The Red-breasted Sapsucker (Sphyrapicus ruber daggetti) and Red-naped Sapsucker (S. nuchalis) are sympatric and hybridize in south-central Oregon, northeastern California, along the California-Nevada border, and in southern Nevada. We examined the overlapping distribution, nature of hybridization, and mate preference in these two taxa. Using a "hybrid index" system, we identified 13 phenotypic classes that represent the range of variation seen in typical parental types and their hybrids. Variation of parental forms in regions of allopatry was used to distinguish parental phenotypes from hybrids in the zone of overlap and hybridization. The percentage representation of various categories of mating (conspecific, 75.8%; backcross, 16.6%; hybrid, 1.4%; and interspecific, 6.2%) shown by 145 nesting pairs in the zone of overlap was used to infer the relative fitness of F 1 and F 2 generation hybrids. Although interspecific matings produce fully viable F 1 offspring in numbers proportional to expectation, hybrid and backcross matings apparently are selected against. We suspect that F 1 individuals and various recombinants have partial sterility barriers. In interspecific matings and in backcrosses, the male nearly always is S. r. daggetti or the redder mate. The data agree most closely with the dynamic-equilibrium model (stable-zone hypothesis), which has been proposed to explain zones of sympatry and hybridization in which gene flow from the extensive regions of allopatry of parental forms is balanced by selection against hybrids. In view of the preponderance of conspecific matings where S. r. daggetti and S. nuchalis occur together, they are regarded as biologic species. This decision is most appropriate for these taxa despite their low degree of hybridization and their near genetic identity as shown by electrophoresis. Received 9 March 1984, accepted 5 September 1984.

Museum of Vertebrate Zoology and Department of Zoology, University of California, Berkeley, California 94720 USA INFORMATION from zones of natural hybrid- ization in animals has been a cornerstone of models of speciation and evolutionary theory (Mayr 1963, Woodruff 1973, Dobzhansky et al. 1977, Endler 1977, Moore 1977, Barton and Hewitt 1981). Analyses of such zones in birds continue to play a significant role in the de- velopment and interpretation of these models (Selander 1971, Meise 1975, Barrowclough 1980). Avian material provides several impor- tant advantages for the study of hybridization and speciation (Miller 1955). First, most bird species can be watched easily as free-living or- ganisms; their habitat preferences, pairing be- havior, mate choice, and reproductive success are clearly revealed in the field. Furthermore, avian reproductive isolating mechanisms em- phasize features of color, pattern, and/or voice that are obvious and relatively easy to analyze. Finally, it is possible to see their responses to the environment, "particularly to the barriers and discontinuities ... important in the pro- cesses of geographic differentiation" (Miller 1955: 15). Avian examples in which very different phe- notypic forms occur in sympatry and inter- breed to varying degrees are especially intrigu- ing. Two forms of specialized woodpeckers, the Red-breasted Sapsucker (Sphyrapicus ruber dag- getti) and the Red-naped Sapsucker (S. nuchalis), clearly illustrate this situation where they lo- cally overlap and hybridize in the western United States (Howell 1952, Short 1982, John- son and Zink 1983). Although the general po- sition of their zone of interaction has been known for many years (Grinnell and Miller 1944: 235-236), Howell (1952) was the first to investigate it in the field during the breeding period. His work was followed by new infor- mation on distribution in Oregon (Browning 1977, Johnson 1978), a review of the status of the "Sphyrapicus varius group" in California (Devillets 1970), and by data that indicated change in the geographic position of S. r. dag- getti in the overlap-hybrid zone (Johnson 1970). Recently, Johnson and Zink (1983) presented data on genic differentiation for all species of Sphyrapicus. Sphyrapicus r. daggetti and S. nu- chalis, despite their profound dissimilarity in appearance, ate virtually identical at 39 pre- sumptive genetic loci revealed by gel electro- phoresis. In this paper, we report on the sympatric dis- tribution, hybridization, and mate choice in the same two taxa. We describe (1) the geography and dimensions of the zone of sympatry and hybridization, (2) the nature of phenotypic variation in allopatry vs. that in sympatry, (3) the degree of intermating of the various phe- riotypes found in the region of overlap and in- terbreeding, based on a very large sample, and (4) we estimate the relative fitness of hybrids as a basis for inferring the evolutionary status of the taxa in the zone of overlap and hybrid- ization. Such data conclusively resolve the de- cades-long debate (Grinnell 1901; A.O.U. 1910, 1957, 1983; Howell 1952; Short 1969, 1982; Mayr and Short 1970; Johnson and Zink 1983) over the taxonomic status of the two forms. In the following discussion, S. r. daggetti and S. nu- chalis will be referred to as "daggetti" and "nu- chalis." MATERIALS AND METHODS Field work was begun in the nesting season of 1963 in the mountains of southern Nevada (Johnson 1965) and continued over 12 of 16 summers through 1978. Initially, the main effort was to define in detail the breeding distributions of daggetti and nuchalis and to collect geographically representative specimens as a basis for interpreting character variation. Later, we concentrated on finding nests in the zone of sym- patry and hybridization in order to determine the phenotypes of mated adults. All important localities of sympatry and hybridization were visited at least twice, but not in succeeding years. This allowed for local population recovery after collecting. Except in a few instances, when we attempted to obtain mated pairs and their young for studies of character inher- itance, only one parent was taken at a nest. Collect- ing was unnecessary for many nests because the pa- rental phenotypes could easily be determined. If one parent was more difficult than the other to charac- terize according to a hybrid index system (see next section), because of plumage wear or damage from tree pitch or because of apparent intermediacy, this individual was the one taken. The surviving parent typically remained to attend to the eggs or young. The major collection studied was that obtained by the first author in Oregon, California, and Nevada (223 specimens). We also had available for compari- son the 14 adult specimens taken in Modoc County, California in 1948 and 1949 by Howell (1952) and the other breeding specimens he mentioned from the Warner Mountains, Modoc County and from the Sweetwater Mountains and Glass Mountain, Mono County. These specimens are all stored in the Mu- seum of Vertebrate Zoology, where many breeding specimens of both forms from regions of allopatry also are available. A total of 499 breeding specimens of both forms and their hybrids was examined. RESULTS HYBRID INDEX SERIES All breeding-season specimens of daggetti and nuchalis, whether they are from regions of al- lopatry or from the more limited areas of sym- patry, can be categorized according to a "hy- brid index series" (Anderson 1949) based on their head patterns. Other differences between these taxa, in degree of spotting and barring, are too slight and too influenced by wear to be of value in distinguishing breeding specimens. Thirteen categories were identified (Fig. 1). These form a visually graded series from the reddest daggetti (score of 0), i.e. specimens with very limited and subdued black-and-white head striping and a very faint shadow of a breast band, to the most extreme examples of nuchalis (score of 12), in which the red is confined to the throat, crown, and nuchal patch, the black breast band is wide, and the black-and-white head striping is well defined. The 13 specimens representing each of the hybrid index scores formed a reference series with which all other Fig. 1. Illustrations of specimens representing the 13 categories (scores 0-12) of the hybrid index series between the Red-breasted and Red-naped sapsuckers. Typical daggetti fall into scores 0-2; typical nuchalis fall into scores 11 and 12. Parental Red-breasted Sapsuckers, especially females, rarely have scores of 3 and parental Red-naped Sapsuckers, especially males, rarely are scored as 10. specimens were compared for determination of their scores. The plumages of the two forms are well de- scribed in the literature (Howell 1952: 240-245, Devillets 1970, Dunn 1978), and only a brief summary is given here. The differing head pat- terns of daggetti and nuchalis are based simply on an interplay between varying proportions of carotenoid and melanin pigments. When ca- rotenoid pigments dominate the barbs of the head feathers, especially their tips, the relative amount of melanin present is reduced and the daggetti phenotype is seen. Conversely, when black melanin pigment is packed into the feathers of the breast and those of the sites of the black head stripes, the nuchalis phenotype emerges. Although only the feathers of the sites of the white head stripes and those of the throat and crown never have black pigment, all sites can have red pigment, especially in the barb tips, except for small lateral stripes at the base of the upper mandible that always remain whitish. However, even in the reddest individ- uals of daggetti, a reduced quantity of melanin pigment still is present as a narrow dark band in the middle and basal portions of the feathers of the head and breast, forming an underlying "shadow" that is reminiscent of the melanistic breast band, cheek patch, and head stripes seen so vividly in nuchalis. This is easily demonstrat- ed by clipping the tips off the barbs of feathers in the relevant areas of the head and breast. This point is of special significance because progressive wear of the head and breast feath- ers during the nesting season, caused primarily in adults when they squeeze through the nest opening, exposes increasing amounts of black- ish or dusky color and causes normal daggetti to be identified incorrectly as hybrids. The pre- sumption of all earlier authors that hybridiza- tion between daggetti and nuchalis is very com- mon has been based on the routine underestimate of the amount of black normally hidden below the red barb tips of the head and breast feathers in daggetti. This problem is per- petuated in even the most recent field guides (National Geographic Society 1983: 269). Because of the simple relationship of carot- enoids and melanins in the expression of col- oration and pattern of the head and breast plumage in daggetti and nuchalis, it was unnec- essary to use a scheme in which scores for "sep- arate" characters (e.g. amount of black in breast band or redness of head plumage) were summed to yield a total hybrid index score for individual specimens. Such an approach is more suitable for studies in which a complex array of characters vary independently. Short (1963) provides an example of such a study in war- biers. Genetic control of head and breast coloration.--It is well established that carotenoid and melanin pigments in feathers differ in their origin (Fox 1976), the former being manufactured by plants and subsequently ingested either directly or indirectly by birds and the latter being synthe- sized by the bird itself. Despite their indepen- dent origin and their probable separate genetic control in all birds, including Sphyrapicus (Howell 1952: 244), these major categories of pigments appear to interact in a simple manner in the sapsucker taxa under consideration. When carotenoid pigment is present in the out- er barbs of a given feather of the head or breast, melanin is confined to a medial position and forms a narrow band. Conversely, when the outer barbs are filled with melanin, carotenoid pigment apparently is absent. At least to the eye, therefore, the outer barbs are filled with either carotenoids or melanins, but not both. However, the feathers of the head and breast have not yet been examined microscopically to determine if the two pigments occur together at the same sites in individual feathers, with one pigment partially or completely masking the other. This condition has been reported in other birds (Brush 1970, Lucas and Stettenheim 1972: 401-402). The extent of both carotenoid and melanin pigmentation in the feather barbs of the head and breast shows a pattern of continuous vari- ation from nuchalis, through the hybrids, to daggetti. Reddish carotenoid pigments gradually subdue the black-and-white head striping and black breast band of typical nuchalis until the head and breast are completely red, as in dag- getti. Continuously varying traits can be con- trolled by a system of additive polygenes. Be- cause carotenoids and melanins probably are under independent genetic control, two dis- tinct series of genes may play a role, one for the deposition of each category of pigment. We suggest further that these two independent se- ries of additive genes may in turn be controlled by a regulatory gene or switch gene that dic- tates the relative amount and position of each pigment type, so that when the carotenoid se- ries is switched on, the melanin series is switched off, and vice versa. Such a regulatory system could facilitate the special interactive mode of color expression already described. We speculate also that the number of gene pairs in each hypothesized series is approximately six. This figure is based on our ability to identify 13 intergrading phenotypic classes of pigmen- tation among the specimens examined. Thir- teen is the number of phenotypic classes of off- spring expected when both parents are heterozygous for the trait in question and when six additive gene pairs are operative (Winches- ter 1977: 196). Admittedly, this genetic model is tentative and exploratory; as such it is offered primarily for its possible heuristic value. But in the absence of the refined breeding data nec- essary to determine the genetics of color traits in sapsuckers, further speculation seems un- warranted. ZONE OF SYMPATRY AND -YBRIDIZATION The breeding distribution of daggetti is fairly continuous through the Sierra Nevada, coast ranges, and mountains of south-central Oregon (Fig. 2). In contrast, in the western portion of its nesting range, nuchalis forms scattered pop- ulations in the mountains of the Great Basin. In this general region, areas of contact between the two species are interrupted and of limited extent, the principal one being in Modoc County, California and in adjacent Lake and Klamath counties, Oregon (Crowder Flat-Steele Meadow, South-Central Oregon, and Warner Mountains in Fig. 2). A fairly local area of sym- patty, where definite hybrids are not yet known, occurs in the vicinity of Desert Creek (Lyon County, Nevada and Mono County, Cal- ifornia) and in the Wassuk Range, Mineral County, Nevada. Five other sites are known where one species is common and the other is rare (Fig. 2). These sites occur along the eastern side of the Sierra Nevada in Sierra County, Cal- ifornia; in the Sweetwater Mountains and on Glass Mountain, both in Mono County, Cali- fornia; and in the Sheep and Spring moun- tains, Clark County, Nevada. Therefore, dag- getti and nuchalis are essentially allopatric in this region. Specimens and sight records from the afore- mentioned areas were plotted in relation to their hybrid index scores (Fig. 2) and tested (Kolmogorov-Smirnov test) for possible differ- ences in the estimation of intrapopulation phe- notypes using specimens vs. sight records. For 4 populations, the distributions were signifi- cantly different: South-Central Oregon (dm. = 0.305, P < 0.01), Crowder Flat-Steele Meadow (dm. = 0.240, P < 0.01), recent Warner Moun- tains (d. = 0.270, P < 0.01), and Desert Creek- Wassuk Range (d. = 0.243, 0.01 < P < 0.02). In 2 samples, the distribution of specimens was not significantly different from that for sight records: Sierra County (d. = 1.10, P > 0.2) and Sweetwater Mountains (d. = 0.075, P > 0.2). Presumably, the statistically significant differ- ences arose in the 4 samples because of oppos- ing sources of bias in the two kinds of records. Because we attempted to collect odd-appearing birds to document the range of variation pres- ent in natural populations, the specimen rec- ords probably are slightly biased toward hy- brids. In contrast, because the subtle hybrid features of some presumptive backcrosses and recombinant types are not always easy to detect in the field, it is likely that the sight records are slightly biased toward pure parental forms. Three main points are evident from the his- tograms of Fig. 2. First, the overwhelming pro- portion of both specimens and sight records are of normal parental phenotypes of either daggetti or nuchalis; definite hybrids are rela- tively scarce. Second, the populations of the Warner Mountains have changed remarkably since the specimens were taken there in 1910 (Kolmogorov-Smirnov test, recent vs. 1910 rec- ords, d. = 0.290; P < 0.01). In the early sam- ple from that region, nuchalis was by far the most common form. Only 2 of the 32 specimens taken in 1910 were daggetti. A few hybrids also were taken. Since 1964, when the first samples for the present study were collected, specimens of daggetti have turned up regularly in the War- ner Mountains. Furthermore, almost 100 sight records obtained in the last 20 yr establish that the Red-breasted Sapsucker is now a common species in the Warner Mountains, where it oc- curs in local sympatry with the Red-naped Sap- sucker. It is of interest that the shape of the histogram for the Warner Mountains in 1910 resembles that for the mountains of southern Nevada. Indeed, they are not different statisti- cally (Kolmogorov-Smirnov test, d. = 0.11; P > 0.2). Third, the histograms and distribu- 0 I 2  4 5 6 7 8 9 I0 II 0  2O Crowder Flat-Steele Mdw., Modoc Co.,Cali[ 47 33 0 I 2 3 4 5 6 7 8 9 I0 II 12 Warner Mtns.,Ore.-Calif. (recent) Warner Mtns,Ore.-Calif. (1910) 32 40 0 80 160 km 0 I 2 3 4 5 6 7 8 9 IO I I 12 ' Co.,Calif. rra 36 31 0 I 2 3 4 5 6 7 8 9 I0 II 12 Sweetwater Mtn.,Mono Co.,Calif. 40 31 Glass Min.,Mona Co.,Calif. 34 3 0 I 2 3 4 5 6 7 8 9 10 II 12 Deser, C and Wossuk Ronge,Nov. 33 IO Sheep and Spring Mtns., Nev. ir5 ____,___ 12. 0 Zone of sympatry and hybridization ß S.r. dagget# commonl ,.,q. nucha#$ rare + S, nuchali$ common;S.r. daelt/ rare X AIIopotric S. ß nuchali$ o tions show that nuchalis maintains large popu- lations at the extreme Western periphery of its breeding range that are in contact with daggetti. This is true despite the major gap of habitat unsuitable for any sapsuckers that runs north to south in western Nevada. This gap isolates these sympatric populations from the main dis- tribution of nuchalis in the mountains across the Great Basin to the east. A prominent gap, Owens Valley, also isolates the allopatric pop- ulations of nuchalis in the White Mountains from the range of daggetti in the Sierra Nevada. In the White Mountains, nuchalis does not seem to be influenced by the latter species; neither hybrids nor typical daggetti have ever been re- corded there. PHENOTYPIC VARIATION Variation in regions of allopatry.--To assess variability in the regions of sympatry and hy- bridization, it is necessary to understand the breadth of normal variation shown by each species in regions of allopatry, away from the influence of the other species. This is demon- strated by the plotting of hybrid index scores of allopatric breeding specimens of 126 daggetti and 88 nuchalis in Fig. 3. Two main points emerge. First, parental populations of each species show a substantial range of variability. Scores for typical daggetti ranged from 0 to 3 (g = 0.97) and those for typical nuchalis from 8 to 12 (œ = 11.44). Second, the sexes differ with- in pure daggetti (2 of males = 0.58, g of fe- males = 1.47) and within pure nuchalis ( of males = 11.23, g of females = 11.77). Thus, the reddest daggetti (specimens with very low scores) are predominantly males and the black- est nuchalis (specimens with very high scores) are mostly females. Stated another way, within pure daggetti most scores of 2 and all of 3 are of females, and within pure nuchalis most scores of 11 and both the single 10 and single 8 are HYBRID INDEX SCORES-ALLOPATRIC SAMPLES 0 I 2 3 4 5 6 7 8 9 ,0 II 12 1 F ø rdaet# ,S'. nuchali$ .97mecn -- II 44rnec]n score score 30- -30 =[] females z 0 0 0 I 2 3 4 5 6 7 8 9 I0 II 12 Fig. 3. Hybrid index scores of breeding speci- mens of daggetti from allopatric localities in Oregon, California, and western Nevada and of nuchalis from allopatric localities in Oregon, California, Nevada, Idaho, Montana, Utah, and Arizona. of males. Within each species, the difference between the sexes in average score was highly significant (Kolmogorov-Smirnov test; for dag- getti, d .... = 0.438, P < 0.01; for nuchalis, dm. = 0.469, P < 0.01). This suggests the presence of a consistent sexual dimorphism within each species that heretofore has not been appreci- ated. Variation in the zone of sympatry and hybridiza- tion.--It is of special interest to determine the phenotypes of mated birds within the zone of sympatry and hybridization. The hybrid index scores of 145 males and their 145 female mates from this zone are plotted in Fig. 4. Kolmogo- rov-Smirnov tests were used to determine the proportion of typical parental specimens of each hybrid index category to be expected in the samples from the zone of overlap and hybrid- ization based on the frequencies seen in the allopatric samples (Fig. 3). Proportions of allo- pattic and sympatric populations were not sig- nificantly different in 3 of 4 comparisons (d .... ranged from 0.077 to 0.117, P > 0.2), those for Fig. 2. Breeding distribution of Red-breasted and Red-naped sapsuckers in portions of Oregon, northern and central California, and in Nevada. The outline in southwestern Oregon, extreme western Nevada, and California describes the essentially continuous nesting range of S. r. daggetti, exclusive of populations in the mountains of southern California. Islands of forest and woodland in south-central Oregon, eastern California, and Nevada, containing the known breeding stations of S. nuchalis (circles), also are outlined. The histograms depict the hybrid index scores of specimens (black bars above the line) and of sight records (striped bars below the line) for nine large samples of the two species and their hybrids from eight restricted geographic regions. Hybrid Index Scores of Sexes of Mated Birds in Zone of Sympatry and Hybridization. 0 I 2 5 4 5 6 7 8 9 I0 It 12 n = 145 MALES 40.7ø/ø , r, daetti  nuchalis n=145 FEMALES 40.7% 5177% Fig. 4. Hybrid index scores of typical daggetti and nuchalis from the zone of sympatry and hybridiza- tion. This analysis includes only birds known to be mated and established for breeding. Shaded bars rep- resent "pure" parental types of either species and clear bars are assumed to be hybrids. female daggetti and for both sexes of nuchalis. However, in males of daggetti, allopatric and sympatric frequencies differed significantly (d .... = 0.290, P < 0.01). The disagreement re- suited from the relative numbers of specimens of hybrid index scores 0 and 1. Zero was a more common score in allopatric than in sympatric populations; the reverse was true for a score of 1. Although the meaning of this difference is unclear, we suspect that it is an artifact of col- lecting date. In general, specimens were taken later in the season in the zone of sympatry and hybridization than in the regions of allopatry, simply because of the fortuitous scheduling of field trips. Thus, the sympatric specimens ap- pear slightly more worn and, therefore, have slightly higher hybrid index scores than those from allopatry. Despite this single instance of nonconcordance, the frequencies of most cate- gories were similar enough according to the Kolmogorov-Smirnov tests to justify the con- clusion that the proportions of typical parental types within the zone of overlap and hybrid- ization were correctly identified. Eighty-eight percent of the males and 92% of the females are either typical daggetti or typical nuchalis (Fig. 4). Thus, only 10% of all mated birds are hybrids. Oddly, 7% more daggetti males than females and 11% more nuchalis females than males were involved in these matings from the zone of overlap. However, this difference only approaches statistical significance (X21 = 2.55, 0.100 < P < 0.250). It is of interest that more males (11.7%) than females (7.6%) of the mated birds are presumed hybrids. Scott et al. (1976) also report an excess of male interme- diates between S. r. ruber and S. nuchalis in their zone of sympatry and hybridization between Kersley and Bouchie Lake, British Columbia. Although hybrid males may have a slight ad- vantage over hybrid females in obtaining mates, the reason for the excess is unclear in either instance. However, John Endler (pers. comm.) has suggested that the excess of males may re- suit from the operation of Haldane's Rule (Maynard Smith 1975), in which hybrid break- down results in greater mortality of the hetero- gametic sex (females in birds). Finally, we compared allopatric (Fig. 3) and sympatric (Fig. 4) frequency distributions for both species with the sexes combined (Kol- mogorov-Smirnov test). The distributions were significantly different (dma. = 0.165, P < 0.005). MATE PREFERENCE IN THE ZONE OF $YMPATRY AND HYBRIDIZATION Mate selection in a zone where diverse phe- notypes interact sympatrically can provide cru- cial evidence bearing on gene exchange, repro- ductive isolation, hybrid fitness, and the systematic status of closely related forms. For daggetti and nuchalis, we have such data for 145 nesting pairs from the principal zone of inter- action in south-central Oregon and northeast- ern California. The mate preferences of these 290 individuals are illustrated in a matrix of hybrid index scores for the sexes of individual pairs (Fig. 5). These data are categorized by mating type in Table 1. Most pairs (110 = 75.9%) mated in a positively assortative manner, i.e. typical daggetti mated to typical daggetti or typ- ical nuchalis mated to typical nuchalis. In most pairs involving a hybrid, the hybrid individual backcrossed with one of the pure parental types (24 matings = 16.6%). Hybrid individuals paired with other hybrids only rarely, in 2 cases (1.4% of the matings). Most interestingly, 9 of the matings (6.2%) were interspecific. These are shown in Fig. 5, either in the upper left corner (8 matings of typical nuchalis females with typ- ical daggetti males) or in the lower right corner (1 pairing of a typ!cal nuchalis male with a typ- ical daggetti female). This asymmetry will be discussed later. Because of the large number of cells in Fig. 5, for purposes of statistical testing we col- lapsed the data into a 4 x 4 contingency table consisting of hybrid index scores 0-3, 4-6, 7- 9, and 10-12. These categories maintain the separation of typical daggetti (scores 0-3) from typical nuchalis (scores 10-12) and from mat- ings involving F hybrids, backcrosses, and re- combinants (scores 4-6 and 7-9). The data are very highly associated or nonrandom (X29 = 100.094, P < 0.005). As would be expected from an examination of Fig. 5, the four corner cells of the contingency table contributed heavily to the very high total x 2 value (18.96, 18.06, 25.09, and 23.01). FITNESS OF HYBRIDS Information on the survival of hybrid off- spring, their relative success in eventual mat- ing, and their production of young is central to interpretations of the status of hybrid zones and of the evolutionary significance of these zones to the taxa in question. Although such data are essentially lacking for the present ex- ample of hybridization in Sphyrapicus, certain useful inferences on hybrid fitness can be gained from further analysis of the phenotypes of mated adults (Fig. 5, Table 1). Twenty-eight hybrids (9.7% of the total population) were re- corded. These fell into two groups: (1) probable F hybrids, or those individuals with hybrid index scores of 5, 6, or 7; and (2) various recom- binants resulting from backcrossing. Ten indi- viduals were presumed to be F hybrids and the remaining 18 hybrids were presumed to be re- Phenotypes of 145 Mated Pairs of $phyropicus in Zone of Sympatry and Hybridization. 12 4 I 3 3 I I I 2 2 i!.16 14 15 II I I I I 37.2%/....*' !'l 12 5  6  2  I  5 u 4 I m 7 30 daetti o ) m ) Score of Males i. 5. atix o matin 2eeence amon mated 2ais o Red-beasted and Red-na2ed sa2suck- es in te zone o sym2at and ybJdization. Te small suaes in te toe left and u22e it co- nes o te daam encompass te scoes o mates uded to be ete ue ett o ue cJs. Be- dcat emales ane om 0 to 3 but tycat males ane om 0 to 2. combinants. [Some backcross matings would be expected to produce phenotypes with hybrid index scores ranging from 5 to 7, but at a low frequency, especially if 3 or more pairs of genes are involved in the control of the traits upon which the scores are based (see p. 5)]. It is of immediate interest to determine if the number of F hybrids present (10) is in agree- ment with the expected number of surviving offspring derived from the frequency of inter- specific crosses occurring (9) in the sample from the zone of overlap and hybridization. Unfor- tunately, data on the relative survival of young from either normal conspecific crosses or from interspecific matings are unavailable. Al- though Howell (1952) presented data from a few interspecific nests that suggested that F fledging success was fully comparable to that from conspecific pairings, a much larger sam- ple is necessary to prove this point. Further- more, we lack information on the welfare of surviving F offspring approximately 1 yr after TABLE 1. ization. Types of matings represented by 145 nesting pairs of Sphyrapicus in zone of sympatry and hybrid- Taxa Number of pairs Percent of total Type of mating S. r. daggetti x S. r. daggetti 56 38.6 Conspecific S. nuchalis x S. nuchalis 54 37.2 Conspecific hybrid x  S. r. daggetti 5 3.5 Backcross hybrid x  S. r. daggetti 2 1.4 Backcross hybrid x  S. nuchalis 4 2.8 Backcross hybrid x  S. nuchalis 13 8.9 Backcross F hybrid x F hybrid 2 1.4 Hybrid S. nuchalis x  S. r. daggetti 8 5.5 Interspecific S. nuchalis x  S. r. daggetti ! 0.7 Interspecific Total 145 100.0 hatching, when they are competing for mates and attempting to raise young of their own. In the absence of such information, we can infer relative production of offspring and their sur- vival from the proportions of the various phe- notypic classes observed in our sample of 290 birds. As a working null hypothesis, assume that the 9 interspecific pairs fledged as many young as conspecific pairs and that the proportion of these young surviving to breed ! yr later was comparable to that from typical matings. As- suming an average clutch size of 4.5 eggs, a reasonable figure for a variety of species of pi- cids (Cody !97!: 501), and 25% survival of young from these 40.5 eggs to the first year of breeding [a figure based on long-term demo- graphic studies by Walter D. Koenig of the Acorn Woodpecker (Melanerpes formicivorus), the only species of woodpecker for which such sur- vivorship data are available], then !0.125 off- spring would be expected from 9 normal pairs of sapsuckers. Considering the many sources of iraprecision in estimates such as these, this fig- ure is remarkably similar to our finding of !0 F hybrids, and the null hypothesis cannot be rejected. Thus, based on these figures, there is no evidence for either F hybrid inferiority or superiority. The interspecific matings account for only 10 of the total of 28 hybrids observed. The re- maining !8 hybrids evidently resulted from backcrosses and intermating of hybrids. Thus it also is useful to estimate the total number of hybrids expected from the number of back- crosses (24) and hybrid matings (2) actually ob- served (Table 1). Because parental phenotypes are produced in addition to hybrids in both backcross and hybrid matings, calculation of expected numbers of different phenotypes is more complex than for interspecific matings, in which all offspring are assumed to be inter- mediate F individuals. The calculation of ex- pected numbers of hybrid and recombinant phenotypes vs. parental phenotypes, for either backcross matings or matings of F hybrids, also requires estimates of the number of gene pairs involved. Earlier, we estimated that six pairs of genes control degree of redness of the head and breast in these sapsuckers. For backcross matings, the formula 0.5", where n represents the number of gene pairs, can be used to compute the expected propor- tion of parental phenotypes. For six gene pairs, !.56% of the offspring would be expected to be parental phenotypes for backcrosses to pure parents homozygous at all six loci (character states 0 and !2). An additional 0.292% of the offspring from backcrosses to character states ! and 11 (heterozygous at a single locus), also assumed to be pure parents in this study, would represent parental phenotypes. Thus, most (98.15%) of the offspring would be hybrid or recombinant phenotypes. For F hybrid matings, the formula 0.256 is appropriate for the calculation of the expected proportion of either extreme parental pheno- type, again assuming that six gene pairs control the red color difference between daggetti and nuchalis. Thirteen phenotypic classes would be expected. Each parental phenotype therefore would occur at a proportion of only 0.0244%. It is clear that under the assumption of six gene pairs, for either backcrosses or hybrid matings, the expected fraction of parental phenotypes in the offspring is so low that it safely can be ig- nored in our calculations. Even if only three pairs of genes control the trait, only 17.2% of backcross offspring and 3.125% of offspring from F hybrid matings would be expected to represent the parental phenotypic classes. The question, therefore, is whether the figure of 18 hybrids or recombinants produced by 26 back- cross and hybrid matings is comparable to what might be expected from an equivalent number of conspecific matings. Again assuming a clutch size of 4.5 eggs, 117 eggs would be expected from these 26 matings. At a survival rate of 25% for the first year, 29.25 1-yr-olds would be ex- pected in the population. Our figure of 18 hy- brids and recombinants is, therefore, low com- pared to expectation. Furthermore, it also is lower than expected when we realize that we have calculated the number of surviving off- spring expected for only a single year. Hybrid and recombinant matings would produce phe- notypes that would be represented for several years in the population. This is in contrast to the interspecific crosses that annually produce only one phenotypic class of offspring, the F hybrids. Therefore, the representation of phe- notypes of hybrids and recombinants in the sample of 290 individuals is much lower than expected. This suggests that survival, and thus fitness, of hybrids is reduced after 1 yr of age. Finally, the apparent greater mortality of fe- male than of male hybrids, possibly resulting from Haldane's Rule, is further evidence for hybrid breakdown. DISCUSSION AND CONCLUSIONS Models proposed to explain hybrid zones.--Moore (1977) reviewed three alternative hypotheses offered to explain hybrid zones in vertebrates. In (1) the "ephemeral zone hypothesis," hy- bridization after secondary contact leads either to complete reproductive isolation (speciation) or to merger (swamping) of the populations. Speciation is probable if the populations have diverged sufficiently so that the parental phe- notypes are more fit than the hybrids; fusion is probable if the reverse is true. Sibley (1957) illustrated both of these outcomes with inge- nious diagrams. In (2) the "dynamic equilibri- um hypothesis," the standard interpretation of hybrid zones (Mayr 1963: 378-379), a balance occurs between gene flow into the hybrid zone, from the extensive areas of allopatry of the pa- rental forms, and selection against hybrids. Bigelow (1965), who has championed this view, felt that introgression of genes from one form into the other would be suppressed by steep gradients of selection on each side of the nar- row hybrid zone and that the development of mechanisms opposing hybridization might be countered by the influx of parental types into the comparatively narrow zone of overlap. Bar- ton (1979a, b, 1980) has developed this hypoth- esis in detail. In addition (3), Moore (1977) of- fered a hypothesis that envisions the superior fitness of hybrid phenotypes over those of either parent in ecologically intermediate hy- brid zones. Finally, Endler (1977) proposed models that allow for differentiation in para- patry by contiguous populations positioned along environmental gradients. Although the results of allopatric vs. para- pattic differentiation cannot be distinguished, certain characteristics of the overlap-hybrid zone described here in Sphyrapicus allow us to discuss the merits for this example of the var- ious hypotheses. Three features of the sapsuck- er hybrid zone are pertinent to the choice of a working hypothesis. First, at least in the short term, the zone is certainly not ephemeral; hy- brids have been known for over a century (Howell 1952). Nonetheless, a zone that gen- erates hybrids for a century could still be ephemeral over a millenium. Second, although parental phenotypes predominate in the hy- brid zone, F hybrids seem to enjoy equivalent viability, and their occurrence is in proportion to the frequency of interspecific matings. We interpret this as evidence that F hybrids are neither selected for nor against (although some of their genes and recombinants may be). F2 hybrids and recombinants, however, do seem to be at a disadvantage. Whether they are at enough of a disadvantage to prevent introgres- sion is unknown. Third, the zone of inter- breeding straddles a steep environmental gra- dient between coastal warm-wet and interior (continental) cool-dry climates (Johnson 1978), which is reflected in the biota present. These features of the hybrid zone in Sphyrapicus would seem to rule out a short-term ephemeral-zone model. Moore's hybrid-superiority model also seems poorly suited to these data; there is no evidence from which to conclude that hybrid individuals are of greater fitness than parental types. We are left with the dynamic-equilibrium model (stable-zone hypothesis), the assump- tions of which are in general agreement with all of the foregoing data. In the sapsucker ex- ample, the "ecotonal" nature and position of the hybrid zone on a climatic gradient, rather than enhancing the fitness of hybrids as ex- pected by Moore's model, seem instead to per- mit coexistence of parental phenotypes in a re- gion compatible with the needs of both parental types and their hybrids. Temporal change in the hybrid zone.--Data on spatial change in hybrid zones are relatively scarce (Selander 1971). Scott et al. (1976) de- scribe the incursion of nuchalis into an area of British Columbia where only S. ruber had nest- ed approximately 25 yr earlier. In the present study we have documented the invasion of daggetti eastward into the Warner Mountains, Modoc County, California, where nuchalis was overwhelmingly the predominant form in 1910. The time of arrival of substantial numbers of daggetti in the Warner region is unknown, but by 1964 both species commonly occurred at several sites (Johnson 1970). The apparently greater numbers of hybrids and of backcross recombinants in the recent Warner sample (Fig. 2) perhaps indicates an increased breakdown of premating isolating mechanisms because of the invasion of daggetti into the formerly ex- clusive domain of nuchalis. Nonetheless, de- spite the apparent increase in hybridization in the Warner region, the two species clearly maintain their integrity. Thus, this document- ed change in distribution may be regarded as evidence against the arguments that there is no selection one way or the other and that the ultimate result of the interbreeding will be complete blending of the two forms. The fac- tors responsible for the eastward expansion of nesting range in daggetti are unknown. Nuchalis also apparently has changed its breeding distribution near the southern por- tion of the zone of contact with daggetti in the present century. The former species was not found in the White Mountains, Mono County, California during the extensive vertebrate sur- veys there in 1917. Yet in 1954, a substantial population of nuchalis was nesting along Cot- tonwood Creek at 2,896 m (Miller and Russell 1956). Premating isolating mechanisms.--Because mat- ing in the zone of sympatry is strongly assor- tative according to phenotype rather than ran- dom, we may assume that some feature of behavior or plumage acts to reduce the num- bers of matings between daggetti and nuchalis. Neither Howell (1952) nor we discovered any differences in display or voice between the two forms; if such exist, they must be subtle. How- ever, it is important to note than no detailed behavioral study of either form has ever been conducted. Presumably, then, the striking plumage differences of daggetti and nuchalis serve as the principal premating isolating mechanisms. Howell (1952: 273) observed that in mixed matings, either of S. r. ruber (or of S. r. daggetti) and nuchalis or of birds of variously interme- diate stock, the male of the pair was the redder individual in all but one instance. This intrigu- ing finding led him to suggest "that the amount of red in the plumage may be a factor in court- ship and mate selection in nuchalis and inter- mediates between nuchalis and daggetti or ru- ber." Our data can be examined in the light of Howell's findings. Of 9 interspecific matings of pure parental phenotypes, the male was dag- getti and the female was nuchalis in 8 instances (Fig. 5, Table 1). Weisser (1973) also reported an interspecific pair of sapsuckers at Lee Vin- ing Creek, Mono County, California where the male was daggetti and the female was nuchalis. Furthermore, in 8 additional matings involv- ing typical female nuchalis and males of scores 3-5, the male was redder. Finally, in 5 exam- ples in which a typical female nuchalis was mat- ed to a male with a score of 7-9, again the male was redder. In 5 backcrosses of a hybrid mated to a pure daggetti, the redder bird was the male. One mating of hybrids, each with a score of 6, was equivocal. However, in 2 backcrosses (male 6 x female 3, male 12 x female 6) the female was the redder mate. In sum, of 36 matings, apart from those involving typical conspecifics, in 33 the male was redder, 1 was equivocal, and in 2 the female was the redder individual. Thus, Howell's hypothesis is strongly supported (Sign test, P < 0.001). Parenthetically, it is of interest to note that in the related species S. varius, in which the male has a red throat and the female a white throat, again the male is the redder sex. However, because most nuchalis mate with their own kind, it cannot be simply a matter of the amount of red that determines mate choice. Howell agreed when he stated (1952: 273), "one can hardly expect mate selection between the races to apply so strictly .... "How, then, can these mixed matings be explained? To this end we offer the following speculative scenario that incorporates both the "lust for red" and other factors into an explanation of interspecific pair- ing in these sapsuckers. It is well established that daggetti arrives on the sympatric nesting areas ahead of nuchalis (Howell 1952, 1953). We speculate that the red- dest males of daggetti dominate the less-red males, win the best territories quickly, and are the earliest to gain mates. However, because the sex ratio in both daggetti and nuchalis prob- ably is biased toward males, the common con- dition in birds (Mayr 1939, Welty 1982: 170), some male daggetti that already hold territories would be unmated at the time the first nuchalis females arrive. Some of these paler male dag- getti possess territories before any nuchalis males. A small percentage of female nuchalis choose established male daggetti as mates because of the superstimulus value of the extensive red. These available mates, although paler than many daggetti, are still redder than any nuchalis male, and if these daggetti males hold territo- ries, this gives them the edge some of the time. Thus, most unmated birds in the summer should be males of nuchalis. This assumption receives support from the proportions of mated birds shown in Fig. 4, which reveals a shortage of male nuchalis. Furthermore, 3 unmated in- dividuals of nuchalis taken during the breeding season at the localities marked "+" in Fig. 2, where daggetti is the commoner form, were males. Postmating isolating mechanisms.--Factors that tend to reduce the success of interspecific cross- es once they occur are not easily identified. All 9 of the interspecific nests, all backcross nests, and the single nest in which both parents were F hybrids contained vigorous young. Whether these nests fledged fewer young than conspe- cific nests is unknown. Because 6.2% of all mat- ings in the zone of sympatry and hybridization were interspecific, a significant crop of F hy- brids must enter the population after each re- productive effort. The fact that F, hybrids were neither more scarce nor more common in the breeding population than this degree of inter- breeding would suggest implies average fitness for this group. In contrast, because F2 hybrids and recombinants were less numerous than ex- pected, postmating isolating mechanisms must be operating to reduce the number of young produced by F and backcross matings. Al- though F hybrids appear to be as fully viable as normal adults of either species, they may be at a disadvantage in gaining mates. Being in- termediate in appearance, these individuals may form pairs less successfully than birds of either parental phenotype (Mayr 1963: 106). How- ever, because a significant number of interspe- cific matings occurs between even more dis- similar phenotypes, the plumage differences among hybrids and recombinants may not se- riously influence their mating success. Another possibility is that partial sterility barriers exist in these viable hybrids. Although the data point to the likely presence of such barriers, a study of comparative reproductive success would be needed to shed light on this possibility. Species limits.--We feel that conclusions on species status in hybridizing taxa such as those discussed here should rest primarily on three criteria: (1) the relative numbers of pure paren- tal types vs. hybrids in the zone of sympatry and hybridization, (2) the degree of assortative mating vs. random mating demonstrated, and (3) the fitness of hybrid offspring. In view of the preponderance of pure parental types that are mating assortatively in the zone and be- cause of the relatively infrequent hybridiza- tion and backcrossing, we conclude that the Red-breasted and Red-naped sapsuckers are best treated as biologic species. The inference of reduced fitness of otherwise viable F hy- brids also argues for species status of these taxa. ACKNOWLEDGMENTS We thank the curators of collections at the follow- ing institutions, who either let us study specimens under their care or sent material on loan: Kenneth C. Parkes, Carnegie Museum of Natural History; Thomas R. Howell, Dickey Collections, University of California, Los Angeles; Amadeo M. Rea, San Diego Society of Natural History; Luis F. Baptista and Rob- ert T. Orr, California Academy of Sciences, San Fran- cisco; Richard F. Johnston and Robert M. Mengel, Museum of Natural History, University of Kansas; and Kenneth E. Stager, Los Angeles County Museum of Natural History. Walter D. Koenig kindly shared with us his calculations of survivorship rates for the Acorn Woodpecker. Some of the data were analyzed with the support of Grant DEB-7920694, awarded to the first author from the National Science Founda- tion. The manuscript was read by George F. Barrow- clough, Kendall W. Corbin, John A. Endler, Ernst Mayr, James L. Patton, James D. Rising, and Lester L. Short. Their collective suggestions led to substantial improvement of both the analysis and final exposi- tion. LITERATURE CITED AMERICAN ORNITHOLOGISTS' UNION. 1910. Check-list of North American birds, 3rd ed. Baltimore, Amer. Ornithol. Union. 1957. Check-list of North American birdsß 5th ed. Baltimore, Amer. Ornithol. Union. 1983. Check-list of North American birdsß 6th ed. Baltimore, Amer. Ornithol. Union. ANDERSON, E. 1949. Introgressive hybridization. New York, J. Wiley and Sons. BARROWCLOUGH, G.F. 1980. Genetic and phenotypic differentiation in a wood warbler (Genus Den- droica) hybrid zone. Auk 97: 655-668. BARTON, N.H. 1979a. Gene flow past a cline. He- redity 43: 333-339. 1979b. The dynamics of hybrid zones. He- redity 43:341-359. 1980. The hybrid sink effect. Heredity 44: 277-278. , & G. H. HEWITT. 1981. Hybrid zones and speciation. Pp. 109-145 in Essays on evolution and speciation in honor of M. J. D. White (W. R. Atchley and D. S. Woodruff, Eds.). Cambridgeß Englandß Cambridge Univ. Press. BIGELOW, R.S. 1965. Hybrid zones and reproductive isolation. Evolution 19: 449-458. BROWNING, M.g. 1977. Interbreeding members of the Sphyrapicus varius Group (Aves: Picidae) in Oregon. Bull. Southern Calif. Acad. Sci. 76: 38- 41. BRUSHß A. H. 1970. Pigments in hybrid, variant and melanic tanagers (birds). Comp. Blochem. Phys- iol. 36: 785-793. COD3/4, M. L. 1971. Ecological aspects of reproduc- tion. Pp. 461-512 in Avian biologyß vol. 1 (D. S. Farher and J. R. King, Eds.). New York, Academ- ic Press. DEVILLERS, P. 1970. Identification and distribution in California of the Sphyrapicus varius group of sapsuckers. California Birds 1: 47-76. DOBZHANSKY, T., F. J. AYALA, G. L. STEBBINS,  J. W. VALENTINE. 1977. Evolution. San Franciscoß W. H. Freeman and Co. DUNN, J. 1978. The races of the Yellow-bellied Sap- sucker. Western Tanager 44: 1-4. ENDLER, J. A. 1977. Geographic variation, specia- tion, and clines. Princeton, New Jersey, Prince- ton Univ. Press. Fox, D.L. 1976. Animal biochromes and structural colors, 2nd ed. Berkeley, Univ. California Press. GRINNELL, J. 1901. Two races of the Red-breasted Sapsucker. Condor 3: 12. --, & A. H. MILLER. 1944. The distribution of the birds of California. Pacific Coast Avifauna 27: 1-608. HOWELL, T. R. 1952. Natural history and differen- tiation in the Yellow-bellied Sapsucker. Condor 54: 237-282. 1953. Racial and sexual differences in mi- gration in Sphyrapicus varius. Auk 70:118-126. JOHNSONß N.K. 1965. The breeding avifaunas of the Sheep and Spring ranges in southern Nevada. Condor 67: 93-124. 1970. The affinities of the boreal avifauna of the Warner Mountains, California. Occ. Pa- pers Biol. Soc. Nevada 22: 1-11. 1978. Patterns of avian geography and spe- ciation in the Intermountain Region. Great Basin Nat. Mere. 2: 137-159. ß & R. M. ZINK. 1983. Speciation in sapsuck- ers (Sphyrapicus): I. Genetic differentiation. Auk 100: 871-884. LUCAS, A.M., & P. g. STETTENHEIM. 1972. Avian anatomy-integument, part 2. Agric. Handbook 362: 341-750. MAYNARD SMITH, J. 1975. The theory of evolution, 3rd ed. Baltimore, Maryland, Penguin Books. MAYR, E. 1939. The sex ratio in wild birds. Amer. Natur. 73: 156-179. 1963. Animal species and evolution. Cam- bridge, Massachusetts, Harvard Univ. Press. , & L. L. SHORT. 1970. Species taxa of North American birds. PubL Nuttall Ornithol. Club 9: 1-127. MEISE, W. 1975. Naturliche Bastardpopulationen und Speziationsprobleme bei Vogeln. Abh. Verh. Naturwiss. Ver. Hamburg 18/19: 187-254. MILLER, A.H. 1955. Concepts and problems of avian systematics in relation to evolutionary processes. Pp. 1-22 in Recent studies in avian biology (A. Wolfson, Ed.). Urbana, Univ. Illinois Press. --, & W. C. RUSSELL. 1956. Distributional data on the birds of the White Mountains of Califor- nia and Nevada. Condor 58: 75-77. MOORE, W.S. 1977. An evaluation of narrow hybrid zones in vertebrates. Quart. Rev. Biol. 52: 263- 277. NATIONAL GEOGRAPHIC SOCIETY. 1983. Field guide to the birds of North America. Washingtonß D.C., Nat. Geographic Soc. SCOTT, D. M., C. D. ANKNEY, & C. H. JAROSCH. 1976. Sapsucker hybridization in British Columbia: changes in 25 years. Condor 78: 253-257. SELANDER, R. K. 1971. Systematics and speciation in birds. Pp. 57-147 in Avian biology, vol. 1 (D. S. Farher and J. R. Kingß Eds.). New York, Academ- ic Press. SHORT, L. L., JR. 1963. Hybridization in the wood warblers Vermivora pinus and V. chrysoptera. Proc. 13th Intern. Ornithol. Congr.: 147-160. 1969. Taxonomic aspects of avian hybrid- ization. Auk 86: 85-105. 1982. Woodpeckers of the world. Green- ville, Delaware, Delaware Mus. Nat. Hist. Mono- gr. Ser. No. 4. SIBLE3/4, C. G. 1957. The evolutionary and taxonomic significance of sexual dimorphism and hybrid- ization in birds. Condor 59: 166-191. WEISSIR, W. 1973. A mixed pair of sapsuckers in the Sierra Nevada. Western Birds 4: 107-108. WEkTY, J.C. 1982. The life of birds, 3rd ed. Phila- delphia, Saunders College Publ. WINCHESTER, A.M. 1977. Genetics, 5th ed. Boston, Houghton Mifflin Co. WOODRUFF, D. S. 1973. Natural hybridization and hybrid zones. Syst. Zool. 22: 213-218. SPECIAL ANNOUNCEMENT The Cooper Ornithological Society announces the reprinting of the following Studies in Avian Biology SAB No.2- SHOREBIRDS IN THE MARINE ENVIRONMENT, Edited by Frank A. Pitelka $12.00 SAB No.6 -- ESTIMATING NUMBERS OF TERRESTRIAL BIRDS, Edited by C. John Ralph and]. Michael Scott $25.00 Price includes shipping and handling. Make checks payable to the Cooper Ornithological Society (U. S. funds only) and mail to: Cooper Ornithological Society James R. Northern, Assistant Treasurer Department of Biology University of California Los Angeles, CA 90024 USA