A Systematic Revision of the New Zealand Gekkonidae
ln this species I recognise 21 species of New Zealand brown geckos (Hoplodactylus), and 7 species of green gecko (Naultinus), bringing the total New Zealand gecko fauna to 28. The only previous comprehensive revision of the New Zealand Gekkonidae was that of McOann (1955), who recognised 4 species of brown geckos (Hoplodactylus), 6 species of South lsland green geckos (Heteropholts), and one species of North lsland green gecko (Naultinus). Bauer (1990) synonymised Heferopholis with Naulftnus. In this thesis I interpret the results of allozyme electrophoresis and morphological examination of specimens of all known taxa, using phylogenetic computer analysis and an Evolutionary Species Concept. Allozyme results were obtained for 686 specimens from 235 populations. I resolved 27 allozyme loci from each of these specimens, using 2 tissue types and 5 buffer systems. The following species are recognised as a result of this study. Newly proposed species are indented below the species to which they have previously been referred: Hoplodactylus chrysosireticus, H. duvaucelii, H. granulatus, H. nebulosus, H. kahutarae, H. maculatus, H. brunneus, H. "Mount Arthur", H. "Kaikouras", H. "Marlborough mini", H. "Southern Alps", H. "Danseys Pass", H. "Cromwell Gorge", H. "Otago", H. "Southern mini", H. pacificus, H. "Matapia", H. "Poor Knights" H. "Three Kings" H. rakiurae, H. stephensi (H. delcourfi - probably from New Zealand; presumed extinct), Naultinus elegans (subspecies N. e. elegans and N. e. punctatus), N. gemmeus, N. grayii, N. manukanus, N. rudis, N. stellatus, N. tuberculatus. Hoplodacfylus will include 21 extant species. Apart from one synonymy suggested by examination of type specimens, I suggest no formal changes to Naultinus, because morphological support for the status quo is strong, but allozymes identify only 5 groups (2 within N.'gemmeus) rather than the 8 currently recognised. I provide a key for the identification of all known New Zealand species. Phylogenetic analysis of allozyme and morphological results suggests 4 major groups in the New Zealand gecko fauna: Naultinus ; a narrow toed group of Hoplodactylus species, including H. granulatus, H. nebulosus, H. kahutarae, H. rakiurae, and H. stephensi ; the H. pacificus complex, including H. pacificus, H. "Matapia", H. "Poor Knights", and H. "Three Kings" ; the H. maculatus complex, including H. chrysosireticus, H. duvaucelii, H. maculatus, H. brunneus, H. "Mount Arthur", H. "Kaikouras", H. "Marlborough mini", H. "Southem Alps", H. "Danseys Pass", H. "Cromwell Gorge", H. "Otago", and H. "Southem mini". However, Naultinus is closely related to the narrow-toed group of Hoplodactylus, and falls between H. stephensi and the rest of the narrowtoed group of Hoplodactylus in some analyses. Within the H. maculatus complex, there are 3 species groups: H. chrysosireficus and H, "Southem mini" ; H. maculatus. H. "Mount Arthur", H. "Kaikouras", and H. "Marlborough mini" ; H. duvaucelii, H. brunneus, H. "Southem Alps", H. "Danseys Pass", H. "Cromwell Gorge", and H. "Otago". H. pacificus is paraphyletic with respect to the sympatric new species H. "Matapia" in all analyses, and H. granulatus is paraphyletic with respect to H. kahutarae in some analyses. I disagree with many character states Bauer (1990) assigned to the New Zealand species in his morphological data matrix for the Carphodactylini. His resulting cladogram is shown to lack strong support, and alternative phylogenetic, and by inference biogeographical, relationships are possible. Some genetically widely divergent species are so morphologically similar that traditional taxonomies had difficulty distinguishing them (e.9., H. stephensi, all species of the H. pacificus complex, and all species of the H. maculatus complex except H. duvaucelfisubsumed under H. pacificusby McCann 1955). Conversely, some morphologically highly distinct species have genetically close, previously unsuspected sister group relationships. Rates of change in allozymes and morphology therefore appear poorly correlated. In the southern group of the H. maculafus complex, the overall pattem of allozyme variation is a stepped or fragmented cline, which continues across species boundaries. There is a marked absence of diagnostic derived allozyme characters in well-defined species. These allele distributions suggest parapatric speciation, with the shared polymorphisms being older than the speciation event. On Matapia island, the sympatry of 2 morphologically distinct and reproductively isolated species, both of which fall within H. pacificus in trees suggest a double colonisation of the i6land, followed.by morphological divergence. There are at least 5 independent cases of coastal populations of very small H. maculafus-complex geckos with larger inland sister-groups, suggesting strong directional selection on body size. There is strong circumstantial evidence for most evolution and speciation having happened in the South lsland, and for all North lsland species except for the H. pacificus complex being geologicalty recent colonists with South lsland ancestors. The Three Kings and Poor Knights lsland groups are the only islands occupied by New Zealand geckos which are believed not to have been connected to the mainland during Pleistocene glacial periods of low sea level, and also the only islands with genetically distinct local endemics. The recognition of these new species will alter conseruation priorities, as some have very restricted known distributions.