The Auk 111(3):661-671, 1994 MITOCHONDRIAL-DNA AND NUCLEAR-GENE DIFFERENTIATION IN NORTH AMERICAN PRAIRIE GROUSE (GENUS TYMPANUCHUS) DARRELL L. ELLSWORTH,L3 RODNEY L. HONEYCUTT, i NOVA J. SILVY,• KEVIN D. RITTENHOUSE,x AND MICHAEL H. SMITH 2 •WildlifeGenetics Laboratory, Department of Wildlifeand Fisheries Sciences, TexasA&M University,College Station, Texas 77843, USA; and 2SavannahRiver Ecology Laboratory, P.O. DrawerE, Aiken,South Carolina 29801, USA ABSTP,•CT.--Thefragmentary effects of Pleistoceneglacial activity have been implicated in speciation among avifauna endemic to the Central Plains of North America. The prairie- grousecomplex (genus Tympanuchus),distributed throughout the central United Statesand Canada,contains three sistertaxa believed to have originated from the expansionof late Pleistocenerefugial populations.We assayedmitochondrial-DNA (mtDNA) restriction-site and allozyme variation in prairie grouseobtained from localitiesthroughout their current range in North Americato examinethe nature and timing of eventspromoting differentiation in Tympanuchus.The geneticdata were not consistentwith Pleistoceneisolation of sufficient duration to allow a taxonomicallyor geographicallystructured pattern of genetic variation to emerge. No clear genetic differencesamong specieswere observed.Allozymes could not distinguishpopulations belonging to different speciesand frequencieswere generally similar acrosstaxa. The mtDNA differentiationwas characterizedby a predominanthaplotype shared by all taxa;the remainder(15) were generallyinfrequent and closelyrelated to the prevalent (and presumablyancestral) haplotype. The presenceof unique mtDNA haplotypeswithin speciesand absenceof certain allozyme alleles from particular taxa implied a degree of isolation and restrictionsto gene flow. However, the mtDNA haplotypesdid not sort phy- logenetically,which suggestsrecent commonancestry of the lineagesand may explain the lack of congruencebetween genetic variation and speciesdesignations. Despite the absence of quantitativegenetic differentiation, considerable morphological and behavioraldifferences are apparent among the putative species.Adult male plumage, vocalization structures,and courtshipbehaviors form the basisfor taxonomicdivisions among prairie grouse,but, con- sidering their closeassociation with reproduction,such charactersmay be subjectto sexual selectionand may evolve rapidly relative to mtDNA or allozymes. Received27 January1993, accepted19 August1993. RANGE FRAGMENTATION and isolation associ- Many subspeciesand speciesin the Great Plains ated with Pleistocene glaciation have been are assumedto have arisen through geographic widely implicated in the differentiation of var- isolationin refugia during Pleistoceneglacial iousclosely related, but largely allopatric,avian maxima (Selander 1965, Hubbard 1973). Con- species(Rand 1948, Mengel 1964, 1970, Selan- versely,the low speciesdiversity of grassland der 1965). The Central Plains of North America, birds and paucity of closely related congeners in particular,have been viewed as an isolating led Mengel (1970) to hypothesizethat there has agent promoting speciation among avian taxa been little pastfractionation of the Central Plains that now occupyeastern deciduous and western into isolated environments. Therefore, it is un- montane forest (Mengel 1964, 1970, Berming- clear as to whether differentiation and specia- ham et al. 1992). However, the role of glacial tion amongbirds of the Central Plains may be activity in the speciationof avifauna endemic attributedto allopatric(due to isolation)or sym- to the Central Plains is not well understood. patric (without geographicsubdivision) mech- anisms. 3 Present address: Center for Demographic and The prairie grouse(genus Tympanuchus) com- Population Genetics,University of TexasHealth Sci- prise an interesting "speciescomplex" for ex- ence Center at Houston, P.O. Box 20334, Houston, amining the competing hypothesesregarding Texas 77225, USA. the fate of the Central Plains during the Pieis- 661 662 ELLSWORTHETAL. [Auk,Vol. 111 tocene and possible modes of differentiation/ gent avian sister taxa believed to have origi- speciationamong birds of the North American nated from the expansion of late Pleistocene grasslands.The prairie-grouse complex con- refugial populations.In this study,we assessed tains three primarily grasslandadapted sister mtDNA andallozyme variation in prairiegrouse taxa distributedthroughout the central United and testedprevious hypotheses regarding dif- Statesand Canada.Two extant formsof T. cupido ferentiation and speciation among the taxa. (Greater Prairie-Chicken) occur primarily in Specifically,we addressedthe following ques- remnant tall-grassprairie in Kansas,Nebraska, tions:What is the nature of genetic differenti- Oklahoma,and SouthDakota (T. c.pinnatus) and ation in the prairie-grouse complex?Are the in smallendangered populations along the Tex- patternsof geneticvariation and estimatedtimes as Gulf Coast (T. c. attwateri; Westemeier 1980). sincedivergence among mtDNA lineagescon- Tympanuchuspallidicinctus (Lesser Prairie-Chick- sistent with late Pleistocene (recent) vicarlance? en) occupies the shrub grasslandsof south- Has sporadic interspecific hybridization in- western Kansassouthward through Oklahoma ducedby human activity influencedthe pattern to eastern New Mexico (Waddell and Hanzlick of differentiationamong taxa?How do levels 1978, Cannon and Knopf 1980, Taylor and of genetic differentiation comparewith mor- Guthery 1980), and T. phasianellus(Sharp-tailed phologicaland behavioraldivergence among Grouse) is resident to the north-central United the putative species? States and central Canada to Alaska (Miller and Graul 1980). METHODS Tympanuchuscupido and T. pallidicinctusare recognized as distinct species(AOU 1983) due Specirnens.--Tissues(brain, heart, liver) were ob- to differences in behavior, habitat affiliation, tained from 86 prairie grouseand frozen on dry ice. and socialaggregation (Grange 1940, Jones 1964, Most birds were harvestedby hunters, but samples Sharpe 1968). However, the differencesare not from the endangered T. c. attwateriwere taken from asgreat asthose seen among other well-defined incidental mortalities at the Attwater Prairie-Chicken grousespecies, causing some researchers (Short National Wildlife Refuge. Collection sites, listed be- 1967, Johnsgard1983) to regard these taxa as low by state and county (USA) or province and mu- nicipality (Canada),were locatedthroughout the cur- allopatricsubspecies. Conversely, T. phasianellus rent range of each taxon: T. c. pinnatus,ILLINOIS, is morphologicallydistinct from the "prairie Jasper(2), Marion (2), KANSAS, Butler (3), Chase(5), chickens." Elongated central and reduced pe- Lyon (2), Shawnee (4), NEBRASKA, Rock (3), Thomas ripheral rectricesas well asrudimentary pinnae (3), OKLAHOMA, Osage(3), SOUTH DAKOTA, Ly- and cervicalapteria distinguish phasianellus from man (6); T. c. attwateri,TEXAS, Colorado(1), Refugio cupidoand pallidicinctus(Ridgway and Fried- (1); T. pallidicinctus,KANSAS, Clark (9), Morton (7); mann 1946). In fact, phasianellus(formerly Pe- T. phasianellus,COLORADO, Routt (2), MANITOBA, dioecetesphasianellus) was once placed in a dis- CANADA, Coldwell (4), Chetfield (2), MINNESOTA, tinct monotypicgenus, but Pedioeceteshas been Aitkin (4), Lake of the Woods (3), NEBRASKA, Tho- synonymized with Tympanuchusto reflect phys- mas(5), Rock (1), NORTH DAKOTA, Billings (1), Bur- leigh (1), Grant (2), McKenzie (1), Morton (1), SOUTH iological (Hudson et al. 1959, 1966) and de- DAKOTA, Campbell (1), Dewey (2), Lyman (2), QUE- mographicsimilarities. BEC,CANADA, Ungava Comt• (3). Wisconsinglaciation has been cited as the Mitochondrial-DNA analysis.--In the laboratory, primary factor promoting subdivisionand di- mtDNA was isolated from frozen tissueand purified vergence among prairie-grouse populations. on cesiumchloride density gradients(Cart and Grif- Hubbard(1973) hypothesized that prairiegrouse fith 1987).Mitochondrial DNA from 51 prairie grouse were continuouslydistributed throughout much (T. c. pinnatus,21; T. c. attwateri,2; T. pallidicinctus,6; of the Great Plains during the Sangamon in- T. phasianellus,22) representativeof localitiesfrom the terglacial,but that considerablerange fragmen- Texas Gulf Coast to Hudson Bay was digested with tation and isolation of populations occurred 12 six-base-recognizing(Barnil I, Cla I, Dra I, EcoRI, EcoRV, Hind III, Nde I, Pst I, SspI, SstI, Stu I, Xba I) during the Wisconsinglacial period. Differen- and 4 four-base-recognizing(Hha I, Msp I, RsaI, Taq tiation within isolated refugia and postglacial I) restriction enzymes. Fragmentswere end-labeled reinvasion of the plains may thus have pro- with 32P-deoxynucleotides,separated by molecular duced the distribution of extant taxa. Therefore, weight on 1.2% agaroseor 4% acrylamide gels, and prairie grouseare useful for characterizingge- visualizedby autoradiography.Fragment sizes were netic variation among morphologicallydiver- estimatedfrom comigrating molecular size standards July1994] DifferentiationinPrairie Grouse 663 composedof lambda DNA and PM2 DNA digested phate dehydrogenase(L) (GAPD; 1.2.1.12);aglycerol- with Hind III. 3-phosphate dehydrogenase (L) (GPD1; 1.1.1.8); Unique fragment patterns produced by each re- sorbitoldehydrogenase (L) (SORD;1.1.1.14); and xan- striction enzyme were designatedalphabetically in thine dehydrogenase(L) (XDH; 1.1.1.204).(2) Con- chronologicalorder of discovery.Restriction sites were tinuous tris