Great Basin Naturalist
Volume 51 | Number 2 Article 4
6-30-1991 Evolutionary differentiation within the northern Great Basin pocket gopher, Thomomys townsendii. II. Genetic variation and biogeographic considerations Mary Anne Rogers Field Museum of Natural History, Chicago, Illinois
Follow this and additional works at: https://scholarsarchive.byu.edu/gbn
Recommended Citation Rogers, Mary Anne (1991) "Evolutionary differentiation within the northern Great Basin pocket gopher, Thomomys townsendii. II. Genetic variation and biogeographic considerations," Great Basin Naturalist: Vol. 51 : No. 2 , Article 4. Available at: https://scholarsarchive.byu.edu/gbn/vol51/iss2/4
This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Great Basin Naturalist by an authorized editor of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. Great Basin Naturalist51{2}, 1991, pp.127-152
EVOLUTIONARY DIFFERENTlA,TION WITHIN TIlE NORTHERN GREAT BASIN POCKET GOPHER, THOMOMYS TOWNSENDII. II. GENETIC VARIATION AND BIOGEOGRAPHIC CONSIDERATIONS
Mary Anne Rogersl
ABsTRACT.~Genetic variation across the known geographic range ofThomomys townsendii was examined by starch gel electrophoresis and nondifferentially stained katyotypes. All specimens examined had a diploid number of76, but some populations possessed one pair ofacrocentric chromosomes in an otherwise biarmed karyotype. Electrophoretic ~alyses of 16 populations revealed Thomomys toWnsendii to be among the least variable ofpocket gopher species studied. Of27 loci examined, 17 were monomorphic and heterozygosity values were low (H = .000-.028, Ii = .012). Genic differentiation between populations was also low (8 = .896-.998,5 = .956) and revealed little concordance with currentsubspecific units. The most marked differentiation is between the Honey Lake Valley/Humboldt River region and the Snake River drainage. Within each ofthese regions population variation is, in some cases, greater within than between currently defined subspecies. F -statistics show that the greatest genic differentiation is seen between populations, and a comparison ofsubspecies within a region shows the greatesthomogeneity. The evolutionary history ofThomomys townsendii is discussed in the context ofboth the phYSiographic ~story ofthe region and the affinities of this species to Thomomys bottae. Results ofthis study and the patterns ofdifferentiation found in the morphological analyses ofthe accompanying paper (Rogers 1991) indicate that the only consistent pattern to be discerned from the overall morphological and genetic homogeneity ofThotnomys townsendii is that between populations ofthe Humboldt and Snake River drainage systems, which are here assigned to Thomomys townsendii nevadensis and Thomomys townsendii townsendii, respectively. Key words: Thomomys townsendii, pocket gophers, genetic variation, evolutionary differentiation, geographic variation, electrophoresis, karyotype.
Since the revisions by Merriam (1895) gophers have been attributed to the selective and Bailey (1915), many studies have focused forces associated with the homogeneous fos~ on defining and.undersbmding Illorphologi~ soria! environment (Nevo et al. 1974, Penney cal variation within the family Geomyidae. and Zimmerman 1976) or to a series of sto Aspects of fossoriality, such as specificity to chastic processes and physiographic changes soil type and low vagility, and enVironmental (Patton 1972, 1981, Patton and Yang 1977). influences, such as climate, topography, and The geographic distribution and phylo food resources, have long been considered genetic affinities of Thomomys townsendii in this context (Bailey 1915, Grinnell 1927, make this species an interesting subject for Davis 1937, 1940, Hall 1946, Ingles 1950). genetic analyses. The presence of T. totvn~ More recently, patterns of chromosomal and sendii in the northern Great Basin is influ electrophoretiC variation have been com~ enced by the distribution of deep fluviatile pared to and correlated with interrelation~ and lacustrine soils and by the physical and ships among habitat preference, physio~ biological barriers that interrupt and restrict graphic features, gene flow, and population these habitats. Patton and Smith (1981) exam size (Patton 1970, 1972, 1981, Nevo et al. ined allozymic and karyotypic reiationships 1974, Penney and Zimmerman 1976, Patton within the bottae group and determined and Yang 1977, Honeycutt and Schmidly T. totvnsendii to be derived from one of the 1979, Tucker and Schmidly 1981, Sudman geographic units of T. bottae. this genetic et al. 1987). To explain the patterns found, evidence and the morphoiogical similarity investigators have presented" different argu found between these two species (Patton and ments for a variety of evolutionary mecha~ Smith 1989) provide an evolUtionary frame nisms. Patterns ofgenetic variation in pocket work within Which to interpret the patterns
lDeparnnentofZoology, Field Museum ofNatural History, Chicago, Illinois 60605.
127 128 M.A. ROGERS [Volume 51
44'
,..
Z t 15 25 0 25 50 .!!! ~ ~ j f miles
Fig. 1. Map of localities sampled for eiectrophoretic and karyotypic analyses. Locality information is given in the Appendix. . of allozymic @d karyotypic vaxiation of T. cies. These components of the biology of townsendii. Given the apparent phylogenetic Thomomys townsendii m'ake it possible to aJf:iliation, results of genetic· analyses of assess its evolutiOnary differentiation and Thomomys bottae' synthesized by Patton the validity ofcurrent taxonomic designations (1981) are especially relevant to the present ofthis species. examination of genetic differentiation within Thomomys townsendii. In general, genic pat MATERIALS AND M~THODS terns in T. bottae do not reflect subspecmc, topographic, or vegetational lJoundaries. ln~ Electrophoretic Procedures stead, broader geographic units ate defined During the spring and Summer of 1981 both karyotypicallY arid genically. This pat~ aIld the. spring of 1982, 252 gophers were tern suggests that environmentalfactors often collected using Macabee gopher traps. Six do not explain patterns of genetic vaxiation, teen populations from areas representing all but historical biogeography and features of seven recognized subspecies and the entire gopher populations such as vagility and effec~ geographic range of Thomomys townsendii tive poplllation size may be infl,uential. were sampled. The localities (numbered as the results of the genetic analyses of this in the accompanYing morphological study study and the patterns ofmorphological varia~ [Rogers 1991]) and sample sizes are detailed tion discussed in the accompanying paper in Figure 1 arid in the Appendix. (Roger 1991) will be considered in the context These animals were preserved as skin of the biogeographic history of the region and skull, skeleton only, or skull only speci~ and the phylogenetic affinities of this spe~ mens and are deposited in the Museum of 1991] GENETIC VARIATION IN THOMOMYS TOWNSENDII lZ9
Vertebrate Zoology, University of California, air-dried before mounting with coverslips. Berkeley. Kidney, liver, and hemolysate and Determination of diploid numbers and con plasma blood fraction samples were prepared struction ofkaryotypes followed standard pro~ and then frozen in liquid nitrogen. Samples cedures (Bender and Chu 1963, Patton 1961). were later stored under ultra~cold conditions Bi~armed chromosomes were categorized (-76 C) in the laboratory. Kidney and liver into eithera metacentricJsubmetacentric class samples were homogenized iJ:J. grinding buffer or a subtelocentric class after Patton and and centrifuged to produce extracts. Extracts Dingman (1970), utilizing the arm ratio crite and plasma samples were then analyzed by ria ofPatton (1967). horizontal starch gel electrophoresis following procedures similar to Selander et al. (1971). Statistical Analyses: The 27 loci scored and the tissue types used Electrophoretic Data (K = kidney, L = liver, P = plasma) are as Alleles were designated alphabetically in follows: en~ymatic proteins: aconitase (E. C. order of irlCreasing mobility, and genotype fre 4.2.1.3, ACON; K), adenosine deaminase quencies were analyzed using the BlOSYS-l (E.C. 3.5.4.4, ADA; K), a~glycerophosphate computer prograrn (Swofford and Selander dehydrogenase (E.C. 1.1.1.8, aGPD; K), 1981). Analyses of geographic differentiation alcohol dehydrogenase (E.C. 1.1.1.1, ADH~2; utilized Wright's (1978) F ~statistics, and trees L), esterase (E.C. 3.1.1.1, EST-4; K), were constructed from coefficients ofgenetic glucosephosphate isomerase (E. C. 5.3.1.9, similarity and genetic distance (Rogers'S, D GPl; K), glutarnate-oxaloacetate transarninase [1972] and modified D [Wright 1978]) using (E.C. 2.6.1.1, GOT-I, GOT-2; K), glycep UPGMA clustering and the distance Wagner aldehyde phosphate dehydrogenase (E. C. method (Farris 1972). 1.2.1.12, GAPDH; K), isocitrate dehydroge nase (E.C. 1.1.1.42, lCD-I, lCD~2; K), lac RESULTS tate dehydrogenase (E.C. 1.1.1.21, LDH-1, LDH~2; K), leucine aminopeptidase (E. C. lntrapopulation Variation 3.4.11 or 3.4.13, LAP; K), malate dehydroge Seventeen of the 27 loci examined were nase (E.C. 1.1.1.37, MDH-l, MDH-2; K), monomorphic. All polymorphic loci are rep map.nose phosphate isomerase (E.C. 5.3.1.8, resentedby dominant alleles withfrequencies MPl; K), peptidase (leucyl-alanine substrate, of 0.95 or less with the exception of E.C. 3.4.11 or 3.4.13, PEP-C; K), peptidase GOT-l (Table 1). This locus is fixed for alter~ (leucyl~glycyl~glycine substrate, E.C. 3.4.11 nate alleles in the populations ofHoney Lake or 3.4.13, PEP-B; K), phosphoglucomutase Valley and the lIumboldt River drainage sys (E.C. 2.7.5.1, PGM-2; K), phosphogluconate tem versus those ofthe Snake River system. dehydrogenase (E.C. 1.1.1.44, 6 PGb; Using the 0.95 criterion of polymorphism, K), purine nucleoside phosphorylase (Eo C. only EST-4 is polymorphic in more than 3 of 2.4.2.1, NP; K), sorbitol dehydrogenase the 16 populations studied. (E. C. 1.1.1.14, SORDH; K), superoxide dis lntrapopulation variation is also low across mutase (E.C. 1.15.1.1, SOD; K); nonenzy loci. Population averages for H (mean pro~ matic proteins: albumin (ALB; K), prealbu~ portion of loci heterozygous per individual, min (PREALB; P), transferrin (TB.F; P). direct count estimate), P (proportion of loci polymorphic, 0.95 criterion), and A (mean Karyotypic Procedures number ofalleles per locus) are presented in Forty-one animals, representing all recog~ Table 1. the meap. number ofalleles perlocus nized subspecies of Thomomys townsendii per population ranges from 1.0 to 1.11, with (Appendix), were examined. Most were pre~ a mean across populations of 1.07. The pro~ treated with yeast (Lee and Elder 1980, with portion of the: loci that is polymorphic per modifications in dosage). Chromosome prepa population ranges from .000 to .111. Mean rations were made from bone rnarrow cells polymorphism, P, calculated across the as described by Patton (1967) except that T. townsendiipopulations studied, is .053. KCI was used as the hypotonic solution. Air Heterozygosity values range from .000 in dried slides were stained in Giemsa (10% Murphy (n = 4) and .002 in Pocatello (n = 19) stock stain solution in phosphate buffer) and to .028 in Grandview~t. (n = 8). The average (jj...... 0 'fkBLE 1. AIIele fFequencies for polymorphicloci', and valuesfor H, P, and'A.
l>Il .1'1 '" '"0 U'"' '" ~ C .... ,'"'., ~ .,~ ~ ~ ~ ., l>Il ,..<:l l:i ~' ::2 .;: '~ ., 0 l:i '0 0 ~ l:i '1:l E ~ :§ >- '1:l :§ 0 ~ '1:: '1:: 8 l:i '";:: ., cl l:i 8' ., os > '~ ., Locus/All'ele 'G0T-l a 'L.OOO' 1'.000 'L.OOO' l!.000 1.000, i1.'OOO' ].000' 1.000' 1.000 .~ Ih ,LOOO LOOO' ].000 LOOO ],000 1.000 ],000' ?" '!;:1j PGM a .625 0 0 b .375 1'.000 1.000' 'L.OOO 1'.000· iliOOO ill.000 LOOO LOOO :1.000' l!.000 1'.'000 l!.000 .375 1.'0@0 LOO@ t:rJ ~ c .'625 '" ADN a .370' .liH 'h 1.000 l.000 l!.000' .250 1'.000 1.000 LOOO .630 .889' 1.000 1.0@0 LOO@ 1..000' LOOO 1..@00' l.OOO' ·c .750' iES1!'-4 a .]84 .023 .0!19 .028 .050 .!132 .l90 .250 .075 b LOOO .929 1:.000 l!.000 l!.000 .8!1i6 .9117 ..981 .972 .950' .868 .8l!0 LOOO l.OOO .750 .925 c .07'L GAPDH a .048 b .048 c 1.000 .904 1.000 1.000, 1.000 1:000 1:000 l.000' l!.000 ill.000' LOOO' LOOO 1..000' LOOO :1.000 1.000 '< >-'0 >::: S (l) PREALB a LOOO 1.000 1.000' 1.000 .800 HOO 1..000 1.000 LOOO' 1.000 LOOO l.OOO 1.000, LOOO 1.000 CJl b .200 l!.000 ...... f-' CD CD G 'FABLE 1 continued. TRF a .026 .125 b 1.000 1.,000 1:.000 1.000 l!.000 .914 1.000 1.000 1.000 1.'000 1.000 .905 1.000 .875 1'.000 1:..000 g{ c .095 Z t"1 ::l () AIDA a .275 b .'053 .023 c 1.000 L000 1.000 LOOO .725 .947 .9n L000 1.000 ill.000 l!.000 1.'000 l!.000 1.000 LOOO .950. ~ d .050 ::l 0z ...... Z .ili67 MPI a 1-3 ib i1.000 l!.000' .833 l!.000 iI!.000 .947 1.000 1..000 ili.000 L000 ili.000 l!.000 ll.000' 1.000 LOOO iI!.000 II: 0 ·c .053 ~ 0 ~ en>< PEP-B a .l!67 .050 ..., b l!.000 1.000 1.000, 1.000 1.000 1.000 L'000 ].000, 1.000 1.000 !L000 l!.000 l!.000 1.000 .833 .950 0 ~ en H .Oil.4 ..009 .0!l2 .0!l9 .03il. .016 .003 ..Olli8 .006 .004 .002' .007 .000 .028 ..0'1'9 .009 t"1 Z tJ l:::: P .037 ..074 .037 .037 .074 .l!ll .000 .037 .037 .037 .037 .074 .000 .074 .074 .l!l1' A 1.04 loU L04 ].04 L07 1l.I5 1.07 L07 1.07 1.04 1.04 L07 L00 L.07 1..01 1.l!l ~ f-' 132 M.A.ROGERS [Volume 51 TABLE 2. Matrix ofbetween-population genetic similarity using Rogers'S (1972). Population 3 1 5 3 Lovelock 1 Valmy .971 5 Quinn River .971 .988 4 Narrows .949 .967 .966 10 Ryndon .959 .977 .976 .955 11 Palisade .965 .986 .985 .961 .973 15 Austin .938 .957 .955 .934 .960 .953 30 Herlong .962 .981 .979 .975 .968 .975 .948 31 Standish .972 .991 .989 .973 .977 .985 .958 .990 19 Amer. Falls .938 .958 .955 .933 .944 .953 .924 .948 .958 20 Pocatello .935 .956 .952 .930 .940 .956 .921 .945 .955 .997 16 Grandview-o. .929 .950 .946 .925 .935 .956 .915 .939 .949 .991 .994 17 Murphy .940 .957 .957 .935 .945 .951 .924 .949 .958 .998 .995 .989 22 Grandview-t. .935 .929 .929 .907 .918 .925 .896 .921 .930 .970 .967 .966 .972 23 Payette .924 .945 .941 .920 .930 .949 .910 .934 .944 .986 .989 .988 .985 .957 24 Ontario .933 .953 .950 .929 .942 .952 .921 .943 .953 .995 .994 .988 .994 .966 .987 3 1 5 4 10 11 15 30 31 19 20 16 17 22 23 TABLE 3. Rogers'S (1972) coefficient ofsimilarity averaged by subspecies and by region. Subspecies # pops. 1 2 3 4 5 6 7 1 bachmani 4 .969 2 elkoensis 2 .970 .973 3 nevadensis 1 .946 .956 **** 4 relictus 2 .978 .976 .953 .990 5 similis 2 .945 .948 .923 .952 .997 6 owyhensis 2 .942 .947 .920 .949 .995 .989 7 townsendii 3 .933 .936 .909 .938 .984 .982 .970 Region # pops. 1 2 3 1 Humboldt River 7 .964 2 HoneyLake 2 .974 .990 3 Snake River 7 .937 .945 .984 ****Only one population included. heterozygosity was calculated across the 12 on Rogers' S (1972) for the 16 populations, populations .2.f n 2:: 10. The value for H was 7 subspecies, and 3 regions examined. not altered (H = .012). The range in similarity values for the popu lations of T. townsendii is .896 to .998, with Interpopulation Differentiation: S = .956. Examination ofthe subspecific and Electrophoretic Analysis regional treatments ofthese datareveals some discordance between subspecific or regional The populations ofT. townsendii examined affinity and degree of genetic similarity. For were arranged within four hierarchical levels: example, in Tables 2 and 3 some populations local population, subspecies, region, and spe of T. t. bachmani exhibit, on average, more cies. The geographic regions (and correspond genetic similarity to T. t. elkoensis and T. t. ing subspecies) included the following: the relictus than to other members of their own Humboldt River drainage system (T. t. bach subspecies. These results are similarly dis mani, T. t. elkoensis, T. t. nevadensis), Honey played in the regional comparison where Lake Valley (T. t. relictus), and the Snake some populations of the :Humboldt River River drainage system (T. t. owyhensis, T. t. region show a higher average similarity to similis, T. t. townsendii). Tables 2 and 3 rep~ Honey Lake than to some ofthe other popu resent the matrices ofgenetic similaritybased lations within the Humboldt River system. 1991] GENETIC VARIATION IN THOMOMYS TOWNSENDII 133 \ ;:~-;'- 112~- - -1- - '-f -- 44· ,.. 49" 25 0 f25 ... i! i miles Fig. 2. Contour diagram derived from UPGMA phenogram based on Rogers' (1972) coefficient of simiiarity, S. Intervals represent a Rogers'S of.01. Goodness"of-fit statistics for phenogram are as follows: Farris (1972) F = 0.977, Prager llI1d Wilson (1976) F = 0.852, percent standard deviation ofFitch and Margoliash (1967) = 1.187, cophenetic correlation coefficient = 0.882. The Snake River subspecies are, however, on River/Boney Lake cluster, the populations of average most similar to the other subspecies geographically peripheral T. t. relictus (Stan: ofthat region. dish and Herlong) show affinities to soine of The unweighted pair group method of the populations of the Humboldt River sub analysis was employed to cluster these data species. The Standish population from Honey and illustrate these same trends in interpopu: Lake Vailey llnd the Valmy, population from lation similarities. Figure 2 presents the north central Nevada are the most similar pair results of this analysis in the form of a con~ of that major cluster and are therefore more tour diagram. Beyond general homogeneity similarto eachotherthaneitheris to any other among the T. townsendii populations (5 = .939), the most obvious feature ofthe contour representative ofits own subspecies. The T. t. diagram is the separation ofthe Honey Lake nevadensis sample from Austin is the most Valley and Humboldt River populations from divergent population from the Humboldt the Snake River populations, forming two River region. Within the Snake River cluster, major clusters (5 > .95). This dichotomy is the greatest genic similarity values are also concordant With subspecmc designations and found between geographically well~separated hydrographic relationships. Within each clus~ populations. Gtandview~t.> aT. t. townsendii ter, however, similarity values do not corre population situated centrally in the Snake spond to current taxonomy or geographic fuvet distribution, is the most differentiated affinities. For instance, within the Humboldt population ofthat cluster. 134 M.A.ROGERS [Volume 51 T. t. bachmanl 0 T. t. etko.nata 6 T. i. nevllijeijsls 0 _-...... ----...,;,,;;.0,;;;80;...----.....~ 0 Lovelock (3) T. to r8l1c~iJs 0 ...... _.;,;;,0~22 0 Valmy (1) T. t. o~yhifiJils i ...... iiiiiiiii.,;·.::;0.3:;;;;!iiiiiiiiiiiiiiiiiii.... 0 Herro~g (30) T. t. a/iiillls .. ~-r .(U.6 T. t. ,own~itldll • .~ .024 .015 - L-,~...... ~.1;,;;0....2 __iiiiiiiii__...... ~...... 0 Narrows (4) L...o...... _iiiiiiiiiiiiiiiiii...... ;.,·;:.;;13~3...... ~ 0 Austin (15) ~r" AriI.. orlcan Falls (19) .089 ~6. Murphy (17) ~ r •Ontario (24) ~. Po~itollo r1 (20) ,...... ;,.;.O.3;;,;2~.....-i .OJ5 • Grai1dYI.~-o. (16) ...I ,:;~~~::;;;;;;;;;=~~;:;:;::;;;;;;;;~~··I I .023 --.• PayoHo (23) ...... ,;.0::.:7.:;.8_---...... ~ 1-...... _• Grandvlow-t. (22) Fig. 3. Wagner tree based on Wright's (1978) modification of Rogers' D. Tree is rooted at midpoint and Was constructed using Swofford's (1981) optimization algorithm. Goodness"of"fit statistics for tree are as follows: Farris (1972) F = 0.631, Prager and Wilson (1916) F = 3.255, percent standard deviation of Fitch and Margoliash (1961) = 9.358, cophenetic correlation coefficient = 0.995. Figure 3 provides a phylogenetic tree ulations are Austin, Narrows, Lovelock, and produced by the distance Wa:gnerprocedure Grandview-t. Each main cluster consists of a of Farris (1972) based on Wright's (1978) group ofpopulations that clusterata relatively modification of Rogers' D. Additional trees, high level ofsimilarity andwhose identity as a constructed with and without the use of group is maintained ill all analyses. Beyond Swofford's (1981) optimization algorithm a,nd the basic dichotomy between the two major using Rogers' D (1972,), give the same general river systems, there is no apparent concor~ topology. The most notable difference among dance between geographic distance or hydro these trees is in the position ofthe divergent graphic relationship and genic similarity. populations within the Humboldt lliver The F -statistics of Wright (1978) were cal cluster. Trees generated using Rogers' D culated to assess how genetic differentiation (1972) show these populations arising frolIl is partitioned within and between the hier within different small clusters (all separated archical levels established in this study. As from adjacent clusters by branch lengths less mentionedabove, the 16populations sampled thap. .01), in contrast to their peripheral posi correspond to the lowest level of the hier~ tion in trees generated with the modffied archy, With subspecies, regions, and the total Rogers' D (Wright 1918). Of the four trees as higher levels. The variance components generated, the one presented in Figure 3 Was and FST values estimating the amount ofvaria chosen for its high correlation coefficient tion accounted for by each ofthe subdivisions (range in cophenetic correlation coefficients relative to another are given inTable 4. These = .986~.995). were calculatedfor the polymorphic loci only. Several features are shared by the contoUr The greatest amoUnt ofdifferentiation is seen diagram derived from the UFGMA pheno in comparing all the populations sampled gram and all four Wagner trees. The most (PST = .672). Ofthe comparisons made, sub notable is the existence of two major clusters species comparedwithin a region are the least that separate the populations into the two differentiated (PST = .151). Within the total, river drainage systems. DiVergent popula~ interpopulation comparisons show greater tions within each ofthese clusters areconsis differentiation than those between subspe~ tently well differentiated in all the algorithm/ cies; these in turn show greater genic diver coefficient combinations utilited. These pop~ gence than do comparisons between regions. 1991] GENETIC VARIATION IN THOMOMYS TOWNSENpn 135 TABLE 4. Variance components and F~statistics HT values (.120~.147) while the five remain~ (Wright 1978) combined across polymorphic loci. ing polymorphic loci are low (.012~.049). Locus~by-locus heter~ Three geographic regions used in hierarchical chi square tests for classification (Humboldt River, ogeneity between populations were also cal Honey Lake Valley, Snake River) culated. Those loci that exhibit the most Variance significant levels of heterogeneity (GOT-I, Comparison COmponent FST PGM-l, ADH-2, and PREALB) are, for the Population-subspecies .252 .392 most part, the same loci that create the great Population-region .366 .{84 est differentiation between the various hier~ Population-total .798 .672 archical levels in the F -statistic analysis. Subspecies-region .114 .1Sl Subspecies-total .547 .460 These loci also contribute to the level of Region=totl!! .433 .364 differentiation .seen in the four divergent Two geographicregions used inhierarchical populations, Austin, Narrows, Lovelock, and classification (Humboldt River/Honey take Valley Grandview-t. and, Snake River) Interpopulation Variatiom Variance . Comparison component KarYotypic Analysis Population=subspecies .252 .392 All individuals kary"Otyped showeda diploid Population=7region .342 .468 number of 76. Five of the seven subspecies Population-total .798 .672 (T. t. elkoensis, r. t. nevadensis, T. t. owyhen Subspecies~region .09:1, .124 sis, T. t. similis, and T. t. totvnsendii) are Subspecies-total .574 .460 meta~ Region-total .456 .384 represented by 16 pairs of or submeta centric chromosomes ranging in size from large to small, and 21 pairs of large to small Similarly, Within a region, interpopulation subtelocentric chromosomes (fundamental differentiation is greater than that measured number [FN] = 148). In contrast, T. t. bach between subspecies. Averaged across all loci, mani and T. t. relictus are characterized by the F ST calculated for all populations is .130. 16 pairs ofmeta~ or submetacentrics, 20 pairs The same statistics were calculated for ofsubtelocentrics, and a single medium-sized a classification scheme With two regions, a.crocentric pair (FN = 146). All seven sub Honey Lake Valiey populations being in~ species have a large submetacentric X cluded With those of the Humboldt River chromosome and a minutel:lcrocentric (dot) drainage (Table 4). 'this set of F~statistics Y chromosome. shows decreased differentiation between the populational ?lld subspecific units within a PHYSIOGRAPHIC B:IStORY OF region and greater differentiation between THE NORTHERN GREATBASIN regions compared to the results ofthe three region scheme. Honey Lake Valley popll1a~ The Great Basin extends from the Sierra tions fit relatively homogeneously into the Nevada and Cascade Range in the west to HuIllboldt River region. The overall pattern the Wasatch Range and Colorado Plateau in still suggests that the greatest ex;tent ofdiffer the east. The northern and southern limits entiation is between the smallest subdivisions are defined by, the Columbia and Colorado and that there is relative homogeneity be platellus and thy Mojave and Sonoran deserts. tween the subspecies ofa region. The Grea.t Basin has internal drainage and is the total limiting variance or total gene made up ofovet 150 smaIler basins, separated diversity (HT ofNei 1973, 1977) perlocus was by more than 160 north=south-=oriented calculated for each ofthe 10 polYIllorphic loci. mountain rangys (Morrison 1965). Many of GOT-l exhibits the highest level ofdiversity, the rivers, lakes, andplayas ofthesebasins are reflecting the frx;ed difference between the remnants ofthe Pleistocene lakes Bonneville Humboldt/Honey Lake region and the Snake and Lahontan. River region (HT = ,492). This locus is proba~ The climate of the Great Basin is semiarid bly responsible for most of the homogeneity to arid. Mean a:r\.nual precipitation varies from Oow FST) found among subspecies within a year to year andincreases with elevation. Pre region. Four polymorphic loci. have moderate cipitation may range from 10 em in the basins ------~--~~---.- ~ 136 M.A. ROGERS [Volume 51 Fig. 4. Map oflate Pliocene and Pleistocene lakes ofthe northern Great Basin (after Morrison 1965 and Wheeler and Cook 1954). The limits ofIdaho Lake, Lake Lahontan, and Lake Bonneville are indicated by diagonal shading. Idaho Lake is oflate Pliocene/early Pleistocene age. Lakes Lahontan and Bonneville and smaller lakes (stippled) ate of Pleistocene age. Present-day lakes are represented by solid black areas. Point of overflow of Lake Bonneville at Red Rock Pass is indicated by arrow. to over 76 cm in some of the higher moun~ basins, the two largest of which were Lake tains. Average January temperatures range Lahontan and Lake Bonneville (Fig. 4). Only from 25 F (-4 C) at Bums, Oregon, to 44.5 F the features ofLake Lahontan will be empha'" (7C)atLasVegas, Nevada. AverageJulytem~ sized here, as the area covered by this peratures range from 67 F (19 C) at Burns to drainage is closely reflected by the range of 86 F (30 F) at Las Vegas (Morrison 1965). As Thomomys townsendii in the northern Great there are no drainages from thebasin, precipi~ Basin. Details ofthe features ofLakes Lahon~ tation is lost only to evaporation and transpira~ tan and Bonneville can be found in the mono~ tion. Evaporation increases and precipitation graphs of Russell (1885) and Gilbert (1890). decreases in a southward trend. This latitudi Broecker and Kaufman (1965), Morrison (1965); nal trend, coupled with the effect ofelevation and Benson (1978) provide more recent dis~ on temperature and evaporation, influences cussions ofthe history ofthese basins. hydrographic features of the present-day The present~day Lahontan drainage in Great Basin, as well as the existence and size cludes parts of eastern California, southeast~ of Pleistocene lakes (Meinzer 1922, Mi£I:lin ern Oregon, and much of northern Nevada. and Wheat 1979). To the west and south it includes the Eagle Lake, Honey Lake, Lake Tahoe, and Walker Pleistocene Lake Lahonton Lake systems. The Reese and Humboldt The Great Basin of the Quarternary was rivers drain into Lahontan Basin from the divided, as today, into many hydrographic east and south while the Quinn River makes 1991] GENETIC VARIATION IN THOMOMYS TOWNSENDII 137 up the northern component. The boundaries Snake River Plain of the lake, at its highest level during the While the Snake River Plain is hydrograph~ Wisconsin, encompassed the.following drain ically notpartofthe Great Basin, itis a natural age basins, Honey Lake, Pyramid Lake, component in terms of fauna, flora, and cli~ Winnemucca Lake, Carson Lake, Walker mate (Davis 1939, Cronquist et al. 1972). As Lake, Smoke Creek Desert, Black Rock much ofthe northern part ofthe distribution Desert, part ofthe Truckee River, Humboldt of T. townsendii lies within this area, the River (eastward to a pointnear Golconda), and physiographic history of the Snake River is Quinn River (Russell 1885, Benson 1978). pertinent to this study. character~ The climate ofthe Wisconsin was The Snake RiVer Plain has been shaped by ized by great fluctuations in temperature and syn~ a IO:J1g history of lava flows. The geography precipitation. Cool, wet times (Pluvials), of the eastern plain is more dominated by chronous with the glacial intervals, resulted remnants ofvolcanic activity than is the west. in high water levels in the Grell.t Basin lakes. These eastern lavas are, in general, younger Each lacustral maximum was followed by (late Pleistocene to Recent) than those ofthe warmer, drier periods during which the lakes western basin. Kuntz et al. (1986) suggest a subsided. A variety of interpretations exists minimum age range of 15,OOO~2,000YBP for for the extent and timing ofthe fluctuation of some of these formations. Pleistocene lavas the level of Lake Lahontan (Broecker and forced the Snake River into its present course Kaufman 1965, Morrison 1965, Benson 1978). along the east plain andformed the surface of Benson (1978) suggests that besides the most the plain there (Malde 1965). recent pluvial episode (spanning roughly The topographically diverse western plain 25,000 to 9,000 years before present [YBP]) Lake Lahontan reached atleast one otherhigh is cut by more tributaries ll.lld canyons ofthe stage before 40,000 YBP. He found no evi~ Snake River and is covered with lacustrine dence for high lake levels between 40,000 and fluviatile deposits. Some of the western and 25,000 YBP but recognized two sub~ lacustrine deposits can perhaps be explained sequent to that time, 25,000=21,500 YBP and bytheformer route ofthe Snake River and the 13,600~1l,000Y13P. The intervening period establishmentofIdaho Lake (Fig. 4). Wheeler representedlowerlake levels, butmostbasins and Cook (1954) suggest a possible late remained connected during this time. Pliocene route to the Pacific, running from Subsequent to the second deep-lake western Idaho to southern Oregon via the period, the climate ofthe Great Basin began Owyhee River. The route continues south~ a warming and drying trend (Antevs 1948). westward along the Steens, Pueblo, and Pine From 9000 to 5000 YBP the Lahontan Basin Forest mountains, across Black Rock and was warm and arid, with all lakes except Pyra Smoke Creek deserts of Nevada to Honey mid Lake desiccated. In the last 5000 years Lake Valley. From here it moves to Chilcoot water levels rose in Pyramid and Walker lakes Pass and westward to the Feather River. until they were tapped for agricultural use In the late Pliocene/early Pleistocene the dUring the past 100 years (Benson 1978). waters of the Snake were impounded and The pre-Wisconsin history of the Great Idaho Lake formed in western Idaho and east~ Basin is not as well documented as that of ern Oregon. In early Pleistocene times Idaho Lake Lahontan time. Since lakes ofthe early Lake found an outlet, and the Snake River and mid~Quaternary probably occupied re began its northward course into the Columbia gions that lakes Lahontan and Bonneville drainage. Alternatively, Taylor (1960) sug occupied subsequently, the majority ofthese gests that the early route of the Snake River deposits have been buried by younger ones. may have emptied into the Pacific via the Exposed deposits, however, suggest two pte Klamath Riverrather than the Feather River. Wisconsin lacustral intervals correlated with Another hydrographic link within the the Kansan and Illinoian glaciations. Early northern Great Basin is seen in the incidence deposits from outside the areas inundated by ofover:l:low ofLake Bonneville into the Snake Lake Lahontanhave beenfound inthefollow River Plain. The Bonneville basin experi= ing regions, along the middle Reese River, enced a Wisconsin pluvial history similar to Pine Valley, Smith Valley, and Carson Valley thatdescribedfor Lake Lahontan. Duringone (Morrison 1965). of the deep~lake periods, Lake Bonneville .. - ._ •.._------~------~ 138 M.A. ROGERS [Volume 51 overflowed at Red Rock Pass (SE ofPocatello, nected drainage basins as their Water levels Idaho; Fig. 4) into a tributary of the Snake rose and fell with climatic fluctuations. River (Morrison 1965). Malde'(1968) discusses Remnants of these Pliocene and Pleistocene the details of the catastrophi:c flood @d the hydrographic features ate the deep and moist confusion over the age ofthe event. it is here lacustrine and flUviatile sOlls of the type that regarded as late Wisconsin, about 12,000 YBP Thotnomys townsendii inhabits today. It was (Broecker and Kaufman 1965, Morrison not until post~Wisconsin times that desert 1965). The flood path extended across the conditiOns prevailed in the Great Basin to Snake River Plain and parts of the bed of the extent that some of these soils became ancient Idaho Lake to a point only seVeral uninhabitable. miles from the Oregon horder. the duration The disjunct nature ofthe current distribu of major flood effects is estimated at a few tion of Thomotnys townsendii reflects the days, although discharge continued for at boundaries ofthese historical drainage basins least a year. and gaps in the distribution of the preferred soil type ofthe species (Rogers 1991: Fig. 1). A major division in the cUrrent distribution HISTORICAL BIOGEOGRAPHY OF THOMOMYS TOWNSENDII of townsendii, and one that is reflected in .. .' ! both genetic and morphological analyses of Today Thomomys townsendii is. found in this and the companion study (Rogers 1991), fluviatile and lacustrine soils ofsouthern Ore lies between populations of the Snake River gon and Idaho, northern Nevada, artd north~ and Humboldt River drainages. These two eastern California (RogerS 199L Fig. 1). At regions ofthe townsendii distribution persist the western edge.of its distribution,. town primarily on the remnants of historical Lake sendii is found in the HOney Lake Valley. To Idaho and Lake Lahontan drainage basins the northeast the species inhabits the Quinn (Fig. 4). Today they are well isolated from River drainage system, Black Rock Desert, eaQh other by mountain ranges and scrub Alvord Desert, artd the basin of Harney and desert. Suitable habitat for townsendii was Malheur lakes. Gophers follow the Humboldt clearly once more abundant and continUous River from Lovelock to the vicinity ofWells, than it is today. Nevada. They are distributed along the Little It is useful here to discuss historical events Humboldt River and other tributaries to the and hydrographi:c features relevaJJ.t to estab north of the FI:umboldt. To the south they lishing where suitable habitat once existed for are found along the Reese River, in Pine and Thomomys townsendii, including where the Huntington Creek valleys, Independence populations ofthe Snake River and Humboldt Valley, and Diamond Valley. Along the Snake Rivermay have once been connected. A his River Plain they inhabit the vicinity ofAmeri torical network of drainage basins extended can Falls Reservoir and, to the west, the from Malheur Lake and the Owyhee River, stretch ofthe Snake River an,d its tributaries south through Honey Lake Valley and Lake from Hammett, Idaho, to the,vicinity ofVale, Lahontan, and east to the limits of the Bon Oregon. neville basin. Remnants of Idaho Lake the physiographic history of the northern (late Pliocene/early Pleistocene) can also be Great Basin explains many aspects ofthe cur~ included here as a potential continuation of rent distribution of Thomomys townsendii. lacustrine deposits extendingfrom southwest The Great Basin in the Miocene and Pliocene ern Oregon into southern Idaho (Fig. 4). was characterized by moderate topographic DaVis (1937) discusses the possibility of relief and a trend toward warming climates. an eastern hydrographic link leading from By the end ofthe Pliocene there Were wide~ the Humboldt River and Pleistocene Lake spread grasslands and well-developed lacus~ Lahontan into the Snake River drainage via trine and riparian environments (Axelrod the historical overflow at Red Rock Pass men 1950). As discussed previously, southern tioned earlier. While this information will be Oregon and Idaho had a history oflava flows. useful in interpreting the general history of The Pleistocene topography of the Great this species, both chronologically and geo Basin was in part chatacteri~ed by lakes graphically, the geological data for the area and rivers that formed a netw:ork ofintercon: under consideration are sufficiently limited 1991] GENETIC VARIATION IN THOMOMYS TOWNSENDII 139 that the correlation of specific historical tan Basin. An~yses of the strengths of the events with specific patternsfound in Thomo barriers of48 drainage basins suggest the fol mys townsendii cannot be undertaken. A re~ lowing: (1) the Bonneville=upper Snake bllr view ofthe distributions ofother Great Basin rier is not a strong one, as evidenced by the organisms willbe used to help reconstruct the Bonneville flood, (2) a FIarney Basin~Snake historically suitable habitats and dispersal River connection is suggested, (3) with the routes for Thomomys townsendii. above two exceptions, colonization between Evidence from fossil and Recentfreshwater the Snake River Basin and the Great Basin fish and invertebrate distributions not only is rare, and (4) the Lahontan Basin shows substantiates the eJeistence of hydrographi~ low internal ba,rrier strength, but the barrier cally related connections through the Great between Lahontan and BonneVille is sub Basin, but also suggests possible corridors stantial. through which T. totvnsendii could have colo~ While T. townsendii is not dependent nized the northern part of its present range. On aquatic habitats, it exhibits a clear prefer In his study ofthe fish ofthe Lahontan basin, ence for lacustrine soils. The importance of Snyder (1918) found that some fish species the fish and mollusk eVidence lies in describ~ are most closely allied to the fauna of the ing a network of intermittently connected southeastern Oregon lakes. He found poten aquatic habitats in an area that has since been tial affinities with the Bonneville basin also. replaced by, Or bisected by, lava flows and Miller (1958) similarly discusses affinities desert. these faun~ studies and the brief between the Lahontan and Bonneville basins geographichistories presented suggest poten~ basedon theirrespective fish faunas and notes tial links via Pleistocene hydrographic fea that there may have been a historical connec tures on both the eastern and western ends tion between the basins in the Pliocene or ofthe CUrrent townsendii distribution. Details early Pleistocene. Miller found the fauna of of the type of stream~to-streamnetwork that the upper Snake River to be sufficiently simi must have eXisted to make gopher dispersal lar to the Bonneville basin to be included in through southeastern Oregon possible are the Bonneville system. presented. by Bisson and Bond (1971). Taylor (1960) presented distributional in Whether the route was as far east as Harney formation for fossil and Recent freshwllter Basin or further west, through northeastern clams and snails that suggests a historical C~ifornia, as suggested by Taylor's (1960) network oflakes and streams leading from the data, remains unclear. However, such a Snake River through southern Oregon and western hydrographic network appears to be northeastern California into the western a more viable corridor for pocket gopher Great Basin. His evidence does not suggest a dispersal than does the eastern Lahontan Lahontan~Bonneville link. the investigation Bonneville~Snake connection described for by Bisson and Bond (1971) into the origin of Great Basin fishes by Miller (1958). In fact, the fish ofHarney Basin in southeastern Ore Thomomys toumsendii is not known from the gon also supports the suggestion ofa dispersal Lake Bonneville region, and, as outlined in corridorfrom the Snake River to southeastern the companion paper (Rogers 1991), there is Oregon. An ancient. connection between no fossil evidence to suggest that this was ever Harney Basin and Malheur RiVer provided a part ofthe species range. route for fish dispers~ and explains the faunal Morphological evidence presented by affinities seen today. Stream connections Davis (1931) suggests affinities between Poca:~ were obstructed, in part, by a late Pleistocene tello (T. t. similis) and northeastern Nevada lava flow, damming the water of Ha,rney (T. t. elkoensis) animals and therefore an east Basin. ern link between northern and southern pop A detailed an~ysis offish distributions and ulations. to explain the apparent similarity, drainage basins of the intermountain region, he proposed that the gophers followed the including the Great Basin, is presented by northern Bonneville shores into Idaho during Smith (1978). His data on the distributions of the flood at Red Rock Pass (late Wisconsin). 83 fish species suggest a well-linked Klamath This study gives no clellr indication ofwhich (southern Oregon)~Snake~Bonneville group route mllY have been taken. Only the color and a separate group representing the Lahon- analysis shows any obvious similarity between 140 M.A.ROGERS [Volume 51 populations of the two river systems, again (Patton 1981). Results of this study are also between T. t. elkoensis and T. t. similis popu~ consistent with the expected effects of these lations (Rogers 1991). However, it hasbeen kinds ofhistorical and population factors. demonstrated that pelage data should be treated carefully in interpreting the historical Patterns ofGenetic Variation in patterns ofpocket gophers. Studies ofT. bot Thomomys townsendii tae (Patton etal. 1979, Smitha,nd Patton 1980, The results ofgenic andkaryotypic analyses Smithetal. 1983) have clearly shoWn the envi~ of Thomomys townsendii reveal a pattern ronmental influence on variation in pocket ofextreme genetic homogeneity across a very gopher coloration. In fact,the correspon~ broad and disjunct geographic range. The dence between pelage color! and vegetation parameters of intrapopulation variation pre zone suggests that dark soils may be a factor sented in this study categorize townsendii as these two subspecies have incommon (Rogers being one of the least variable of the pocket 1991). gopher species studied. this species shows A further suggestionthat ,Red Rock Pass less variability (63% monomorphic) thanother did not serve as a dispersal corridor for members ofthe bottae-group or Thomomys in T. townsendii is presented by Wells's (1983) general (T. bottaei 9% monomorphic, Patton discussion ofmontane conifer dispersal in the and Yang 1977; T. umbrinus: 17%, Hafner et al. Great Basin. Red Rock Pass is one ofthe two 1987; T. talpoides: 29%, Nevo etal. 1914). topographically high conneQtions between Thomomys townsendii is also low in het the high central plateau of the Great Basin erozygosity values and in average proportion and the western Rocky Mountains. The 1820~ of polymorphic loci relative to other rodents meter contour that Wells usesto outline these and to mammals in general (mean H = .039 high mountain connections is graphically for 50 mammals species examined, Avise and complementary to the distribution of T. Aquadro 1982; also, see reviews by Selander townsendii and delineates historical and et al. 1974, Powell 1975, Nevo 1978, Kil~ present~dayboundaries to eastward dispersal. patIick 1981). While the ranges in population values for P and H in T. townsendii overlap DISCUSSION with those values in other gopher species, the means across populations in this species are Causes of genic homogeneity in fossorial among the lowest known in pocket gophers rodents have been of particjllar interest to (P = .053 and Ii = .012). The means and some authors. Nevo and Shaw (1972) and ranges ofvalues for closely related species are Nevo et al. (1974) have applIed a model for P = .334 (.130~.565) andIi = .093 (.030-.169) selection for homozygosity in uniform sub for Thomomys bottae (Patton and Yang 1977) terranean environments to data for Spalax and P = .183 (.043~.391)and Ii = .051 (.008 ehrenbergi and Thomomys talpoides, which .100) for Thomomys umbrinus (Hafner et al. both show very low levels of genic variation 1987). Of the pocket gopher species exhibit~ within and among populations. Other authors ing lower levels of within~populationvariabil have found this selection model inappropriate ity (Selander et aL 1974, Honeycutt and for explaining the genetic patterns seen in Williams 1982, Rafner and Barkley 1984), Thomomys, arguing that features of pocket none maintains these low levels so uniformly gopher population biology and stochastic over such a broad geographic range as does factors explain their genetic divergence (Pat Thomomys townsendii. ton and Yang 1917, Patton 1981, Patton and Analyses ofgenic variation have also shown Feder 1981). tn contrast to. the pattern of a. high level ofsimilarity (S = .956) among the genic homogeneity seen in talpoides, varia populations oftownsendii. While these values tion within and among populations of bottae are not unique among rodent studies (Avise is high. Small, effective population size, non 1974), they are extreme among Thomomys random breeding, and migration rate in con species (T. bottaei S = .81, Patton and Yang junction with historical changes in habitat, 1977; T. umbrinus: S = .80, Patton and Feder distribution, and population connectedness 1978, and Hafner etal. 1987; T. talpoides: S= are cited as variables that have affected the .84, Nevo et al. 1974). Whereas values in genetic patterns of T. bottae and its relatives townsendii are high in comparison to bottae 1991] GENETIC VARIATION INTHOMOMYS TOWNSENDII 141 as a whole, comparable levels of similarity type is the presence of a single acrocentric are seen among the Great Basin populations pair ofchromosomes in samples ofT. t. bach ofbottae (Patton and Yang 1977). mani and T. t, relictus. This pattern differs The nature of allozymic variation between from the karydtypes reported by Wentworth populations in this species reflects the effects and Sutton.(1969) and Thaeler (1913), who ofrare variants at a small number ofthose loci group the chromosomes of all seven subspe examined. Wright's F ~statistics show that cies in the following way: 12 pairs ofmetacen populations are the units ofgreatest differen trics; 22 pairs ofsubmetacentrics, and 3 pairs tiation (highest FST) ofany ofthe hierarchical ofacrocentrics (FN = 142); theX chromosome levels compared. Within a geographic region, is a large submetacentric, and the Y is a small populations are more differentiated than are acro~ or subtelocentric. The lack ofagreement the subspecies of that region. However, all with the published results can be explained, populations are united at a high level ofgenic in part, by the'lack ofresolution in standard similarity, S = .939. FST values for Thomomys karyotypes. Without the aid ofbanded mate bottae (Patton and Yang 1977) similarly show rial, one is forced to use an arbitrary morpho interpopulation differentiation representing a logical classification. This may explain the dis large component of the variation. This was agreement in the numbers of chromosome true throughout the populations sampled in pairs in each ofthe biarmed categories. this study and within each of the geographic Although the distinction between uni~ areas. Variation within this group of pocket armed and biarmed chromosomes is more gophers is generally greatest between the easily made, Wentworth and Sutton (1969) local units ofany given area. The subspeci£ic and Thaeler (1913) assigned three pairs of or regional categories within these species chromosomes to the uniarmed group, while are thus internally quite heterogeneous. one ornone was so assigned in this study. The The distinction between the two geo classification of the single uniarmed pair in graphic regions is the onlyclear division atany T. t. bachmani and T. t. relictus remains level in T. townsendii and is due mostly to the somewhat ambiguous. However; the exis effect ofthe single locus GOT-I, which is fixed tence of such a variant in an otherwise kary for alternate alleles in populations of the otypically homogeneous species is almost to Humboldt River/Honey Lake Valley versus be expected. Variation seen in related species the Snake River. The divergent populations is often of thisna.ture; arm number may be within each of the regions are differentiated polymorphic or polytypic, while the diploid slightlyfrom the remainderoftheirrespective number remains constantwithin a given taxon clusters by shifts in allele frequencies or by (Patton 1972, Patton and Feder 1978, Patton possession of an allele not found elsewhere. and Sherwood 1982, Hafner et al. 1987). In These few variants are of little absolute sig~ addition, it is significant that the alternative ni£icance in distinguishing these populations karyotype is found in the Honey Lake Valley from the remainder oftheir cluster, however. and in the neighboring T. t. bachmani popu Patton and Feder (1981) showed in studies of lations which togetherform the western com Thomomys bottae at Hastings Reservation ponent of the Humboldt River region. Since that genic differentiation between fields can southeastern Oregon samples of T. t. bach be as high as the differentiation found be mani were not karyotyped in this study, it tween subspecies ofa region in T. townsendii cannot be determined whether this pattern (T. bottae FST = .142, T. townsendii F ST = extends to those populations as well. .151 and .124). Consideredin this context, the The discrepancy found in the morphology genic differentiation represented among the ofthe Ychromosome is less easily explained as populations and subspecies within the two preparation artifact. The Y chromosomes pre geographic units is remarkably low. sented by Wentworth and Sutton (1969) are Results of the karyotypic analysis corrobo~ substantiallylargerthan the dot chromosomes rate the findings of Wentworth and Sutton found in this study. As all other bottae-group (1969) and Thaeler (1973) that the diploid karyotypes exhibit the minute acrocentric number is 76 in all Thomorn,ys townsendii morphology in·, the Y chromosome (Patton subspecies. The only differentiation found and Dingman 1970, Patton 1972, Patton and within the otherwise totally biarmed karyo- Feder 1978, Hafner et al. 1987), it is probable 142 M.A.ROGERS [Volume 51 that Thomomys townsendii also shows this nipted, and therefo:re genetic variability pattern. has been maintained. Historically, isolated . The amount of genic variation found in populations ofbottaewere notchromosomally geomyids is often accompanied. by greater differentiated to the point ofbeing reproduc~ k:nyotypic differentiation than that seen in tively isolated from other populations. There Thomomys townsendii a.s well. Average simi fore, recontact with populations could have larity values within Geomys h1.ltsarius ;md meant interbreeding, and bottlenecks that G. persol1atus (data of Y. J. Xi:Q1 presented occurred due to isolation could have been by Nevo et a1. 1974 and Selander et al. 1974) overcome. are .16 and .86, :respectively, andboth species are karyotypically polymo:rphic. Ofthe speci Evolutionary and Biogeographic History mens ~alyzed allozYmically,by Kim (Selan" ofThomomys townsendii der et al. 1974), G. bursarius wa.,s represented While Thornornys townsendii does not by seven chromosomal forms (2n = 70, 72, 74 share the genetic diversity chara.cteristic of and FN = 68=74), and G. personatus was the bottae~group, its distributional history is represented by five (2n = 68, 70 and FN = mo:re similar to that of bottae, undisturbed 68-74). Cmomosom;ti variation in these spe~ by gla.ciation, than to t4lpoides, whose boreal cies is due to Robertsonian changes (centric distribution .was sculpted by glacial influ" fusions and fissions), inversions, and translo ences. Furthermore,· the genic pattern in cations. Examination ofcontact zones in other totvnsendii is :remarkably similar to the kind of studies of G. butsarius has shown tha.t barri- variation seen in the Great Basin populations o ers to genetic exchange exist between some ofbottae (Patton and Yang 1977). The concor cmomoso:Q1al forms (Honeycutt and Schmidly dance between genic patterns of these two 1979, Tucker and Schmidly 1981, Bohlin species in the sa:Q1e geographic area suggests and Zimme:rman 1982). Within the Thomo that a si:Illilar history of biogeographic and mys talpoides complex (Neyo et a.1. 1974), stochastic eVents may have influenced their average genic similarity for the six karyotypic evolution in the Great Basin. variants examined is .874. Thaeler believes Given that historical biogeography may that these mo:rphs represent a.t least five have affected the genetic patterns of bottae separate species (Nevo et al, 1974). Within and talpoides, it is useful to understand the this complex diploid numbers range from evolutionary history and affinities ofT. town 40 to 60, and chromosomal changes are typi sendii before trying to interpret the patterns cally due to Robertsonian changes and peri seen in this species. Studies by Bailey (1915), centric inversions. Thaeler (1980), and Patton and Smith (1981) Patton and Yang (1971) discussed differ support a dichotomy within Thomomys. Mor~ ences in genetic variability of T. bottae and phologica.l, electromorphic, and karyological members of the T. 1;4lpoides complex. The analyses place Thomomys totvnsendii iJ1 the patte:rn seen in talpOid~sis on,e ofgenic homo bottae-group, which also contains T. bottae, geneity and chromosomal differentiation to T. urnbrinus, and T. bulbivorus. This group, the point ofrep:rodUctive isolation. Historical characterized as the "heavy~rostrum" g:roup fragmentation and isolation, of populations, by Bailey (1915), can be distinguished kary~ bottlenecks, and founder effects may have otypically as having diploid numbe:rs ranging resulted in the fix;ation of chromosomal re from 74 to 88 (Tha.eler 1980, Patton and Sher" arrangements and fix;ation for alternate alleles wood 1982). These characters distinguish the where loci we:re once polymo:rphic. It is sug~ bottae=gtoup from T. talpoides and other gested that if :reproductive isolation were members ofBa.iley's (1915)"slender~rosttum" attained, genic homogeneity would be IIlm.n~ . group (with diploid numbers ranging from tamed in these isolates. In contrast, chromo 40 to 60 [Thaeler 1980]), including T. clusius, SO:Q1al variability of T. bottae does not seem T. idahoensis, T. mazama, and T. monticola. to affect reproductive isolation ~d interpopu Thaeler (1980) formalized this dichotomy by lation genic variability and Within~poptilation recognizing the slender- and heavy"rostrum heterozygosities are high. E~tensive range groups as separate subgenera, Thomomys fragmentation is not suggested for bottae, and Megascapheus, respectively. Patlon and gene flow is believed to ha.ve been less inte:r" Smith (1981) did an electrophoretic ana,lysis of 1991] GENETIC VARIATIQN iN THOMOMYS TOWNSENDII 143 six Thomomys species (talpoides and m,onti~ The evolution;iry affinities between town cola in addition to the bottae-group species) sendiiand bottae atefurther reflectedin Wor= and tre~ted T. tou,m~endii as pa,rt ofthe bottae mation from studies ofthe distribl.!tion of go= unit; all alleles found in townsendii Were also pher lige. Pocket gopher spe9ies often serve found in bottae. The same study determilled as hosts to mote than one species oflol.!se, and that bottae and townsendii were a sister a single louse species may be fOl.!nd On more group relative to umbrinus, that towilseridii than one spePies of pocket gopher (Emerson stemmed from Within one of the major geo and Price 1981). T. towns?ndii is known to be graphic units of T. bottae, and therefore the host of one lOl.!se species, Geomydoecus that bottae was paraphyletic relative to tOW1]r idahoensis, e4cept in two zones ofhybridiza= sen,dii. .An analysis of cranial shape also tion with bottae (Patton et al. 1984). In Honey supports this relationship (Patton and Smith Lake Valley T. town,sendii relictus hybridizes 1989). with T. bottae saxqtilis and T. bottae canus. The evolution~ relationships of hottae, Inthese hybridzones the bottae 10l.!se is found townsendii, and um,brinus are manifested in on some ofthe "pure" townsendii as well gs On comparisons of karyotypic morphology and some of the hybrids. G. idahoe11$is belongs cellular DNA content, and :i.n hybrid zone to the oregonus species grOl.!p. Another mem studies. The whole=-arm heterochromatin ber of this group, Geomydoecus shastensis, found ill townsendii resembles the pattern is morphologically very similar to idahoensis seen ill bottae and is distinct from the inter" and, interestingly, is the l@se found on T. stitial heterochromat:i:r:l. characteristic of um bottqe sqxattlis (Price and HeUenthal1980).. brinus (Patton and Sherwood 1982). Simi= the level of chromosomal and geni9 simi larly, the amOl.!nt of cellular DNA (C -vall.!e) larity of. townsendii and bottae and the evi fOl.!fid ill townsendii is with:i.n the range fOl.!nd dence for paraphyly(Patton and Smith 1981, ill bottae populations, whereas urnbrinu,s has 1989) suggest that Thomomys townsendii Was a higher C-value than would be e:x:pected for derived from thomomys bottae. In addition; T. bottae With a similar karyotype (Sherwood Bailey (1915) and Davis (1937) recogniZed the and Patton 198~). morphological ~ties between the two spe The karyotypic similarity between botta? cies. Further examination of the nature of and townsendii can be contrasted to the this morphological similaritY de:r:nonstrates meiotically problematic strUctural rearrange= that there is basically no difference in cranial ments found between bottae and umbrinus shape between the two species. It has been (Patton 1973, Patton 1981), although studies sl.!ggested that the difference in body size of hybrid zones indi9gte thgt genetic intro~ between the two species may have resulted gtession does not. ocPl.!r ill either 9gse. in differences in demographic features that Both town,sendii and umbrinus a,re knOWll to eventually created genetic isola,tion (Patton hybridize minimally with bottae (12% and @d Smith 1989). 11-15% hybrids, respe9tively). In bottae= What can be determilled from the results umbrinus hybrids, PI males are sterile, ofthis studyregarding the nawre ofthe distri female fertility is reduced, and there is no bution of this species from the time 'of its genetic introgression (Patton 1973). While fOl.!ndillg popula,tion. to the present? Was some F1 individuals fro:r:n crosses of bottae the range oftownsendii once continl.!OUs over and townsen,dii show normal reproductive the northern Great BasIn.and subsequently 9hara9tei-istics, there is again no eVidence for fragmented by unsuitable habitats? Or is the genic or morphologi9al introgression illto the distribl.!tion of townsendii the result of rela, pgrental populations (Patton etal. 1984). Both tively reCent dispersal fro:r:n one segment of pairs of hybridizing species remain geneti the species range? cally isolated species, bl.!t only the bottae The concordance of genetic data with his um,brinus case is likely to be chromosomally tori()al and biogeographic patterns has been mediated. The lack ofgenetic:i.ntrogression :i.n emphasized in other p09ket gopher studies. the presence ofkaryotypic similarity between Patton (1981) discusses the influenges of his bottae and townsendii suggests other influ tory and features ofpocket gopher population ences. The signi:ficance ofbehaviorill this caSe biology ill the cOntext of the agreement has been suggested (P;Jtton et al. 1984). of karyotypic 'and geniG data for Thomomys 144 M.A.ROGERS [Volume 51 bottae. Morphological patterns of differentia distribution has always been characteristic of tion, however, ate often best explained by this species and also suggests in which areas environmental factors such as quality and a more continuous distribution may have color ofthe soil and nutritional value ofavail- existed historically. i able food and probably do not provide mUyh For instance, analyses of the genetic data information regarding the eVolutionary his indicate that populations of the Sn~ke River tory of T. bottae (Smith and Patton 1980). belong to a very cohesive cluster (5 = .984) Further studies (Patton and Btylski 1987, despite the geographical separation of the Smith and Patton 1988) have demonstrated T. t. similis populations by more than 100 that while pelage coloration and body size ate miles. Fossil evidence from Wilson Butte strongly influenced by environmental factors, Cave (Gruhn 1961, Rogers 1991) suggests cranial shape variation is more likely to repre~ that continuity between the western (T- t. sent genetically based changes. In the ab; owyhensis and T. t. townsendii) and eastern sence ofan analysis focusing on cranial shape (T. t. similis) parts ofthe Snake River distribu in Thomomys townsendii, the results of the tion may have existedfairly recently. Only the genetic analyses will be emphasized in devel eastern T. t. townsendii population (22) is oping a historical scenario for this species. clearly differentiated from the other Snake While the nature ofthe historical distribu River populations. Most of the samples from tion of Thomomys townsendii is not easily this region differ from each other in allele determined by the results ofthis study andits frequencies and, ill one case, in the presence companion paper (Rogers 1991), eVidence of a rare ADA allele in low frequency. The does exist that eliminates some possibilities magnitude ofthe differentiation ofthe Grand and suggests others. A description ofthe his view T. t. townsendii population (22) is pri; torical biogeography of the northern Great matily due to the appearance of the PGM c Basin that focuses <;>n the chahges in and the allele in relatively high frequency (.625). This availability of suitable habitat for Thomomys allele was not encountered in other popula townsendii has been presented. It indicates tions. The reaSOn for the divergence of this that the wetter climates of the Pleistocene of population is not cleat, butgeologic evidence the northern Great Basin produced a network suggests that historically this area may have oflacustrine andfluviatile soils thatWere occa beenisolated bylavaflows (Malde and Powers sionally continuous until desert conditions 1962, Malde 1965). prevailed, rendering sOme of these soils un While variation in the Humboldt River inhabitable. The continuity of these habitats region is very limited, more genetic differen could have provided historical corridors for tiation is seen among these populations than dispersal. among those of the Snake River region. the However, while these historical connec only karyotypic differentiation documentedill tions can be reconstructed, it is unlikely that this study is seen in representatives of two the species could have maintained a continu of the Humboldt River subspecies. Analyses ous range over most of the northern Great ofthe populations ofthis region resulted in a Basin at one time. Rather, it seems feasible cluster ofsix genetically Very similar (5 > .97) that after the origin from T. bottae, segments populations and three more differentiated ofthe range ofT. townsendii were connected populations: Lovelock, Narrows; and Austin. to various degrees at different points in its The members of the more cohesive cluster distributional history. Barriers to the continu include T. t. elkoensis; two of the four T. t. ity of the range probably arOse at different bachmani populations, and the Honey Lake times in different places. As climates in Valley (T. t. relictus) populations. These sam creased in aridity and lakes receded, some ples are differentiated from One another by areas became unable to support gopher popu the presence of one to four alleles in low fre~ lations; and townsendii remained along rivers quency that are notfound inotherpopulations and receding lake shores. Further fragmenta ofthat cluster. tion and isolation ofsome sections ofthe range Despite their peripheral location, Honey of T. townsendii probably occurred with Lake Valley populations display affinities the intrusion of lava flows. Information from to the Humboldt tliver populations as they this study supports the ideathat a fragmented did in the morphological analyses of the 1991] GENETIC VARIATION INTHOMOMYS TOWNSENDII 145 companion paper (Rogers 1991). Examination this valley. Isolation by inadequate habitat of the F ~statistics further supports the inclu may have contributed to separation and sion ofHoney Lake Valley popl.llations in the divergence of this population. Furthermore, Humboldt River geographic region. Subunits T. talpoides is currently found in this valley within regions show less differentiation when only 24 miles north ofAustin. The presence of Honey Lake Valley is regarded as a part ofthe this species may have subsequently enforced Humboldt River regiOn. Greater genic simi; existing physical barriers. larity is found between Honey Lake Valley Another possible explanation for the differ~ popl.llations and certain Humboldt River entiation eXists if this area was one of the populations than among some Humboldt earliest colonized subsequent to the deriva River populations. The same is true at the tion of townsendii from bottae. Davis (1931) subspecific level. Historically, Honey Lake remarked that, ofthe townsendii subspecies, Valley was part of Lake Lahontan, so the T. t. nevadensis is structllrally most similar affinity to these populations is not surprising. to bottae. Ifthe Reese River valley were colo~ Only the' karyotypic analysis shows Honey nized from the SOl.lth rather than from the Lake Valley populations deviating from most Humboldt River to the north, it is possible other populations in that T. t. relictus and that distribution through this area was never T. t. bachmani possess a single acrocentric continuous with the remainder of the Illlm; pair among an otherwise biarmed comple; boldt Rlver distribution, or that this area was isolated relatively early. . ment ofchromosomes. The genic divergence of two of the T. t. Much of the information presented above regarding genic differentiation among T. bachmani populations (Lovelock and Nar" toWnsendii populations has a morphological rows) is due to the prevalence ofa single allele counterpart from the analyses of cranial and not found elsewhere. the T. t. nevadensis pelage differentiation (Rogers 1991). How sample from Austin is the most differentiated ever, relative to the genic data, morphological population of the Humboldt River region. characters reflect a general pattern of more atttibu~ The magnitude of this divergence is overlap with fewer strongly differentiated table almost entirely to a fixed difference for a populations. As one would expect, some PREALB allele. This allele is otherwise only aspects of the biogeographic history of T. present, in moderately low frequency (.200), toivnsendii populations seem to be more in a single T. t. elkoensis population. Austin closely reflected in the genic data. For in" and Narrows animals are consistently differ; stance, the four most divergent populations entiated from the other populations in both (Austin, Narrows, Grandview-t., and Love genetic and morphological analyses (Rogers lock) all have an alternative allele as the 1991). This may reflect their peripheral posi; predominant one at SOme locus. This kind of tion relative to the Lahontan drainage system. pattern is sometimes associated with isola AJthough located in a Pleistocene lake basin, tion, whether it be a result ofa founder event Narrows is actually outside the Lahontan or range fragmentation. For some of these hydrographic basin, but it was probably once populations the features maintaining biogeo connected to the main basin. the divergence graphic isolation are obvious, although the of this population from other units. of this mechanism by which isolation began is diffi; region probably reflects its isolation due to a cultto determine. the factors that mighthave combination of encroaching desert regions contributed to the isolation of the Austin, and the influence ofthe drainage divide. Narrows, and Grandview-t. populations have Austin is located in the Reese River valley, already been discussed. Explainingthe extent south of the Humboldt River. The Reese of differentiation seen in the T. t. bachmani River was a tributary ofthe Humboldt River population at Lovelock is more problematic. dl.lring Lahontan times. but was not repre Since this areais partofthelarge and continu; sented by a pluvial lake (Morrison 1965), ous Pleistocene Lake Lahontan, reasons for suggesting relative aridity in this basin even isolation and differentiation are not clear. during pluvial times. Today the Reese River However, itis peripherallylocated in the spe reaches the Humboldt only intermittently cies range, and the nearest neighboring speci so that habitable areas for T. townsendii mens examined in this study are from about may be restricted to the southern portions of 75 miles up the HumboldtRiver. 146 11. A. ROGERS [Volume 51 Thomomi]s townsendii is characterized not divergence from T. bottae. Derivation from a only by genetic homogeneity Within pop"Qla single highly homozygous founding popula tions, but by shifts in allele frequencies and tion would proVide the genetic background appearance ofrare alleles in high frequencies necessary for dominant alleles to be main :unong populations. In addition; patterns of tained across the distribution despite discon", genic similanty between populations do not tinuities that inevitably arose after T. town correspond to geographic dist@ce, and popu= sendii colonized the area. High interpopula lations of different subspecieS or geographic tion similarity suggests thatlow levels ofgenic sub"Qnits are often more similar than are other Variability were maintained as the founding members ofthe same subspedes. These fea population eventually dispersed through the tures do not suggest a broadly continuous Great Basin. The remnants of Pleistocene species distribution historically. It is not sur lakes served as a corridor for the dispersal of prising that this pattern instead suggests the the species through northern Nevada, north effects ofdrift and :independent evolution on eastern California, and southeastern OregotL population isolates that arose from the colo= and possibly into lake beds to the West ofthe nization ofnew areas Or by the fragmentation present Oregon distribution,· judg:ing from of once more continuous habitat. Pocket fossil evidence (Allison 1966). Colonization gophers are sl::tongly susceptible to drift due of this area probably proceeded as a series to .small; effective population size and their offounder events with portions of the distri inability to cross various environmental barri~ bution becoming fragmented. the back ers, which may include other gopher species. ground of genetic hOmogeneity was main", A reasonable explanation of the events ill taiIled With diversification in the form of the origin and distribution of 'thomomys allelic variants. The fixed difference at the townsendii begtns. with a relatively recent GOT"l locus and the homogeneity seen deriv~tion from Thomomys bottae. In their among the Snake :River populations suggest analysis of T. bottae genic evolution, :Patton thatthis region was probably coloni:?:ed by a and Smith (1981) discuss what appears to be single founding population from the Hum'" a decreased rate of allo~mic -change :in boldtRiver region. Subsequentto dispersal of T. townsendii. However, they suggest that the species across the Snake River plain, dis'" their resmts probably do not represent a.ctual continuities to the range arose and popula rate changes; but reflect low levels ofwithin= tions became slightly differentiated. population variation in T. townsendii. they The relative homogeneity of the Snake hypothesize that this homogeneity is the re River populations also suggests this region sult ofa fO"Qnder event and subsequent isola has had less time to differentiate than has tion, and hl.ter morphological studies suggest the Humboldt River region. To demonstrate that the deriva:!ion from boffg.e was a rela,,,, the relative difference in age among and be'" tively recent event (Patton and Smith 1989). tween the Snake River and Humboldt River While sufficient Mormationis notavailable to areas, estimated divergence times between pinpoint when and where townsendii arose populations were calculated and averaged for from boffaej an origin from the southern part these two major geogra.phic regions. Values of the current townsendii distribution has generated for Nei's genetic distance and the been suggested. Davis (1937) noted the simi methods ofNei (1971) as described by Patton larity between bottae and the most southern and Yang (1977) Were used for these esti", subspecies, T, t. nev(Ldensis j Md later stated mates. An average divergence time between that· townsendii had probably moved :into pop"Qlations of these two geographic units Idaho from the south (Davis 1939). In addi", is 255,400 years before present (maximum, tion; in comparing T. townsendii to T. bottae 475,000 YBP; minimUm, 190;000 YBP). The populations; Patton and Smith (1989) show same value, calculated for pop"Qlations of that townsendii is genetically most similar to the Humboldt River region, is 92,083 years bottae populations from central and sO"Qthetn (maximum, 300,000 YBP; minimum, 5,000 California. YBP), while that for the Snake River popula The low level ofvariation seen throughout tions is 27,143 years (maxhnum, 95,000 YBP; T. townsendii suggests that stages ofisolation minimUm, 0 YBP). As additional reference and subsequent genetic drift followed the points, these divergence times were adjusted I 1991] GENETIC VARIAtION INTHOMOMYS TOWNSENDII 147 using the albumin immunological distance I a more homogeneous unit. The variants ob~ method (SaIich and Cronin 1976, Sarich served in this study are usually quite subtle 1971), again following Patton and Yang (1977). and do not warrant taxonomic distinction. These estimates show average divergence. The population from Austin deserves spe" time between populations of the two river cial consideration here. These individuals are systems to be 1,072,680 YBP; among Bum" clearly differentiated morphologically (Rogers boldt lliver populations the average diver 1991) and genically. The historicalbiogeogra gence time is 386,749 YBP, and within the phy of the T. t. nevadensis distribution and Snake lliver system it is 114,001 YBP. While the presence ofT. talpoides to the north sug" these values present very different approxi~ gest effective and possibly early barriers to mations of divergence tilnes within Thomo= genefloW with Thomomys townsendii popula mys totvnsendii, they do indicate a general tions to the nOrth. The intermediate level of time frame and what the sequence of events differentiation seen in the Austin sample is might have been. These estimates suggest noteworthy in the context of understanding that divergence among the Snake River popu evolutionwithin this species butdoesnotwar~ lations occurred subsequent to the isolation rant the level of distinction given to the two between the northern aIld southern parts of major geographic regions. the distribution andafter diversification ofthe The absence ofgenetic variation in Thomo" Humboldt River populations had begun. mys townsendi,i is most remarkable consider Physical aspects of the two regions may ing the phylogenetic affinities between have also contributed to the different degrees townsendii and T. bottae. The patterns of of morphological and genetic differentiation genetic variability found in T. townsendii and seen between the Humboldt River aIld Snake T. bottae are strikingly dissimilar. Thomomys River regions. The greater biogeographic bottae is characterized by extreme interpopu=. diversity and increased aridity of the Hum~ lation differentiation in allozymes (Patton and boldt River region may have promoted isola= Yang 1977), With an average Rogers' S be tion;md differentiation in this area. Features tween populations of .81. Similarly, karyo ofthe habitat, particularly soil color and qual typic variability is great, consisting primarily ity, may be mOre homogeneous along the of variation in arm number. These variants Snake lliver drainage and influential in main" are the result ofadditions and/or deletions of taining morphological homogeneity. heterochromatic arms (Patton and Sherwood 1982). AlloZymic Patterns of differentiation Systematic Conch:tsions are concordant with geographic patterns The genic data are the strongest indicators of change in.karyotype (Patton and Yang of differentiation that is greatest between 1977, Patton and Smith 1981). ICaryotypic and the regions represented biogeogrliPhically by allozymic patterns are also concordant in the Humboldt River aIld the Snake lliver T. totvnsendii in that they show homogeneity drainage systems. One fixed difference across geography. The type of chromosomal (GOT-I) and the appearance of additional polymorphism found in T. townsendii is char" alleles at other loci sep Thomomys townsendi nevadensis Bailey, 1915, North in an otherwise completely biarmed comple American Fauna 39:44. ment. The variation found in the genic data is Thomomys relictus Grinnell, 1926, University ofCalifor primarily based on shifts in allele frequencies nia Publications in Zoology 30(1):2. Type from in some populations and the presence ofrare valley of Susan River two miles south of Susan alleles in others. F -statistics indicate that ville, Lassen County, California; Museum ofVer_ tebrate Zoology no. 35271. genic differentiation is greatest between pop Thomomys townsendiirelictus Grinnell, 1933, University ulations and slightest between the subspecies ofCalifornia Publications in Zoology 40(2): 137. within each geographic region. Subspecmc Thomomys townsendii bachmani Davis, 1937, Journal of integrity is not maintainedbyany units within Mammalogy 18(2): 150. Type from Quinn River the two regions. Given this high level of Crossing, 4100 feet altitude, Humboldt County, Nevada; Museum ofVertebrate Zoology no. 7855. homogeneity, a conservative interpretation Thomomys townsendii elkoensis Davis, 1937, Journal of of the details of differentiation is necessary Mammalogy 18(2): 151. Type from Evans, Eureka in assessing meaningful taxonomic categories County, Nevada; Museum ofVertebrate Zoology within Thomomys townsendii. no. 70583. Thomomys umbrinus bachmani Hall, 1981, The Mam The only pattern that is concordant among mals ofNorth America 1: 477. the datasets presentedhere and inthe accom~ Thonwmys umbrinus elkoensis Hall, 1981, The Mammals panying paper (Rogers 1991) is the break be~ ofNorth America 1: 482. tween the Snake River populations and those Thomomys umbrinus nevadensis Hall, 1981, The Mam of Honey Lake Valley and the Humboldt mals ofNorth America 1: 488. Thomomys umbrinus relictus Hall, 1981, The Mammals River. Discriminant function analyses based ofNotth America 1:492. on cranial characters show some differentia tion between these two gr0upS, although thomomys townsendii townsendii (Bachman) overlap exists (Rogers 1991). A fixed differ Geomys townsendii Bachman, 1839, (from Richardson's ence at the GOT~1 locus provides a clear split manuscripts) Journal of the Academy of Natural between these two regions despite their gen Sciences, Philadelphia8(1):105. Typelocality "Co erally high genic similarity (5 = .94). lumbia River" is incorrect according to Bailey (191~:42), who feels the locality is near Nampa, The evidence from this study suggests that Idaho; Academy ofNatural Sciences Philadelphia only two geographical units can appropriately no. 147. be given subspecmc distincti()n. The ranges Thomomys townsendiiAllen, 1893, Bulletin oftheAmeri of these units correspond to' (1) the Snake can Museum ofNatural History 5(5): 61. River Plain of Idaho and southeastern Ore~ Thomomys nevadensis atrogriseus Bailey, 1914, Proceed ings of the Biological Society of Washington 27: gon and (2) the Humboldt and Quinn River 118. Type from Nampa, Idaho; United States drainage systems in Nevada, southeastern National Museum, Bioiogical Survey Collection Oregon (vicinity of Malheur Lake and Lake no. l81, 196. Alvord), and Honey Lake Valley, California. Thomomys townsendi townsendi Bailey, 1915, North American Fauna39: 42. Within each of these regions there is local Thonwmys townsendii townsendii Whitlow and Hall, differentiation in some characters, but partic~ 1933, University of California Publications in ular populations or subregions are not con Zoology 40(3): 255. sistently delineated across many characters Thomomys townsendii owyhensis Davis, 1937, Journal of and modes ofanalysis (Rogers 1991). The cur~ Mammalogy 18(2): 154. Type from Castle Creek, eight miles south of Oreana, Owyhee County, rently recognized subspecies T. t. bachmani, Idaho; Museum ofVertebrate Zoology no. 67490. T. t. elkoensis, T. t. nevadensis, and T. t. Thonwmys townsendii similis Davis, 1937, Journal of relictus are herein regarded as synonyms of Mammalogy 18(2): 155. Type from Pocatello, Thomomys townsendii nevadensis Merriam, Bannock County, Idaho; Museum of Vertebrate Zoology no. 46507. 1891, which has priority. The Snake RiVer Thomomys umbrinus owyhensis Hall, 1981, The Mam populations, representing T. t. owyhensis, mals ofNorth America 1: 489. T. t. similis, and T. t. townsendii, are treated Thonwmys umbrinus similis Hall, 1981, The Mammals as synonyms of Thomomys townsendii town ofNorth America 1: 494. . (Bachman 1839). Thomomys umbrinus townsendii Hall, 1981, The Mam sendii mals ofNorth America 1: 495. Thomomys townsendii nevadensis Merriam Thonwmys nevadensis Merriam, 1897, Proceedings of ACKNOWLEDGMENTS the Biological Society of Washington 11: 213. Typefrom Austin, Nevada; United States National I give special thanks to J. L. Patton for Museum, Biological Survey Collection no. =. his support, gUidance, and encouragement 1991] GENETIC VARIATION IN THOMOMYS TOWNSENDII 149 through all phases of this project and for descriptions of other species of quadrupeds the iinportant conttibutions he made to the found in North America. Journal ofthe Academy ofNatural Sciences, Philadelphia 8: 75-105. various drafts of this paper. D. B. Wake, BAILEY, V. 1914. Eleven new species and subspecies of W. Clemens, W. E. Bemis, B. D. Patterson, pocket gophersofthe genus Thomomys. Proceed~ and W. T. Stanley read this a!1d earlier drafts ings of the Biological Society of Washington 27: of the manuscript and provided useful criti~ 115-118. cism. R. D. Sage, M. F. Smith, M. Frelow, =-=-. 1915. Revision of the pocket gophers of the genus Thomomys. North American Fauna No. 39: and E. Tang provided advice and assistance 1-136. during the electrophoretic analyses. I thank BENDER, M. A., AND E. H. Y. CHU. 1963. The chromo D. S. Rogers for his guida,nce through my somes ofPrimates. Pages 261-310 in J. Btiettner chromosome work and for his v~uable Janusch, ed., Evolutionlit)' and genetic biology of Primgtes. Vol. i. Academic Press, NewYork. support. N. Jo, M. A. :Barros, a!1d S. W. BENSON, L. V. 1978. Fluctuations in the level of pluvial Sherwood also gave helpful suggestions on Lake Lahontan during the last 40,000 years. Qua chromosome preparations. Many ranchers ternary Research 9: 300-318. and l;mdowners of the northern Great :Basin BISSON, :p. A., AND C. E. BOND. 1971. Origin and distribu_ were helpful and cooperative in my collecting tion ofthefishes ofHarneyBasin, Oregon. Copeia 2: 268~281. efforts. For their support and confidence BOHLIN, R. G., AND E. G. ZIMMERMAN. 1982. Genic differ I thank my parents, Owen and Hally Rogers. entiation oftwo <::hrom~some races oftheGeomys I extend deep a,ppreciation to W. E. Bemis burs4rius complex. JOurnal of Mammalogy 63: for providing invaluable assistance in the field 218-228. BROECKER, W. S., AND A. ~UFMAN. 1965. Radiocarbon and CO!1sistent support throughout this pro chronology ofLake Lahontan andLake Bonneville ject. This study a,nd its companion paper are II, Great Basin. Geological Society of America based on a thesis completed in partial fulfill Bulletin 76: 537-566. ment ofrequirements for the M.A. degree in CRONQUIST, A., A. H. HOLMGREN, N. H. HOLMGREN, AND J. L. REVEAL. :1.972. Intermountain flora~vascular the Department of Zoology, University of plilllts of the Intermountain West, VSA. Vol. 1. California, Berkeley. FiIla,ncial support for Hafner Publishing Co., NewYork. 270 pp. this research was provided by the Depart~ DAVIS, W. B. 1937. Variations in Townsend pocket go ment of Zoology;· Musetnll of Vertebrate phers. Joutt\al ofMammalogy 18: 145-158. Zoology (two Louise M. Kellogg Grants~in~ ~. 1939. The Recent mammals of Idaho. Caxton :printers, C Zygogeomys (Rodentia: Geomyidae). Journal of MORRISON, R. ]3. 1965. Quaternary geology of the Great Mammalogy 65: 474-479. Basin. Pages 265-285 in H. E. Wright, Jr., and HAFNER, M. S., J. C. HAFNER, J. L. PATION, AND M. F. D. C. Frey, eds., The Quaternary ofthe United SMITH. 1987. Macrogeographic patterns ofgenetic States. Princeton University Press, Pnnceton, c:lifferenti ~. 1989. Population structure and the genetic and SHERWOOD, S. W., AND r. L. PATTON. 198~. Genome evolu morphologic divergence among pocket gopher tion in pocket gophers (genus Thomomys). n. species (genus Thomomys). Pages g84-304 in D. Variation in: cellular DNA content. Chromosoma Otte andJ. Endler, eds., Speciation and its conse" (Berlin) 85: 163-179. .. . quences. Sinauer Associates, Inc., Sunderland, SMITH, G. R. 1978. :I3iogeography ofintermountaih fishes. Massachusetts. Pages 17-42 in K. T. llaiper and J. L. Reveal, PATTON, I. L, AND S, Y. YANG. 1977. Genetic variation ill eds., Intermountain biogeography: a symposium. Thomomys bottae pocket gophers: macrogeo~ Great Basin Naturalist Memoirs g: 1-268. graphic patterns. EVolution 31: 697-720. SMITH, M. F., AND I. L. PAtrON. 1980. Relationships of PATTON, I. L., I. C. HAFNER, M. S. HAFNER, AND M. P. pocket gopher (Thomomys bottae) populations of SMITH. 1979. Hybrid zones in Thomomys boNae the lower Colorado River. Journal ofMammalogy pocket gophers: genetic, phenetic, and ecologic 61: 681-696. concordance patterns. Evolution 33: 860-876. ~. 1988. Subspecies ofpocket gophers: causal bases PATTON, I. L., M. F. SMITH,R. D. Pro;CE,AND R. A. HJ;:LLEN for geographic differentiation in Thomomys bot THAL. 1984. Genetics ofhybti([jzation between the tae. Systematic Zoology 37: 163-178. pocket gophers Thomomys bottae anCl Thomomys SMITH, M. P., J. L. PATTON, I. C. HAFNER, AND D. I. townsendiiinnortheastern California. GreatBl!Sin HAFNER. 1983. Thomomys bottae pocket gophers Naturalist 44: 431~440. of the cential Rio Grande Valley, New Mexico: PRAGER, E. M., AND A- C. WILSON. 1976. Congruency local differentiation, gene flow, and historical bio~ of phylogenies derived from different proteins. geography..Occasional Papers of the Museum of A molecular anelysjs ofthe phylogenetic position Southweste~Biology 2: 1~16. ofcracid birds, Journal ofMolecul:ir Evolution 9: SMITH, M. fl., R. K SELANDER, AND W. E. JOHNSON. 1973. 45-57. :I3iochemica). polymorphism and systematicsin the PENNEY, D. E., AND E. G. ZIMMERMAN. 1976. Genic c!iver~ genus Peromyscus. Ill. Variation in the Florida gence and local population differentiation by ran deer mouse (Peromyscus fioridanus), a Pleis dom drift in the pocket gopher genus Geomys. tocene relict. Journal ofMammalogy 54: 1-13. Evolution 30: 4'/3-483. SNYDER, J. O. 1918. The fishes ofthe Lahontan system of POWELL, J. R. 1975. Protein variation in natural popula~ Nevada and northeastern California. Btilletin of tions ofanimals. Pages '/9-119 in T. Dobzh~sky, the United States Bureau ofFisheries 35: 33-86. M. K. Hecht, and W. C. Steere, eds., Evolution SUDMAN, P. D.,I. R. CHOATE, AND E. G. ZIMMERMAN. 1987. ary biology. Vol. 8. Plenum Press, New York. Taxonomy of chromosomal races of Geomys bur" PRICE, R. D., AND R. A. HELLENTHAL. 1980. The Geomy~ sarius lutescens Merriam. Journal ofMammalogy doecus oregonus complex (Mallophaga: Tricho 68: 526=543. dectidae) of the western United States pocket SWOFFORD, D. 1. 1981. On the utility of the distance gophers (Rodentia: Geomyidae). Proceedings of Wagner procedure. Pages g5-43 in V. A. Funk the Entomological Society of Washington 8g: and D. R. Brooks, eds., Advances in cladistics. 25-38. Proceedings ofthe first meeting ofthe Willi Hen ROGERS, I. S. 19'/2. Measures of genetic similarity and nig Society, N.Y. Botanical Garden, Bronx, New genetic distance. Studies in Genetics VII. Univer" York. xii + 250 pp. sity ofTexl!S Publication '/213: 145-153. SWOFFORD, D. L., AND R. B. SELANDER. 1981. BIOSYS-l. ROGERS, M. A. 1991. Evolutionary differentiation within A computer program for the analysis of allelic the northern Great Basin pocket gopher, Tho" variation in genetics. Dep(l.ttment ofGenetics and momys townsendii. I. Morphological variation. Development, University ofIllihois, Urbana. Great Basin Naturalist 51: 109-126. TAYLOR, D. W. 1960. Distribution ofthe freshwater clam RUSSELL, I. C. 1885. Geological historyofLake Lahontan, Pisidium ultrqmontanum: a zoogeographic in a Quaternarylake ofnorthwestern Nevada. Mono~ quiry. American Joumal ofScience, Bradley Vol. graphs ofthe United States Geological Survey 11: 258~A: 325=334. 1-288. THAELER, C. S., JR. 1973. The karyotype of Thomomys SARICH, V. M. 19'/'/. Rates, sample sizes, and the neutral" townsendii similis and comments on the karyo ity hypothesis for electrophoresis in evolutionaty types of Thomomys. Southwestern Naturalist 1'/: studies. Nature 265: 24-0-28. 3g7-331. SARICH, V. M., AND I. E. CRONIN. 1976. Molecular sys~ ~. 1980. Chromosome numbers and systematic rela tematics of the Primates. Pages 141-170 in tions in thegenus Thomomys (Rodentia: Geomyi" M. Goodman, R. E. Tashian, and J. H. Tashian, dae). Journal ofMamtnalogy 61: 414-4g~. eds., Molecular anthropology. Plenum Press, TUCKER, P. K., AND D. J. SCHMIDLY. 1981. Studies of a New York. contact zone among three chromosomal races of SELANDER, R. K., D. W. KAUFMAN, R. I. :I3AKER, AND Geomys butsarius in east Texas. Journal ofMam S. L. WILLIAMS. 1974. Genic and chromosomal malogy 62: 258=g72. differentiation in pocket gophers of the Geomys WELLS, P. V.1983. Paleobiogeogrllphy ofmontane islands bursarius group. Evolution g8: 557=564. in the Great Basin since the last glaciopluvial. SELANDER, R. K., M. H. SMITH, S. Y. YANG, W. E. JOHNSON, Ecological Monographs 53: 341-382. AND I. B. GENTRY. 1971. Biochemical polYmor WENTWORTH, F. A., AND D. A. SUTTON. 1969. Chromo phism andsystematics in the genus Peromyscus. I. somes oftheTownsend pocketgopher, Thomomys Variation in the old~field mouse (Peromyscus townsendii..Southwestern Natunilist 14: 157-161. polionotus). Studies in Genetics VI. University WHEELER, H. E., AND E. F. COOK. 1954, Structural ofTexas Publication 7103: 49-=90. and stratigraphic significance of the Snake River 152 M. A. ROGERS [Volume 51 capture, Id:iho"Oregon. IoutIl~ of Geology 62: Thol1),omys townsendii nevadensis 525-536. CALIFORNIA: Lassen Co., [30], Bird Fillt Ranch, 3 mi. WHITLOW, W. B., AND E. R. HALL. 1933. Mllmmals ofthe S, 2.7 mi. W Herlong, 4080 ft. (n = g7,5); [31],4 mi. W Pocatello region of southern Idaho. University Stllndish, 4110 ft. (n = 18,2). California Publications in Zoology 40: g35-276. NEVADA: :Elko Co., [10], 0.5 mi. SW Ryndon, 5100 ft. WRIGHT, S. 1978. Evolution. and the genetics ofpopula (n = 20,4); Eureka Co., [11], Hay Ranch, 17 mi. SE tions. Vol. 4. Variability within and llmong natural falisade, 5160 ft. (n = 19,3); Ilumboldt Co., [5], Quinn popullltions. University of ChiCllgo fress, Chi River Crossing, 4100 ft. (n = 9,3); Lander Co., [15], 5.5 cago, illinois. 580 pp. mi. W Austin, 5700 ft. (n = ~2,5); Pershing Co., [3], Big Meadow Ranch, Lovelock, 4000 ft. (n = 16,3); [1], 2 mi. Received16January 1990 NWValmy, 4450 ft. (n = 21,5). OREGON: Ilarney Co., [4],4 mi. SW Narrows, 4200 ft. Revised15January 1991 (n= 6,0). . Accepted20 February 1991 Thomomys townsendii townsendii IDAHO: Bannock Co., [20], FloydIohnson Ranch, 4 mi. APPENDIX NWfocatello, 4500ft. (n = 19,3); Elmore Co., [2~], 3 mi. Specimens Examined E Grandview, 2800 ft. (n = 8,0); Owybee Co., [16], 7.5 mi. SE Grandview, 2600 ft. (n = 21,3); [17], 6 mi. SE Listed below ate specimens used in electrophoretic Murphy, 3000 ft. (n = 4,1); Payette Co., [g3], 1.5mi. NE and karyotypic analyses. Boldfllce numbers preceding Payette, g~oo ft. (n = 9,1); Power Co., [19], g.5 mi. NW localities represent corresponding composite locality American Falls, 4500 ft. (n = 10,g); Washington Co., [23], numbers. Sample sizes in parentheses represent the g.9 mi. SE Weiser, glOO ft. (n = 1,0). numberusedfor eiectrophoretic anaiysis and the number OREGON: Mlllbetir Co., [24], g.5 mi. N Ontario, glOO karyotyped, respectively. ft. (n = gO,l). .