novel type. Species of the Chinese genus Leptolalax have this type of tadpole, and L. oshanensis is shown in Fig. 2. The oral disc is subterminal, and the ODW/HBW ratio ranges from 0.453 to 0.485 (median = 0.469) (Table 3). Marginal papillae are complete, small, and numerous except for a narrow median gap on the upper lip. Submarginal papillae are large, mostly fewer than ten, and without labial teeth. Tooth rows are re-duced and short, labial tooth row formula of < 5/4. The distal anterior labial ridge and tooth row on each lip are undivided and almost equal to adjacent divided rows. The large areas of the oral disc are without tooth rows. Jaw sheaths are robust with serrations. The TOT is moderate, ranging from 50.8 to 53.7 mm (median = 52.3 mm). The eyes are dorsolateral. The snout is conical in lateral view. Integumentary glands are present and posi-tioned dorsolaterally. The buccopharyngeal area is slightly elongate, and the belly is flattened. The HBH/HBW ratio ranges from 0.692 to 0.781 (median = 0.736), and the HBL/TL ratio ranges from 0.461 to 0.544 (median = 0.503). The spircular tube is long with a free distal end. The tail muscle is massive; and the DTM/TL ratio ranges from 0.117 to 0.153 (median = 0.135). The fin is low. The forepart of the fin is the highest, and the tip of the fin is pointed (Fig. 2). 2.3 Type C: Brachytarsophrys type

Another novel type. Only species of the genus Brachytarsophrys have this type of tadpole, and B. platyparietus is shown in Fig. 3. The oral disc is um-belliform, but its direction is antero-upward. The oral disc is smaller than that of other Megophryiinae, and the

1 Nylander JA, 2004. MrModeltest v2, Program distributed by the author, Evolutionary Biology Centre, Uppsala University.


Table 3 Comparative morphologies of four larva types in the Megophryidae

Characters Leptobrachiini type A (19 species)

Lalax type B (2 species)

Brachytarsophrys type C (1 species)

Megophryini type D (8 species)

TOT (mm) Range: 46.2-106.3 Median: 68.8±17.8

Range: 50.8-53.7 Median: 52.3±2.1 49.9 Range: 32.2-44.3

Median: 39.1±3.9

HBH/HBW Range: 0.742-0.905 Median: 0.826±0.04

Range: 0.692-0.781 Median:0.736±0.06 0.791 Range: 0.824-0.877

Median: 0.848±0.02

ODW/HBW Range: 0.440-0.591 Median: 0.490±0.04


Median: 0.835±0.11

HBL/TL Range: 0.511-0.733 Median: 0.587±0.06


Median: 0.407±0.04

DTM/TL Range: 0.125-0.207 Median: 0.146±0.02


Median: 0.110±0.01

Oral disc orientation Anteroventral Subterminal Antero-upward Dorsal

Oral disc shape Typical oral disc Typical oral disc Small umbelliform oral disc Large umbelliform oral disc

Marginal papillae Complete, small, and numerous Complete, small, and numerous Absent Absent

Submarginal papillae Numerous and always with labial teeth Few and without labial teeth Numerous and without

labial teeth Numerous and without labial teeth

Labial tooth row formula (LTRF)

Tooth rows are multiple and long, LTRF > 4/4

Tooth rows are reduced and short, LTRF < 5/4 Absent Absent

Jaw sheaths Robust with hypertrophied serrations

Robust with hypertrophied serrations

Thin and small with small serrations

Thin and small with small serrations

Eye position Dorsal Dorsolateral Dorsolateral Lateral

Body shape Belly is flattened slightly The buccopharyngeal area is slightly elongate, and the bellyis flattened

The buccopharyngeal area is slightly elongate, and the belly is flattened

Trunk is cylindrical, the buccopharyngeal area is noticeably elongate

Integumentary glands Absent Present Absent Absent

Tail muscle Most massive Massive Massive Slender

Fig. 1 Oral disc (scale bar = 1 mm) and body form (scale bar = 10 mm) of Vibrissaphora boringii tadpole (CIB 575303) TOT = 81.6 mm, ODW = 8.0 mm. ODW/HBW ratio is 0.588 (Table 3). Marginal papillae are absent, submarginal papillae are long and numerous. Jaw sheaths are thin and small with small serrations. Tooth rows are absent. The TOT is moderate, 49.9 mm, and the HBL/TL ratio is 0.478. The eyes are positioned

Fig. 2 Oral disc (scale bar = 1 mm) and body form (scale bar = 10 mm) of Leptolalax oshanensis tadpole (CIB 571847) TOT = 52.4 mm, ODW = 4.1 mm dorsalaterally. The snout is round in lateral view. The buccopharyngeal area is slightly elongate, and the belly is flattened. The HBH/HBW ratio is 0.791. The tail muscle is massive, and the DTM/TL ratio is 0.135. The fin is low. The middle-part of the fin is the highest; the


tip of the fin is pointed (Fig. 3). Tadpoles often found among interstices of small rocks.

Fig. 3 Oral disc (scale bar = 1 mm) and body form (scale bar = 10 mm) of Brachytarsophrys platyparietus tadpole (CIB 80I0329) TOT = 51.6 mm, ODW = 5.1 mm

2.4 Type D: Megophryini type

Species of the genera Atympanophrys, Ophryophryne, and Xenophrys share this type of tadpole, and X. boett-geri is shown in Fig. 4. These tadpoles inhabit rivulets, floating right below the water surface. The oral disc is enlarged and umbelliform with a dorsal orientation. The ODW is the widest among the tadpole types, and the ODW/HBW ratio ranges from 0.722 to 0.957 (median = 0.835) (Table 3). Marginal papillae are absent, submar-ginal papillae are fingerlike, long, and numerous. Jaw sheaths are thin and small with small serrations. The TOT is the smallest among the four tadpole types, ranging from 33.2 to 44.3 mm (median = 39.1mm). The eyes are positioned laterally. The buccopharyngeal area is noticeably elongate. Snout is conical in lateral view. Trunk is cylindrical, and the HBH/HBW ratio ranges from 0.824 to 0.877 (median = 0.848). Body is strongly depressed, and the HBL/TL ratio ranges from 0.384 to 0.486 (median = 0.407). The tail muscle is the thinnest of the four tadpole types, and the DTM/TL ratio ranges from 0.098 to 0.130 (median = 0.110). The tail is long and slender with a low fin. The terminal part of the fin is the highest, the tip of the fin is round (Fig. 4). 2.5 Phylogenetic analysis

The combined 16S and cyt b sequences produced a dataset with 28 taxa and 727 characters. Among the characters, 369 were variable and 310 were phyloge-netically informative. Fig. 5 presents the phylogenetic hypothesis from the partitioned Bayesian analysis, and most nodes received strong support with BPP higher than 0.95. All species were grouped into two clades; the

Megophryinae clade included Atympanophrys, Brachy-tarsophrys, Ophryophryne, and Xenophrys, and the Leptobrachiinae clade included Leptobrachium, Lepto-lalax, Vibrissaphora, Scutiger, and Oreolalax. Within the former clade, the genus Xenophrys was paraphyletic with genera Atympanophrys, Brachytarsophrys and Ophryophryne. In the latter clade, the genera Oreolalax and Scutiger formed a monophyly, albeit receiving only weak statistal support (BPP=0.5), and Vibrissaphora is most closely related to Leptobrachium, albeit with weak support (BPP=0.57). Leptolalax was the sister taxon to the clade consisting of Leptobrachium, Oreolalax, Scu-tiger and Vibrissaphora (BPP=1.0).

Fig. 4 Oral disc (scale bar = 1 mm) and body form (scale bar = 10 mm) of Xenophrys boettgeri tadpole (CIB 64I3795) TOT = 46.8 mm, ODW = 5.5 mm 2.6 Evolution of larvae

Fig. 5 illustrates the Bayesian tree with tadpole types and ancestral state mapped also. The state of most re-cent common ancestor (MRCA) of megophryids was ambiguous: type A or B. In both scenarios, three evolu-tionary changes were required to explain the data. We further resolved ambiguity using branch length informa-tion, i.e., a change is more likely to occur on a longer branch than a short one. Accordingly, type A is most likely to be the ancestral type. If type D represents the tadpole type of the MRCA of the subfamily Megophry-inae, one evolutionary change is needed (type D→C) to explain the data. However, if type C is plesimorphic, you will need two evolutionary changes to explain the data. Following the parsimony principle, the MRCA of megophryids was most likely type A tadpoles. Type D represents the tadpole type of the MRCA of the sub-family Megophryinae. Both type B and D are likely derived type A, and type C is a specialized form of type D (Fig. 5).


Fig. 5 50% major rule consensus tree for megophryid toads obtained from partitioned Bayesian analysis based on mito-chondrial 16S rRNA and cytochrome b gene fragments Numbers above lines or beside the nodes are Bayesian posterior probabilities. The result of a parsimony-based ancestral character state reconstruc-tion is mapped on the tree, showing the evolution of four tadpole types.

3 Discussion

Two types of tadpoles described here, the Lalax type and the Brachytarsophrys type, are novel in the family. Historically, the Lalax type was included in the Lepto-brachiinae type and the Brachytarsophrys type was in-cluded in the Megophryinae type. However, even though the oral disc of the Lalax type is similar to that of Leptobrachiinae, its body shape is similar to that of Megophryinae. In contrast, the oral disc of Brachytarso-phrys type is similar to that of Megophryinae, its body shape and eye position is similar to that of Leptobrachiinae.

Our phylogenetic analysis suggests that type A is an-cestral and both type B and D are derived type A. Fur-thermore, type C evolved from type D. From this phy-logenetic perspective, several hypotheses can thus be formulated. First, the umbelliform oral disc is apomor-

phic in the larval evolution of megophryids, which is consistent with the hypothesis proposed by Liu and Hu (1961) and Dubois and Ohler (1998). Second, the slen-der bodies and expanded lips of type B are likely de-rived from spheroidal bodies and numerous rows of labial teeth on unexpanded lips of type A. Inger (1983) first noticed these two evolutionary trends and our study supports his hypothesis. Interestingly, the slender bodies of type B and D are independently evolved (parallel evolution), and the body shape and eye position of type C are likely the result of a reversal from type D.

Considering these evolutionary scenarios, the larval evolution of megopharyids is likely driven by ecological divergence. Liu and Hu (1961) first recognized that the Chinese megophryid tadpoles live in three distinctive habitats, and their mouth shape is closely associated with habitat. Altig and Johnston (1989) also suggested


that the microhabitat that tadpoles occupy reflects the degree of adhesion and oral complexity because tad-poles adhere to substrates using their mouths. The mor-phological variation in megophryid tadpoles demon-strated a progressive adaptation to a changing habitat from fast to slow moving water. Within fast moving water, the typical oral disc with multiple tooth rows of type A tadpoles are correlated with lotic-suctorial, ben-thic feeders with an anteroventral oral disc and large body. In slower moving water, the lotic-adherent feeders of type B tadpoles have tube-shaped labium with re-duced tooth rows. Furthermore, the enlarged umbelli-form oral disc in type D tadpoles which inhabit micro-habitats of nearly still water pools of rivers, are indica-tive of adaptive traits of lotic-neustonic surface feeders. Although type C tadpoles have a smaller umbelliform oral disc and are normally considered lotic-neustonic surface feeders, an important difference between type C and D tadpoles should be noted. Larvae of type C prefer the interstices of bottom gravel in small streams, which is similar to the habitats of type B tadpoles; they shuttle using their massive muscular tail among the bottom of interstitial spaces. Conversely, type D tadpoles cling to the water surface using their large umbelliform oral disc and hang vertically. In summary, the interplay between megophryid tadpoles and their microhabitats closely follows the prediction of Altig and Johnston (1989).

Acknowledgments We are grateful to Jian Li (Chengdu Institute of Biology, CAS) for his kind help with drawing tad-poles. We thank J. Fu of Guelph University, J. P. Jiang and J. T. Li of Chengdu Institute of Biology for their constructive comments. This research was supported by the Knowledge Innovation Program of the Chinese Academy of Sciences (KSCX2-YW-Z-0905, 08B3021) and the National Natural Science Foundation of China (30470252, 30700062).

References Akaike H, 1974. A new look at the statistical model identification.

IEEE T. Automat. Contr. 19: 716–723. Altig R, Johnston GJ, 1989. Guilds of anuran larvae: Relation-

ships among development modes, morphologies, and habitats. Herp. Monogr. 3: 81–109.

Brandley MC, Schmitz A, Reeder TW, 2005. Partitioned Bayesian analyses, partition choice, and the phylogenetic relationships of scincid lizards. Syst. Biol. 54: 373–390.

Dubois A, Ohler A, 1998. A new species of Leptobrachium (Vi-brissaphora) from northern Vietnam, with a review of the tax-onomy of the genus Leptobrachium (Pelobatidae, Megophryi-nae). Dumerilia 4: 1–32.

Frost DR, 2010. Amphibian species of the World: An online ref-erence. V5.4 (8 April, 2010). Electronic database available at http://research.amnh.org/vz/herpetology/amphibia/ American

Museum of natural History, New York, USA. Frost DR, Grant T, Faivovich J, Bain RH, Haas A et al., 2006. The

amphibian tree of life. B. Am. Mus. Nat. Hist. 297: 1–370. Fu J, Weadick CJ, Bi K, 2007. A phylogeny of the high-elevation

Tibetan megophryid frogs and evidence for the multiple ori-gins of reversed sexual size dimorphism. J. Zool. 273: 315–325.

García-París M, Buchholz DR, Parra-Olea G, 2003. Phylogenetic relationships of Pelabatoidea re-examined using mtDNA. Mol. Phylogenet. Evol. 28: 12–23.

Gosner KL, 1960. A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16: 183–190.

Gutell RR, Fox GE, 1998. A compilation of large subunit RNA sequences presented in a structural format. Nucleic Acaids Res. 16 (Suppl.): r175–r269.

Huang YZ, Fei L, Ye CY, 1991. Studies on internal oral structures of tadpoles of Chinese Pelobatidae. Acta Biol. Plat. Sinica 10: 71–99 (In Chinese with English abstract).

Inger RF, 1983. Larvae of Southeast Asian species of Lepto-brachium and Leptobrachella (Anura: Pelobatidae). In: Rhodin A, Miyata K ed. Advances in Herpetology and Evolutionary Biology. Columbia: University of Missouri Press, 13–32.

Liu CC, Hu SC, 1961. The Tailless Amphibians of China. Beijing: Science Press (In Chinese).

Maddison DR, 1994. Phylogenetic methods for inferring the evo-lutionary history and processes of change in discretely valued characters. Annu. Rev. Entomol. 39: 267–292.

Nylander JA, Ronquist F, Huelsenbeck JP, Nieves-Aldrey JL, 2004. Bayesian phylogenetic analysis of combined data. Syst. Biol. 53: 47–67.

Rao DQ, Wilkinson JA, 2008. Phylogenetic relationships of the mustache toads inferred from mtDNA sequences. Mol. Phy-logenet. Evol. 46: 61–73.

Ronquist F, Huelsenbeck JP, 2003. MrBayes 3: Bayesian phy-logenetic inference under mixed models. Bioinformatics 19: 1572–1574.

Ronquist F, 1994. Ancestral areas and parsimony. Syst. Biol. 43: 267– 274.

Swofford DL, 2002. PAUP*, Phylogenetic Analysis Using Parsi-mony (and other methods), Version 4.10. Champaign, Illinois: Illinois Natural History Survey.

Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG, 1997. The Clustal X windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25: 4876–4882.

Zhao EM, Inger RF, Wu GF, Shaffer HB, 1994. Morphological variation and ecological distribution of co-occuring larva forms of Oreolalax (Anura: Pelobatidae). Amphib-reptil. 15: 109–121.

Zheng YC, Mo BH, Liu ZJ, Zeng XM, 2004a. Phylogenetic rela-tionships of megophryid genera (Anura: Megophryidae) based on partial sequences of mitochondrial 16S rRNA gene. Zool. Res. 25: 205–213 (In Chinese with English abstract) .

Zheng YC, Zeng XM, Yuan YZ, Liu ZJ, 2004b. Phylogenetic position of Ophryophryne and four Leptobrachium group gen-era in Megophryidae (Anura). Sichuan J. Zool. 23: 290–295 (In Chinese with English abstract).

tadpole types of chinese megophryid frogs (anura: mego- phryidae

Documents