Although designating the Night Parrot (Geopsittacus occidentalis) and the Ground Parrot (Pezoporus wallicus) as each other's closest taxonomic relatives is generally accepted, placing this group with respect to other Australo-Pacific parrots has proven problematical. To examine the phylogenetic relationships of these two species, a 924-bp fragment of the cytochrome-b gene was sequenced from single representatives of the following genera: Geopsittacus, Pezoporus, Neophema, Melopsittacus, Platycercus, Polytelis, Strigops, and Calyptorhynchus. Maximum-parsimony, maximum-likelihood, and distance trees all supported a close association between Geopsittacus and Pezoporus. These two genera were also found to be closely linked with Neophema and Melopsittacus. Despite superficial morphological similarities, Geopsittacus and Strigops (Kakapo) were found not to be closely related. Received 21 December 1992, accepted 6 October 1993.

Department of Ornithology, Museum of Victoria, 71 Victoria Crescent, Abbotsford, Victoria 3067, Australia; 2Department of Genetics and Human Variation, LaTrobe University, Bundoora, Victoria 3083, Australia; and 3Division of Vertebrate Zoology, Australian Museum, 6 College Street, Sydney, NSW 2000, Australia THE PSITT^CIFORMES (parrots, lorikeets, and cockatoos) are distributed throughout the Neo- tropical, Ethiopian, Oriental and Australasian regions. Studies based on morphological (Brer- eton 1963, Smith 1975, Homberger 1980), bio- chemical (Christidis et al. 1991a), and chro- mosomal (Christidis 1990, Christidis et al. 1991b) data have shown that most diversity is centered within the Australasian region. Although these studies have identified major groupings among the Australasian psittaciforms, the placement of several genera and species is still poorly re- solved. Two such species are the Night Parrot (Geopsittacus occidentalis) and the Ground Parrot (Pezoporus wallicus). The monotypic genus Geopsittacus often has been considered to be allied to, or congeneric with, the monotypic genus Pezoporus (Brereton 1963, Ford 1969, Schodde and Mason 1980). The paucity of knowledge about the behavior and biology of Geopsittacus has hampered attempts to assess its links with Pezoporus. Until 1990, when a dead specimen was found (Boles et al. 1991, 1994), Geopsittacus had not been reliably documented since 1912. It is represented in mu- seum collections by only 22 skins (excluding the 1990 specimen) and some bones (Forshaw et al. 1976). Schodde and Mason (1981) regarded Geopsittacus and Pezoporus as closely related vi- cariants, replacing each other in inland and wet coastal habitats of southern Australia, respec- tively. There is some similarity in the structure of their preferred low scrubby/heathland hab- itats. These two parrots share similar green plumage crossed with bands of black and yel- low. Both are primarily terrestrial. Geopsittacus receives its common name from its nocturnal habits, and Pezoporus is essentially crepuscular (Forshaw 1981). Superficially, Geopsittacus and Pezoporus re- semble diminutive versions of the flightless New Zealand Kakapo (Strigops habroptilus), a fact com- mented on by Gould (1861). Like the two Aus- tralian taxa, Strigops is nocturnal and terrestrial, and has at least superficially similar green-barred plumage. Furthermore, Strigops and Geopsittacus possess enlarged fleshy ceres ringed by fine "hairlike" feathers. These are pronounced in the former, but small in the latter. Forshaw (1981, 1989) proposed that Geopsittacus, Pezoporus, and Strigops were closely related and constituted an ancient relict group. He also suggested that the Budgerigar (Melopsittacus undulatus) and Neo- phema linked this group with the "typical" pla- tycercines or broad-tailed parrots. In the strict- est sense, the broad-tailed parrot assemblage comprises the following Australian genera: Platycercus, Barnardius, Northiella, Purpureice- phalus, Lathamus, Psephotus, Pezoporus, Geopsit- tacus, and Neophema. Cyanoramphus from New Zealand and the South Pacific is generally in- cluded (Forshaw 1989), and some authors also include other South Pacific genera such as Eu- nymphicus (Smith 1975, Hornberger 1980) and Prosopeia (Hornberger 1980). Melopsittacus shares with Pezoporus, Geopsittacus, and Strigops a yel- low-green plumage marked with dark barring on the dotsurn. Brereton (1963) placed Geopsit- tacus and Pezoporus with Melopsittacus in their own family, the Pezoporidae. Biochemical stud- ies (Christidis et al. 1991a) do not support a close association between Melopsittacus and Pezopo- rus. Smith (1975) included Melopsittacus, Geop- sittacus, and Pezoporus within the platycercines and kept Strigops in its own monotypic tribe. Several authors (e.g. Forshaw 1989, Sibley and Monroe 1990) have placed Geopsittacus, Pezo- porus, and Strigops adjacent to each other in lin- ear sequence, at least implying a relationship. In an effort to resolve the systematic affinities of Geopsittacus and Pezoporus, we collected se- quence data for a 924-base-pair fragment of the mitochondrial cytochrome-b gene from eight psittaciform genera, including Strigops and Mel- opsittacus. MATERIALS AND METHODS The following specimens were examined (with ac- cession number, year of collection, and DNA source, in parentheses): Geopsittacus occidentalis, Night Parrot (36246, 1876, feather); Pezoporus wallicus wallicus, Ground Parrot (R8794, 1924, feather); Strigops habrop- tilis, Kakapo (R2402, 1907, feather); Neophema petro- phila petrophila, Rock Parrot (MV254, 1989, frozen liv- er); Melopsittacus undulatus, Budgerigar (MV1058, 1992, frozen liver); Platycercus icterotis xanthogenys, Western Rosella (MV347, 1989, frozen liver); Polytelis antho- peplus westralis, Regent Parrot (MV340, 1989, frozen liver); and Calyptorhynchus banksii graptogyne, Red- tailed Black-Cockatoo (MV1616, 1989, feather). Spec- imens are lodged in the Museum of Victoria, Mel- bourne, Australia. The choice of an outgroup was compounded by the lack of conclusive evidence con- cerning the nearest relatives of Psittaciformes (sum- marized in Sibley and Ahlquist 1990). Consequently, two outgroups were chosen. A gruiform, Grus rubi- cunda (the Brolga; MV790, 1991, frozen liver) was sequenced. In addition, published cytochrome-b se- quences from the passeriform Pitta sordida (the Hood- ed Pitta; Edwards et al. 1991) were used. The cyto- chrome-b sequences collected in the present study have been deposited in GenBank with the accession numbers U13620 to U13628. Where feather tips were used, DNA extraction was performed as described in Leeton et al. (1993). For frozen liver samples, total DNA was extracted by grinding 0.I to 0.3 g of tissue in 500 1 of digestion solution (50 mM Tris HC1, 10 mM EDTA, I00 mM NaC1, pH 8.0) to which I0 1 proteinase K (10 g/ml) was added. Digestion was performed at 37øC for 2 h. DNA was then extracted through two washes with equal volumes of phenol :chloroform (1:I), and pre- cipitated following the addition of 2.5 volumes of ice- cold absolute ethanol. After drying, the DNA was resuspended in I00/d TE (10 mM Tris, 0. I mM EDTA, pH 8.0). The cytochrome-b sequence was amplified as a se- ries of small segments (300-400 bp) using the primer pairs LI4841/HI5149 (Kocher et al. 1989), L15114/ H15547, and LI5424/HI5767 (Edwards et al. 1991). Letters refer to light and heavy strands and numbers to positions of the 3' nucleotides in the human mtDNA sequence (Anderson et aL 1981). Double-stranded PCR amplifications were performed in a 50 1 reaction mix (25 /M of each appropriate primer, 2.5 mM dNTPs [Boehringer], i unit of Tth DNA polymerase [Toyobo], adjusted to appropriate specifications using I0 x Tth buffer [Toyobo]). Thirty-five rounds of amplification were performed, with each cycle consisting of de- naturation at 92øC for 60 s, reannealing at 52øC for 60 s, and extension at 72øC for 60 s. The final product was cleaned and concentrated using a microcentri- fuge filter tube (Millipore), and resuspended in 50 1 of TE (pH 8.0). Asymmetric PCRs were carried out under similar conditions, in a reaction volume of I00 1 using I 1 of double-stranded PCR product and 0.25 g of the appropriate primer (the addition of a limiting primer was not deemed necessary, due to small amounts of "carry over" following filtration). The final asym- metric PCR product was purified and concentrated following ethanol precipitation. Asymmetric prod- ucts were sequenced directly by the protocol of Sang- er et al. (1977) using a commercial kit (Sequenase, United States Biochemical). Sequences were obtained for the entire light strand and a portion of the heavy strand. There was complete congruence between the heavy- and light-strand sequences. DNA sequences were aligned relative to the human mtDNA sequence (Anderson et al. 1981). No base insertions or deletions were detected. Sequence di- vergence, transition: transversion ratios, and evolu- tionary weightings were calculated from the data set. Following Brown et al. (1982), the transition:trans- version ratio calculated for the least-diverged species pairs was used for all the taxa examined because the saturation of transitions for distantly related com- parisons obscures the real transition: transversion bias. Maximum-parsimony trees were constructed using the exhaustive algorithm in PAUP version 3. I. 1 (Swof- ford 1993). Initial analyses were based on all the char- acters without weighting (global parsimony). A series TABLE 1. Sequence divergence measured between species pairs examined over the 924-bp cytochrome-b fragment. Above diagonal, Kimura (1980) two-parameter distance; below diagonal, frequency of sequence divergence. Taxon 1 2 3 4 5 6 7 8 9 10 1 Geopsittacus -- 0.0903 0.1139 0.1735 0.1105 0.1348 0.1770 0.2340 0.2140 0.2266 2 Pezoporus 0.0840 -- 0.1346 0.1682 0.1163 0.1297 0.1779 0.2378 0.2048 0.2312 3 Melopsittacus 0.1050 0.1220 -- 0.1363 0.1101 0.1065 0.1587 0.2164 0.2076 0.2136 4 Strigops 0.1540 0.1490 0.1230 -- 0.1586 0.1276 0.1405 0.2470 0.2193 0.2439 5 Neophema 0.1020 0.1070 0.1020 0.1420 -- 0.1162 0.1631 0.2341 0.2164 0.2339 6 Platycercus 0.1220 0.1180 0.0980 0.1160 0.1070 -- 0.1312 0.2129 0.2067 0.2222 7 Polytelis 0.1560 0.1570 0.1410 0.1270 0.1450 0.1190 -- 0.1943 0.2300 0.2574 8 Calyptorhynchus 0.1990 0.2010 0.1860 0.2070 0.1990 0.1830 0.1690 -- 0.2704 0.2475 9 Grus 0.1840 0.1770 0.1800 0.1880 0.1860 0.1800 0.1960 0.2250 -- 0.2002 10 Pitta 0.1930 0.1960 0.1830 0.2060 0.1980 0.1890 0.2140 0.2420 0.1730 -- of conditional data sets was also analyzed: (1) weight- ing for differential evolutionary rates at the three codon positions (weighted parsimony); (2) excluding third positions; (3) transversions only (transversion parsimony). The resolving power of the sequence data was assessed using the bootstrap procedure (Felsen- stein 1985) as implemented in PAUP and the jack- knifing procedure as described by Lanyon (1985). A maximum-likelihood analysis (Felsenstein 1981) using the DNAML option of PHYLIP version 3.5 (Fel- senstein 1993) was performed including 100 bootstrap replicates. The data set was analyzed both with and without evolutionary weightings for the three codon positions. A transition: transversion bias of 4:3, gen- erated from a consideration of all positions in the Pezoporus-Geopsittacus comparison, was included. A distance matrix was constructed using the DNADIST program of PHYLIP employing Kimura's (1980) two-parameter method (Table 1). From the ma- tfix, Fitch-Margoliash (Fitch and Margoliash 1967) and neighbor-joining (Saitou and Nei 1987) trees were generated using the FITCH and NEIGHBOR options of PHYLIP. Bootstrap resampling was implemented for both trees. RESULTS Within the 924-bp region of cytochrome b (Appendix), 301 positions were variable over all psittaciform taxa examined; 59 were at the first-codon position, 56 at the second, and 186 at the third. These converted to evolutionary weightings for the first-, second-, and third- codon positions of 3:3:1. The percent sequence divergence between psittaciform species pairs (Table 1) ranged from 8.4% (Geopsittacus and Pe- zoporus) to 20.7% (Calyptorhynchus and Strigops). Kimura (1980) two-parameter distances, within the psittaciforms, ranged from 0.0903 (Geopsit- tacus and Pezoporus) to 0.2470 (Calyptorhynchus and Strigops), the pattern of divergences being similar to that of the percent differences (Table 1). Average percent divergences and two-pa- rameter distances between the psittaciforms and the outgroups (Grus, Pitta) were 19.6% and 0.2403, respectively. Transition: transversion ratios between spe- cies pairs varied considerably ranging from 1:1 (Calyptorhynchus : Strigops) to 4:3 (Geopsittacus : Pezoporus). The highest transition: transversion ratios occurred between those species that had the lowest percent divergences and Kimura (1980) two-parameter distances. The choice of either Pitta or Grus as the outgroup affected the topology and resolution of some of the analyses and, therefore, these are described separately below. A global-parsimony analysis, without weighting and using Grus as the outgroup, pro- duced two equally most-parsimonious trees of length 668 steps, which differed from one an- other only in the relative positioning of Calyp- torhynchus and Strigops. In one, Polytelis and Ca- lyptorhynchus were sister taxa and, in the other, Strigops and Polytelis were sister taxa. Bootstrap analysis supported only the former topology. In the bootstrap consensus topology (Fig. 1), Geopsittacus and Pezoporus were sister taxa, and these were linked with Neopherna, Melopsittacus, and Platycercus in a stepwise pattern. Bootstrap values at each of these nodes were 94, 77, 72, and 73%, respectively. A consideration of near- parsimonious trees revealed 4 with 671 steps or less and their strict consensus was identical in topology to the above. Weighted parsimony provided less resolution; the Geopsittacus-Pezo- porus-Neophema clade was present, and Polytelis was aligned with Strigops, but with a low boot- _ Geopsittacus Pezoporus 72 [ Neophema 73 I Melopsittacus Platycercus 77 Strigops Polytelis Calyptorhynchus Grus Fig. 1. PAUP maximum-parsimony analysis based on unweighted characters for psittaciform genera ex- amined and using Grus as the outgroup. Length of two equally most-parsimonious trees, before boot- strapping, was 668 and consistency index was 0.669. Bootstrap values greater than 50% indicated at nodes. strap value of 56%. Global parsimony using Pitta as the group produced a single most-parsimo- nious tree with 683 steps (not shown). This to- pology differed from the above (Fig. 1) in that Platycercus was now linked with Strigops, al- though this node was not supported following bootstrapping. The Pezoporus-Geopsittacus clade was supported in all of the bootstrap replicates, while nodes linking Neophema and Melopsitta- cus to this clade were well supported with boot- strap values of 91 and 71%, respectively. The Polytelis-Calyptorhynchus clade was supported by a bootstrap of 78%. A consideration of near- parsimonious trees revealed 7 with 687 steps or less, and their strict consensus was identical in topology to the above. Weighted parsimony, with Pitta as outgroup, differed in that Polytelis was aligned with Strigops, but this node had a low bootstrap value (55%). Following the jack- knife procedure, only the Geopsittacus-Pezopo- rus-Neophema clade was identified in all repli- cates, regardless of choice of outgroup. More- over, in all replicates that included Neophema, both Melopsittacus and Platycercus were aligned with the Geopsittacus-Pezoporus-Neophema clade. When first- and second-codon positions only were considered, with Pitta as the outgroup, Calyptorhynchus was separated from the remain- ing parrots, but only with a bootstrap value of 51%. The only other resolution was a node link- ing Polytelis with Strigops with a low bootstrap value of 62%. When Grus was used as the out- 72 72 -- Geopsittacus Pezoporus Neophema Melopsittacus Platycercus Strigops 93 [ Polytelis  Calyptorhynchus Grus Fig. 2. Maximum-likelihood tree based on un- weighted characters for psittaciform genera examined and using Grus as the outgroup. Log likelihood before bootstrapping was -4355.78900. Bootstrap values greater than 50% indicated at nodes. group, no resolution was obtained following bootstrapping. Transversion parsimony using Pitta as the outgroup (not shown) produced three clades: (1) Geopsittacus-Pezoporus; (2) Neophema-Platy- cercus-Melopsittacus; and (3) Polytelis- Strigops. The first two were sister clades and Calyptorhynchus pulled out as the most divergent lineage. Fol- lowing bootstrapping, however, only the Geop- sittacus-Pezoporus clade (75%) was supported. When Grus was used as the outgroup, Calyptor- hynchus was again the most divergent lineage, Geopsittacus and Pezoporus were linked with Neophema, Melopsittacus and Platycercus, and Polytelis was linked with Strigops. Following bootstrapping, the node separating Calyptor- hynchus from the rest had a bootstrap of 63%; the only other nodes with values higher than 50% were Geopsittacus-Pezoporus (68%) and Stri- gops-Polytelis (55%). The maximum-likelihood tree (Fig. 2), using Grus as the outgroup, had a topology compa- rable to that obtained with global parsimony (Fig. 1). Polytelis and Calyptorhynchus were linked with Strigops, although this node had a boot- strap value of only 40%. All other nodes were supported by bootstrap values of 72% and high- er. A consideration of evolutionary weightings did not alter the topology or resolution of the maximum-likelihood tree. When Pitta was used as the outgroup, the position of Platycercus was not resolved following bootstrapping, but Stri- gops was linked with Polytelis-Calyptorhynchus with a bootstrap of 64%. The Fitch tree, using Grus as the outgroup (not shown), was identical in topology to the maximum-likelihood tree, and there was only one topological difference between the Fitch and neighbor-joining (Fig. 3) trees. In the latter, Strigops was aligned with the Geopsittacus-Platy- cercus clade, and this node was supported by a bootstrap of 51%. When Pitta was used as the outgroup, the neighbor-joining tree was iden- tical in topology to the maximum-likelihood tree, while in the Fitch tree, Platycercus was aligned with the Strigops-Calyptorhynchus-Poly- telis clade and this node was supported by a bootstrap of 51%. DISCUSSION Phylogenetic position of Geopsittacus and Pe- zoporus.--Cytochrome-b data are unequivocal and consistent in their linking of Geopsittacus with Pezoporus. The two genera differ by 8.4% sequence divergence (Table 1), which is com- parable to divergences for other congeneric avi- an species. Edwards et al. (1991) reported values ranging from 6.2 to 12.0% between congeners in Pomatostomus, while Smith et al. (1991) re- corded values of 5.7 to 13.0% within the genus Laniarius. The present data are consistent with the conclusions of Ford (1969) and Schodde and Mason (1980) that Geopsittacus and Pezoporus should be treated as congeneric species. Geop- sittacus occidentalis, therefore, becomes Pezoporus occidentalis. The pattern of distribution and pre- ferred habitats for these species parallels those in the genus Neophema, which also has both inland (bourkii and splendida) and coastal species ( chrysogaster, chrysostoma, pulchella, petrophila, and elegans). In concert with the shared patterns of distri- bution, Geopsittacus and Pezoporus were consis- tently allied with Neophema by the cytochrome-b data. Melopsittacus also was aligned with this assemblage in all analyses. On the basis of mor- phological characters, such as lack of a nape spot and presence of barred feathers, Brereton (1963) linked Melopsittacus with Pezoporus and Geopsit- tacus; however, lack of a nape spot appears to be the ancestral condition within the Psittaci- formes. Only the other Australian platycercine genera and Cyanoramphus possess a nape spot. Although Smith (1975) listed Eunymphicus as possessing a nape spot, our examination found that it in fact lacks one. If the Geopsittacus-Mel- opsittacus assemblage, as identified in our study, 100 80 64 92 0.052 0.050 0.049 0.075 78 O.015 0.132 0.138 Geopsittacus  Pezoporus Neophema Melopsittacus Platycercus Strigops P01ytelis  Calyptorhynchus Grus Fig. 3. Neighbor-joining tree based on the Kimura (1980) two-parameter distances for psittaciform gen- era examined, using Grus as the outgroup. Bootstrap values greater than 50% indicated above nodes. Num- bers below nodes refer to branch lengths. is part of the platycercines, then two scenarios are possible to explain the occurrence of a nape spot in the various genera. In the first (Fig. 4A), the nape spot evolved early in the evolution of the platycercines and has since been lost twice, in Melopsittacus and in Geopsittacus-Pezoporus. The alternative scenario (Fig. 4B) is that the nape spot evolved in the "core" platycercines (sensu Christidis et al. 1991a) after their divergence from the lineage, giving rise to the Geopsittacus- Melopsittacus assemblage, and also evolved in- dependently in Neophema. There are no data at present to favor one hypothesis over the other. The distribution of barred feathers within the platycercines also has two possible explana- tions. Barred feathers may have arisen early in the evolution of the Geopsittacus-Melopsittacus assemblage and then been lost secondarily in Neophema (Fig. 4A). Alternatively, barred feath- ers may have evolved independently in both Melopsittacus and Geopsittacus-Pezoporus (Fig. 4B). To differentiate between these two hypotheses, we examined in detail the barred feathers in Pezoporus, Geopsittacus, Strigops, Melopsittacus, and Psittacella. Other than their darker green color, the barred feathers of Pezoporus are very similar to those of Geopsittacus, agreeing in details of pattern on each part of the body. The most ob- vious difference is the barring of underpart feathers, which in Pezoporus extends to the bor- ders of the webs, whereas in Geopsittacus it does not. Melopsittacus does not exhibit this same agreement of feather patterns. On no part of the body does its pattern of barring appear ho- mologous with that in Pezoporus-Geopsittacus. a -- Geopsittacus Pezoporus , Neophema Melopsittacus "typical" platycercines b i I -- Geopsittacus Pezoporus Neophema , Melopsittacus "typical" platycercines Fig. 4. Possible pathways for evolution of nape spots and barred feathers in platycercines with (A) character loss, or (B) characters arising more than once. Closed rectangles represent gain of nape spot; open rectangles represent loss of nape spot. Closed circles represent gain of barred feathers; open circles rep- resent loss of barred feathers. This adds support to the idea that barred pat- terns evolved independently in each group. The barred feathers of Strigops show even more sub- stantial differences from those of Geopsittacus- Pezoporus, and are more likely to have been in- dependently derived. New Guinean Psittacella is similar to Geopsittacus and Pezoporus in its un- derparts and to Melopsittacus in its upperparts. It was shown to be linked with the platycercines by Christidis et al. (1991a), but was not treated by Brereton (1963) or included in this study. The relationship of Psittacella to these other genera needs to be assessed and further studies should include it. Neophema, Melopsittacus, and Geopsittacus share a Type A-1 carotid formula (no data are avail- able for Pezoporus), whereas the other platycer- cines (including Cyanoramphus) are Type A-2 (Garrod 1873, 1874, 1876, Glenny 1957). Apart from Nestor, Psittrichas, and Prosopeia (which are Type A-2), other Australasian and Asian genera examined have a Type A-1 carotid formula. Bed- dard (1898) listed Eunymphicus (which he called Nymphicus cornutus) as possessing a Type A-2 carotid formula, citing Garrod (1873, 1874,1876) as the source. This appears to be in error, as Garrod (1873, 1874, 1876) did not examine Eu- nymphicus. Similarly, Smith (1975) listed Psit- tacella as possessing a Type A-1 carotid formula, but his cited sources (i.e. Garrod 1874, Beddard 1898, Glenny 1957) did not mention Psittacella. The African Psittacini and all South American genera examined are Type A-2. It can be argued that the Type A-1 carotid formula is ancestral for the Australasian genera and that a Type A-2 formula evolved in the "core" platycercines af- ter this lineage diverged from the one leading to the Geopsittacus-Melopsittacus assemblage. Based on 924 bp of cytochrome-b sequences, no association was apparent between Strigops and either Geopsittacus or Pezoporus (cf. Forshaw 1989). Thus, the similarities in appearance be- tween Strigops and Geopsittacus are convergent, presumably brought about by similar life hab- its. Predominantly green plumage crossed with bands of black and yellow may serve as cam- ouflage for these ground-dwelling species, which shelter in tussocks or clumps of vege- tation by day. The "hairlike" feathers present in both Geopsittacus and Strigops have also evolved convergently, possibly acting as tactile sensors when the birds are moving through thick ground cover. Deeper relationships within Psittaciformes.--The platycercine assemblage is generally consid- ered to include Neophema, Pezoporus, and Mel- opsittacus (Smith 1975, Hornberger 1980). Al- though the qualitative and distance analyses given above generally aligned Platycercus with the Pezoporus-Melopsittacus clade, bootstrap analysis revealed that the position of Platycercus could not be resolved confidently by the cyto- chrome-b sequences. Moreover, resolution of the position of Platycercus was affected by the choice of outgroup. To resolve this inconsistency re- garding the position of Platycercus, sequence data are required from other typical platycercines such as Northiella, Psephotus, and Barnardius. Such studies are currently in progress. If one assumes that the Melopsittacus to Pezoporus clade is part of the platycercine assemblage, it probably rep- resents a highly divergent group of taxa. Chris- tidis et al. (1991a), based on allozyme studies, argued for two successive radiations among the platycercines. The first involved the divergence of Pezoporus (and presumably Geopsittacus, which was not included in that study), Neophema, and Melopsittacus from the core platycercine group. The second, more recent one, gave rise to Platy- cercus, Barnardius, Purpuricephalus, Northiella, La- thamus, and Psephotus. Christidis et al. (1991a) also reported a close association between the polytelitines and platycercines, but this was not apparent in the present study. Although Poly- telis was aligned with Calyptorhynchus in most analyses, the relatively high sequence diver- gence between the two (17.1%) suggests that an examination of additional taxa is required to accurately assess their relationships. Moreover, bootstrapping and jackknifing indicated that the relationships of Polytelis, Strigops, and Calyptor- hynchus could not be resolved from the present data set. The patterns of resolution obtained in this study indicate that the cytochrome-b gene is particularly useful in determining relation- ships between closely related genera, but that for deeper branches, transition bias needs to be considered. The cockatoos, represented in our study by Calyptorhynchus, are generally consid- ered to be the sister group to a large assemblage that includes all the Australian genera exam- ined here (Smith 1975, Homberger 1980, Chris- tidis et al. 1991a, b). Only when the transition bias was taken into consideration was such a relationship apparent from the present se- quence data. Moreover, the levels of sequence divergence between Calyptorhynchus and the other psittaciforms examined were of similar magnitude to those recorded between the psit- taciforms and the two outgroups Grus and Pitta, indicating that the level of transition diver- gence had reached saturation in those compar- isons. ACKNOWLEDGMENTS We thank the Department of Conservation and Land Management of Western Australia for permission to collect materiM for this study. We also thank Belinda Gillies for assistance in the field and Toula Galari- niotis (Museum of Victoria) for typing the various drafts of the manuscript. This work was partly funded by an Australian Research Council Grant (A19031161) to L.C. and M.W., and by an Australian Geographic Grant to W.E.B. LITERATURE CITED ANDERSON, S., A. T. BANKIER, B. G. BARP,LL, M. H. L. BRUIJN, A. R. COULSON, J. DROUIN, I. C. EPERON, D. P. NIERLICH, B. A. ROE, F. SANGER, P. H. SCHREIRER, A. J. H. SMITH, R. STADEN, AND I. G. YOUNG. 1981. Sequence and organization of the human mitochondrial genome. Nature 290:457- 465. BEDDARD, F.E. 1898. The structure and classification of birds. Longmans Green, London. BOLES, W. E., N. W. LONGMORE, AND M. C. THOMPSON. 1991. The fly-by-night parrot. Aust. Nat. Hist. 9: 689-695. BOLES, W. E., N. W. LONGMORE, AND M. C. THOMPSON. 1994. A recent specimen of the Night Parrot, Geopsittacus occidentalis. Emu 94:37-40. BRERETON, J. LE GAY. 1963. Evolution within the Psit- taciformes. Pages 411-501 in Proceedings XIII In- ternational Ornithological Congress (C. G. Sib- ley, Ed.). Ithaca, New York, 1962. American Ornithologists' Union, Washington, D.C. BROWN, W. M., E. M. PRAGER, A. WANG, AND A. C. WInSON. 1982. Mitochondrial DNA sequences of primates: Tempo and mode of evolution. J. Mol. Evol. 18:225-239. CHRISTIDIS, L. 1990. Animalcytogenetics--Aves. Ge- briider Borntraeger, Berlin. CHRISTIDIS, L., R. SCHODDE, D. D. SHAW, AND S. F. MAYNES. 1991a. Relationships among the Aus- tralo-Papuan parrots, lorikeets, and cockatoos (Aves: Psittaciformes): Protein evidence. Condor 93:302-317. CHRISTIDIS, L., D. D. SHAW, AND R. SCHODDE. 1991b. Chromosomal evolution in parrots, lorikeets and cockatoos (Aves: Psittaciformes). Hereditas 114: 47-56. EDWARDS, S. V., P. ARCTANDER, AND A. C. WILSON. 1991. Mitochondrial resolution of a deep branch in the genealogical tree for perching birds. Proc. R. Soc. Lond. B. Biol. Sci. 243:99-107. FELSENSTEIN, J. 1981. Evolutionary trees from DNA sequences: A maximum likelihood approach. J. Mol. Evol. 17:368-376. FELSENSTEIN, J. 1985. Confidence limits on phylog- enies: An approach using the bootstrap. Evolu- tion 39:783-791. FELSENSTEIN, J. 1993. PHYLIP, manual version 3.5. Department of Genetics, Univ. Washington, Se- attle. FITCH, W. M., AND E. MARGOLIASH. 1967. Construc- tion of phylogenetic trees. Science 155:279-284. FORD, J. 1969. Distribution and taxonomic notes of some parrots of Western Australia. S. Aust. Or- nithol. 25:99-105. FORSHAW, J. M. 1981. Australian parrots, 2nd rev. ed. Lansdowne, Melbourne. FORSHAw, J.M. 1989. Parrots of the world, 3rd rev. ed. Lansdowne, Willoughby, Sydney. FORSHAW, J. M., P. J. FULLGAR, AND J. I. HARMS. 1976. Specimens of the Night Parrot in museums throughout the world. Emu 76:120-126. GARROD, A.H. 1873. On the carotid arteries of birds. Proc. Zool. Soc. Lond. 1873:457-472. GARROD, A.H. 1874. On some points in the anatomy of the parrots which bear on the classification of the suborder. Proc. Zool. Soc. Lond. 1874:586- 598. GARROD, A.H. 1876. Notes on the anatomy of certain parrots. Proc. Zool. Soc. Lond. 1876:691-692. GLENNY, F. H. 1957. A revised classification of the Psittaciformes based on the carotid artery ar- rangement patterns. Ann. Zool. Agra 2:47-56. GOULD, J. 1861. On a new genus and species of par- akeet from Western Australia. Proc. Zool. Soc. Lond. 1861:100-101. HOMIERGER, D. G. 1980. Funktionell-Morpholo- gische Untersuchungen zur Radiation der Er- nahrungs-und Trinkmethoden der Papageien (Psittaci). Bonn. Zool. Monogr. 13:1-192. KIMURA, M. 1980. A simple method for estimating evolutionary rate of base substitution through comparative studies of nucleotide sequences. J. Mol. Evol. 116:111-120. KOCHER, T. D., W. K. THOMAS, A. MEYF, S. V. ED- WARDS, S. PAnO, F. X. VILLA BLANC, A, AND A. C. WILSON. 1989. DynamicsofmitochondrialDNA evolution in animals: Amplification and sequenc- ing with conserved primers. Proc. Natl. Acad. Sci. USA 86:6196-6200. LAN3/4ON, S.M. 1985. Detecting internal inconsisten- cies in distance data. Syst. Zool. 34:397-403. LEETON, P. R., L. CHRISTIDIS, AND M. WESTFMAN. 1993. Feathers from museum skins--A good source of DNA for phylogenetic studies. Condor 95:465- 466. SAITOU, N., AND M. N. 1987. The neighbour-join- ing method: A new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406-425. SANGER, F., S. NICKLEN, AND A. R. COULSON. 1977. DNA sequencing with chain-terminating inhib- itors. Proc. Natl. Acad. Sci. USA 74:5463-5467. SCHODDE, R., AND I. J. M^SON. 1980. Nocturnal birds of Australia. Lansdowne, Melbourne. SIBLEY, C. G., AND J. E. AHLQUIST. 1990. Phylogeny and classification of birds. Yale Univ. Press, New Haven. SIBLEY, C. G., AND B. L. MONROE, JR. 1990. Distri- bution and taxonomy of birds of the world. Yale Univ. Press, New Haven. SMrrH, G. A. 1975. Systematics of parrots. Ibis 131: 18-68. SMITH, E. F. G., P. ARCTANDER, J. FJELDS., AND O. G. AMIR. 1991. A new species of shrike (Laniidae: Laniarius) from Sorealia, verified by DNA se- quence data from the only known individual. Ibis 133:227-235. SWOFFORD, D.L. 1993. PAUP manual version 3.1:1. Illinois Natural History Survey, Champaign. 6 : : : : :6;; %o  ...... oo ......... IE .........  r .........  r ......... O ......... O O ......... O O ......... o ......... t0, ......... -, ......... o ......... i r ......... I r ......... 0E ......... 4 O ......... E E-* ......... o .........  O ß .r . r -  O ...... E * ß 0E ......... 4 O ......... I, O ......... o * -E ......  O ...... EE ß ; O -EE,,E,, - ,E .........  E ......... m E-* ......... i,E ......... H  ......... ,- O ...... E - - 40 - - .....   O ........ E - O ......... r .........  O .........  r ......... ;, ........ r ; O, .... E ß ß  OE ...... E - E .........  E ......... N' ...... E ß - _r . -E ...... - O ......... E E ......... r ......... O ......... N, ......... ,- rE ........ 0E - - ß . . - E E ......... r r ......... IO ...... ,,  O ...... E - -  O, - - -,E, - ,O ß - ...... ,- E ......... 1 E ......... E, .........  , .........  r ......... eO -E - .r ....  O ...... E ß ß  O ....  * ß *  ......... Nr ......... , EO ........ i ......... g ......... 'g : : : :.,4 :d : : o .........  E .........  O ..... rr ß ß 40 ......... , ......... 4 O ......... 1/2=r ...... E ß ß 1/22 E ......... 1/22 O ......... E, ......... I, E ......... E , ......... IO .........  O ......... 1  ....... O - N, ......... , E .........  O ......... EE ......... I, E .........  r ......... O .........  O ....... E - -O  ..... E - - - ß ,4, ......... , E .........  E ......... O ......... I, E .........   ..... OO ß ß iE -O,O,,O,, i  -O - - .r . . . ; O ..... EE - - ,O - -E -E ....  O .........  E ......... E, ....... r . 0 E ......... ,- O ......... 10 ..... E, - - 10 - - ....  ß  O ......... r  : : : : : : : :  : : : : : : : : : o  : : :.4 : : :::::::: ß  ....    :,-; :1/2 :,-;   : : : : : : : :d  : : : : : : :d  : : : : : : : : : '1/2 : : : : : : : : : No ::d::::  ......... . : : : : : : : : :  ......... ......... ! !... ! ! !2!! ......... =, ....  .   : : : : : :,,,4 : :  iq i i ! i i!  .........  ......... ß o ooiiiiiiiii  : ::::::  .........  . -4 '  :dd : : :ddd o :::: : :  ......... ,",o   : : : : : : : : : ......... - : : : : : - : : ,. : : : : : : : : : ff ......... '-' : : : : : : : :  : : : : : : : : : ø : : : : : : : : : ,.-.,o : :o..o.o.. :õo.. 1/2-',..R ' ' ' ' '   ......... ='  =' ......... too o ......... ' ...... . o : : : : : :,-; : : N : : : : : :d ,-, o : : : : :.,,4.,,4.,,4 =  :.,,4.,,4 :.,,46 : .4 r ......... 1E .........  E ......... r .........  r ......... ,- O .........  : : : : : : : : ,-,   : :d : :d :d o ø . .. ....  o ø ........  No ......  ß o ......... Nr .........  r .........  ......... o. ......... oo ......... r. ......... , O ......... N' ......... rr ......... E , .........  E ......... ß E ß .o - -ooo -  r . .E .... EE i O ......... ,E ......... f  ......... N' ......... I,E .........   ß .o -oo -oo E E ......... 842 LEETON ET ^L. [Auk, Vol. 111 October 1994] Parrot Affinities 843