The Dynamics of Two Hybrid Zones in Appalachian Salamanders of the Genus Plethodon

Evdzuion. 46(4). 1992, pp. 930-938

THE DYNAMICS OF TWO HYBRID ZONES IN APPALACHIAN SALAMANDERS OF THE GENUS PLETHODON

/ NELSON G. HNKXON, SR, R. HAVEN WILEY, AND CIEARLES K SMRI’ Department of Biology. University of North Carolina, Chapel Hill, NC 27599 US AND J / KENNERiAK.NEIDEL ,?’ Charlotte Latin School, 9502 Providence Road, Charlotte, NC 28226 UkA t

Abstract.-Two zones of i&gradation between populations of PIethocfon have been studied for 18 and 20 years, respectively. The data consist of systematic scores of colors, made at least twice annually. Near Heintooga Overlook in the Balsam Mountains (Great Smoky Mountains National Park), the salamanders’ cheeks are gray. Proceeding north &ward the Smokies, there is increasing iizquency and intensity of red color at two, four, and six miles. There has been no change in the scores at any location. The width of the zone and our Eaiiure to detect any change can be explained by assuming neutrality of the character and random diffusion during the probable time since contact between the two in&grading forms, which most likely took place after the Hypsithermal Interval, 8,000-5,000 BP. At Coweeta Hydrologic Laboratory in the Nantahala Mountains, Plethodon jordani and P. glutinosus hybridize at intermediate elevations. The lateral white spots of gZutinosus demease and the red on the legs of jordani increases witl+vation &om 685 m to 1,052 m. At the higher elevation, the proportion of animals scored as ~fiure’* joraizni declined si&kantly from 1974 to 1990, an indication that the hybrid zone is qn+eading upward. The rate of spread is too great to be explained by random diffusion, so selection for gZutinosus charac&s is the best explanation. The rate of spread of the hybrid zone indicates that hybridization began 60-65 years ago, * at the end of the time of intense timbering. Such human dishubances have caused hybridization in other organ+Ils.

Key words.-Disturbance, hybrid zone, North Carolina, paleoecology, spreading rate.

Received February 4, 199 1. Accepted November 22, 199 1.

Beginning with the period of the evolu- 1981a, 1981b;.Harrison and Arnold, 1982; tionary synthesis (Mayr, 1942,1963), zones Harrison, 1985, 1986; Arntzen and Wallis, of intergradation have been regarded as im- 199 1; Szymura and Barton, 199 l), as well - portant in understanding speciation and as in theory (Bazykin, 1969; Crosby, 1970; other aspects of the evolutionary process. It May et al., 1975; Endler, 1977; Barton, was recognized that steep clines could not 1979a, 19796, 1983; Barton and Charles- be maintained without either selection worth, 1984; Slatkin, 1973, 1975, 198 1, against-the hybrids or some other selective 1985). This recent literature, reviewed by force, such as that involving ecotones be- Barton and Hewitt (1985), and more exten- tween ecologically di&ent habitats to which sively by Harrison (1990), has stressed the the two intergrading populations were sep- conditions necessary for the stability of hy- arately adapted (Mayr, 1963). Without these brid zones, especially the importance of se- conditions, normal gene flow would spread lection against hybrids. The evidence has the different genes of the two populations included the absence or ratity of F2 and until the cline becomes nearly flat, its slope backcross individuals (Arntzen andWallis, depending on the rate of gene flow and the 199 l), or apparent linkage disequilibrium number of generations since the process be- in the absence of chromosomal linkage, in- gan. volving excess representation of parental In recent years, there have been signifi- ‘genomes in the hybrid zone (Szymura and cant advances in empirical studies (Little- Barton, 199 1). john, 1965; Littlejohn et al., 1971; Little- As noted by Harrison (1990), relatively john and Watson, 1985; Barton and Hewitt, few hybrid zones have been observed for more than brief periods. Most studies thus ’ Present address: Biology Department HP-2, High lack the time dimension, and interpretation Point College, High Point, NC 27261 USA. has involved an assumption of stability. 930 931 Nearly all of the studies that have spanned long periods have involved early and late observations, with few or none between (e.g., Rising, 1983; Moore and Buchanan, 1985; Hillis and Simmons, 1986; Turner, 1971; Amtzen and Wallis, 199 1). Szymura and Barton (199 1) provide an exception in the case of Bombina in eastern Europe, with data extending over a period of nearly 60 Y-- We report observations made twice annually for 18 to 20 years on two zones of intergradation between large Plethodon sal- FICA 1. Map of western North Carolina, showing amanders in the southern Appalachians (Fig. the locations of the two hybrid zones. Solid lines are 1). One of the zones is between two gee- riverx broken lines are state boundaries. graphic representatives of Pkthodon jordani that adjoin one another near the junc- ulation stability, population structure, in- tion of the Great Smoky Mountains and the terspecifrc competition, microdistribution Balsam Mountains of North Carolina. In and individual movements (Hairston, 1973, the Great Smoky Mountains, a range 77 km 1980a, 1980b, 1983a, 1983b; Hairston et long, the P. jordani population consists al- al., 1987; Kneidel, pers. obs.; Nishikawa, most entirely of red-checked individuals. 1985, 1990). For those u&miliar with Where the Smokies and Balsams come to- Pkthodon, we emphasize the fact that all gether, there is a narrow zone of intergra- species are completely terrestrial, with di- dation, beyond which northeastward to the rect development on land, and thus do not limits of the distribution of P. jordani, a migrate to and f?om water. maximum distance of 375 km, no specimens can be found with any red on their METHODS cheeks. Thus, the change between the two The two zones of intergradation were ob- forms is relatively abrupt. The other zone served as exercises,for undergraduate class- involves P. jordani and P. gh&wsu.s in the es. Two or three classes of 15 to 18 students Nantahala Mountains; one of five restricted visited the field sites each year. Most visits areas in which hybrids between these two were in September; one was in June; and species have been found. No hybridization’ threewereinApril.Thecl.&eintheGreat has been detected in most areas of contact Smoky Mountains-Balsam Mountains area between these species throughout the south- was found in 1973 and studied three times ern Appala&ians @I&&ton and Henry, in 1973 and 1974, during which time the 1970; Peabody, 1978). It is the exceptional classiCcation of colors was established. Ob- nature of the hybridization that has caused servations started at Heintooga Overlook in most taxonomists to continue to’ recognize the Great Smoky Mountains’National Park them as separate species (Highton and Hen- and proceeded at odometer-measured two- ry, 1970; Peabody, 1978; Larson, 1984). mile intervals along the service road toward Highton (1983) has recognized the geo- Round Bottom Camping Area (Fig. 2). The graphic representative of gZutinosus in the observations were made between 20~00 and Nantahalas as P. teyahalee. We believe that 23:00 at night, when salamanders are active the designation is unjustified (Hairston, in above ground, so as to avoid disturbing the press). habitat by turning logs and stones. At each These two hybrid zones have been under stop, a class attempted to collect 10 or more our observation for .18 and 20 years, re- specimens of Pkthudon jordani (12 in 1990), spectively, and identical measurements have and with one or two of us then scored the been made at least twice annually for 16 amount of red pigment on the cheeks of years. Other studies of the two species dur- each specimen. We recognized five catego- ing the same period have given us further ries of color. 0 for none, 1 for small spots, information about population density, pop- 2 for a continuous line along the top of the 932 NELSON G. HAIRSTON, SR ET AL.

TABLE 1. Proportional distribution of color of Pfeth- odon jordani along the Heintooga-Round Bottom Road. Data from all field trips combined.

Distance fium Hcintocga (miles) 0 2 4 6 N: 366 373 450 470 Amount of red: 0 0.997 0.86 1 0.120 0.006 : 0.003 0.019 0.121 0.324 0.278 0.215 0.072 3 0.238 0.402 4 0.040 0.304

collected and scored in the same way as at Heintooga, except that two scores for coloration were recorded for each specimen. White spots, a characteristic of P. glutino- su.s, were recorded as 0 (none), 1 (very few spots along the sides or cheeks), 2 (spots FIG. 2A. Contour map of the vicinity of Heintooga. along the sides but none on the back), and Contour interval 1,000’ (305 m). Brolccn line outlines 3 (many spots on the sides and at least a the hybrid zone.’ In this area, Pkfhodon jorduni has few on the back). Red legs, a characteristic not been found below 4,000’ (1,220 m) and is uncom- P. jordani in the Nantahalas, were mon below 4,500’ (1,372 m), thus limiting the length of re- of the hybrid zone. O+n circle: Round Bottom; filled corded as 0 (none), 1 (very small spots of circles: cokcting stopq’x major peak 2B. Profile of red on the forelegs), 2 (red on the forelegs theareaasviewedfromthewestverticalexaggeration but little ifany on the hindlegs), and 3 (the 6: 1. Hatched area: hybrid zonq dotted line: approxi- proximal parts ofboth forelegs and hindlegs mate lower limit of P. jordani. fdly or mostly red). During each visit, stops cheek but little else, 3 if the whole cheek were made according to elevation. We be- was covered, and 4 ifthered color extended gan at 685 m and used an ,aneroid altimeter onto the throat and could be seen from be- to determine each subsequent stop at 91.5 low. All specimens were returned to the for- m (300’) intervals until we reached 1,052 est after examination. Stops were made at m, making a total of five observations. On 0, 2.0, 4.0, ‘and 6.0 miles, at elevations of only two ‘occasions out of 35 did we f%.il to 1,6 17 down to 1,434 m. The direct distance obtain the planned 50 specimens. ‘from the start to mile 6.0 on the road is The methods used to obtain other infor- 4,230 ti. The plan was thus to observe 40 mation on these salamanders are described specimens ori each visit to the area. Qn only in the papers cited above. 3 of 34 visits did we fail to tid the requisite OBSERVA~ONS number, because of early cold weather, and we .never obtained fewer than 32. Two Geographic Representatives of Observations in the Nantahala Moun- Plethodon jordani tains were conducted at the Coweeta Hydrol North of Hemtooga, our studies revealed logic Laboratory of the U.S. Forest Service. no consistent change in scores for coloration This area consists of two main east-facing at any collecting location. Regressions per- watersheds. A forest service road ascends formed on the arcsine percent of each score the mountain along each watershed. Hy- for red at each location on the number of bridization of the two species at Coweeta years since the start of the study revealed was first observed by NGH in 1971 and .a only one statistically significant relationship classification of colors was established in (slope = -0.011 and fl = 0.13). As only 1972-l 974. Our class exercises were carried one specimen out of 360 (0.003) at the Bal- out along BaJl Creek Road, in the southern sams end of the transect had a tiny fleck of of the two watersheds. Salamanders were red on one cheek, and only 3 out of 470 HYBRJD ZONES l-N APPALACHIAN SALAMANDERS 933

T~u 2. Changes in color characters with altitude along the Ball Creek Road, Coweeta Hydrologic Laboratory (38 visits, 1974-1990). Numbers in the body of the tabie are proportions of spe&las gth e&l combination of colors at the respective altitude.

COh

“f”;”m N R3>w R2>w RI >W R-W Wl>R WZ>R W3>R > 1,067 37 0.378 0.297 0.216 0.108 1,052 447 0.340 0.348 0.227 0.075 0.002 961 521 0.155 0.27 1 0.274 0.253 0.038 0.008 869 464 0.069 0.168 0.297 0.331 0.100 0.034 778 339 0.014 0.108 0.280 0.345 0.25 1 685 319 0.012’ 0.046 0.271 0.670 NotrR3>W:Red3,WhiteO:RZ>W:RcdfWhieO,aRed3,Whitel;Rl~W:Redl WhiteO.aRal3.Whi~7,,~R&~~&~- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ o:w2>R:white3.Rcd1,orwhitczRcdo;w3>B:whitc3.Rcdo.

(0.006) at the Smokies end of the transect amanders rated as either 3 red 0 white or 2 lacked red, we are confident that our sta- red 0. white (“pure” jordani) at 1,052 m tions at 0 and 6 miles are close to the edges (3,450’) declinedsign&antly between June of the zone of intergradation, the width of 1974 and September 1990 (df = 33, P < which is thus close to 4,230 m. It is clear 0.000 1, slope = -0.0166, rZ = 0.504). The that if there has been any change in the dis- proportion-rated 3 white b red or 2 white 0 tribution of colors during the 18 years of red (“pure” ghtin+.s) at 685 m (2,250’) our observations, we have failed to detect also declined over the same period (df= 33, it* Table 1 summarizes the ‘proportions of P= 0.13, slope = -0.009, rZ = 0.071). The scores at each location for all years com- data for the latter trend are too variable to bined. give a statistically significant result; so we cannot ,claim that there was a real change Hybridization between P. ghGnosus at the lowest elevation. This analysis shows and P. jordani at Coweeta that the hybrid zone spread at least upward A transect along Ball Creek Road in 1972 during 16 ‘/z years of intensive study. Un- revealed a cline. f?om “pure” ghtinosus to conscious systematic bias of the partici- “pure” jorahi throughout the range‘of el- pants in this study from year to year seems evations between 685 m. and 945 m, above especially unlikely, in view of the consis- which no further change was found (Hair- tency at the other location. ston, 1973). The subsequent detailed anal- In 1971, a transect was made in the ad- yses of this transect have confirmed these jacent watershed along Shope Creek Road initial observations (Table 2). In every year, at elevations from 685 m to 1,281 m. We there has been a gradual change in color did not find a regular change from “pure” with elevation, with the greatest variation glutinosus to “pure”jordani as we ascended. among individuals at intermediate eleva- White spots d ecreased and red on the legs tions, as expected in a hybrid zone. Peabody increased up to an elevation of 915 m, but (1978) reported a transect of collections from there to i,ll3 m the trends were re- along the Ball Creek Road, which were an- versed. Above that elevation, the original alyzed for genetic variation in the seven pro- trend was again found. teins that distinguish between the two spe- Some observations made by one of us cies. The change in &quencies was closely (&UC) in 1977 suggest that environmental similar to the changes in color reported here. factors a&ct the distribution of glutinosqs The summation of scores for colors ob- along Shope Creek Road transect The re- scures an interesting point. Despite the data versal of trends at intermediate elevations being taken by the same people as on the along the road was confirmed. Interpreta- Heintooga-Round Bottom Road, the results tion of this transect is complicated because were not constant from the start of the class the road leaves the moist creekside habitat exercises in 1977 to the present. Regressions to follow a dry ridge at intermediate ele- show two trends, one ofwhich is statistically vations. To separate the effects of habitat significant (Fig. 3). The proportion of sal- and elevation, four new transects each fol- 934 NELSON G. -TON, SR. ET AL.

tal populations. Of these, the first involves the simplest mechanisms and thus consti- tutes an appropriate null hypothesis. The most obvious cause for the postulat- ed separation is climatic change during the Pleistocene or Holocene. Molecular data suggest that the differentiation of jordani into geographic races apparently occurred at, least as early as lower Pleistocene. The red-checked form, currently con5ned to the 1.0’ l ...... e . . Great Smoky Mountains, is as different in . . l . 0 allozymes from the other forms of jordani . l * . . 2 . . . i . .* . as jordani is Corn the local form of the glu- 5 . . tinosus complex (Peabody, 1978). Using . Peabody’s data, and assuming a molecular hi clock, Larson (1984) estimated that the two . 4 . r species separated 2-3 x lo6 years ago. Thus, D.2.I there have probably been many isolations and contacts among the geographic forms 1. o .n. . . . .m. * . * .a. . . . .¶o, of jordani. We need be concerned only with the most recent isolation and contact. FIG. 3. Change in proportions of “pure” Pfethodbn jordani at 1,052 m and “pure” P. glutinosm at 685 m The location of this zone of intergrada- on tlie Ball Creek Transct, Coweeta Hydrologic Lab- tion on a ridge joining higher peaks (Pig. oratory, between 1974 and 1990. Solid symbols rcp- 2B) suggests that isolation on those separate resent 2 identical observations. See text for explanation peaks during a warmer period was followed of “pure,” anda for statistical analysis of the data. by spread along the ridge by both forms. Contact would thus have resulted sometime lowed a single habitat from low to high el- after the begmning of a cooler and moister evation. The scores for coloration along each period The requisite cooler period is, how- trgmsect showed regular changes towardjor- ever, unlikely to have been a glacial maxi- dani at successively higher locations, but on mum. There is substantial geological evi- the dry ridges scores were more like g&i- dence that the timberline was no higher than - nosus than on the transects in wetter habi- 1,300 m as far south as the Great Smoky tats at comparable elevations. Thus gluti- Mountains during the maximum Wisconsin nosustraits increase on dry ridges away from glaciation 18,000-14,000 BP (Clark, 1968; streams for any given elevation. Michalek, 1968). The woodland salaman- ‘. I>IscussrON ders would have been confined below that level, and the mountain ranges would have The Pro&& Duratibn ‘of the Clines been barriers, rather than the refugia that Heintooga-Round Bottom. --The ‘steep- they are now (Deevey, 1949). An alternative ness and apparent stability ofthis cline could possibility is that secondary contact a short be explained in several possible ways distadce North of Heintooga began quite (Moore, 1977; Barton and Hewitt, 1985; recently. It is possible that jordani was con- Harrison, 1990). (1) If the traits involved fined above its present zone of intergrada- are neutral ,or nearly so, the width of the tion during the Hypsithermal Interval be- dine would reflect the rate of gene diEusion tweeti 8,000 and 5,000 BP. The warmer because the two geographic races of P. jor- climate apparently forced the lower limit of da@ caq19 intO contact after a period of sep- the boreal spruce& forest nearly 400 m aration. (2) Hybrids could be confined to a higher than it now exists (Whittaker, 1956), narrow zone of intermediate habitat where and the lower limits of other distributions they have greater fitness than parental types. would also have been raised. As a conse- Or (3) the hybrid zone could represent a quence, the two forms ofjordani could have dynamic equilibrium between hybrid infe- been isolated from each other on the peaks riority and migration from adjacent paren- higher than l,750-1,800 m. With the cool- HYBRID ZONES IN APPALACHIAN SALAMANDERS 935

ing trend after 5,000 BP (Delcourt and 4,230 m, the time is consistent with contact D&court, 1980), the red-cheeked and gray- following isolation during the Hypsither- cheeked isolates would have come into con- mal. tact eventually, on the ridge between Soco As the nearest peaks high enough (2,0 19 and Pin Oak Caps. The null hypothesis m or 1,901 m to the North, 1,919 m to the above assumes random difFusion of genes SOuth) to serve as refuges are 11.9 km and for red cheeks since then. 16.1 km, respectively, from the center of Is this hypothesis consistent with the the zone of intergradation (Fig. 2), it might probable spread of the two populations as be objected that there would not have been they interbred? Under the assumption of sufficient time for the two color morphs to neutrality of characters, the time since the reach each other by random walk. However, two forms came into contact can be calcu- intraspecific competition ‘would presum- lated from the equation derived by Endler ably slow movement compared with the rate (1977): of spread before contact. For example, the present distribution of P. ghtinosus extends T = 0.35(w/L)2 to 240 km north of the maximum extent of where T is the number of generations to the Wisconsin ice sheet, and must have create a cline w meters wide with a rate of reached that limit since the appearance of gene flow of L meters per generation. The forests 14,000 BP. me rate of spread into width of a cline is defined as the distance habitat unoccupied by conspecil%s was thus between the gene frequencies 0.2 and 0.8. 17 m per year. At that rate, the two forms Lower and higher Frequencies constitute the of jordani would have made contact after tails of the distribution, the use of.which at most 900 years, leaving more than 4,000 might invalidate the use of the equation. years to achieve the present width of the Hairston (19833) calculated the mean gen- cline. eration time of P. jbrdani as 9.78 years; and These calculations thus suggest that the the gene flow rate can be estimated from the features ,of this hybrid zone are consistent diffusion constant, D. yith neutral difhrsion following secondary Berg (19 8 3) gives the following equation contact. We cannot conclusively exclude for one-dimensional random walk: other mechanisms in this case. However, s2 = 2 Dt the abundance of all intermediates in coloration in the center of the zone suggests where s2 is the variance in movement per many F2 and backcrossed individuals and time unit; D is the dif&sion constant; and argues against any heterozygote disadvan- t is time. The variance in movement found tage. In addition, there is no readily appar- by Nishikawa (1985, 1990), using a mark- ent environmental gradient or trough in release-recapture method over an entire population densities that might constrain season of activity,.and averaged for four age the location of this zone. We thus conclude and sex classes was 2.9 m for average in- tentatively that these two forms of P. jor- tervals of 17.2 days, or 117.85 of the active dani exemplify neutral or nearly neutral dif- season of 135 days, as determined by Hair- fusion since contact some 4,000 years BP. ston (pers. obs.). Solving, D is 0.084 m per Coweeta.-Although the cline at Hein- day, or 11.38 m per season of activity. Gene tooga and our failure to detect any change flow per generation would thus be 111.3 m. in it can be accounted for by the assump- We can obtain maximum estimates for w tions of neutral diflitsion, such is not the and T if we assume that the cline extends case at Coweeta. The upward spread of the the entire way from 0 to 6 miles, the direct zone of hybridization has been much too distance between which is 4,230 m. In that rapid to have been accomplished solely by case, the solution to EndleTs equation gives movements of the salamanders. The obser- the time required to reach the present width vations ofNishikawa (1985,199O) show that of the zone of intergradation as 4,950 years. glutbzosus is even more sedentary than jor- Allowingforthe~~thatthellnknownwidth dani. me mean‘distance between captures from gene frequencies between 0.2 and 0.8 was 1.095 m, and ranged from 0.8 1 m for (Endler’s requirement) is surely less than two-year-olds to 1.69 m for adult females 936 N-&SON G. HAIRSTON, SR. ET AL.. over a mean period of 14.4 days. Variances ton and Henry (1970). Records at the Cow- ranged from 0.245 to 3.15, and averaged eeta Hydrologic Laboratory show that al- 1.73 m. Extrapolation of the regression in though settlement of the area began as early Figure 3 gives 1997 as the year in which the as 1840, heavy lumbering occurred mostly proportion of “pure” jurdkzni at 1,052 m is in the period between 1903 and 1923. More expected to have declined to the overall av- exact estimates of the coincidence are not erage at 96 1 m. The direct distance between possible because we do not know the alti- points at elevations of 961 and 1,052 m is tudinal width of the zone of overlap prior 477 m, which would require the average to hybridization. It is quite likely that the glutinosus 854 years to cover by the as- hybrids found between Plethodon glutino- sumption ofrandom walk. This time is about sus and other isolates of P. jordani are of 43 times longer than the observed change equally recent origin. Recent hybridization predicts. The assumption of neutrality is between well-differentiated species of plants therefore rejected, and we conclude that the (oaks, Muller, 1952); (fish, Hubbs, 1955), glutinosus characters have increased their (snow geese, Cooke et al., 1988) and other proportion at the higher elevation as a result groups have also been associated with hu- of either hybrid superiority or a selective man disturbance. advantage for glutinosus. These analyses have demonstrated that Our studies in the Shope Creek watershed in spite of a superhcial similarity, the two also suggest that selection might influence clines that we have studied have different the distribution of glutinosw characters in origins and very di&rent dynamics: Care- this area. In the hybrid zone, the drier ridges fully repeated observations over many years favor gZutinosw traits, which are found at were essential to the demonstration. higher elevations- there than elsewhere. These results are consistent with the general distribution of the two species, withjurdani We thank successive officials of the Great on higher, cooler, and moister mountains, Smoky Mountains National Park for per- and glutinosw in lower, warmer, and drier mission to carry out the study at Heintooga, valleys. and we thank especially D. DeFoe for show- Thus, natural selection in the Nantahalas ing NGH the road and other cooperation. appears to explain the upward spread of hy- We thank Dr. W. Swank for permission to brid traits at a rate faster than the animals work at Coweeta. The O’Bryan family kind- themselves could move. An unexplained ly welcomed us to their summer home after feature of these observations is the apparent the night sessions, making the field ‘trips a linkage of color characters with physiolog- pleasure for us and hundreds of students. ical traits that would presumably be subject Some of the classes used the facilities of the to selection. It is a fortunate circumstance Highlands Biological Station. Finally, we for us, be&se without the linkage, detect- thank the many teaching assistants, who pa- ing the effects would not be possible with tiently recorded the scores. the methods that we have used. The rate of spread of giutinosus traits fbr- LlTERATlJREcrrED AIQTIZN,J.W.,ANDG.P.WALLIS. 1991. Restricted thermore suggests that hybridixation itself gene flow in a moving hybrid zone of the newts is of very recent origin. Assuming that the Triturus cristatus and T. manrwmtus in western hybridization began at about the midpoint France. Evolution 45805-826. of the present hybrid zone, which is ap- BARTON, N. H. .1979u. Gene flow past a cline. He- proximately 3,000 m wide at Coweeta, and redity 43:333-339. -. 19796. The dynamics of hybrid zones. He- assuming that the traits spread at the same redity 43:341-359. rate as we have observed, the present width -. 1983. Multilocus clines. Evolution 37:454- of the zone would have been reached in 63 471. years (477 m in 20 years or 1,500 m in 63 BARTON, N. H., AND B. cHAwswoRTH. 1984. Ge- netic revolutions, founder effects, and speciation. years). The calculated time for achieving the Annu. Rev. Ecol. Syst. 15: 133-l 64. present hybrid zone indicates that its begin- BARTON, N. H., AND G. M. HEwrrr. 198 la. The ge- ning coincided approtimately with the end netic basis of hybrid inviability in the grasshopper of intense lumbering, as suggested by High- Podisma pedestris. Heredity 47~367-383. HYBRID Z0NE.S IN APPALACHIAN SALAMANDERS 937

-. I 98 I h. A chromosomal cline in the grass- interactions between species of North Amer&m hopper Pdisma peabtris Evolution 35: 1008-10 18. ~amanders of the genus Pfethodon. Evol. Biol. 4: -. 1985. Analysis of hybrid zones. Annu. Rev. 2 1 l-256. Ecol. Syst. 16:113-148. HruJs, D- M-9 AND J. E. SIMMONS. 1986. Dynamic BAZYIUN, A. D. 1969. Hypothetical mechanism of chaws of a zone of parapatry between two species speciation. Evolution 23~685-687. Of ~~~~idobolus (Acura: Gymnophthalmidae). J. BERG, H. C. 1983. Random Walks in Biology. Prince- HerpetoL 20~85-87. ton University Press, Princeton, NJ USA. HUBBS, C- L. 1955. Hybridization between fish spe. CLARK, G. M. 1968. Sorted patterned ground New ties in nature. syst. zool. 41-20. Appalachian localities south of the glacial border. LARSON, A. 1984. Neontological inferences of evo- Science 161:355-356. lutionary pattern and process in the salamander COOKE , F., D. T. Pm, AND R F. ROCK~FLL 1988. Family Plethodontidae. Evol. Biol. 17: 119-2 17. Evidence of former allopatry of the two color phases LITIUIOI-IN, M. J. 1965. Premating isolaticiii ii~ the of lesser snow geese (Chen caeruiescens caendes- HJ&z avingi complex (Anuraz Hylidae). Evolution tens). Auk 105:467-479. 19~234-243. CRGSBY, J. L 1970. The evolution of genetic discon- Lrrizxrori~, M. J., AND G. F. WATSON. 1985. Hybrid tinuityz Computer models of the selection of bar- zones and homogamy in Austmhan tiogs. Annu. riers to interbreeding between subspecies. Heredity Rev. Eool. Syst. 16:85-l 12. 25253-297. LITTLEJOHN, M. J., G. F. WATSON, AND J. J. DEWEY, E. S. 1949. Biogeography of the Pleistocene. LO-HILLS.197 1. Contact hybridization in the Bull. GeoL Sot. Am. 60~13 15-1416. Cfinia &rev& complex (Anuraz Leptodactylidae). DELOXJRT, P. k, AND H. R. Deu?ouRT. 1980. veg- Aust. J. Zool. 19:85-100. etation maps for eastern North America: 40,000 Yr MAY, R M., J. A. ENDI.=, AND R. E McMuu’rntx. B.P. to the present, pp. 123-l 65. In R C. Romans 1975. Gene frequency clines in the presence of (ed.), Geobotany II. Plenum Press, N-Y., USA selection opposed by gene flow. Am. Nat. 109:659- ENDLER, J. A 1977. Geographic Variation, Specia- 676. tion, and Clines. Princeton University Press, MAYR, E. 1942. Systematics and the Origin of Spe- . Princeton, NJ USA. ties. Columbia University Press, N-Y., USA. ~~AIRSON, N. G. 1973. Ecology, selection and sys- -. 1963. Animal Species and Evolution. Har- tematics. Breviora 414~1-21. vard university Press, Cambridge, MA USA -. 198Oa. The experimental test of an analysis MICHALEK, D. D. 1968. Fanlike features and related of field distributions: Competition in terrestrial sal- phenomena of the southern Blue Ridge. Ph.D. Diss. amanders. Ecology 6 1:8 17-826. UniversityofNorthtkdina,Cha~Hill,NCUSA -. 19806. Evolution under interqx%%c com- MOORE, W. S. 1977. An evaluation of hybrid zones petition: Field experiments on terrestrial salaman- in vertebrates. Q. Rev. Biol. 52263-277. ders. Evolution 34~409-420. MOORE, W. S., AND D. B. BUCHANAN. 1985. Stability -. 1983a. Alpha selection in competing sala- ofthe northern flicker hybrid zone in historical timesz mandersz Experimental ver&ation of an a priori Implications for adaptive speciation theory. Evo- hypothesis. Am. Nat. 122:105-l 13. lution 39:135-151. -. 19833. Growth, survival and reproduction MULIZR, C H. 1952. Ecological control of hybrid- of Pkthoabn jotdank Trade-oflh between selective ization in Quercz A factor in the mechanism of pxsswes. Copeia 1983:1024-1035.. evolution. Evolution 6~147-16 1. -. 1992. The status of PLethodm jordani teya- N~~HXAWA, K. C. 1985. The ecology and evolution hafee Hairston. Brimleyana. In press. of aggressive behavior in two species of terrestrial HAIRSTON, N. G., K. c. NISEIUWA, AND s- L. salamanders. Ph.D. Diss. University ofNorth Car- Swezy 1987. The evolution.of com&+ng olina, Chapel Hill, NC USA temstd salamanders: Niche parhtzoib -. 1990. Intmspeczc spatial relationships of kg or interference? Evol. ECOL 1:247-262. two species ofterrestrial salamanders. Copeia 1990: HARRISON, R G. 1985. .Barriers to gene exchange 41&426. between closely related cricket species. IL Life cycle PEABoDY, R B. 1978. Eleetrophoretic analysis of geo- variation and temporal isolation. Evolution 39:244- graphic variation and hybridization of two. Appa- 259. la&an salamanders, Pkthodon jordani and Pleth- -. 1986. Pattern and process in a narrow hybrid 0&7ngtutirwus.Ph.D.Diss.UnivexsityofMarylaIl~ zone. Heredity 56:337-349. College Park, MD USA -. 1990. Hybrid zonesz Wmiows on evolution- RISINQ, J. D. 1983. The progress of oriole hybridixa- ary process, pp. 69-128. In D. J. Futuyma and J. tion in Kansas. Auk 100~885-897. Antonovics (e&s.), oxford Studies in Evolutionary SLATKIN, M. 1973. Gene flow and selection in a cline. Biology. oxford Univ. Press, oxford, UK. Genetics 75:733-756. HMZRISON, R G., AND J. ARNOLD. 1982. A narrow -. 1975. Gene flow and selection in a two-locus hybrid zone between closely related cricket species. system. Genetics 8 1:209-222. Evolution 36:535-552. -. 1981. Estimating levels of gene flow in nat- HI~.ITO~J, R 1983. A new species of woodland sal- ural populations. Genetics 99:323-335. amander of the Pkthodon gktinosus group from -. 1985. Gene flow in natural populations. A.unu. the southern Appr&~chian Mountains. Brimleyana Rev. Ecol. Syst. 16:393-430. 9: l-20. SZYMURA, J. M., AND N. H. BARTON. 199 1. The ge- HIGHTON, R., AND S. A HENRY. 1970. Evolutionary netic structure of the hybrid zone between the iire- 938 NELSON G. HAIRSTON, SR. ET AL.

bellied toads Bornbina bornbina and B. varkgata: WISITAKER, R H. 1956. Vegetation of the Great Comparisons between transects and between loci. Smoky Mountains. Ecoi. Monogr. 26:1-80. Evolution 45~237-26 1. TURNER, J. R G. 197 1. Two thousand generations ofhybridization in a Heliconius butterlIy. Evolution 25:471-482. Corresponding Editor. R. Han&n