Journal of East African Natural History 94(1): 191–201 (2005) GENETIC AND MORPHOLOGICAL EVIDENCE FOR TWO SPECIES IN THE UDZUNGWA FOREST PARTRIDGE XENOPERDIX UDZUNGWENSIS Rauri C.K. Bowie Evolutionary Genomics Group, Department of Botany and Zoology, Stellenbosch University Private Bag X1, Matieland 7602, South Africa bowie@sun.ac.za Jon Fjeldså Zoological Museum, University of Copenhagen, Universitetsparken 15 Copenhagen, Denmark jfjeldsaa@zmuc.ku.dk ABSTRACT The Udzungwa forest partridge, Xenoperdix udzungwensis is known from only three forests within the Udzungwa and Rubeho Highlands of the Eastern Arc Mountains. Given the phenotypic differences between the Udzungwa (X. u. udzungwensis) and Rubeho Highland (X. u. obscurata) populations, it seems unlikely that there remains recurrent gene flow between these two populations. We used a combination of mitochondrial (1041 bp of NADH Dehydrogenase subunit 2) and nuclear DNA (569 bp of Fibrinogen intron 5 and 387 bp of Glyceraldehyde-3-phosphate Dehydrogenase intron 11) markers to investigate the degree to which these two taxa are separated. In mtDNA, 0.5% sequence divergence, with five-fixed mutational differences (two amino acid changes) was recovered between X. u. udzungwensis and X. u. obscurata. One fixed difference was found for Fib5 and none for Gadph11. Coalescent models suggest that no gene flow is taking place between the Udzungwa and Rubeho Highlands and that divergence between the two taxa took place about 200 000 years before present. The presence of fixed mutational differences in mtDNA and in one of the two nDNA markers analysed, the lack of gene flow, and diagnosable morphological differences (including potential display signals) between X. u. udzungwensis and X. u. obscurata, suggest that X. u. obscurata be accorded full species status, Xenoperdix obscurata, for which we put forth the common name ‘Rubeho forest partridge’. INTRODUCTION The Udzungwa forest partridge, Xenoperdix udzungwensis Dinesen, Lehmborg, Svendsen, Hansen & Fjeldså, 1994, is one of Africa's rarest birds, considered threatened ("Vulnerable", see BirdLife International, 2000) because of its very small range. The species is known from only three forests within the Udzungwa and Rubeho Highlands of the Eastern 192 R.C.K. Bowie & J. Fjeldså Arc Mountains in Tanzania (figure 1). Whereas population numbers may be stable as these forest patches are relatively pristine, any indication of decline should result in uplisting to "Endangered" species status (BirdLife International, 2000). Figure 1. Distribution and collecting sites of Xenoperdix u. udzungwensis and X. u. obscurata populations sequenced in this study; black areas are remaining evergreen forest tracts. C = Chugu Hill in the Mafwemiro Forest, Lu = Luhombero Forest, Nd = Ndundulu Mts, Ny = Nyumbanitu Mts. This species has drawn special attention because it is a relict form, whose discovery in 1991 (Dinesen et al., 1994) has been characterised as one of the major surprises for the ornithological world in recent years (McGowan, 1994). Its nearest relatives are primitive forest partridges (Arborophila, Rollulus and others) of the Oriental Region (Dinesen et al., 1994; Crowe et al., in press). Thus, Xenoperdix likely diverged from its nearest living relatives in the early Miocene when the Tethys Sea closed and there was a brief opportunity for oriental forest birds to colonise Africa. Consequently, the survival of small relict populations in the Eastern Arc Mountains is remarkable. This, together with the concentration of other relict taxa in the same mountains (Burgess et al., 1998; Burgess et al., in press), lends support to studies that suggest that as a consequence of the highly predictable ecological conditions that these mountains experience from nearly continuous orographic rain, exceptionally low extinction rates occur (Fjeldså & Lovett, 1997; Jetz et al., 2004). Within the Udzungwa Highlands, Xenoperdix is known only from elevations above 1300 m in the Nyumbanitu (7°44'–50'S, 36°19'–26'E; 55 km2) and Ndundulu/ Luhombero forests (7°36'–49'S, 36°25'–42'E; 240 km2) (figure 1; Dinesen et al., 1994; Butynski & Ehardt, 2003), with an estimated population of 3700 individuals (Dinesen et al., 2001). Surprisingly, a small population described as a subspecies, X. udzungwensis obscurata Fjeldså & Kuire 2003, was recently found in the northern part (Chugu Hill) of the Mafwemiro (Mafwomero) Forest (6°49'S, 36°35'E), 100 km farther north in the Rubeho Highlands (Fjeldså & Kiure, 2003). This population most likely only numbers a few hundred individuals. In spite of extensive surveys of other forest tracts in the Udzungwa, Uvidunda, Rubeho and Wota Evidence for two species in Udzungwa forest partridge 193 Highlands in central Tanzania (Fjeldså et al., in press), no other populations have been found. The species may, therefore be relictual even within its present range (figure 1). Given the phenotypic differences between the Udzungwa and Rubeho Highland populations of Xenoperdix (table 1; see also Fjeldså & Kuire, 2003), as well as the presence of relatively dry, low-lying, unsuitable habitat in the region between these montane highlands, it seems unlikely that there remains recurrent gene flow between these two populations. This suggests that these two populations should be managed as separate taxonomic entities. In this study, we include tissue samples from all seven specimens collected since the species was first discovered in 1991. We use a combination of mitochondrial and nuclear DNA markers to establish whether gene flow is taking place between these two isolated populations, estimate genetic diversity, and evaluate any signal in the data for recent demographic change (population expansion versus decline). Finally, we comment on the taxonomic status of X. udzungwensis obscurata restricted to the Rubeho Highlands. Table 1. Description of morphological differences between Xenoperdix u. udzungwensis and X. u. obscurata. Key plumage differences are depicted in Fjeldså and Kuire (2003). Sexes are similar based on specimens examined: X. u. udzungwensis, 2 adult females, 1 adult male and 1 alcohol preserved specimen which was not sexed. X. u. obscurata: 3 specimens not sexed, 1 is almost certainly a female based on the presence of a vascularised brood-patch. Character Wing-length Maximum tail-length Face X. u. udzungwensis 137.5–149.0 mm 68.0–73.0 mm Mostly orange-brown Neck Necklace of mostly white feathers with variable black spots Under-tail coverts Ochraceous Secondaries Wing-coverts Prominent barring Black and ochraceous barring Retrices broad 15–20 mm X. u. obscurata 130.8–137.0 mm 57.5–61.8 mm Orange-brown with dense dusky speckles No necklace, but rather an arc of black spots overlaid on the olivegrey boundary between throat and breast Only a faint trace of ochre towards feather tips Less distinct barring Distinctive grey to whitish distal margins giving a scaly effect Narrow, 11–14 mm, and the bars of the central feathers less distinctly black METHODS Sequence data were obtained from all seven specimens of Xenoperdix collected since the species was discovered in 1991 (table 2). This consisted of four X. u. udzungwensis individuals sampled from two different localities within the Udzungwa Highlands and three individuals of X. u. obscurata sampled from Mafwemiro Forest at Chugu Hill in the Rubeho Highlands. The hill partridge Arborophila javanica Gmelin 1789 was used as an outgroup. The vouchers, including the type specimens of both forms are deposited at the Zoological Museum of the University of Copenhagen, Denmark, except for one specimen, which is in the Institute of Zoology and Marine Biology, University of Dar es Salaam, Tanzania. Taxon X. u. udzungwensis X. u. obscurata Arborophila javanica Voucher ZMUC114233 ZMUC114249 ZMUC114257 ZMUC115139 ZMUC129213 ZMUC129233 ZMUC129234 ZMUC JBK01-290103 Locality Ndundulu Mts Ndundulu Mts Nyumbanitu Mts Nyumbanitu Mts Mafwemiro Forest at Chugu Hill Mafwemiro Forest at Chugu Hill Mafwemiro Forest at Chugu Hill Captive NADH2 DQ093797 DQ093798 DQ093799 DQ093800 DQ093801 DQ093802 DQ093803 DQ093804 Gadph11 DQ093805 DQ093806 DQ093807 DQ093808 DQ093809 DQ093810 DQ093811 DQ093812 194 Table 2. List of all specimens collected since the discovery of Xenoperdix in 1991, as well as the outgroup taxon Arborophila javanica. Genbank numbers are listed for the three DNA regions sequenced in this study. Fib5 DQ093813 DQ093814 DQ093815 DQ093816 DQ093817 DQ093818 DQ093819 DQ093820 R.C.K. Bowie & J. Fjeldså Evidence for two species in Udzungwa forest partridge 195 Laboratory procedures DNA was extracted from frozen tissues using a Puregene DNA isolation kit (Gentra Systems, Minneapolis, Minnesota) following the manufacturer’s animal tissue protocols, but with an overnight proteinase K digestion at 55oC. The mitochondrial NADH Dehydrogenase subunit 2 gene was amplified using primers L5216 and H6313 (Sorenson et al., 1999) under standard PCR conditions. Two nuclear DNA introns, Beta-Fibrinogen 5 (primers: Fib 5: 5`CGCCATACAGAGTATACTGTGACAT-3` and Fib6: 5`-GCCATCCTGGCGATTCTGAA3`) and Glyceraldehyde-3-phosphate Dehydrogenase intron 11 (primers: GapdL890 and GapdH950; Friesen et al., 1997) were amplified using annealing temperatures of 54oC and 64oC, respectively. PCR products were electrophoresed on 1.5% low-melting point agarose gels (FMC Bioproducts) stained with ethidium bromide and visualized under UV light. Amplicons of the appropriate length were cut out of the gel and purified using GELase™ (Epicentre Technologies, Madison, Wisconsin). The purified product was cycle-sequenced using Big Dye terminator chemistry (Applied Biosystems, Inc [ABI]). Cycle–sequencing reactions were precipitated with 3 M ammonium acetate, rinsed in ethanol, dried and resuspended in formamide-EDTA solution and run on an ABI 3100 automated DNA sequencer. All sequences were checked using the programme Sequencher 3.0 (Gene Codes Corp.) and NADH2 aligned to the chicken Gallus gallus Linnaeus 1758 mtDNA sequence (Desjardins & Morais, 1990) to test for the presence of any insertions, deletions or stop codons. All sequences have been submitted to Genbank (table 2). Coalescent-based estimates of gene flow and divergence times The program MDIV (Nielson & Wakeley, 2001; Nielson, 2002) was used to simultaneously estimate a variety of parameters between the Udzungwa and Rubeho Highland populations of Xenoperdix: theta (θ = 2Nefμ), migration (gene flow) rate (M = 2Nefm), time of population divergence (T = t/Nef), and time to most recent common ancestor (TMRCA = tμ), where Nef is the female-effective population size, t is the generation time, and μ is the per locus mutation rate. Six different pairwise simulations were conducted: one run each of 10 million, 25 million, and 50 million generations with priors set as Tmax = 15 and Mmax = 10, or Tmax = 50 and Mmax = 10. Multiple runs with increasing sampling of the posterior distribution were conducted to determine if convergence in the mode of the posterior distribution was being reached. A finite-sites (HKY) model was used in all analyses. Values for theta, M and T were plotted, and the mode of the posterior distribution was accepted as the best estimate. Where possible, 95% credibility intervals also were estimated for each parameter. Standardized estimates for T and TMRCA are provided in which each estimate is multiplied by theta to account for different effective population sizes and allow for direct comparison among estimates. These values were converted into years before present using the widely assumed mutation rate of 2% per million years for mtDNA (see García-Moreno, 2004), or 1.915 x 10-5 substitutions per site per year, and an estimated generation time of 1 year as occurs for most galliformes (McGowan, 1994). RESULTS Sequence characteristics and divergence In total 1041 bp of mtDNA (NADH2) and 956 bp of nDNA (569 bp Fib5 and 387 bp Gadph11) were amplified from all seven individuals. Fib5 was particularly difficult to sequence in both directions due to a homo-polymerous region of 22 adenines in the middle of 196 R.C.K. Bowie & J. Fjeldså the intron. Five fixed nucleotide differences, two of which resulted in amino acid changes were documented in NADH2 between X. u. udzungwensis and X. u. obscurata (table 3). One fixed difference was recovered in Fib5 and none for Gadph11 (table 3). These correspond to a sequence divergence of 0.5% between X. u. udzungwensis and X. u. obscurata for NADH2, and 0.2% for Fib5. Sequence divergence between Xenoperdix and Arborophila was much larger: 20% for NADH2, 9–10% for Fib5, and 11% for Gadph11. This was also accompanied by a number of indels between Xenoperdix and Arborophila in the nDNA markers: five indels (1–8 bp in length) for Fib5, and 10 indels (1–18 bp) for Gadph11, respectively. Table 3. Sequence characteristics, highlighting the genetic differences between X. u. udzungwensis and X. u. obscurata. NADH 2 (1041 bp) Fib5 (569 bp) Gadph11 (387 bp) Position Mutation Characteristics 120 144 464 821 843 236 55 T/C T/C T/C A/G C/T A/T C or T Fixed Fixed Fixed Fixed Fixed Fixed Heterozygote Codon position 3rd 3rd nd 2 2nd rd 3 N/A N/A Amino acid change Ile to Ile His to His Ile to Thr Glu to Gly Ile to Ile N/A N/A Almost no genetic diversity was detected within either of the populations of Xenoperdix. For both populations, only one haplotype for NADH2 and Fib5 was recovered (although differing by five and one fixed substitution, respectively). The only detectable within population diversity was a heterozygote at position 55 in the Gadph11 intron, and even this was found in only one of the seven individuals sampled. Coalescent-based estimates of gene flow and divergence times All coalescent models suggested that there is no current gene flow between populations in the Udzungwa and Rubeho Highlands (table 4, figure 2). Estimates of gene versus population divergence were more difficult to determine as the posterior distribution of T (population divergence) even after 50 million generations did not converge (table 4, figure 2). Depending on the prior set for Tmax, gene divergence, or time to most common recent ancestor (TMRCA) of the haplotypes recovered was estimated to have occurred between 181 000 and 196 000 years before present. DISCUSSION The location of known Xenoperdix populations in mountains flanking the Great Ruaha Basin to the south and north (figure 1) could be related to long-term climatic predictability. Throughout the Pleistocene, these highlands may have been under a constant influence of warm humid winds from the Indian Ocean (Fjeldså & Lovett, 1997; Burgess et al., 1998) and conditions may be particularly favourable in humid highlands bordering towards the inter-montane basin between the two highlands, which acts as a local high pressure centre with warm air rising up the mountain slopes. Several narrowly endemic bird taxa Evidence for two species in Udzungwa forest partridge 197 10 million generations 50 million generations M = 10 M = 10 T = 15 T = 50 Theta 0.008 Theta 0.008 0.007 0.007 0.006 0.006 0.005 0.005 0.004 0.004 0.003 0.003 0.002 0.002 0.001 0.001 0 0 0 0.5 1 1.5 2 2.5 3 0 3.5 0.5 1 1.5 2 2.5 3 3.5 M: Gene Flow M: Gene Flow 0.12 0.12 0.1 0.1 0.08 0.08 0.06 0.06 0.04 0.04 0.02 0.02 0 0 0 1 2 3 4 5 6 7 8 9 10 0 T: scaled population divergence time 0.003 1 2 3 4 5 6 7 8 9 10 T: scaled population divergence time 0.0025 0.0025 0.002 0.002 0.0015 0.0015 0.001 0.001 0.0005 0.0005 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 5 10 15 20 25 30 35 40 45 50 Figure 2. Pairwise estimates of theta, migration rates or gene flow (M = Nefm) and time since population divergence (T) based on analysis of mtDNA sequence data using MDIV under two different models of Tmax for 10, 25 and 50 million generation runs, respectively. All analyses supported a model of no gene flow between Xenoperdix u. udzungwensis and X. u. obscurata. (Xenoperdix, Bubo vosseleri Reichenow 1908, Batis sp. nov. Fjeldså, Bowie and Kuire unpublished, Arcanator orostruthus Vincent 1933, Swynnertonia swynnertoni Shelley 1906, Sheppardia aurantiithorax Beresford, Fjeldså & Kuire 2004, S. lowei Grant and MackworthPraed 1941, Apalis chariessa Reichenow 1879, Scepomycter winifredae Moreau 1938, 198 R.C.K. Bowie & J. Fjeldså Nectarinia moreaui Sclater 1933, N. rufipennis Jensen 1983, Ploceus nicolli Sclater 1931, Serinus whytii Shelley 1897 and S. melanochrous Reichenow 1900) are found in this area, some of them restricted to one side of the Great Ruaha Basin, others inhabiting highlands on both sides (Bowie et al., 2004; Fjeldså et al., in press). Table 4. Pairwise estimates of theta (because the mutation rate is the same for each population, differences in theta correspond to differences in Nef), migration rates or gene flow (M = Nefm), time since population divergence (T), and time to most recent common ancestor (TMRCA) between X. u. udzungwensis and X. u. obscurata based on analysis of mtDNA sequence data using MDIV. The highest posterior probability scores for theta and M are given with their 95% credibility intervals. Values for T were within undefined bounds due to likelihood estimates for the parameter not converging (see figure 2 and text for further details). 10 mil generations Tmax = 15/Mmax = 10 Tmax = 50/ Mmax = 10 25 mil generations Tmax = 15/Mmax = 10 Tmax = 50/ Mmax = 10 50 mil generations Tmax = 15/Mmax = 10 Tmax = 50/ Mmax = 10 Theta M (Gene Flow) TMRCA T 0.49 (0.11-2.11) 0.35 (0.10-2.15) 0.0 (0.00-0.01) 0.0 (0.00-0.01) 7.21 10.08 Undefined Undefined 0.50 (0.11-2.11) 0.38 (0.09-2.12) 0.0 (0.00-0.01) 0.0 (0.00-0.01) 7.23 10.00 Undefined Undefined 0.49 (0.11-2.11) 0.37 (0.09-2.14) 0.0 (0.00-0.01) 0.0 (0.00-0.01) 7.24 9.98 Undefined Undefined Coalescent models and fixed mutational differences in mtDNA, and one of the two nDNA markers sequenced, strongly suggest that no gene flow is currently taking place between populations of Xenoperdix in the Udzungwa (X. u. udzungwensis) and Rubeho Highlands (X. u. obscurata). Using the widely accepted 2% sequence divergence per million years in mtDNA (see García-Moreno, 2004), the 0.5% divergence in NADH2 suggests that the Xenoperdix populations on either side of the Great Ruaha Basin diverged about 250 000 years ago. Coalescent estimates of gene divergence suggest a slightly earlier divergence date of between 196–181 000 years BP. We were unable to estimate the date when gene flow ceased or population divergence occurred (T in figure 2 and table 4), but it likely followed shortly after the initial gene divergence given that both taxa are reciprocally monophyletic for two of the three markers analyzed (NADH2 and Fib5, not Gadph11). The estimated divergence data of ca. 200 000 years BP is considerably more recent than the final uplifting of the Udzungwa and Rubeho Highlands around 7 million years BP (Griffiths, 1993). However, the relatively recent divergence between X. u. udzungwensis and X. u. obscurata is in agreement within ongoing studies of several forest bird species, which have demonstrated a sharing of haplotypes between birds in the Udzungwa and Rubeho Highlands, suggesting that birds do move (if rarely) across the Great Ruaha River Gorge (Bowie et al., 2004, Fjeldså et al., in press). Sample sizes are small, but neither the population in the Udzungwa nor the population in the Rubeho Highlands had high levels of genetic diversity. Thus, based on the present dataset, it is not possible to identify whether the much larger Udzungwa Xenoperdix population acted as a source for the Rubeho population or vice versa. Both Fib5, a novel application to galliform systematics, and Gadph11 were variable, although Gadph11 was not as variable in Xenoperdix as it is in other galliform taxa (Holder et al., 1999). Almost no genetic diversity was detected within either of the populations of Xenoperdix. This raises the important question of whether Xenoperdix is inbred, or has managed to Evidence for two species in Udzungwa forest partridge 199 successfully purge deleterious alleles. Due to the extremely small distribution of this species, it is critical that levels of genetic diversity and extent of local population movements within each of the highlands be accurately documented. Although we have sampled all seven specimens that have been collected for this species to date, the numbers are not sufficiently large to be statistically confident that X. u. udzungwensis and/or X. u. obscurata are genetically depauperate. However, given that two of the five fixed differences detected in NADH2 are at second codon positions, usually the most conservative because changes always result in an amino acid substitution, it is more parsimonious to assume that the Xenoperdix populations have been through bottlenecks, leading to low population numbers and fixation of haplotypes/alleles by genetic drift. We suggest that conservation managers would be well advised to establish a program of continued genetic assessment, obtaining samples with non-invasive methods, e.g. collecting all feathers left at dust-bathing sites in both the Udzungwa and Rubeho Highlands for further genetic analyses in order to better document genetic diversity and help meet long term objectives of conserving this unique African bird lineage. Many biologists associate the degree of genetic diversity with extent of reproductive isolation. The rationale underlying this stems from the idea that the accumulation of genetic differences at many loci occurs over a long period of time. However, the long time periods required for sufficient genetic divergence to accumulate between two populations in order to be termed two species, are probably not often required for strong pre-mating isolation to occur (Ferguson, 2002). Indeed, the genetic basis of reproductive isolation varies so much between taxa that a predictive rate about degree of genetic divergence required for recognition of separate species is not possible (Ferguson, 2002). From a genetic perspective, a more robust statement about species status of a particular population/taxon can be made if one can demonstrate fixed genetic differences between the population/taxon in question and other populations/taxa. Such differences imply that not only has genetic differentiation taken place, but that there is a lack of gene flow as well (Ferguson, 2002). Assignment of species rank is further supported if genetic divergence is corroborated with morphological differences between populations/taxa. The two Xenoperdix taxa differ in facial ornaments and tail dimensions (table 1), both of which are potential display signals. Xenoperdix u. obscurata is distinct from X. u. udzungwensis, with fixed mtDNA (NADH2, five fixed differences, two of which are at second codon positions resulting in amino acid substitutions) and nDNA (Fib5, one fixed) differences (table 3), and morphological characters (table 1). Thus, following Ferguson’s (2002) arguments outlined above, it seems that the 0.5% mtDNA divergence between X. u. obscurata and X. u. udzungwensis is of little consequence when assigning species rank, and merely suggests that the two taxa diverged within the last 200 000 years. Further, coalescent models strongly support a model of no gene flow (figure 2, table 4). Within the framework of operational species concepts, such as the Phylogenetic Species Concept (Cracraft, 1983), and the ‘Comprehensive Biological Species Concept’ (Johnson et al., 1999), the consistent diagnosable differences suggest that X. u. obscurata and X. u. udzungwensis can be accorded species status. Proponents of a traditional Biological Species Concept (Mayr, 1942) may argue that were the two taxa not allopatric they would interbreed. This may be the case, but we do not know, and in view of the dense human settlement, severe habitat degradation, and absence of forest cover in most parts of these highlands, it is unlikely that they will ever meet. The uncertainty of how to treat allopatric populations reflects one of the great difficulties with applying the biological species concept as an operational definition (Zink, 1997). There is no doubt that species are the currency of conservation (Rojas, 1992; Zink, 1997). 200 R.C.K. Bowie & J. Fjeldså Although an objective taxonomy should not build on non-biological criteria such as conservation policy, species rank for the Rubeho population will probably lead to higher conservation attention. Otherwise we fear that this distinctive population would be neglected as "a marginal isolate". Such taxonomic neglect has already led to the extinction of other small and isolated populations (e.g. Hazevoet, 1996). The genetically and morphologically distinctive Rubeho forest partridge X. obscurata is an endangered species in need of conservation attention. ACKNOWLEDGMENTS The fieldwork was supported by grant no. 21-00-0178 from the Danish Research Council to JF. Laboratory work was supported by a grant from Stellenbosch University to RCKB. We thank Tom Butynski, Benny Bytebier, John Bates and Tim Crowe for helpful reviews of this manuscript. This project is a contribution from the Field Museum’s Pritzker Laboratory for Molecular Systematics and the Evolutionary Genomics Group at the University of Stellenbosch. REFERENCES BirdLife International (2000). Threatened Birds of the World. Lynx Edicions and BirdLife, Barcelona and Cambridge. Bowie, R.C.K., J. Fjeldså, S.J. Hackett & T.M. Crowe (2004). Systematics and biogeography of double-collared sunbirds from the Eastern Arc Mountains, Tanzania. Auk 121: 660–681. Burgess, N., J. Fjeldså & R. Botterweg (1998). Faunal importance of the Eastern Arc Mountains of Kenya and Tanzania. Journal of East African Natural History 87: 37–58. Burgess, N., N. Cordeiro, N. Doggart, J. Fjeldså, L.A. Hansen, K. Howell, F. Kilahama, F. Nashanda, M. Menegon, D. Moyer, D. Perkin, W. Stanley & S. Stuart (in press). The biological importance of the Eastern Arc Mountains of Tanzania and Kenya. Biological Conservation. Butynski, T.M. & C.L. Ehardt (2003). Notes on ten restricted-range birds in the Udzungwa Mountains, Tanzania. Scopus 23: 13–28. Cracraft, J. (1983). Species concepts and speciation analysis. Current Ornithology 1: 159–187. Crowe, T.M., P. Bloomer, E. Randi, V. Lucchesi, R. Kimball & J.G. Groth (in press). Supra-generic cladistics of landfowl (Order Galliformes). Proceedings of the 23rd International Ornithological Congress. Desjardins, P., & R. Morais (1990). Sequence and gene organisation of the chicken mitochondrial genome. A novel gene order in higher vertebrates. Journal of Molecular Evolution 212: 599–634. Dinesen, L., T. Lehmberg, J.O. Svendsen, L.A. Hansen & J. Fjeldså (1994). A new genus and species of perdicine bird (Phasianidae, Perdicini) from Tanzania: a relict form with Indo-Malayan affinities. Ibis 136: 3–11. Dinesen, L., T. Lehmberg, M.C. Rahner & J. Fjeldså (2001). Conservation priorities for the forests of the Udzungwa Mountains, Tanzania, based on primates, duikers and birds. Biological Conservation 99: 223–236. Ferguson, J.W. (2002). On the use of genetic divergence for identifying species. Biological Evidence for two species in Udzungwa forest partridge 201 Journal of the Linnean Society 75: 509–516. Fjeldså, J. & J.C. Lovett (1997). Geographical patterns of old and young species in African forest biota: the significance of specific montane areas as evolutionary centres. Biodiversity and Conservation 6: 325–246. Fjeldså, J. & J. Kuire (2003). A new population of the Udzungwa forest partridge, Xenoperdix udzungwensis. Bulletin of the British Ornithologists Club 123: 52–57. Fjeldså, J., J. Kiure, N. Doggart, L.A. Hansen & A. Perkin (in press). Distribution of highland forest birds across a potential dispersal barrier in the Eastern Arc Mountains of Tanzania. Steenstrupia. Friesen, V.L., B.C. Congdon, H.E. Walsh & T.P. Birt (1997). Intron variation in marbled murrelets detected using analyses of single-stranded conformational polymorphisms. Molecular Ecology 6: 1047–1058. García-Moreno, J. (2004). Is there a universal mtDNA clock for birds. Journal of Avian Biology 35: 465–468. Griffiths, C.J. (1993). The geological evolution of East Africa. In J.C. Lovett & S.K. Wasser (eds), Biogeography and Ecology of the Rain Forests of Eastern Africa. Cambridge University Press, Cambridge. Pp 9–21. Hazevoet, C.J. (1996). Conservation and species lists: taxonomic neglect promotes the extinction of endemic birds, as exemplified by taxa from eastern Atlantic islands. Bird Conservation International 6: 181–196. Holder, K., R. Montgomerie & V. Friesen (1999). A test of the glacial refugium hypothesis using patterns of mitochondrial and nuclear DNA sequence variation in rock ptarmigan (Lagopus mutus). Evolution 53: 1936–1950. Jetz, W., C. Rahbek & R.K. Colwell (2004). The coincidence of rarity and richness and the potential signature of history in centers of endemism. Ecology Letters 7: 1180–1191. Johnson, N.K., J.V. Remsen Jr. & C. Cicero (1999). Resolution of the debate over species concepts in ornithology: a new comprehensive biological species concept. In N.J. Adams & R.H. Slotow (eds), Proceedings of the 22nd International Ornitholological Congress. Durban, BirdLife South Africa. Pp. 1470–1482. Mayr, E. (1942). Systematics and the Origin of Species. Columbia University Press. McGowan, P.J.K. (1994). Family Phasianidae. In J. del Hoyo, A. Elliott & J. Sargatal (eds), Handbook of Birds of the World. Lynx Edicions, Barcelona. Pp. 434–553. Nielson, R. (2002). MDIV software. Available at http://www.biom.cornell.edu/Homepages/Rasmus_Nielson/files.html. Nielson, R., & J. Wakeley (2001). Distinguishing migration from isolation: a Markov Chain Monte Carlo Approach. Genetics 158: 885–896. Rojas, M. (1992). The species problem and conservation: what are we protecting? Conservation Biology 6: 170–178. Sorenson, M.D., J.C. Ast, D.E. Dimcheff, T. Yuri & D.P. Mindell (1999). Primers for a PCR-based approach to mitochondrial genome sequencing in birds and other vertebrates. Molecular Phylogenetics and Evolution 12: 105–114. Zink, R.M. (1997). Species concepts. Bulletin of the British Ornithologists Club. 117: 97–109.