The Mediterranean Alectoris (including A. rufa, A. graeca, A. chukar, and A. barbara) comprise a group of closely related and morphologically uniform partridges with largely allopatric distributions and instances of natural hybridization in parapatric contact zones. Their taxonomic status and evolution are controversial. We have used multilocus protein electrophoresis to estimate the extent of genetic divergence among nominal Alectoris species and within A. chukar, A. graeca and A. rufa. The average Nei's (1978) genetic distance among conspecific populations (D[bar] = 0.008; range 0.003-0.021) was 26 times smaller than among species (D[bar] = 0.208; range 0.071-0.312). The most genetically similar species were A. rufa and A. graeca (D[bar] = 0.081); A. barbara and A. chukar were the most divergent (D[bar] = 0.303). The Fst, values among species (Fst, = 0.75) were more than eight times larger than among conspecific populations (Fst = 0.09). The gap in D and Fst, values for intraspecific and interspecific comparisons indicates a prolonged interruption of gene flow among species and independent evolution of their gene pools. Dendrograms summarizing genetic distance matrices and cladistic analyses of discrete character states suggested that A. rufa and A. graeca are sister species of recent origin, followed by the most distantly related and ancient A. chukar and A. barbara. Because protein electrophoresis results are concordant with biogeographical and paleontological information, we construct a hypothesis for the evolution of the Mediterranean Alectoris. Received 2 May 1991, accepted 13 January 1992.

lIstituto Nazionale de Biologia della Selvaggina via, Cd Fornacetta, 9, 40064 Ozzano dell'Emilia (Bo), Italy; 2Dipartimento di Biologia Animale, Universitd di Pavia, Piazza Botta, 9, 27100 Pavia, Italy; and 3Jacob Blaustein Institute for Desert Research, Ben-Gurion University of the Negev, Sede Boqer Campus 84990, Israel THE ALCTORIS partridges (Galliformes, Pha- sianidae) are distributed widely in the Palaearc- tic (Fig. 1). They present intriguing and chal- lenging questions with regard to taxonomic and evolutionary relationships. Peters (1934) in- cluded graeca, chukar, and magna as subspecies of A. graeca, a taxonomy that was followed by Dementiev and Gladkov (1952). Voous (1960) largely accepted this classification, but ques- tioned the separate species rank assigned to A. rufa. From ethological evidence, Menzdorf (1984) agreed that graeca, chukar, and rufa had not yet attained true species status. Vaurie (1959), on the contrary, argued that these forms com- prised separate species owing to diagnostic dif- ferences in facial plumage and vocalizations. This view was supported by Watson (1962a, b), who offered additional evidence of species-spe- cific plumage characters, and of behavioral and ecological separation among parapatric popu- lations. The Vaurie (1959) and Watson (1962a, b) view of seven closely related Alectoris species is now widely accepted (Cramp and Simmons 1980, Johnsgard 1988). Most Alectoris species are very similar mor- phologically, differing only with respect to sub- tle but diagnostic face and throat plumage pat- terns (Johnsgard 1988). Their natural ranges (Fig. 1) are largely allopatric, except for sympatry between melanocephala and philbyi in southern Arabia. Parapatric contact zones exist between chukar and graeca at the border of Greece and Bulgaria, between rufa and graeca in the French Alps, and probably between magna and chukar in central China (Watson 1962a, Bernard-Lau- rent 1984). Only two zones of overlap and hy- bridization (sensu Short 1969) have been re- ported: a well-documented rufa and graeca hybrid zone in the southern French Alps (Ber- nard-Laurent 1984); and an unconfirmed graeca and chukar hybrid zone in Thrace south of the Rhodope Mountains (Dragoev 1974). Extant Alectoris populations of Mediterranean and North Atlantic islands probably resulted from human introductions (Watson 1962b, Blondel 1988). The present pattern of only one species per island may represent the outcome of com- petitive exclusion among two or more species following repeated introductions (Blondel 1988). Therefore, several lines of evidence, including allopatric distributions, natural hybridization, ecological exclusion, and morphological simi- larity have been used in support of hypotheses of a recent radiation, and perhaps incomplete speciation, of the Mediterranean Alectoris par- tridges (Voous 1960, Watson 1962a, b, Blondel 1988). Two conflicting models have been recently proposed to explain evolution and speciation in Alectoris (Fig. 1). Watson (1962a) argued that Alectoris comprises: (1) the "superspecies" grae- ca (encompassing graeca, and magna); (2) the "superspecies" chukar (encompassing chukar, philbyi, barbara, and tufa); and (3) melanocepha- la, a separate and more distantly related species. From morphological and biogeographical evi- dence, Watson argued that graeca was the an- cestor of the chukar lineage, and that rufa evolved recently in southwestern Europe after barbara crossed the Straits of Gibraltar. Span0 (1975) criticized Watson's model and suggested close relationships of chukar with graeca and rufa, but not with barbara and philbyi. Bernard-Laurent (1984) and Blondel (1988) cited extant zones of overlap and hybridization for including graeca, rufa, and chukar in a single superspecies of which graeca was the ancestral form. According to these authors, barbara never crossed Gibraltar to Eu- rope and is only distantly related to graeca. We used multilocus protein electrophoresis to estimate the extent of genetic divergence among the four Mediterranean species of Alec- toris: Red-legged Partridge (A. rufa), Rock Par- tridge (A. graeca), Chukar (A. chukar), and Bar- bary Partridge (A. barbara). In particular, we examined the following hypotheses posed by previous authors: (1) the Mediterranean Alec- toris speciated very recently (Watson 1962a) and, perhaps, incompletely (Voous 1960, Blondel 1988); (2) graeca was the ancestral form of a su- perspecies from which chukar and rufa origi- nated at the eastern and western range bound- aries, respectively (Bernard-Laurent 1984, Blondel 1988); or, alternatively, (3) chukar, bar- bara, and rufa constitute a superspecies, with rufa originating from barbara after crossing Gi- braltar (Watson 1962a). Genetic distances were used to estimate levels of divergence among conspecific populations and among species, and to obtain dendrograms showing phenetic and phylogenetic relationships among species. A tentative calibration of the average rate of pro- tein evolution was used to estimate divergence times, which have been related to current evo- lutionary and biogeographical hypotheses. MATERIAL AND METHODS We analyzed 117 specimens belonging to the fol- lowing populations: A. tufa population 1 (n = 12, wild, SW Spain); A. tufa 2 (n = 20, captive-reared, Italy); A. graeca 1 (n = 10, wild, E Alps, Italy); A. graeca 2 (n = 20, captive-reared, Italy); A. chukar 1 (n = 20, wild, China); A. chukar 2 (n = 5, captive-reared, Bulgaria); and A. chukar 3 (n = 20, wild, central Israel); A. barbara (n = 20, wild, Sardinia, Italy). Captive specimens were obtained from pure-bred, [arm-reared stocks. Given the recent [oundation of the captive stocks, and their derivation from more than 20 breeding pairs, there was little likelihood that the source populations were subject to substantial inbreeding. Wild birds were col- lected [rom localities at which there have been no restocking with reared birds. We also analyzed tissues of Gray Partridge (Perdix perdix, n = 2) and Ring- necked Pheasant (Phasianus colchicus, n = 2) as out- groups for rooting phylogenetic trees. Liver and heart samples were dissected from fresh- ly-killed birds, stored at -20øC for several hours after death, and then stored at -80øC until processing. We separately homogenated about 0.5 g of each tissue in 1 ml of 0.01 M Tris/HC1 pH 7.5, 0.001 M Na2.EDTA, and 0.001 M B-mercaptoethanol buffer and centri- fuged for 15 rain at 13,000 rpm. Supernatants were diluted in one volume o[ a 40% glycerol solution, aliquoted in Microtiter plates, and frozen at -80øC until used. Vertical polyacrylamide gel electropho- resis (concentration of 7.5% monomers in the contin- uous systems) was used to resolve 33 loci. Staining recipes were adapted from Harris and Hopkinson (1976). Electromorphs were presumed to have a sim- ple genetic basis, and were considered as alleles. A1- leles were coded by their mobility from the starting line, with the most anodal allele coded as "a." The BIOSYS-1 program (Swofford and Selander 1989) was used to compute percent polymorphic loci (P) and heterozygosity (H) values. Agreement with Hardy-Weinberg expectations was tested using chi- square analysis (Sokal and Rohlf 1981). Other statis- tical procedures included: contingency tests of allelic heterogeneity among populations (Workman and Niswander 1970); Nei's (1978) and Rogers' (1972) ge- netic distance matrices, UPGMA phenograms (Sneath and Sokal 1973); and Wagner networks (Swofford 1981). We have computed F-statistics (Wright 1978) within and among nominal species. These provided two lines of evidence (i.e. genetic distances and F,) on the extent of genetic divergence at different tax- onomical levels (Corbin 1987). Cladistic trees were constructed using the program PAUP (Swofford 1985) after coding alleles as present or absent according to the independent-allele model (Buth 1984). The SPSS Fig. 1. Distribution of partridges of genus Alectoris (adapted from Watson 1962a and Blondel 1988). Con- tinuous lines indicate the evolutionary relationships among species as suggested by Watson (1962a, b). Broken lines indicate evolutionary relationships as hypothesized by Blondel (1988). (Nie et al. 1975) package was used to compute a Mann- Whitney U-test of difference of heterozygosity among populations. RESULTS We were able to resolve 33 presumed genetic loci among eight populations of Mediterranean Alectoris, and the Perdix perdix and Phasianus col- chicus outgroups (Table 1). Observed hetero- zygosity ranged from 0.018 (rufa 1) to 0.085 (chu- kar 3), and percent polymorphic loci ranged from 6.1 (barbara) to 39.4 (chukar 2, Table 2). Similar levels of genetic variability have been observed in many other bird species (Corbin 1987, Evans 1987). Heterozygosity did not differ between wild and captive populations of the same spe- cies (Mann-Whitney U-test, P < 0.05). All poly- morphic loci were at Hardy-Weinberg equilib- rium, excepting sME in rufa 2 and PEP-2 in chukar 3 (P < 0.01; x2-test with Levene's [1949] correction for small sample size, and exact prob- ability test). A positive fixation index F (Wright 1965) indicated a significant deficiency of het- erozygotes in both cases. Allele frequencies over all polymorphic loci differed significantly among all species (P < 0.01; contingency x2-test), and among the three populations of chukar (P < 0.01) and two pop- ulation of graeca (P < 0.05). The average F among species was 0.75, more than eight times higher than the average F, among conspecific populations (0.09). There were no fixed allelic differences be- tween rufa and graeca. We found that chukar had fixed allelic differences at 12% of loci from rufa and graeca, while there were 24% fixed allelic differences between barbara and the rufa-graeca pair. There were 27% fixed differences between chukar and barbara. Allele frequencies of rufa and graeca differed significantly (single-locus con- tingency x2-test) at six (18%) of their polymor- phic loci. Major differences in allele frequencies between these two species occurred at EST-2, sGOT, and sME. The mean F, between them was 0.572, as compared to only 0.063 and 0.076 among the two rufa and the two graeca popu- lations, respectively. Intraspecific heterogene- ity was greater among chukar populations (ist = 0.159). Nei's (1978) standard unbiased genetic dis- tances averaged 26 times larger among nominal Alectoris species (b = 0.208; range 0.071-0.312) than among conspecific populations (/ = 0.008; range 0.003-0.021; Table 3). Interspecific ge- netic distances were lowest between A. rufa and A. graeca, but even these values (9 = 0.081; range 0.071-0.091) were 18 times larger than the av- erage interpopulation genetic distance within the two species (b -- 0.0045; range 0.003-0.006). The largest genetic distance of the study was between barbara and chukar (/ = 0.303), which is one of the highest D-values obtained between what are considered to be congeneric bird spe- cies (Zink 1988, Gill and Gerwin 1989). A phenogram was generated with the un- weighted pair-group method using arithmetic averages (UPGMA) of Nei's distances (Fig. 2A). A Wagner network (Fig. 2B) was derived using Rogers' (1972) distances. Nei's D, a nonmetric distance measure, has been widely applied in ornithological research and, therefore, enabled us to compare our results with previous work. Moreover, Nei's D is intended to estimate the proportion of mutational divergence among pairs of lineages, and is related to the time of divergence from the last common ancestor in case of regular rates of molecular evolution (Nei 1978). The two dendrograms were topologically identical. They indicated that rufa and graeca are most similar, linked at 0.081, followed by chukar at 0.168, and barbara at 0.282. To construct the Wagner tree, we used Perdix perdix and Phasi- anus colchicus as outgroups to root the multiple- addition-criteria network. We optimized branch lengths to maximize goodness-of-fit statistics. Similar lengths among sister branches suggest- ed a regular rate of protein divergence among the lineages. A relative rate test (Beverley and Wilson 1984) was performed using Perdix perdix and Phasianus colchicus as outgroups. The aver- age ratio of branch lengths among lineages was 1.0133 + 0.0039, confirming the idea of a reg- ular rate of protein evolution. Parsimony trees (not shown) were derived by cladistic analysis (PAUP) of allele distribution among species. These agreed in topography with the dendro- grams shown in Figure 2. DISCUSSION Applications of multilocus protein electro- phoresis in the study of avian evolution have consistently indicated a comparatively slow rate of genetic divergence among avian taxa. Avise and Aquadro (1982) derived an average Nei's (1972) D of 0.08 among 173 congeneric species of birds, an order of magnitude lower than the average of other vertebrates. Variation of pair- wise interspecific genetic distances is large, ranging from 0.00 to 0.39 (Gill and Gerwin 1989, Zink and Avise 1990). Many conspecific bird populations are little differentiated, and it is difficult to find local populations or subspecies with F,-values greater than 0.05 and D greater than 0.02 (Barrowclough 1983). Exceptions have been found recently in some Neotropical birds, which had comparatively large genetic dis- tances: (a) among conspecific populations oc- cupying separate Amazon banks (e.g. b = 0.040 between trans-Amazonian populations of Pipra coronata; Capparella 1988); (b) among subspe- cies (e.g. / = 0.066 among subspecies of Chi- roxiphia pareola; Capparella 1988); and (c) among species (Hackett and Rosenberg 1990). These phenomena have been attributed to mecha- nisms of geographic variation and speciation that may be peculiar to South American birds (Capparella 1988). Hackett and Rosenberg (1990) have suggested the findings may indicate that a reconsideration of the taxonomy of Neotrop- ical birds is warranted. While theory provides no absolute thresholds of genetic distance for ranking bird taxa (John- son and Zink 1983), the large body of empirical evidence provides useful guidance. Corbin (1987) described the occurrence of different slopes of the regression line of D on Ft between conspecific populations or between species. Large, abrupt differences in D and Ft between intraspecific and interspecific levels suggest historical interruption of gene flow among taxa and reorganization of the genomes (e.g. found- er effect, random drift, natural selection), and indicate speciation. The average genetic distance among Alectoris (b = 0.208) was substantially higher than values obtained for most other congeneric bird species (Gill and Gerwin 1989). However, the low av- erage difference among conspecific populations of our study (t3 = 0.008) is typical of birds in general, and of galliforms in particular. Inter- specific genetic distance between Lagopus lago- pus and L. mutus in Scandinavia was 0.046, as compared to an average of only 0.0035 and 0.0009 among their conspecific populations, respec- tively (Gyllensten et al. 1985). The average ge- netic distance among seven populations of Cal- lipepla californica was 0.005 (Zink et al. 1987), and their estimated level of gene flow of 5.5 birds per generation was nearly identical to the rate estimated among three populations of Coli- nus virginianus (Ellsworth et al. 1989). Intraspecific genetic distances among the Alectoris populations of this study ranged from 0.003 to 0.021, and were only about 5% of the average interspecific genetic distance of 0.208. The Fs,-values among Alectoris species were more 362 RAr E'r At. [Auk, Vol. 109 vv vv TABLœ 2. Genetic variability at 33 loci in Mediter- ranean Alectoris partridges. Percent poly- Heterozygosity morphic Population loci Observed Expected A. rufa I 12.1 0.018 0.017 A. rufa 2 18.2 0.029 0.039 A. graeca I 21.2 0.052 0.051 A. graeca 2 21.2 0.042 0.053 A. chukar I 15.2 0.050 0.047 A. chukar 2 12.1 0.048 0.042 A. chukar 3 39.4 0.085 0.100 A. barbara 6.1 0.030 0.028 than eight times higher than among conspecific populations. We found rufa and graeca to be the least divergent species, with an average genetic distance of 0.081, 18 times larger than their av- erage interpopulation distance of 0.0045. Ge- netic distances and F, within and between spe- cies of Alectoris are comparable to the values found in Palaearctic and North American birds (Corbin 1987, Evans 1987). The relatively large gap between the population and the species lev- els suggests an extended interruption of gene flow and an independent evolution among the various Alectoris gene pools. The genetic dis- tances (/ = 0.303) between barbara and chukar populations are very high for congeneric birds. The extent of genetic heterogeneity among chu- kar populations (F, = 0.159) is three to five times larger than the values usually observed among conspecific bird populations, and indicates the existence of significant geographic divergence within the chukar range. The pattern of phylogenetic relationships de- picted by our dendrograms and cladistic trees (Fig. 2) suggests that Alectoris species did not result from contemporaneous episodes of spe- ciation (i.e. as consequence of the fragmenta- tion of an ancestral population). Rather, they arose from at least three waves of speciation. Therefore, the two reported hybrid zones in- volve sister taxa (rufa and graeca) and nonsister taxa (chukar and graeca). The ability to hybridize is most probably attributable to the conserva- tive morpho-anatomical evolution and conse- quent retention of ancestral characters, rather than to incomplete speciation (McKitrick and Zink 1988, Cracraft 1983, 1989). The rare in- stances of natural hybridization (which do not compromise the evolutionary independence of the rufa, graeca, and chukar genomes), the large TABLE 3. Nei's (1978) genetic distances (lower left) and Rogers' (1972) genetic distances (upper right) among Mediterranean Alectoris partridges and the outgroups Perdix perdix and Phasianus colchicus. A B C D E F G H I L A A. rufa 1 -- 0.019 0.114 0.108 0.181 0.159 0.183 0.249 0.605 0.572 B A. rufa 2 0.003 -- 0.107 0.100 0.177 0.155 0.176 0.245 0.601 0.566 C A. graeca 1 0.091 0.074 -- 0.035 0.210 0.194 0.188 0.252 0.592 0.563 D A. graeca 2 0.087 0.071 0.006 -- 0.197 0.177 0.179 0.253 0.605 0.563 E A. chukar 1 0.181 0.171 0.211 0.193 -- 0.042 0.059 0.284 0.628 0.567 F A. chukar 2 0.154 0.144 0.184 0.162 0.008 -- 0.065 0.270 0.626 0.566 G A. chukar 3 0.161 0.151 0.161 0.147 0.021 0.017 -- 0.287 0.628 0.571 H A. barbara 0.272 0.263 0.266 0.267 0.312 0.297 0.300 -- 0.599 0.568 I Perdix perdix 0.926 0.920 0.898 0.924 0.994 0.990 0.993 0.917 -- 0.394 L Phasianus colchicus 0.849 0.838 0.831 0.830 0.839 0.836 0.839 0.843 0.501 -- gap between interpopulation and interspecies genetic distances, and the existence of diag- nostic phenotypic characters indicate that the Mediterranean Alectoris are composed of what can be considered good evolutionary species. A chukar 1 chukar 2 chukar 3 graeca 1 graeca 2 rua 1 Perdix Phasianu$ 1.00 0.83 0,67 0.50 0,33 0.17 0.Nei'sD chukar 1 chukar 2 chukar 3 graeca 1 graeca 2 rufa 1 rufa 2 barbara Perdx Phaslanu$ 0'.00'0107  0113' 0.20  0,27' 0133' 0.40 Distance from the Root Rogers' D Fig. 2. (A) UPGMA dendrogram of Nei's standard unbiased genetic distances (Table 3) among Mediter- ranean Alectoris and outgroups Perdix perdix and Phasi- anus colchicus. Cophenetic correlation is 0.994. Time scale computed using calibration 1 D = 22.9 myr. (B) Wagner tree obtained from Rogers' genetic distances (Table 3). Cophenetic correlation is 0.999. Comparisons of sister-branch lengths in UPGMA and Wagner dendrograms (Fig. 2), as well as the relative rate test, indicate a rela- tively constant rate of protein evolution in Alec- toris. Therefore, we can attempt a calibration of the molecular clock in order to date the diver- gence times of Alectoris species. Based on a gal- liform fossil (the odontophorin North Ameri- can quail Cyrtonyx cooki), Gutierrez et al. (1983) proposed Nei's values of 1 D = 23.6 million years (myr). Marten and Johnson (1986) pro- vided a similar estimate of 1 D = 19.7 myr. These rates have proved useful in estimating times of evolutionary branchings in several bird taxa (Johnson and Zink 1983, Zink and Johnson 1984, Randi et al. 1991b), and also approximate the rate of mtDNA evolution in Ammodramus (Zink and Avise 1990). From multiple comparisons of nDNA (Helm-Bychowski and Wilson 1986) and enzyme phylogenetic trees, as well as several informative fossils, Randi et al. (1991a) derived the calibration 1 D = 22.9 myr for phasianid birds. Assuming this rate, we derived the fol- lowing divergence times (Fig. 2A); (1) an initial splitting of the ancestral Alectoris into the pres- ent barbara and chukar lineage about 6.4 million years ago (mya); (2) a second splitting of the chukar and graeca-rufa lineages about 3.8 mya; (3) a final splitting of graeca and rufa only 1.8 mya. Our results indicate that genetic divergence among the Mediterranean Alectoris is great and that they can be considered good evolutionary species, as proposed by Vaurie (1959) and Wat- son (1962a, b). Phylogenetic relationships among the species as indicated by our studies do not support the evolutionary scenarios proposed by Watson (1962a) and Blondel (1988): chukar is not strictly related to barbara and rufa; graeca is not the ancestral form of a superspecies encom- passing rufa and chukar; and barbara is not the stem species of rufa. Evolution and speciation of Alectoris are not recent Pleistocene events as supposed by Watson (1962a); based on our es- timated divergence times, speciation events could span from the Miocene-Pliocene bound- ary to early Pleistocene. The conservative mor- phology and the small plumage difference are consistent with a slow rate of morphologic evo- lution and not with recent origins of these spe- cies. We propose the following model of Alectoris evolution, substantially concordant with Span0 (1975), and characterized by at least three waves of speciation. About 6 mya, at the Miocene-Plio- cene boundary, and ancestral species split into barbara and in the chukar lineage. This division occurred during a period of climatic warming and aridity that resulted in the closure of the Mediterranean Sea at Gibraltar and its subse- quent dessication (Voous 1974). Concurrently, uplift of the Carpathian Mountains created an east-west divide of the central European plains, and separated the Sarmatic Sea from the Med- iterranean (Voous 1974). The climate probably favored the spread of birds adapted to arid and steppe habitats, such as barbara and chukar. We suppose that geologic events resulted in an east- west splitting of the ancestral populations in which the chukar lineage spread eastward around the Sarmatic Sea, while barbara spread westward along the Mediterranean littoral of the Middle East and Africa. Following speciation, barbara eventually crossed the Straits of Gibraltar, and then spread eastward along the European Med- iterranean coasts, thereby explaining the pres- ence of fossil barbara in France up to the middle Pleistocene (Mourer-Chauvir 1975). Then, about 4 mya (early Pliocene), the ancestral chu- kar populations, spreading westward, gave rise to the graeca-rufa lineage. Finally, at the start of the Pleistocene glaciations, about 1.8 mya, chu- kar populations in western Europe survived in fragmented populations, which probably un- derwent repeated contraction and expansion according to climatic changes. The fossil record indicates that barbara disappeared from Europe during the middle Pleistocene, while both rufa and graeca occurred in France (Mourer-Chau- vir 1975). The origin of graeca and rufa was probably fostered by climate-driven contrac- tions of steppe habitat during the early Pleis- tocene (Blondel 1988). Postglacial warming and extensive deforestation during the Holocene throughout the Mediterranean basin favored expansion of the European and Middle-Eastern populations of rufa, graeca, and chukar. They be- came parapatric, and originated the actual zones of overlapping and hybridization. The model is consistent with available fossil, biogeographic and genetic information on the Mediterranean Alectoris. It also leads to the fol- lowing hypothetical predictions for other Alec- toris species: (1) First, melanocephala and philbyi, the only two sympatric Alectoris (Watson 1962a), arose during the first wave of speciation from anciently separated lineages. They came into extensive contact in the Arabian Peninsula, and evolved character displacement and ecological compatibility (Watson 1962a), allowing their sympatry. From plumage characters (Watson 1962a), we hypothesize that melanocephala is re- lated to the barbara lineage, and that philbyi is related to the chukar lineage. (2) Second, magna arose from ancient fragmentation of chukar pop- ulations in China. Plumage similarities be- tween magna and graeca resulted from conver- gence rather than from common ancestry. This is in contrast to Watson's (1962a) argument that magna is a relic population of a widespread ancestral graeca lineage. Our model indicates that graeca never reached central Asia. These hypotheses are amenable to testing by expand- ing biochemical genetic research to all seven extant species of Alectoris. ACKNOWLEDGMENTS We thank Ariane Bernard-Laurent and T. Crowe for the comments and suggestions that greatly im- proved an early version of the manuscript. We are indebted to J. Blondel and S. J. Hackett for the ex- cellent reviews of the paper. LITERATURE CITED AVISE, J. C., AND C. F. AQUADRO. 1982. A comparative summary of genetic distances in the vertebrates. Evol. Biol. 15:151-158. BARROWCLOUGH, G. F. 1983. Biochemical studies of microevolutionary processes. Pages 223-261 in Perspectives in ornithology (A. H. Brush and G. A. Clark, Jr., Eds.). Cambridge Univ. Press, Cam- bridge. BERNARD-LAURENT, A. 1984. Hybridation naturelie entre Perdrix bartavelle (Alectoris graeca saxatilis) et Perdrix rouge (Alectoris rufa rufa) dans les Alpes- Maritimes. Gibier Faune Sauvage 2:79-96. BEVERLEY, S. M., AND A. C. WILSON. 1984. Molecular evolution in Drosophila and the higher Diptera. II. A time scale for fly evolution. J. Mol. Evol. 21: 1-13. BLONDEL, J. 1988. Biogographie volutive/ diff- rentes chelles: L'histoire des avifaunes mditer- ranennes. Pages 155-187 in Acta XIX Congr. Int. Ornithol. (H. Ouellet, Ed.). Natl. Mus. Nat. Sci., Ottawa, Ontario. BUTH, D.J. 1984. The application of electrophoretic data in systematic studies. Annu. Rev. Ecol. Syst. 15:501-522. CAPPARELLA, A. P. 1988. Genetic variation in Neo- tropical birds: Implications for the speciation process. Pages 1658-1664 in Acta XIX Congr. Int. Ornithol. (H. Ouellett, Ed.). Natl. Mus. Nat. Sci., Ottawa, Ontario. CLAYTON, J. W., AND D. N. TRETIAK. 1972. Amino- citrate buffers for pH control in starch gel elec- trophoresis. J. Fish. Res. Board Can. 29:1169-1172. CORBIN, K.W. 1987. Geographic variation and spe- ciation. Pages 321-353 in Avian genetics. A pop- ulation and ecological approach (F. Cooke and P. A. Buckley, Eds.). Academic Press, London. CRCRFr, J. 1983. Species concepts and speciation analysis. Curr. Ornithol. 1:159-187. CRCRFr, J. 1989. Speciation and its ontology: The empirical consequences of alternative species concepts for understanding patterns and pro- cesses of differentiation. Pages 60-81 in Specia- tion and its consequences (D. Otte and J. A. End- let, Eds.). Sinauer Associates, Sunderland, Mas- sachusetts. CRMP, S., AND K. E. L. SImmONS (EDs.). 1980. Hand- book of the birds of Europe, the Middle East and North Africa. Vol. 2, The birds of the western Palearctic. Oxford Univ. Press, Oxford. DAviS, B.J. 1964. Disc electrophoresis. II. Method and application to human serum proteins. Ann. N.Y. Acad. Sci. 121:404-427. DEMENTIEV, G. P., AND N. A. GLADKOV (EDs.). 1952. Birds of the Soviet Union. Translated in 1967 by the Israel Programme for Scientific Translations, Jerusalem. DRAGOEV, V. 1974. On the population of the Rock Partridge (Alectoris graeca Meisner) in Bulgaria and methods of census. Acta Ornithol. 14:251- 255. ELLSWORTH, D. L., J. L. ROSEBERRY, AND W. D. KLIMSTRA. 1989. Genetic structure and gene flow in the Northern Bobwhite. Auk 106:492-495. EVANS, P. J.H. 1987. Electrophoretic variability of gene products. Pages 105-162 in Avian genetics. A population and ecological approach (F. Cooke and P. A. Buckley, Eds.). Academic Press, Lon- don. GILL, F. B., AND J. A. GERWIN. 1989. Protein relation- ships among Hermit Hummingbirds. Proc. Acad. Nat. Sci. Phila. 141:409-421. GUTIeRReZ, R. J., R. M. ZINK, AND S. Y. YANG. 1983. Genetic variation, systematic, and biogeographic relationships of some galliform birds. Auk 100: 33-47. GYLLENSTEN, U., N. RYMAN, AND T. SAETHER. 1985. Genetic divergence between Willow Grouse (La- gopus lagopus L.) and Rock Ptarmigan (Lagopus mu- tus L.) and the genetic structure of Scandinavian grouse populations. Hereditas 102:47-55. HACKETT, S. J., AND K. V. ROSENBERG. 1990. Com- parison of phenotypic and genetic differentiation in South American antwrens (Formicariidae). Auk 107:473-489. HARRIS, H., AND D. A. HOPKINSON. 1976. Handbook of enzyme electrophoresis in human genetics. North Holland Publishing Co., Amsterdam. HELM-BYCHOWSKI, K. M., AND A. C. WILSON. 1986. Rates of nuclear DNA evolution in pheasant-like birds: Evidence from restriction maps. Proc. Natl. Acad. Sci. USA 83:688-693. JOHNSGARD, P.A. 1988. The quails, partridges, and francolins of the world. Oxford Univ. Press, Ox- ford. JOHNSON, N. K., AND R. M. ZINK. 1983. Speciation in sapsuckers (Sphyrapicus): I. Genetic differen- tiation. Auk 100:871-884. JOLLE"V, W. B., AND H. W. ALLEN. 1965. Formation of complexes between basic proteins of leucocytes and plasma globulins. Nature 208:390-391. LEVENE, H. 1949. On a matching problem arising in genetics. Ann. Math. Stat. 20:91-94. MACLELLAN, T., AND J. A.M. RAMSHAW. 1981. Serial electrophoretic transfers: A technique for the identification of numerous enzymes from single polyacrylamide gels. Biochem. Genet. 19:647-653. MARTEN, J. A., AND N. K. JOHNSON. 1986. Genetic relationships of North American cardueline finches. Condor 88:409-420. McKITRICK, M., AND R. M. ZINK. 1988. Species con- cepts in ornithology. Condor 90:1-14. MENZDORF, A. 1984. Zur Kenntnis des Sozialver- haltens und der Lautiu0erungen einiger Feld- huhnarten (Phasianidae, Perdicinae: Alectoris spp.). Vogelwelt 105:9-21. MOURER-CHAUVIR, C. 1975. Les oiseaux du Pleis- tocene moyen et suprieur de France. Thse d'Etat, Univ. Claude-Bernard, Lyon. NE,M. 1972. Genetic distance between populations. Am. Nat. 106:283-292. NEI, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of in- dividuals. Genetics 89:583-590. NIE, N. H., C. H. HULL, J. G. JENKINS, K. STEINBRENNER, AND D. H. BENT. 1975. Statistical package for the social sciences. McGraw-Hill, New York. PETERS, J.L. 1934. Check-list of birds of the world, vol. 2. Harvard Univ. Press, Cambridge. RANDI, E., G. FUSCO, R. LORENZINI, AND T. M. CROW. 1991a. Phylogenetic relationships and rates of allozyme evolution within the Phasianidae. Bio- chem. Syst. Ecol. 19:213-221. RANDI, E., G. FUSCO, R. LORENZINI, AND F. SPINA. 1991b. Allozyme divergence and phylogenetic relation- ships within the Strigiformes. Condor 93:295- 301. ROGERS, J. S. 1972. Measures of genetic similarity and genetic distance. Studies in Genetics, Univ. Texas Publ. 7213:145-153. SHORT, L.H. 1969. Taxonomic aspects of avian hy- bridization. Auk 86:84-105. StmA?H, P. H., AD R. R. SOV, AL. 1973. Numerical taxonomy. W. H. Freeman and Co., San Francisco. SOV, AL, R. R., AD F. J. ROHF. 1981. Biometry, 2nd ed. W. H. Freeman and Co., San Francisco. SPAC, S. 1975. Considerazioni biogeografiche sul genere Alectoris Kaup, 1829 (Galliformes, Phasi- anidae). Ann. Mus. Civ. Stor. Nat., Genova 80: 286-293. SWOFFORV, D.L. 1981. On the utility of the distance Wagner procedure. Pages 25-43 in Advances in cladistics: Proceedings of the first meeting of the Willi Hennig Society (V. A. Funk and D. R. Brooks, Eds.). New York Botanical Garden, Bronx. SWOFFORV, D.L. 1985. PAUP, Phylogenetic analysis using parsimony, version 2.4. Illinois Nat. Hist. Surv., Champaign. SWOFFORV, D. L., AV R. K. SEDER. 1989. BIOSYS- 1. A computer program for the analysis of allelic variation in population genetics and biochemical systematics, release 1.7. Illinois Nat. Hist. Surv., Champaign. VAURIE, C. 1959. The birds of the Palearctic fauna. Non-Passeriformes. H. F. and G. Witherby, Lon- don. Voovs, K.H. 1960. Atlas of European birds. Nelson, London. Voovs, K.H. 1974. The birds of the tropical 'Middle Seas,' past and present. Pages 697-704 in Proc. XVI Congr. Int. Ornithol. (H. J. F. Frith and J. H. Calaby, Eds.). Canberra, Australia. WATSON, G.E. 1962a. SympatryinPalearcticAlectoris partridges. Evolution 16:11-19. WATSON, G. E. 1962b. Three sibling species of Alec- toris partridges. Ibis 104:353-367. WIrrs, D. E., AV R. A. REISFV. 1964. Disc elec- trophoresis in polyacrylamide gels: Extension to new conditions of pH and buffers. Ann. N.Y. Acad. Sci. 121:373-381. WORKMAN, P. L., An J. D. NISWANDER. 1970. Pop- ulation studies on southwestern Indian tribes. II: Local genetic differentiation in the Papayo. Am. J. Hum. Genet. 22:24-29. WRYtilT, S. 1965. The interpretation of population structure by F-statistics with special regard to sys- tems of mating. Evolution 19:395-420. WRIGHT, $. 1978. Evolution and the genetics of pop- ulations. Vol. 4, Variability within and among natural populations. Univ. Chicago Press, Chi- cago. ZINc, R.M. 1988. Evolution of Brown Towhees: A1- lozymes, morphometrics and species limits. Con- dor 90:72-92. Zc, R. M., A J. C. AvsE. 1990. Patterns of mi- tochondrial DNA and allozyme evolution in the avian genus Ammodramus. Syst. Zool. 39:148-161. Zc, R. M., AV N. K. JOHnSOn. 1984. Evolutionary genetics of flycatchers. I. Sibling species in the genera Empidonax and Contopus. Syst. Zool. 33: 205-216. ZIc, R. N., D. F. Lor, AD D. W. AVRSO. 1987. Genetic variation, population structure, and evo- lution of California quail. Condor 89:395-405.