The Journal of Heredity 2001:92(4)
© 2001 The American Genetic Association 92:309-314
A Narrow Hybrid Zone Between Two Cottus Species in Wills Creek, Potomac Drainage
From the Department of Biology, Frostburg State University, Frostburg, Maryland.
Address correspondence to Andrew P. Kinziger, Department of Biology, St. Louis University, 3507 Laclede Ave., St. Louis, MO 63103-2010, or Kinziger{at}slu.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsMaterial ExaminedAllozyme MaterialResultsDiscussionReferences |
|---|
We describe a narrow hybrid zone between the mottled sculpin (Cottus b. bairdi) and the Blue Ridge sculpin (C. caeruleomentum). Seven characters (dorsal fin rays, pectoral fin rays, caudal base band condition, male spawning coloration, and one frequency and two fixed allozyme differences) distinguish the two taxa in the hybrid zone. C. caeruleomentum and C. b. bairdi diverged in these characters in allopatry as indicated by their distribution on opposite sides of the Atlantic–Ohio divide. However, a stream capture placed these two taxa in secondary contact in Wills Creek, Potomac drainage (Atlantic slope). Allozyme data indicate the presence of post-F1 hybrids in the zone of secondary contact. Changes in allozymes, morphology, and spawning coloration along a transect in Wills Creek reveal the hybrid zone is less than 20 river kilometers in length. Estimates of root mean square dispersal and gene flow tentatively suggest that selection is operating in the Wills Creek hybrid zone. C. b. bairdi and C. caeruleomentum are maintaining their identity in seven distinguishing characters on opposite ends of the hybrid zone revealing these two taxa are independent evolutionary lineages.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsMaterial ExaminedAllozyme MaterialResultsDiscussionReferences |
|---|
We describe a hybrid zone between Cottus b. bairdi (mottled sculpin) and a recently described species, C. caeruleomentum (Blue Ridge sculpin; Kinziger et al. 2000). Cottus b. bairdi and C. caeruleomentum are allopatric throughout most of their geographic ranges; C. b. bairdi occurs primarily west of the Atlantic slope and C. caeruleomentum is found in Atlantic slope drainages (Kinziger et al. 2000). However, syntopic populations of these two taxa occur on the Atlantic slope in Wills Creek, a tributary to the Potomac River. It is hypothesized that this area of syntopy originated when the higher gradient Wills Creek captured approximately 37 km2 of Blue Lick Creek, Youghiogheny (Ohio) drainage (Howard and Morgan 1993; Thompson 1939; Figure 1; referred to as the Wills/Blue Lick stream capture). In addition to the transfer of C. b. bairdi from Blue Lick Creek into Wills Creek, we collected two other predominantly Ohio River basin species (Etheostoma nigrum and E. b. blennioides) in upper Wills Creek, providing further evidence of stream capture in this area (but see Schwartz 1965). In this study we use allozymes, morphology, and spawning coloration to determine diagnostic characters and the distance between parental populations of C. b. bairdi and C. caeruleomentum in the Wills/Blue Lick stream capture arena. Further, we use allozymes to identify the presence of hybrids between C. caeruleomentum and C. b. bairdi.
|
|
Methods |
|---|
|
TopAbstractIntroductionMethodsMaterial ExaminedAllozyme MaterialResultsDiscussionReferences |
|---|
Allozymes
Procedures for allozyme electrophoresis are described in Kinziger et al. (2000). Allozyme data were collected from 20 individuals from Blue Lick Creek (Ohio basin) and from each of three localities in Wills Creek (Atlantic slope; see Material Examined). Our initial analysis revealed the presence of C. b. bairdi x C. caeruleomentum hybrids in Wills Creek, thus two additional populations were assessed at two diagnostic loci (Ck-A and Ldh-B) to aid in determining the width of the hybrid zone (Figure 1).
Morphology
Twenty morphological characters including fin spines, fin rays, pores of the cephalic lateralis canal system, palatine teeth, chin pigmentation, and caudal base band condition were enumerated from each individual as described by Kinziger et al. (2000). Morphological characters were recorded from 66 Youghiogheny River and 200 Potomac River specimens (see Material Examined). When available, data from Robins (1954) were used to supplement meristic counts. In some cases it was not possible to record every character for a single specimen because of physical deformities or missing fins.
Spawning Coloration
Surveys of spawning male breeding coloration were conducted on 10 dates between 5 March and 15 May 1998 in the Youghiogheny and Potomac drainages (see Material Examined). Spawning males of each taxon were also held in aquaria for color comparisons.
Data Analysis
We partitioned the hybrid zone area into three regions (see Results) and used contingency chi-squared tests of allozyme and morphological data to determine diagnostic characters for the parental populations of C. b. bairdi and C. caeruleomentum in Wills Creek. Categories were combined when expected counts were less than five. In addition, allozyme data were entered into the BIOSYS-1 program (Swofford and Selander 1981) to calculate mean heterozygosity (Nei 1978), percent polymorphic loci, conformance to Hardy–Weinberg proportions, and gene frequency differences among populations. POPGENE (Yeh and Boyle 1997) was used to estimate Burrow's linkage disequilibrium between pairs of loci and perform chi-squared tests for significance (Weir 1979).
Hybrid Zone
Due to inconsistent use of the term hybrid zone, we use the definition proposed by Arnold (1997): hybrid zones occur where "two populations of individuals that are distinguishable on the basis of one or more heritable characters overlap spatially and temporally and cross to form viable and at least partially fertile offspring."
Designation of a hybrid zone as narrow or wide depends on the root mean square (RMS) dispersal distance between parents and offspring (Barton and Hewitt 1985, 1989). RMS dispersal is analogous to a diffusion coefficient (Barton and Hewitt 1989), thus we calculated RMS dispersal distance using
 | (1) |
is the RMS dispersal distance per generation, x is the dispersal rate (km/generation), and n is the number of samples (Brown et al. 1997). Generation time for C. bairdi is 2 years (Jenkins and Burkhead 1994). Thirteen dispersal rate values were obtained from McCleave (1964; Table 1).
|
The RMS dispersal estimate was supplemented by comparing gene flow in the Wills Creek hybrid zone to estimates of gene flow derived from other Cottus allozyme studies (e.g., Albertson 1995; Zimmerman and Wooten 1981). We made this comparison to determine if there were any barriers to gene flow in the Wills Creek hybrid zone. Gene flow estimates were generated using allozyme data following Slatkin (1981). Estimates of gene flow are generated by plotting the average frequency of alleles occurring in i populations [conditional average frequency (p(i))] against the number of populations in which an allele is present [occupancy number (i)]. Concave curves from these plots are characteristic of populations with high gene flow and convex curves are characteristic of populations with low gene flow.
|
Material Examined |
|---|
|
TopAbstractIntroductionMethodsMaterial ExaminedAllozyme MaterialResultsDiscussionReferences |
|---|
Specimens were collected using a backpack electrofisher (Smith-Root model 15-A electrofisher, variable voltage) or seine (3 m x 1.5 m seine with 3.1 mm mesh). Specimens used for morphological analysis were fixed in 10% formalin, stored in 50% isopropanol and are currently housed in the University System of Maryland Frostburg (USMF) fish collection; those used for electrophoresis were placed on dry ice in the field and stored at -80°C.
|
Allozyme Material |
|---|
|
TopAbstractIntroductionMethodsMaterial ExaminedAllozyme MaterialResultsDiscussionReferences |
|---|
C. b. bairdi: Yough: Blue Lick Creek at SR 2026, W of Berkleys Mill, Somerset County, PA (N = 20); Upper Wills: Wills Creek at T 377, 1 km E of Mance, Somerset County, PA (N = 20); Wills Creek at SR 2015, Glencoe, Somerset County, PA (N = 20).
C. b. bairdi, C. caeruleomentum and C. b. bairdi x C. caeruleomentum: Middle Wills: Shaffers Run at the confluence with Wills Creek, Fairhope, Somerset County, PA (N = 20); Wills Creek at the confluence with Gooseberry Run at SR 3004, Hoblitzell, Bedford County, PA (N = 20).
C. caeruleomentum: Lower Wills/Potomac: Little Wills Creek at T 350, N of Hyndman, Bedford County, PA (N = 20).
Morphological Material
C. b. bairdi: Yough/Upper Wills: USMF 9288, 30 specimens; USMF 96141, 13 specimens; USMF 96137, 13 specimens; USMF 9251, 10 specimens; USMF 9683, 29 specimens; unnamed tributary to Wills Creek at the intersection of T 377 and SR 2017, Mance, Somerset County, PA [DAN96-22 (N = 16]; USMF 9703, 11 specimens; USMF 9684, 40 specimens.
C. b. bairdi, C. caeruleomentum, and C. b. bairdi x C. caeruleomentum: Middle Wills: USMF 9651, 9 specimens.
C. caeruleomentum: USMF 96143, 34 specimens; USMF 92111, 30 specimens; USMF 9363, 20 specimens; USMF 9364, 7 specimens; PSU 1231, 4 specimens.
Spawning Material
C. b. bairdi: Youghiogheny: Big Shady Run at State Hwy 495, S of Grantsville, Garrett County, MD (April 6, 1998); Upper Wills: Wills Creek at the intersection of Witt and Gameland Rds, S of Mance, Somerset County, PA (March 23, 1998; May 3, 1998).
C. caeruleomentum: Lower Wills/Potomac: Little Wills Creek along State Hwy 96, 4 mi N of Hyndman, Bedford County, PA (March 23, 1998); Town Creek at Hwy 144, E of Flintstone, Allegany County, MD (May 1, 1998); Blue Lick Run along Blue Lick Rd, 1.75 air km SW of Avilton, Garrett County, MD (May 15, 1998); Evitts Creek at I68, E of Cumberland, Allegany County, MD (April 7, 1998; May 15, 1998); Murley's Branch along Williams Rd, S of Flintstone, Allegany County, MD (April 6, 1998); Wolf Run at confluence with Wills Creek, Madley, Bedford County, PA (April 9, 1998).
|
Results |
|---|
|
TopAbstractIntroductionMethodsMaterial ExaminedAllozyme MaterialResultsDiscussionReferences |
|---|
Results of the allozyme and morphological data analysis revealed that our study area could be divided into three regions. The first region, Yough/Upper Wills, consisted of the Youghiogheny River and upstream Wills Creek from 0 to 19.47 river kilometers (rkm) and was inhabited by C. b. bairdi (distances measured from upstream to downstream using U.S. Geologic Survey topographic maps). The second region, Middle Wills, consisted of Wills Creek from 19.48 to 38.94 rkm and was inhabited by C. b. bairdi, C. caeruleomentum, and C. b. bairdi x C. caeruleomentum hybrids. The last region, Lower Wills/Potomac, consisted of Wills Creek below 38.95 rkm and was inhabited by C. caeruleomentum.
Allozymes
Of the 20 gene loci resolved, 13 were monomorphic (sAat-A, Ada-A, Ak-A, Pep-A, Est-1, Gpi-B, Gapdh-1, Ldh-A, mMdh-A, Pgm-A, sSod-A, Tpi-1, Xdh-1) and 7 were polymorphic (Table 2). Of the seven variable loci, Ck-A, G3pdh-1, and Ldh-B were useful in differentiating C. b. bairdi in the Yough/Upper Wills region from C. caeruleomentum in the Lower Wills/Potomac. Chi-squared tests revealed significant heterogeneity at the G3pdh-1 locus between C. b. bairdi in the Yough/Upper Wills and C. caeruleomentum in Lower Wills/Potomac (Yough versus Lower Wills/Potomac, 
2 = 14.24, P < .05; Upper Wills versus Lower Wills/Potomac, 2 = 9.64, P < .05). In addition, C. b. bairdi in the Yough/Upper Wills and C. caeruleomentum in the Lower Wills/Potomac were fixed for alternate alleles at Ck-A and Ldh-B (Table 3). The length of the hybrid zone as indicated by the distance between pure parental forms is maximally 19.47 rkm (Table 3).
|
|
Populations in Middle Wills contained a mixture of genotypes at Ck-A and Ldh-B, indicating the presence of C. caeruleomentum x C. b. bairdi hybrids (Table 3). Individuals heterozygous at both diagnostic loci were presumed to be F1 hybrids; however, to ensure a high probability of distinguishing a probable F1 from a post-F1 hybrid, six diagnostic loci are desirable (Campton 1990). Some individuals had combinations of diagnostic alleles from both parental species (e.g., heterozygous at Ck-A and homozygous at Ldh-B) indicating the presence of post-F1 hybrids. Genetic variability measures for populations where all 20 loci were examined revealed that hybrid populations had higher percentages of polymorphic loci and higher levels of heterozygosity than either parental population (Table 4). All loci conformed to Hardy–Weinberg proportions except Ldh-B at 26.22 rkm in Wills Creek [this population was heterozygote deficient (P < .05; Fis = 0.499; Table 3)]. No significant linkage disequilibrium was detected.
|
Morphology
Chi-squared analysis of the 20 morphological characters indicated that the number of dorsal fin spines (DFS), pectoral fin rays (PTFR), and the caudal base band condition were useful in differentiating C. b. bairdi in the Yough/Upper Wills from C. caeruleomentum in Lower Wills/Potomac (tables 5, 6, and 7). The number of DFS and PTFR of C. b. bairdi in Yough/UpperWills was significantly higher than that of C. caeruleomentum in Lower Wills/Potomac (DFS,
2 = 82.89, P < .05; PTFR, 2 = 167.50, P < .05). The caudal base band of C. b. bairdi in Yough/Upper Wills was notched significantly more than the caudal base band of C. caeruleomentum in Lower Wills/Potomac (2 = 210.01, P < .05). Changes in morphological characters from those diagnostic of C. b. bairdi to those of C. caeruleomentum indicate the hybrid zone spans the length of the Middle Wills region (19.47 rkm).
|
|
|
Spawning Coloration
Spawning males of C. b. bairdi in the Yough/Upper Wills region have blackish chins. Conversely, spawning males of C. caeruleomentum in Lower Wills/Potomac have a blue to bluish-green coloration on their chins, mouths, branchiostegal membranes, and at the insertion of pectoral and pelvic fins. It is not known if blue chins are a primary or secondary male coloration.
Hybrid Zone
Using 13 dispersal rates for C. bairdi provided by McCleave (1964), we calculated an RMS dispersal of 4.63 km/generation (Table 1). Removal of one individual that dispersed 153 m downstream in 7 days (perhaps due to tagging injury) reduced the RMS dispersal to 1.42 km/generation. The width of the Wills Creek hybrid zone is maximally 19.47 rkm, which is narrow compared to the RMS dispersal of 1.42–4.63 km/generation.
We found three datasets in two studies that were suitable for estimating gene flow (Table 8). A plot of p(i) against i/d, where d is the number of populations, for nonhybrid zone Cottus studies resulted in convex curves, indicating high levels of gene flow. Conversely, the plot for the Wills Creek hybrid zone was almost linear, indicating intermediate levels of gene flow (Figure 2). The arrow in Figure 2 indicates a fluctuation that is typical for these types of plots and may be due to small sample sizes (Slatkin 1981).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsMaterial ExaminedAllozyme MaterialResultsDiscussionReferences |
|---|
Hybrid Zone Origins
Hybrid zones can originate from allopatric divergence and subsequent sympatry (secondary contact) or from in situ divergence (primary intergradation; Harrison 1990). C. caeruleomentum and C. b. bairdi diverged in allopatry, as indicated by their distribution in streams on opposite sides of the Atlantic–Ohio drainage divide. Subsequently they came into contact and hybridized when C. b. bairdi were transferred to the Atlantic slope via the Wills/Blue Lick stream capture. Thus the hybrid zone is the result of secondary contact. Frequently cited evidence for the origin of hybrid zones by secondary contact is concordant change in several characters (Harrison 1990). Seven characters change from those characteristic of C. b. bairdi to those indicative of C. caeruleomentum over a distance of less than 20 rkm, supporting secondary contact as an explanation for the origin of the Wills Creek hybrid zone (Table 9).
|
Defining a Narrow Hybrid Zone
Whether a hybrid zone is characterized as narrow or wide depends on RMS dispersal distance between parents and offspring (Barton and Hewitt 1985, 1989). We estimated RMS dispersal for C. bairdi as 1.42–4.63 km/generation. Allozyme, morphological, and spawning coloration characters reveal the hybrid zone is less than 20 rkm in length, suggesting the length of the Wills Creek hybrid zone is narrow relative to our RMS dispersal estimate. However, our estimate of RMS dispersal distance is tenuous because it is based on a small sample size and a C. bairdi population from a different geographic locality. Nevertheless, our assertion that the Wills Creek hybrid zone is narrow compared to RMS dispersal is strengthened because our value is probably an underestimate of dispersal ability. First, because long-distance dispersers are often difficult to detect and failure to detect long-distance dispersers can drastically change dispersal values, long-distance dispersers contribute significantly more to RMS dispersal estimates than many short-distance dispersers (Moore and Dolbeer 1989). Second, our RMS dispersal estimate does not account for episodic dispersal that may occur during extreme flooding.
Other dispersal data for Cottus also suggest the Wills Creek hybrid zone is narrow relative to dispersal. Studies of C. bairdi have detected maximum movements of 54.8 m in 18 months (Hill and Grossman 1987) and 143 m in 12 months (Bailey 1952). Anecdotal accounts of dispersal have recorded movement of female C. bairdi at least 12 km upstream from Green Bay into the lower Fox River (Brown County, WI) in approximately 6 months (UWMZ 10237; Cochran PA, personal communication, 1998), and movement of two C. cognatus approximately 2 km upstream from Lake Michigan into Sucker Creek (Ozaukee County, WI) in approximately 1 month (Lyons J, personal communication, 1998).
Hybrid Zone Maintenance
There are two main hypotheses that attempt to explain the maintenance of narrow hybrid zones. The bounded hybrid superiority model suggests that hybrids have equal or greater fitness than both parental types within a narrow zone along some ecological gradient (Moore 1977), while the tension zone model indicates that hybrid zones are maintained by a balance between selection against hybrids and dispersal of parental types into the hybrid zone (Barton and Hewitt 1985, 1989). In the bounded hybrid superiority model, selection is environment dependent because it involves interactions between the environment and the genome (e.g., selection along an ecological gradient), whereas in the tension zone model, selection is environment independent because it involves "hybrid inviability and/or sterility and is due to disruption of parental gene combinations" (Arnold 1997).
Two lines of evidence suggest that selection is operating in the Wills Creek hybrid zone. First, the hybrid zone seems to be narrow relative to dispersal; if selection were not operating, one would expect the hybrid zone to be much wider. Second, gene flow in the Wills Creek hybrid zone seems to be less than that predicted from similar Cottus studies. The amount of gene flow in the Wills Creek hybrid zone would be expected to be in accord with nonhybrid studies if there was no selection. Although selection is possibly operating in the Wills Creek hybrid zone, the type of selection (environment dependent or environment independent) that is operating is unknown. Thus neither model of hybrid zone stability can be invoked as an explanation.
Taxonomic Considerations
C. b. bairdi and C. caeruleomentum have been in contact in Wills Creek from the date of the Wills/Blue Lick stream capture to the present and allozyme data indicate hybridization within a narrow zone between these two taxa. Nevertheless, C. b. bairdi and C. caeruleomentum have remained distinct in seven distinguishing characters (Table 9), supporting the assertion that these two taxa are independent evolutionary lineages.
|
Acknowledgments |
|---|
We thank S. Davis, J. Jennings, B. Margolis, E. Raesly, A. Rettig, and J. Trimpy for help with field work. R. Gregg provided advice and guidance with electrophoresis and D. Buth furnished allozyme nomenclature advice. M. Bowe, J. Howard, and D. Neely provided useful insights and commented on drafts of this manuscript. F. Kessler prepared the figure. This project was done by A. P. Kinziger in partial fulfillment of a Master of Science in applied ecology and conservation biology at Frostburg State University. This work was funded by the Maryland Department of Natural Resources Natural Heritage Program.
|
Footnotes |
|---|
Corresponding Editor: Robert Wayne
|
References |
|---|
|
TopAbstractIntroductionMethodsMaterial ExaminedAllozyme MaterialResultsDiscussionReferences |
|---|
-
Albertson LE, 1995. Genetic variation and geographic isolation within a species of Cottus from the Potomac River drainage (MS thesis). Frostburg, MD: Frostburg State University.
Arnold ML, 1997. Natural hybridization and evolution. New York: Oxford University Press.
Bailey JE, 1952. Life history and ecology of the sculpin Cottus bairdi punctulatus in southwestern Montana. Copeia 1952:243–255.
Barton NH and Hewitt GM, 1985. Analysis of hybrid zones. Annu Rev Ecol Syst 16:113–148.[Web of Science]
Barton NH and Hewitt GM, 1989. Adaptation, speciation and hybrid zones. Nature 341:497–503.[Medline]
Brown TL, LeMay HE Jr, and Bursten BE, 1997. Chemistry: the central science. Englewood Cliffs, NJ: Prentice Hall.
Campton DE, 1990. Application of biochemical and molecular markers to analysis of hybridization. In: Electrophoretic and isoelectric focusing techniques in fisheries management (Whitmore DH, ed). Boca Raton, FL:CRC Press; 241–264.
Harrison RG, 1990. Hybrid zones: Windows on evolutionary process, vol 7. In: Oxford surveys in evolutionary biology (Futuyma D and Antonovics J, eds). Oxford: Oxford University Press; 69–128.
Hill J and Grossman GD, 1987. Home range estimates for three North American stream fishes. Copeia 1987:376–380.
Howard JH and RP Morgan, 1993. Allozyme variation in the mottled sculpin (Cottus bairdi): a test of stream capture hypotheses. Copeia 1993:870–875.
Jenkins RE and Burkhead NM, 1994. Freshwater fishes of Virginia. Bethesda, MD: American Fisheries Society.
Kinziger AP, Raesly RL, and Neely DA, 2000. New species of Cottus (Teleostei: Cottidae) from the middle Atlantic Eastern United States. Copiea 2000:1007–1018.
McCleave JD, 1964. Movement and population of the mottled sculpin (Cottus bairdi Girard) in a small Montana stream. Copeia 1964:506–513.
Moore WS, 1977. An evaluation of narrow hybrid zones in vertebrates. Q Rev Biol 52:263–277.
Moore WS and Dolbeer RA, 1989. The use of banding recovery data to estimate dispersal rates and gene flow in avian species: case studies in the red-winged blackbird and common grackle. Condor 91:242–253.
Nei M, 1978. F-statistics and analysis of gene diversity of subdivided populations. Ann Hum Genet 41:225–233.
Robins CR, 1954. A taxonomic revision of the Cottus bairdi and Cottus carolinae species groups in eastern North America (Pisces, Cottidae) (PhD dissertation). Ithaca, NY: Cornell University.
Schwartz FJ, 1965. The distribution and probable postglacial dispersal of the percid fish, Etheostoma b. blennioides, in the Potomac River. Copeia 1965:285–290.
Slatkin M, 1981. Estimating levels of gene flow in natural populations. Genetics 99:323–335.
Swofford DL and Selander R, 1981. BIOSYS-1: a FORTRAN program for the comprehensive analysis of electrophoretic data in population genetics and systematics. J Hered 72:281–283.
Thompson HD, 1939. Drainage evolution in the southern Appalachians. Bull Geol Soc Am 47:1323–1356.
Weir BS, 1979. Inferences about linkage disequilibrium. Biometrics 35:235–254.[Web of Science][Medline]
Yeh FC and Boyle TJB, 1997. Population genetic analysis of co-dominant and dominant markers and quantitative traits. Belg J Bot 129:157.
Zimmerman EG and Wooten MC, 1981. Allozymic variation and natural hybridization in sculpins, Cottus confusus, and Cottus cognatus. Biochem Syst Ecol 9:341–346.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||