*Corresponding author. Tel.:#81-824-24-7482; fax:#81-824-24-0739. E-mail address:[email protected] (M. Sumida)
Biochemical Systematics and Ecology 28 (2000) 721}736
Evolutionary relationships among 12 species
belonging to three genera of the family
Microhylidae in Papua New Guinea revealed
by allozyme analysis
Masayuki Sumida
!
,
*, Allen Allison
"
, Midori Nishioka
!
!Laboratory for Amphibian Biology, Faculty of Science, Hiroshima University, Higashihiroshima 739-8526, Japan
"Division of Vertebrate Zoology, Bishop Museum, Honolulu, Hawaii 96817, USA
Received 10 September 1998; received in revised form 14 October 1999; accepted 18 October 1999
Abstract
To elucidate the potential of electrophoretic analysis for understanding relationships among microhylid frogs in Papua New Guinea, an allozyme analysis was conducted. A total of 119 individuals from nine species ofCophixalus, two species ofSphenophryneand one species of Barygenys, all of which belong to the family Microhylidae, were studied. Fourteen enzymes extracted from skeletal muscles and livers were analyzed by starch-gel electrophoresis. These enzymes were encoded by genes at 20 loci. There were 2}15 phenotypes produced by 2}12 alleles at these loci. The mean proportion of heterozygous loci per individual, mean proportion of polymorphic loci per population, and mean number of alleles per locus in 12 species were 6.1%, 17.1% and 1.17a on average, respectively. The NJ and ML trees constructed from Nei's genetic distances showed that the genusSphenophrynecan be distinguished biochemically from CophixalusandBarygenys, and that the species groups of Cophixalus, which are similar in
external morphology, can be divided biochemically into several species. ( 2000 Elsevier
Science Ltd. All rights reserved.
Keywords:Allozyme analysis; Anura; Microhylidae; Sphenophryne;Cophixalus;Barygenys; Papua New Guinea
1. Introduction
Microhylid frogs are a diverse group that exhibit considerable ecological and morphological diversity in the rainforests of New Guinea (Zweifel, 1972), comprising about 47% of the species of frogs recognized in the region (Zweifel and Tyler, 1982). The microhylids of New Guinea are arranged in two subfamilies, Asterophryinae and Genyophryninae ("Sphenophryninae) (Zweifel, 1972). In this region, 39 asteroph-ryine species belonging to seven genera (Asterophrys, Barygenys, Hylophorbus,
Pherophapsis, Callulops, XenobatrachusandXenorhina) and 49 named genyophrynine
species belonging to six genera (Choerophryne, Cophixalus, Copiula, Genyophryne,
Oreophryneand Sphenophryne) have been reported (Frost, 1985; Zweifel and Tyler,
1982). The taxonomic diversity is related to the ecological diversity of the family (Burton and Stocks, 1986). Menzies (1976) conveniently divided the family Micro-hylidae into four habitat groups; fossorial, terrestrial, scansorial and arboreal.
Two allozyme analyses of New Guinean frogs have resolved the genetic relation-ships among several species within the generaLitoria(Dessauer et al., 1977) andRana
(Donnellan et al., 1989). Nishioka and Sumida (1990) examined the genetic relation-ships betweenPlatymantis papuensisfrom Papua New Guinea and twoRanaspecies from East Asia by allozyme analysis. The genetic relationships and phylogeny of New Guinean hylid frogs were studied by allozyme analysis using 11 species of the family Hylidae (Sumida et al., 1998). However, allozyme analyses on taxa of New Guinean microhylid frogs have yet to be conducted using several species belonging to di!erent genera.
In the present study, the phylogenetic relationships of 12 species belonging to three genera of the family Microhylidae were investigated by allozyme analysis to assess the potential of electrophoretic analysis for understanding relationships among micro-hylid frogs in Papua New Guinea.
2. Materials and methods
A total of 119 specimens from 12 species belonging to three microhylid genera were used (Table 1, Fig. 1). Fourteen enzymes extracted from muscles and livers were analyzed by horizontal starth-gel electrophoresis (Nishioka et al., 1980,1992) (Table 2). Enzymes were visualized following the method outlined by Harris and Hopkinson (1976). Multiple bu!er systems have not been examined for each enzyme, so the present results probably underestimate the total genetic variation. The genetic distan-ces were calculated following Nei (1972,1975,1987). Two di!erent methods, the NJ method (Saitou and Nei, 1987) and the maximum-likelihood (ML) method (Felsen-stein, 1973), were employed to infer the phylogenetic relationships among taxa on the basis of genetic distances, using the programs included in version 3.5c of PHYLIP (Felsenstein, 1993).
Table 1
Specimens of Papua New Guinean microhylid frogs used in the present study
Species Sample size Body length (mm) Date of collection Individual numbers!
Range Mean
Sphenophryne palmipes 10 30.0}50.4 41.3 Nov. 1982 6510}6512, 6583}6586, 6587*}6589*
Sphenophryne rhododactyla 2 24.4, 60.2 42.3 Nov. 1982 6576*, 6577*
Cophixalus parkeri 5 21.0}29.0 26.6 Nov. 1982 6501}6503, 6575*, 6578*
Cophixalus kaindiensis 12 23.6}30.0 26.2 Nov. 1982, Jan. 1985 6505, 6545}6553, 6666, 6667 Cophixalusvariegatusgroup sp. 1 14 15.0}20.6 18.3 Nov. 1982, Jan. 1985 6513, 6514, 6516}6519, 6538, 6561,
6565*}6567*, 6595, 6610, 6622 Cophixalusvariegatusgroup sp. 2 3 13.0}18.0 15.4 Nov. 1982, Jan. 1985 6515, 6594, 6620
Cophixalusvariegatusgroup sp. 3 2 15.6, 17.2 16.4 Nov. 1982 6563*, 6564*
Cophixalus cryptotympanum 50 19.6}38.2 28.1 Nov. 1982, Jan. 1985 6526, 6527, 6554, 6600}6603, 6607, 6609, 6615}6619, 6621, 6623, 6648}6652, 6655, 6659, 6661}6665, 6668, 6671}6675, 6680}6692, 6778}6780
Cophixalus riparius 10 37.0}53.4 43.3 Nov. 1982 6555}6560, 6579*}6582*
Cophixalus pansus 8 16.5}27.8 21.8 Nov. 1982 6562, 6571*}6574*, 6591}6593
Cophixalus sphagnicola 1 19.8 19.8 Nov. 1982 6570*
Barygenysyavigularis 2 21.8, 22.8 22.3 Jan. 1985 6597, 6641
Total 119
!Asterisked specimens were provided by M. Kuramoto.
Fig. 1. Papua New Guinean microhylid frogs. Scale bars equal 2.5 mm. (A,B)Cophixalusvariegatusgroup sp. 1. (C, D)Cophixalusvariegatusgroup sp. 2.Cophixalusvariegatusgroup sp. 1 and 2 closely resemble each other in external characters, buta are distinct in abdominal color pattern.
Table 2
Enzymes analyzed in the present study
Enzyme Abbreviation E.C.No.! Tissue Bu!er system"
Aspartate aminotransferase AAT 2.6.1.1 Skeletal muscle T}C pH 7.0
Adenylate kinase AK 2.7.4.3 Skeletal muscle T}C pH 7.0
Creatine kinase CK 2.7.3.2 Skeletal muscle T}B}E pH 8.0
a-Glycerophosphate dehydrogenase a-GDH 1.1.1.8 Skeletal muscle T}C pH 6.0 Glucose phosphate isomerase GPI 5.3.1.9 Skeletal muscle T}B}E pH 8.0 Isocitrate dehydrogenase IDH 1.1.1.42 Skeletal muscle T}C pH 7.0 Lactate dehydrogenase LDH 1.1.1.27 Skeletal muscle T}C pH 6.0 Malate dehydrogenase MDH 1.1.1.37 Skeletal muscle T}C pH 6.0
Malic enzymes ME 1.1.1.40 Skeletal muscle T}C pH 7.0
Mannose phosphate isomerase MPI 5.3.1.8 Skeletal muscle T}C pH 7.0
Peptidase PEP 3.4.3.1 Liver T}B}E pH 8.0
Phosphogluconate dehydrogenase PGD 1.1.1.44 Skeletal muscle T}C pH 7.0
Phosphoglucomutase PGM 2.7.5.1 Skeletal muscle T}B}E pH 8.0
Superoxide dismutase SOD 1.15.1.1 Skeletal muscle T}B}E pH 8.0
!Enzyme Commission numbers (Nomenclature Committee of Biochemistry, 1992).
"T}C, Tris-citrate bu!er. T}B}E, Tris-borate-EDTA bu!er.
3. Results
3.1. Electrophoretic patterns and allelomorphs
The electrophoretic patterns of 14 enzymes analyzed are encoded by genes at 20 loci. Electrophoretic bands corresponding to multiple alleles at each locus are named A, B, C, etc., in the order of mobility from fast to slow, and alleles are indicated by a, b, c, etc. The CK and MDH-2 loci are the least variable: two phenotypes are produced by two alleles. The MPI locus is the most polymorphic; 15 phenotypes are produced by 12 alleles. At these 20 loci, there are an average of 8.7 phenotypes produced by an average of 7.0 alleles (Table 3).
3.2. Allele frequencies
Table 4 presents the gene frequencies for each species at 20 loci. At three loci, a-GDH, LDH-A and PEP-D, each genus has one or more alleles not found in either of the other genera. At seven other loci, AK, CK, IDH-1, LDH-B, MDH-1, MDH-2 and PEP-A, alleles in Sphenophryne are not found in Cophixalus and Barygenys
allowing alleles at these loci to identifySphenophryne. Thus, a total of 10 loci can be interpreted as diagnostic forSphenophryne.
3.3. Geneticvariation
Table 3
The numbers and kinds of alleles and phenotypes at 20 enzyme loci in 12 species belonging to three genera of the family Microhylidae respectively (Table 5). There are no distinct di!erences between the actual and expected values, expect for species represented only by one or two specimens.
3.4. Genetic distances
The genetic distances among the 12 species examined are shown in Table 6. The distance is 1.501 between two Sphenophrynespecies. Among nine species of genus
Cophixalus, the genetic distances are 0.553}2.069 (x6"1.284), with the smallest value
obtained betweenC.variegatusgroup sp. 1 andC.variegatusgroup sp. 2 (0.553), and then betweenC. pansusandC. sphagnicola(0.732).Cophixalus cryptotympanumandC.
parkeri show the largest distance value (2.069) within the genus. The intergeneric
Table 4
Gene frequencies at 20 enzyme loci in 12 species belonging to three genera of the family Microhylidae. If single letter, frequency"1.00
Species (Sample
size)
AAT-1 AAT-2 AK CK a-GDH GPI IDH-1 IDH-2
Table 4*continued
Species (Sample
size)
LDH-A LDH-B MDH-1 MDH-2 ME-1 ME-2 MPI
Table 4*continued
(0.08) (0.92) (0.14) (0.86) (0.07) (0.93)
C.variegatusgroup sp. 2 ( 3) c c c d f f
(0.50) (0.50)
C.variegatusgroup sp. 3 ( 2) b f h b f
C. cryptotympanum (50) c c d e g e g f
(0.17) (0.83) (0.21) (0.79) (0.23) (0.77)
Table 5
Genetic variabilities at 20 enzyme loci in 12 species belonging to three genera of the family Microhylidae
Species Sample per individual!(%) per population (%)
Sphenophryne palmipes 10 7.5 ( 5.8) 15.0 1.15
Sphenophryne rhododactyla 2 12.5 ( 8.1) 20.0 1.20
Cophixalus parkeri 5 0 ( 1.6) 5.0 1.05
Cophixalus kaindiensis 12 5.4 ( 7.4) 25.0 1.25
Cophixalusvariegatusgroup sp. 1 14 8.9 (10.3) 35.0 1.35
Cophixalusvariegatusgroup sp. 2 3 3.3 ( 3.9) 10.0 1.10
Cophixalusvariegatusgroup sp. 3 2 0 ( 0) 0 1.00
Cophixalus cryptotympanum 50 8.2 ( 8.0) 35.0 1.35
Cophixalus riparius 10 6.0 ( 8.0) 20.0 1.20
Cophixalus pansus 8 8.8 ( 9.1) 20.0 1.20
Cophixalus sphagnicola 1 5.0 ( 2.5) 5.0 1.05
Barygenysyavigularis 2 7.5 ( 5.6) 15.0 1.15
Average 9.9 6.1 ( 5.9) 17.1 1.17
!Parentheses show an expected value.
Table 6
Genetic distance (D) among 12 species belonging to three genera of the family Microhylidae
Species S. pal S. rho C. par C. kai C.var C.var C.var C. cry C. rip C. pan C. sph
C. cry 2.543 1.791 2.069 1.158 1.275 1.296 1.011
C. rip 2.876 1.323 1.098 0.952 1.252 1.602 1.072 1.193
C. pan 2.022 2.542 1.084 1.271 1.270 1.136 1.232 1.100 1.223
C. sph 2.953 4.327 1.366 1.335 1.017 1.171 1.884 0.958 1.555 0.732
B.ya 2.219 1.478 1.860 1.076 1.642 1.561 1.218 0.803 1.016 1.209 1.507
1.478}2.219 (x6"1.849) between Sphenophryne and Barygenys, and 1.323}4.327
(x6"2.695) betweenSphenophryneandCophixalus.
3.5. Phylogentic relationships
The NJ tree show that the genera examined are largely divided into two groups, one containing the genusSphenophryneand the other containing the generaBarygenysand
Cophixalus(Fig. 2A). The ML tree also show that the genera examined are largely
divided into two groups, one containing two species of the genusSphenophryneand the other containing 10 species belonging to the generaBarygenys and Cophixalus
(Fig. 2B). In contrast to the ML tree, the NJ tree show the genusBarygenysto have a sister-group relationship with the genusCophixalus.
4. Discussion
Microhylid genera have been diagnosed largely on the basis of skeletal morpho-logy. Two genera,CophixalusandSphenophryne, are distinguished by features of the pectoral girdle:Cophixaluslacks the clavicles and procoracoid cartilages exhibited by
Sphenophryne(Zweifel, 1962). Tyler (1971, 1972) showed that supplementary slips of
large (1.323}4.327,x6"2.695). These values roughly correspond to the intergeneric genetic distances between two genera,RanaandPlatymantis, of the subfamily Raninae obtained by Nishioka and Sumida (1990) (3.128}3.200,x6"3.164). On the other hand, the genetic distances between the genusBarygenys of the subfamily Asterophryinae and the generaCophixalus andSphenophryneof the subfamily Genyophryninae are smaller (0.803}2.219, x6"1.337) than those between the latter two genera (1.323}4.327,x6"2.695). These results may imply the inappropriate assignments of the genera to the two subfamilies, or may suggest that allozyme analysis is not a prefer-able technique for elucidating the relationships between these genera or subfamilies. Zweifel and Allison (1982) revealed that Cophixalus pansus di!ered more from typical scansorial Cophixalus chie#y in ways associated with a ground-dwelling as opposed to an arboreal way of life: the complete absence of toe discs and terminal grooves, relatively small eyes, and short legs.Cophixalus pansusmay represent the extreme of adaptation to terrestrial environments seen in theCophixalusevolutionary line. The only other species ofCophixalusthat lacks expanded terminal discs on both "ngers and toes isC. sphagnicola(Green and Simon, 1986; Zweifel and Allison, 1982). Large size, longer"rst"nger and absence of terminal grooves on digits distinguish
C. pansusfromC. sphagnicola. Zweifel and Allison (1982) argued that the similarity of
the two species does not necessarily re#ect their close a$nity, althoughC. sphagnicola
has been compared withC. pansusin determining generic status. They suggested that their common and presumably derived states of characters may represent parallel adaptations to a terrestrial life style. However, the present study showed that the genetic distance between the two species was relatively small (0.732) in comparison with those betweenCophixalus species examined in the present study, and that C.
pansusis closely related toC. sphagnicola. On the other hand, Kuramoto and Allison
(1989) reported thatC. pansusis a tetraploid with 2N"52 chromosomes, and that its haploid set is very similar to that ofC. ripariusbased on a statistical examination of karyotypes. The present study showed that the genetic distance betweenC. pansusand
C. ripariusis 1.223, and that these two species are not closely related genetically.
The present study revealed that theC.variegatusgroup contains three species (sp. 1}3), of which sp. 1 and 2 closely resemble each other in external characters (Fig. 1A,C), but are distinct in abdominal color pattern (Fig. 1B,D). The genetic distance betweenC.variegatusgroup sp. 1 and 2 (0.553) was the smallest one as the interspeci"c value for the genus obtained in this study. Nevertheless, they were distinguished by diagnostic alleles ata-GDH, IDH-2, ME-1, MPI, PEP-A and PGM loci. Thus, they were considered to be independent sibling species. According to karyological observations by Kuramoto and Allison (1989),C.variegatusgroup`sp. 1ais unique in that it has many small pairs with a high arm ratio. Karyotypes of
C.variegatusgroup`sp. 1aandC.variegatusgroup`sp. 2a(C.variegatusgroup sp. 3 in
the present study) di!er remarkably (Kuramoto and Allison, 1989), and they have di!erent acoustic features (Allison, unpublished). The present study showed that
C.variaegatusgroup sp. 3 is distinct formC.variegatusgroup sp. 1 and 2 by alleles at
AAT-2, AK, LDH-B, ME-2 and PEP-D loci. The generic distances between C.
variegatusgroup sp. 3 and the latter two were comparatively large, being 1.429 and
separated from the other two into a di!erent species group. There are many unde-scribed species that very closely resembleC. variegatus(Allison, unpublished data), and the taxonomic revision of this complex requires further study. Burton and Zweifel (1995) thought it appropriate to formalize the recognition of thevariegatusgroup as a genus distinct fromCophixalus, and created a new genus,Albericus, to accommodate three species, including the variegatus group, removed from the genus Cophixalus. However, the present study showed that thevariegatusgroup could not be distin-guished from other species of the genusCophixalus.
Zweifel (1979) described one new species, Cophixalus kaindiensis, found on Mt. Kaindi, Morobe District, Papua New Guinea. This species is virtually identical with its sympatric congenerC. parkeriin size and proportions, although the two taxa di!er slightly in color patterns and greatly in mating calls. The present study showed that the genetic distance betweenC. kaindiensisandC. parkeri(1.237) is larger than that betweenC. ripariusandC. kaindiensis(0.952). Both NJ and ML trees showed thatC.
kaindiensisis closest toC. riparius. These two taxa di!er greatly in size;C. ripariusmay
grow up to 50 mm in body length, which makes it the largest known species in the genus, whereasC. kaindiensisis a small species not growing much over 30 mm in body length.
According to Zweifel and Tyler (1982), two subfamilies, Asterophryinae and Genyophryninae, share direct embryonic development, and thus are considered monophyletic. Kuramoto and Allison (1989), after examining the karyotypes of 15 species of New Guinean microhylid frogs belonging to"ve of these genera from both subfamilies, suggested that the Asterophryinae and Genyophryninae are closely related to each other and are karyologically conservative. Results of the present allozyme analysis, which was the "rst attempt at a molecular genetic approach to Papuan microhylid systematics, do not support the monophyly of the two subfamilies, Genyophryninae and Asterophryinae, due to the relatively close a$nity ofBarygenys
withCophixalus. Monophyly of the generaCophixalusandSphenophryneis supported
by our results. The present study also revealed that nine species of the genus
Cophixalusseem to be remotely related to each other (Fig. 2). This result could suggest
that a common ancestor diverged into several species during a comparatively short period to adapt to the ecologically variable environments, and thereafter considerable parallel evolution of adaptive types may have taken place, leading to great diversity in external appearance as suggested by Zweifel and Allison (1982) and Zweifel and Tyler (1982).
The genetic distances between taxa of various levels has been reviewed mainly in vertebrates (Avise and Aquadro, 1982; Thorpe, 1982; Nei, 1987). According to these studies, the interspeci"c genetic distances vary from 0.22 to 1.60 in the majority of cases of congeneric species. If birds and some mammals are excluded, the interspeci"c genetic distances are about 0.05 or larger and can be as large as 3.00 or more, and intergeneric distances are larger than interspeci"c distances (Nei, 1987). The present study showed that the genetic distances between congeneric species of genus
Cophixaluswere 0.553}2.069 (x6"1.284), and those between three confamilial genera
Fig. 2. NJ tree (A) and ML tree (B) for 12 species of Papua New Guinean microhylid frogs belonging to three genera of the family Microhylidae based on Nei's genetic distances. Scale bars represent branch length in terms of Nei's distances for the NJ and ML trees.
distances than those of any non-amphibian vertebrates shown by Nei (1987) and Avise and Aquadro (1982). These results may imply that genera and families in the di!erent classes of vertebrates are not equivalent in levels of structural gene divergences as suggested by Avise and Aquadro (1982).
Acknowledgements
supported by a Grant-in-Aid for Overseas Scienti"c Survey from the Ministry of Education, Science and Culture, Japan (No. 59041025).
References
Avise, J.C., Aquadro, C.F., 1982. A comparative summary of genetic distances in the vertebrates. Evol. Biol. 15, 151}185.
Burton, T.C., 1984. A new character to distinguish the Australian microhylid generaCophixalusand Sphenophryne. J. Herpetol. 18, 205}207.
Burton, T.C., Stocks, R., 1986. A new species of terrestrial microhylid frog from Papua New Guinea. Trans. Roy. Soc. S. Aust. 110, 155}158.
Burton, T.C., Zweifel, R.G., 1995. A new genus of genyophrynine microhylid frogs from New Guinea. Am. Mus. Novit. 3129, 1}7.
Dessauer, H.C., Gartside, D.F., Zweifel, R.G., 1977. Protein electrophoresis and the systematics of some New Guinea hylid frogs (genusLitoria). Syst. Zool. 26, 426}436.
Donnellan, S.C., Adams, M., Aplin, K.P., 1989. A biochemical and morphological study ofRana(Anura: Ranidae) from the Chimbu Province, Papua New Guniea. Herpetologica 45, 336}343.
Felsenstein, J., 1973. Maximum likelihood and minimum-steps methods for estimating evolutionary trees from data on discrete characters. Syst. Zool. 22, 240}249.
Felsenstein, J., 1993. PHYLIP (Phylogeny Inference Package), Ver. 3.5c. University of Washington, Seattle. Frost, D.R., 1985. Amphibian Species of the World: A Taxonomic and Geographical Reference. Allen Press,
Lawrence, Kansas.
Green, D.M., Simon, M.P., 1986. Digital microstructure in ecologically diverse sympatric microhylid frogs, generaCophixalusandSphenophryne(Amphibia: Anura), from Papua New Guinea. Aust. J. Zool. 34, 135}145.
Harris, H., Hopkinson, D.A., 1976. Handbook of Enzyme Electrophoresis in Human Genetics. North-Holland, Amsterdam.
Kuramoto, M., Allison, A., 1989. Karyotypes of microhylid frogs of Papua New Guinea and their systematic implications. Herpetologica 45, 250}259.
Menzies, J.I., 1976. Handbook of Common New Guinea Frogs. Wau Ecology Institute, Papua New Guinea.
Nei, M., 1972. Genetic distance between populations. Amer. Natur. 106, 283}292. Nei, M., 1975. Molecular Population Genetics and Evolution. North-Holland, Amsterdam. Nei, M., 1987. Molecular Evolutionary Genetics. Columbia University Press, New York.
Nishioka, M., Ohtani, H., Sumida, M., 1980. Detection of chromosomes bearing the loci for seven kinds of proteins in Japanese pond frogs. Sci. Rep. Lab. Amphibian Biol., Hiroshima Univ. 4, 127}184. Nishioka, M., Sumida, M., 1990. Di!erentiation ofRana limoncharisand two allied species elucidated by
electrophoretic analyses. Sci. Rep. Lab. Amphibian Biol., Hiroshima Univ. 10, 125}154.
Nishioka, M., Sumida, M., Ohtani, H., 1992. Di!erentiation of 70 populations in theRana nigromaculata group by the method of electrophoretic analyses. Sci. Rep. Lab. Amphibian Biol., Hiroshima Univ. 11, 1}70.
Saitou, N., Nei, M., 1987. The neighbor-joining method: A new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406}425.
Sumida, M., Allison, A., Nishioka, M., 1988. Genetic relationships and phylogeny of Papua New Guinean hylid frogs elucidated by allozyme analysis. Jpn. J. Hepetol. 17, 164}174.
Thorpe, J.P., 1982. The molecular clock hypothesis: Biochemical evolution, genetic di!erentiation and systematics. Ann. Rev. Ecol. Syst. 13, 139}168.
Tyler, M.J., 1971. The phylogenetic signi"cance of vocal sac structure in hylid frogs. Univ. Kansas Publ. Mus. Nat. Hist. 19, 319}360.
Tyler, M.J., 1972. Super"cial mandibular musculature, vocal sacs and the phylogeny of Australo-Papuan leptodactylid frogs. Rec. S. Aust. Mus. 16, 1}20.
Zweifel, R.G., 1962. A systematic review of the microhylid frogs of Australia. Am. Mus. Novit. 2113, 1}40. Zweifel, R.G., 1972. Results of the Archbold Expeditions 97. A revision of the frogs of the subfamily
Asterophryinae, family Microhylidae. Bull. Am. Mus. Nat. Hist. 148, 411}546.
Zweifel, R.G., 1979. A new cryptic species of microhylid frog (genusCophixalus) from Papua New Guinea, with notes on related forms. Am. Mus. Novit. 2678, 1}14.
Zweifel, R.G., Allison, A., 1982. A new montane microhylid frog from Papua New Guinea, and comments on the status of the genus Aphantophryne. Am. Mus. Novit. 2723, 1}14.
Zweifel, R.G., Tyler, M.J., 1982. Amphibia of New Guinea. In: Gressitt, J.L. (Ed.), Biogeography and Ecology of New Guinea. W. Junk, The Hague, pp. 759}801.
