Evolution. 43(5), 1989, pp. 1085-1096

I

HIGH LEVELS OF GENETIC VARIABILITY IN THEHAPLOID MOSS PLAGIOMNIUM CILIARE

ROBERT WYATT, IRENEUSZ J. ODRZYKOSKI,' AND ANN STONEBURNERDepartment of Botany, University of Georgia, Athens, GA 30602

Abstract. —Horizontal starch-gel electrophoresis was used to measure variability at 14 enzyme locifrom 13 natural populations of the dioecious moss Plagiomniwn ciliare. Overall levels of geneticpolymorphism were unexpectedly high for a haploid organism. Using a 1% frequency criterion,71% of the loci surveyed were polymorphic for the species as a whole. The number of alleles perpolymorphic locus for the species as a whole was 2.82 ± 0.34 (mean ± standard error), and meangene diversity per locus was 0.078 ± 0.035. While total gene diversity (//T = 0.178) was similarto that observed for highly outcrossed diploid plants such as pines, the variance within (/fs = 0.098± 0.027) and among (DST = 0.080 ± 0.033) populations was more evenly distributed than thatreported for populations of conifers. Genetic distances between populations ranged from 0.0002to 0.2064, with mosses from the Piedmont region of the southeastern United States showing lessdifferentiation among populations than did mosses from the Appalachian Mountains. Gene di-versity was much reduced in populations from disturbed, secondary forests in the Piedmont (0.058± 0.018) relative to those from minimally disturbed, primary forests in the mountains (0.146 ±0.048). Intensive sampling within populations revealed heterogeneity even within small ( 5 x 5cm) clumps. The discovery of high levels of genetic variability in a plant with a dominant haploidlife cycle challenges the traditional view of bryophytes as a genetically depauperate group. Multiple-niche selection is proposed as a possible explanation for this anomaly, but the data are also consistentwith the view that allozyme polymorphisms are selectively neutral.

Received February 1, 1988. Accepted March 20, 1989

Traditional views of the genetic structureof bryophyte populations hold that mossesand liverworts are genetically depauperateorganisms that underwent adaptive radia-tion long ago and today are limited to amodest role in natural communities. Theview that bryophytes evolve more slowlythan flowering plants and have remainedrelatively unchanged for millions of yearshas been expressed by many authors (Gem-mell, 1950; Steere, 1954; Anderson, 1963,1980; Schuster, 1966; Crum, 1972). Suchthinking is based on the facts that mostbryophytes are functionally haploid and thatthe genotype is, therefore, subjected directlyto natural selection. The widespread occur-rence of asexual reproduction and presum-ably high levels of self-fertilization are alsoexpected to contribute to low levels of ge-netic variability. Rates of evolution havebeen assumed to be slow, because fossilbryophytes usually are morphologicallysimilar to extant taxa. On the other hand,some researchers have argued that geneticvariation in bryophyte populations may be

1 Permanent address: Department of Genetics, In-stitute of Biology, Adam Mickiewicz University, 165Dabrowskiego, Poznan 60-594, Poland.

extensive (Khanna, 1964; Longton, 1976;Smith, 1978; Wyatt, 1982,1985). Their viewis based on recent discoveries of extremediversity in biochemical, physiological, andecological properties of mosses and liver-worts.

Few workers have attempted to assesslevels of genetic variability in natural pop-ulations of bryophytes using the techniquemost commonly used in similar studies ofother plants and animals: electrophoresis ofproteins. Existing evidence suggests thatlevels of electrophoretically detectable ge-netic variation in mosses are much higherthan predicted by the traditional view(Cummins and Wyatt, 1981; Daniels, 1982,1985; de Vries et al., 1983; Wyatt et al.,1988). Similarly, electrophoretic analyses ofliverwort populations have uncovered asurprising abundance of genetic variation(KrzakowaandSzweykowski, \977a, 1977b,1979; Szweykowski and Krzakowa, 1979;Szweykowski et al., 198la, 1981*; Odrzy-koski and Szweykowski, 1981; Odrzykoskietal., 1981;Yamazaki, 1981, 1984;Dewey,1989). Wyatt et al. (1989) have discussedthe implications of these findings in termsof bryophyte population structure and evo-lution.

1085

1086 R. WYATT ET AL.

100 km

FIG. 1. Locations of the 13 populations of Plagiomnium ciliare sampled in the southeastern United States.Abbreviations for populations (indicated by dots)'are given in Table 2. Abbreviations for physiographic provincesare: AP = Appalachian Plateaus, VR = Valley and Ridge, BR = Blue Ridge, P = Piedmont, and CP = CoastalPlain.

As part of a larger study of evolutionaryrelationships among haploid-polyploidspecies pairs in the Mniaceae, we assessedlevels of electrophoretically detectable ge-netic variation in natural populations of thedioecious moss Plagiomnium ciliare (C.Muell.) Kop. Gametophytes of this speciesare haploid (n = 6) and occur in coloniesconsisting of plagiotropic sterile shoots anderect fertile shoots 0.5-2 cm tall. An en-demic North American species, P. ciliaregrows abundantly in mesic woods in theeastern United States and adjacent Canada,with its center of distribution in the Ap-palachian Mountains (Koponen, 1971).

MATERIALS AND METHODSPopulation Samples. —We sampled a to-

tal of 13 populations from throughout therange of P. ciliare in the southeastern UnitedStates (Fig. 1). These populations were lo-cated in several physiographic provinces. At

each site, we collected 5-cm x 5-cm clumpsfrom within discrete colonies, placing thesesamples into small plastic pots. We collect-ed along stream banks until we had sampleda total of 36 discrete colonies or until wehad covered a distance of approximately 1km. Samples were returned to the lab, anda single shoot from each pot was selectedfor electrophoresis. Plants from the Botan-ical Garden (BG) population, which weremonomorphic at nearly all loci, were usedas "standards" for comparing enzyme mo-bilities.

To test for the possibility of microscalegenetic variation, we sampled the 36 clumpsfrom the Morning Star (MS) population in-tensively. From each 5-cm x 5-cm clump,we removed five erect shoots, one from thecenter and one from each corner of thesquare pots, and analyzed each shoot byhorizontal starch-gel electrophoresis. Wethen tabulated the percentage of clumps

within which two or mphoretic phenotypes oc

Electrophoretic Proadures for horizontal st.resis were similar to thozykoski and Gottlieb (snoots were homogeniztraction buffer (0.1 Mcontaining 10 mM K(6H2O, 1 mM EDTA (NX-fOD, and [added jus42 mM 2-mercaptoethathen filtered through aacloth onto 4-mm x 8-per wicks. All steps of hidone over crushed ice.

Saturated wicks weretical slot (the origin) cutgel, and enzymes werethree buffer systems. Bsolved malate dehydrogiphosphate isomerase (1glucomutase (PGM), cccitric acid, titrated to(3-aminopropyl) morpruwas prepared by dilutingbuffer with 964 ml watzymes also can be separ(43 mM trisodium citra:pH 7.0 with citric acid)this system consisted ofHC1 titrated to pH 7.0buffer gave similar pherbut yielded slightly btPGM. Buffer S was use(mate oxaloacetate trans;dolase (ALD), esterase (coisomerase (PGI), mali<peptidase (PEP). The ethis system was 190 m]\mM LiOH-H2O (pH 8was a mixture of 900 mmM citric acid, (pH 8.electrode buffer. After nof the gel buffer decreas

Gels were run in a re(4°C) for four hours in 1a constant amperage ofhours in buffer S at a co45 mA. By the end of thephenol blue marker hacin buffers M and H, andfront" had migrated 80 iter separation, enzymes

GENETIC VARIABILITY IN A HAPLOID MOSS 1087

IB southeastern United States,is for physiographic provincesPiedmont, and CP = Coastal

ed 5-cm x 5-cm clumpse colonies, placing theseplastic pots. We collect-iks until we had sampled:te colonies or until weince of approximately 1returned to the lab, andi each pot was selectedPlants from the Botan-

population, which wereearly all loci, were usedcomparing enzyme mo-

tossibility of microscalee sampled the 36 clumpsStar (MS) population in-';h 5-cm x 5-cm clump,pet shoots, one from the;om each corner of themalyzed each shoot byjel electrophoresis. Wej percentage of clumps

within which two or more distinct electro-phoretic phenotypes occurred.

Electrophoretic Procedures. —Our proce-dures for horizontal starch-gel electropho-resis were similar to those described by Odr-zykoski and Gottlieb (1984). Single mossshoots were homogenized in 50-100 jul ex-traction buffer (0.1 M Tris HC1, pH 7.5,containing 10 mM KC1, 10 mM MgCl2-6H2O, 1 mM EDTA (Na2 salt), 0.1% TritonX-100, and [added just before extraction]42 mM 2-mercaptoethanol). The extract wasthen filtered through a small strip of Mir-acloth onto 4-mm x 8-mm Beckmann pa-per wicks. All steps of homogenization weredone over crushed ice.

Saturated wicks were placed into a ver-tical slot (the origin) cut across a 10% starchgel, and enzymes were separated in one ofthree buffer systems. Buffer M, which re-solved malate dehydrogenase (MDH), triosephosphate isomerase (TPI), and phospho-glucomutase (PGM), consisted of 40 mMcitric acid, titrated to pH 6.1 with N-3(3-aminopropyl) morpholine. The gel bufferwas prepared by diluting 36 ml of electrodebuffer with 964 ml water. These three en-zymes also can be separated using buffer H(43 mM trisodium citrate • 2H2O, titrated topH 7.0 with citric acid). The gel buffer forthis system consisted of 5 mM DL-histidineHC1 titrated to pH 7.0 with NaOH. Thisbuffer gave similar phenotypes to buffer Mbut yielded slightly better resolution ofPGM. Buffer S was used to separate gluta-mate oxaloacetate transaminase (GOT), al-dolase (ALD), esterase (EST), phosphoglu-coisomerase (PGI), malic enzyme (ME), andpeptidase (PEP). The electrode buffer forthis system was 190 mM boric acid and 60mM LiOH-H2O (pH 8.3). The gel bufferwas a mixture of 900 ml of 50 mM Tris, 6mM citric acid, (pH 8.3), and 100 ml ofelectrode buffer. After mixing, the final pHof the gel buffer decreased to 8.2.

Gels were run in a refrigerated chamber(4°C) for four hours in buffers M and H ata constant amperage of 35 mA and for fivehours in buffer S at a constant amperage of45 mA. By the end of these runs, the bromo-phenol blue marker had migrated 90 mmin buffers M and H, and the brown "boratefront" had migrated 80 mm in buffer S. Af-ter separation, enzymes were visualized us-

ing standard colorimetric methods of stain-ing (Shaw and Prasad, 1970; Harris andHopkinson, 1976) with only slight modifi-cations. Except for EST and GOT, whichwere stained in liquid assay, all enzymeswere stained for 1-3 hours, using the agar-overlay method. Staining was done in anincubator at 37°-40°C.

Coded data for the 14 loci in the 13 pop-ulations were analyzed using BIOSYS-1(Swofford and Selander, 1981) and a pro-gram developed in the laboratory of ]. L.Hamrick (Department of Botany, Univer-sity of Georgia).

RESULTSElectrophoretic Patterns. —Two enzymes,

ALD and ME, showed only one region ofactivity on the gels (see Appendix). ALDwas monomorphic in all populations, whileME existed as two mobility variants. GOTand PGI showed one intensely staining re-gion and another, less active isozyme, whichwe chose not to score. From 2-5 regions ofEST activity, we scored the most intenselystaining zone, which appeared to accountfor more than half of the total activity. Thisenzyme always appeared as a two-bandedphenotype. Mobility differences invariablyinvolved both bands changing in concert. Afew plants showed no EST activity and werescored as carrying null alleles. We foundthree isozymes of PEP that could use DL-leucyl-phenylalanine as a substrate. WhilePEP-1 and PEP-3 used L-valyl-L-leucine asa substrate, only PEP-2 was able to useL-leucyl-glycyl-glycine. This substrate speci-ficity convinced us that these enzymesshould be treated as products of separategenes.

PGM activity was detected in three re-gions, the fastest of which was monomor-phic in all plants that expressed it. Manyplants, however, lacked this isozyme orshowed reduced PGM activity. Therefore,we did not include it in our analysis. Theother PGM isozymes were consistentlyscoreable, each with three or four mobilityvariants. We found seven two-isozymecombinations of these variants, and there-fore, we assumed that the enzymes are prod-ucts of two separate genes. In three regionsof the gel, we found bands of MDH activity.The fastest single band, which was usually

1088 R. WYATT ET AL.

monomorphic, was unstable and, therefore,was omitted from our analysis. Since in allcases changes in mobility affected all threebands of MDH-1 activity, this three-band-ed phenotype possibly represents posttrans-lational modification of a single gene. Phe-notypes of MDH-2 also were three-banded.We treated two variants detected in this re-gion as two alleles of a single gene. TPI ac-tivity consisted of five isozymes in two sep-arate regions, but only three of these hadhigh activity. The first region contained threetwo-banded phenotypes, which we inter-preted conservatively as allelic variants ofa single gene (TPI-1). From the second re-gion, only the slowest band had high activityand was found to exist as three allelic vari-ants (TPI-2). Photographs and further dis-cussion of these loci in closely related speciesof Plagiomnium section Rosulata are pro-vided by Wyatt et al. (1988, 1989).

Levels of Genetic Variation.— Of the 14enzymes screened by electrophoresis, onlythree (GOT-1, ALD-1, and PGI-1) weremonomorphic in all populations (Table 1).PEP-3 was polymorphic in only one pop-ulation, while MDH-2 was polymorphic inonly two. Using a 1% frequency criterion,71% of the loci surveyed were polymorphicfor the species as a whole. Even using themore stringent 5% frequency criterion,polymorphism in P. ciliare was 36%. Onaverage, 31.1% of the loci were polymorphicper population, with a range from 0% forthe Broad River population to 64% for theCoweeta and Morning Star populations. Themean number of alleles per locus rangedfrom 1.00 to 1.79, with a mean of 1.35.Considering only polymorphic loci, thenumber of alleles per locus for the speciesas a whole was 2.82 ± 0.34 (mean ± stan-dard error). Mean intrapopulational genediversity (//s; Nei, 1973) for all loci rangedfrom 0.000 to 0.138 with a weighted meanof 0.078 (Table 2).

Total gene diversity (HT; Nei, 1973, 1975)based on mean allelic frequencies of poly-morphic loci over all populations was 0.178(Table 3). For all polymorphic loci exceptMDH-1 and ME-1, the largest proportionof this variance is due to diversity withinpopulations (//s), rather than between pop-ulations CDST). Differences between the twomeasures were generally small, however,

with Hs averaging 0.098 ± 0.027 and DST

averaging 0.080 ± 0.033. Nei's( 1973, 1975)Dm is an absolute measure of gene differ-entiation which estimates the minimum netcodon differences between populations in-dependent of gene diversities within sub-populations. For our moss populations, Dm

ranged from 0.000 to 0.299, with a meanof 0.086 ± 0.036 (Table 3). <7ST, which mea-sures diversity between populations relativeto total diversity (Nei, 1973, 1975), aver-aged 0.248 ± 0.070, while JRsT, the ratio ofbetween- to within-population diversity,averaged 0.565 ± 0.229 (Table 3). This in-dicates that there is approximately half asmuch variation between populations as thereis within populations.

As indicated by gene-diversity statistics(Table 4), individual populations of P. cil-iare differ strongly in levels of genetic poly-morphism. Populations from the Piedmontare clearly less polymorphic than those fromother physiographic provinces in terms ofthe proportion of polymorphic loci (16.5%vs. 44.9%), the mean number of alleles perlocus (1.17 vs. 1.52), and gene diversity(0.058 vs. 0.146). Three loci (Mdh-1, Pgm-2, and Me-1) were far more variable in pop-ulations from the mountains than from thePiedmont. On the other hand, Pgm-1showed high levels of gene diversity in bothareas, although more of the variation inPiedmont populations was due to differ-ences between populations (GST = 0.548 forthe Piedmont, GST = 0.272 for the moun-tains). Overall gene differentiation, as mea-sured by Nei's (1975) GST, is similar in Pied-mont and mountain populations (0.161 vs.0.171). On average, populations from thePiedmont are also more similar genetically:.Dm = 0.039 for six pairs of Piedmont pop-ulations; Dm = 0.082 for seven pairs ofmountain populations.

Genetic distances between pairs of pop-ulations ranged from 0.0002 (between Bo-tanical Garden and Broad River) to 0.2064(between Watson's Mill and Pond Drain)(Table 5). Generally, there was less differ-entiation among populations from the Pied-mont than among populations from themountains. A phenogram summarizing ge-netic similarities among the populationsgrouped all of the Piedmont samples plusthe sample from Alabama together before

TABLE 1. Allele frequenciesoutheastern United States.glutamate oxaloacetate trans= malate dehydrogenase, PGPEP = peptidase. Plants thai

LocusGot-1Ald-1Pgi-1Est-1

Mdh-1

Mdh-2

Pgm-1

Pgm-2

Tpi-1

Tpi-2

Me-1

Pep-1

Pep-2

Pep-3

Allele

aaaabcnullabcababcabcdabcabcababcabab

BG

1.001.001.001.00

—_

—1.00—

—1.00

—0.930.07

—1.00

—_

—1.00

——1.00

——1.00—

1.00—

—1.00—

1.00—

GMC

l.OCl.OCl.OC0.870.13_

—1.00

——

1.00—

0.310.69_

0.940.06_

—1.00

——

1.00——

1.00—

0.840.16

—0.810.191.00

—

these joined any of the(Fig. 2). Populations fr<were much less similar t(were, however, no obvictween geographical distaulations and their geneti

Patterns of geographicdividual loci generally nfrom gene-diversity stadistances. Me-lb is absepopulations of P. ciliare 1as the most common alkpopulations except Tan3A). Piedmont populatii

GENETIC VARIABILITY IN A HAPLOID MOSS 1089

.098 ± 0.027 and Dsr

033. Nei's (1973, 1975)leasure of gene differ-aates the minimum net:tween populations in-iiversities within sub-• moss populations, Dm

to 0.299, with a meanble 3). GST, which mea-•en populations relativefei, 1973, 1975), aver-while RST, the ratio of

i-population diversity,.229 (Table 3). This in-approximately half as

een populations as thereis.gene-diversity statisticsil populations of P. cil-n levels of genetic poly-ions from the Piedmont-norphic than those from; provinces in terms ofolymorphic loci (16.5%in number of alleles per52), and gene diversityhree loci (Mdh-1, Pgm-ar more variable in pop-lountains than from thee other hand, Pgm-1of gene diversity in bothore of the variation inons was due to differ-ilations (GST = 0.548 for= 0.272 for the moun-differentiation, as mea-

5) OrST, is similar in Pied-n populations (0.161 vs.;, populations from thenore similar genetically:pairs of Piedmont pop-

OS 2 for seven pairs ofons.:s between pairs of pop-m 0.0002 (between Bo-Broad River) to 0.2064

; Mill and Pond Drain)ly, there was less differ-jpulations from the Pied-

ig populations from thelogram summarizing ge-among the populationsPiedmont samples plus

Alabama together before

TABLE 1. Allele frequencies for 14 enzyme loci sampled in 13 populations of Plagiomnium ciliare in thesoutheastern United States. Abbreviations for populations are given in Table 2. Codes for enzymes: GOT =glutamate oxaloacetate transaminase, ALD = aldolase, PGI = phosphoglucoisomerase, EST = esterase, MDH= malate dehydrogenase, PGM = phosphoglucomutase, TPI = triose phosphate isomerase, ME = malic enzyme,PEP = peptidase. Plants that showed no EST activity were scored as carrying null alleles.

Locus

Got-lAld-1Pgi-1Est-1

Mdh-1

Mdh-2

Pgm-1

Pgm-2

Tpi-1

Tpi-2

Me-]

Pep-1

Pep-2

Pep-3

Allele

aaaahu

nullabcababCabcft14abcabcababcabab

BG

1.001.001.001.00

—1.00

——1.00

—0.930.07

1.00——

1.00

——

1.00

——

1.00—

1.00——1.00

—1.00

—

OMC1.001.001.000.87n i ^\j, i j

—

1.00——

1.00—

0.310.69

0.940.06

—

1.00

——1.00——

1.00—

0.840.16

—0.810.191.00

—

WM1.001.001.001.00

—1.00

——

1.00—

0.070.93

0.900.10

1.00——

1.00——1.00

—1.00

——1.00

—1.00

—

EM

1.001.001.001.00

—0.920.08

—0.810.19

—1.00

1.00——

1.00

——

1.00——1.00

—1.00_

—1.00

—1.00—

BR

1.001.001.001.00

—1.00

——1.00

—1.00

—

1.00

——

1.00——1.00_

—1.00

—1.00

——1.00

—1.00

—

IER

1.001.001.001.00

—0.940.06

—1.00

—0.390.61

1.00——

0.830.17

—0.920.08

—1.00

—1.00_

—1.00

—1.00

—

'opulaliorCOW

1.001.001.000.94

0.060.060.860.081.00

—0.750.25

0.190.73

—n nx0.780.22

—0.970.03

—0.940.060.97

—0.031.00

—0.970.03

IPC

1.001.001.001.00

—0.450.55

—1.00

—0.320.68

0.910.09

—

1.00

——1.00

——0.45

0.551.00_

—1.00

—1.00

—

OCS1.001.001.001.00

—0.360.64

—1.00

—1.00_

0.440.56

—

0.860.110.031.00

——

0.190.811.00

——1.00

—1.00

—

TC

1.001.001.001.00

—0.400.60

—1.00

—0.800.20

0.570.43_

0.700.30

—1.00——

1.00—

0.93—0.07

0.970.031.00

—

PD1.001.001.001.00

—0.030.97

—1.00—

0.810.19

0.610.170.22

1.00——

0.94—0.06

0.110.891.00

——1.00

—1.00—

MS

1.001.001.000.97

0.030.970.03

—0.940.060.420.55n rn\j.\jj0.720.250.03

0.94

—0.060.970.03

—0.250.751.00

——0.970.031.00

—

HLD

1.001.001.000.92

n nsu.uo

1.00——

1.00—

0.290.71

0.920.08

—

1.00——

1.00

——0.620.380.960.04

—0.830.171.00

—

these joined any of the mountain samples(Fig. 2). Populations from the mountainswere much less similar to each other. Therewere, however, no obvious correlations be-tween geographical distances between pop-ulations and their genetic distances.

Patterns of geographical variation for in-dividual loci generally reinforce the picturefrom gene-diversity statistics and geneticdistances. Me-lb is absent from Piedmontpopulations of P. ciliare but is present, oftenas the most common allele, in all mountainpopulations except Tamassee Creek (Fig.3A). Piedmont populations also are nearly

monomorphic for Mdh-1", while mostmountain populations have Mdh-lb as themost common allele (Fig. 3B). Mountainpopulations also are more variable amongthemselves, and Mdh-lc is restricted to theCoweeta population in the Blue RidgeMountains. Nearly all Piedmont popula-tions are fixed for Pgm-2" (Fig. 3C). Pgm-2b occurs in higher frequency in the moun-tains, and two rare alleles (Pgm-2c andPgm-2tf) are restricted to mountain popu-lations. Finally, there is no clear pattern tovariation at Pgm-1 (Fig. 3D). Even closelyadjacent populations, such as Broad River

1090 R. WYATT ET AL.

TABLE 2. Sample sizes (N), gene diversities (Hs), and standard errors (SE) for 14 enzyme loci surveyed in 13populations of Plagiomnium ciliare from the southeastern United States. The first six populations are from thePiedmont.

PopulationBGGMCWMEMBRERCOWPCOCSTCPDMSHLD

Locality

Botanical Garden, Athens, GAGoldmine Creek, Braselton, GAWatson's Mill State Park, GAEchol's Mill, Lexington, GABroad River, Elberton, GAEno River, Durham, NCCoweeta Hydrologic Lab, NCPanther Creek, Toccoa, GAOconee Station Cove, SCTamassee Creek, Walhalla, SCPond Drain, Mt. Lake, VAMorning Star, Basye, VAHolt Lock and Dam, Holt, AL

N

72322926153636223630363624

US ± SE

0.009 ± 0.0090.097 ± 0.0400.023 ± 0.0160.033 ± 0.0240.000 ± 0.0000.073 ± 0.0390.127 ± 0.0410.116 ± 0.0550.110 ± 0.0510.138 ± 0.0540.088 ± 0.0440.124 ± 0.0470.113 ± 0.045

Grand weighted mean: 430 0.078 ± 0.035

and Watson's Mill from the Piedmont ofGeorgia, frequently have very different al-lele frequencies. Again, the rare allele Pgm-lc is found only in one mountain popula-tion. Intensive sampling within the 36clumps of P. ciliare from the Morning Starpopulation detected five clumps that weregenetically heterogeneous (i.e., consisting oftwo or more plants that differed in multil-ocus electrophoretic phenotypes). Three ofthese five clumps showed variability formore than two enzyme loci.

DISCUSSION

Gottlieb (1982) proposed a basic modelfor plant isozymes that suggests that num-

bers and subcellular locations of isozymesare highly conserved in diploid floweringplants. Most plants have two isozymes foreach specific enzyme, one of which is activein organelles and one of which catalyzes thesame reaction in the cytosol. Most of theenzymes we surveyed in the moss P. ciliareare consistent with this model. Some non-specific enzymes, such as esterases and pep-tidases, showed additional isozymes, as didMDH, which often exists as 3-4 differentisozymes in flowering plants (Gottlieb,1982). MDH-2 displayed an unusual phe-notype consisting of three bands of equalactivity. We treated the two variants thatwe detected as corresponding to two alleles

TABLE 3. Total gene diversity (Hi) and gene diversities within (Hs) and between (Z>ST) populations of Pla-giomnium ciliare for the polymorphic loci. Also represented are indexes of gene differentiation between popu-lations (Dm), between-population diversity relative to within-population diversity (.RST). and between-populationdiversity relative to total diversity (Gsr)- Enzyme codes are given in Table 1.

Locus

Est-1Mdh-1Mdh-2Pgm-1Pgm-2Tpi-1Tpi-2Me- 1Pep- 1Pep-2Pep-3

Mean:SE:

H-f

0.04130.42030.03200.45730.35530.13070.03210.38300.04120.05430.0046

0.17750.0554

US0.03810.14470.02760.27060.23090.11080.03060.13450.03730.04720.0045

0.09790.0268

DST0.00320.27560.00450.18680.12440.01980.00150.24850.00390.00710.0001

0.07960.0329

Dm

0.00340.29850.00480.20230.13480.02150.00160.26920.00420.00760.00010.08620.0356

GST0.07640.65570.13920.40840.35020.15190.04720.64890.09460.13010.0255

0.24800.0701

*ST0.08962.06290.17520.74780.58390.19400.05362:00210.11320.16210.0283

0.56480.2291

at a single locus. It is pothese three-banded prepresent fixed heterozenzyme. Additional wetermine which of thecorrect.

We found three isozciliare and five isozymefrom the usual two ((creases in isozyme nun-gene duplication or frgenomes via polyploid}viewed the evidencezymes in diploid plantduplication have beenin any bryophyte. Like^gested that any moss sjlikely to be polyploid.gested that repeated cyclgene silencing have occrous ferns, yielding isoical of diploid plants (1Soltis [1986] and Soltifor alternative explanattherefore, that isozymeing from an ancient polancestors of P. ciliare rall but a few loci. Becatdetected at least threebeen duplicated, it seeithe genome of P. ciliareplicated by polyploidy alenced at the majority cently gathering addincluding studies of istother species ofPfagiomallow us to reject eithesingle-gene duplicationtion hypothesis.

The most importantdrawn from our analystion in P. ciliare are thzmaintains an unexpecte

TABLE 4. Estimates of genemountains. TVpop = number •putative loci; PLP = averageGST, and £>m are denned in 1

Population A'enzPiedmontMountain

Total:

67

13

GENETIC VARIABILITY IN A HAPLOID MOSS 1091

or 14 enzyme loci surveyed in 13; first six populations are from the

HS ± SE

))j4

0

0.009 ± 0.0090.097 ± 0.0400.023 ± 0.0160.033 ± 0.0240.000 ± 0.0000.073 ± 0.0390.127 ± 0.0410.116 ± 0.0550.110 ± 0.0510.138 ± 0.0540.088 ± 0.0440.124 ± 0.0470.113 ± 0.0450.078 ± 0.035

illular locations of isozymesiserved in diploid floweringlants have two isozymes forizyme, one of which is activend one of which catalyzes thein the cytosol. Most of the

irveyed in the moss P. ciliarewith this model. Some non-

ics, such as esterases and pep-•d additional isozymes, as didoften exists as 3-4 differentflowering plants (Gottlieb,2 displayed an unusual phe-ting of three bands of equal

treated the two variants thats corresponding to two alleles

I between (£>ST) populations of Pla-f gene differentiation between popu-ersity C/?ST), and between-populatione 1.

GST KST

\

\>

: jj

; t•1 2' '•>: \23

0.07640.65570.13920.40840.35020.15190.04720.64890.09460.13010.0255

0.24800.0701

0.08962.06290.17520.74780.58390.19400.05362.00210.11320.16210.0283

0.56480.2291

at a single locus. It is possible, however, thatthese three-banded phenotypes actuallyrepresent fixed heterozygosity of a dimericenzyme. Additional work is necessary to de-termine which of the two explanations iscorrect.

We found three isozymes of PGM in P.ciliare and five isozymes of TPI, an increasefrom the usual two (Gottlieb, 1982). In-creases in isozyme numbers can result fromgene duplication or from the addition ofgenomes via polyploidy. Gottlieb (1982) re-viewed the evidence for duplicated iso-zymes in diploid plants. No cases of geneduplication have been reported previouslyin any bryophyte. Likewise, no one has sug-gested that any moss species with n = 6 islikely to be polyploid. Haufler (1987) sug-gested that repeated cycles of polyploidy andgene silencing have occurred in homospo-rous ferns, yielding isozyme numbers typ-ical of diploid plants (but see Haufler andSoltis [1986] and Soltis and Soltis [1989]for alternative explanations). It is possible,therefore, that isozyme multiplicity result-ing from an ancient polyploidization in theancestors of P. ciliare has been silenced atall but a few loci. Because we have alreadydetected at least three loci that may havebeen duplicated, it seems quite likely thatthe genome of P. ciliare may have been du-plicated by polyploidy and subsequently si-lenced at the majority of loci. We are pres-ently gathering additional evidence,including studies of isozyme numbers inother species of Plagiomnium, which shouldallow us to reject either the hypothesis ofsingle-gene duplication or the polyploidiza-tion hypothesis.

The most important conclusions to bedrawn from our analyses of genetic varia-tion in P. ciliare are that this haploid mossmaintains an unexpectedly high amount of

1 EM

1 1 HLD| 1 WM

0.80 0.83 0.87 0.90 0.93 0.97 1.00Genetic Similarity

FIG. 2. Phenogram expressing overall levels of ge-netic similarity among 13 populations of Plagiomniumciliare based on Rogers's (1972) coefficient of geneticsimilarity using 14 putative gene loci. Abbreviationsfor populations are given in Table 2.

variation and that populations from thesoutheastern United States display strongpopulation differentiation. Also of majorimportance is the discovery that geneticvariability is severely reduced in the dis-turbed Piedmont region versus the relative-ly undisturbed mountain regions.

Our results agree with those of most pre-vious electrophoretic studies of bryophytepopulations: more genetic variation existsthan is predicted by the traditional view ofbryophyte variation and evolution. Highlevels of polymorphism and mean numbersof alleles per locus were detected by de Vrieset al. (1983) in two species of Racopilum.In fact, average gene diversities within pop-ulations of/?, spectabile and R. cuspidiger-um were closely comparable to those forwind-pollinated, highly outcrossed pines(Guries and Ledig, 1982; Loveless andHamrick, 1984). With the exception of Ya-mazaki's (1981, 1984) studies, genetic vari-ation reported in liverworts appears to beless than that in mosses, as predicted byKhanna(1964).

TABLE 4. Estimates of genetic variation in populations of Plagiomnium ciliare from the Piedmont and themountains. A^p = number of population samples; Nen2 = number of enzymes screened; A^i — number ofputative loci; PLP = average percentage of loci polymorphic; k = mean number of alleles per locus. HI, HS,GST, and An are denned in Table 3.

Population

PiedmontMountain

Total:

Nfov

61

13

Ncal

99

9

Moci141414

PLP16.544.931.1

k

1.171.52

1.35

H-r

0.0910.216

0.178

US

0.0580.146

0.098

GST0.1610.171

0.248

Dm

0.0390.0820.086

FIG. 3. Geographic patterns of variation in allele frequencies at four loci of Plagiomnium ciliare: A) Me-1;B) Mdh-1; Q Pgm-2; D) Pgm-1. See Figure 1 for more information regarding the sample locations.

TABLE 5. Nei's (1972) genetic distances (below the diagonal) and genetic identities (above the diagonal) for 13populations of Plagiomnium ciliare in the southeastern United States. Abbreviations for populations are givenin Table 2. Geographical locations of the populations are shown in Figure 1.

Population

BGGMCWMEMBRERCOWPCOCSTCPDMSHLD

BG_

0.04190.06330.00240.00020.03100.13360.09240.12620.05500.15020.08080.0530

GMC

0.9590

0.00830.04910.04690.00640.15660.05680.17480.08330.18470.05850.0121

WM

0.93870.9918

—0.07130.06890.01080.17320.05770.19650.10200.20640.06430.0156

EM

0.99760.95200.9312

—0.00220.03740.13080.09490.12250.05320.14680.08650.0604

BR

0.99980.95420.93340.9978_

0.03550.13530.09760.12600.05620.15140.08530.0581

ER

0.96950.99360.98920.96330.9651

—0.14190.05220.15990.06690.16960.05650.0152

cow

0.87490.85500.84090.87740.87340.8677

—0.11440.07160.02650.10510.17100.1640

PC

0.91180.94480.94390.90950.90700.9491

.0.8919—

0.07460.07170.06030.03350.0320

Plagiomnium ciliaredominant life cycle, fa:values of dicots for themorphic loci per poinumber of alleles perversity (Hamrick et alP. ciliare with diploidbasis of polymorphic Icmoss is above averagetion diversity and be^population diversityHamrick, 1984). Over;versity in P. ciliare areto those measured in pida) by Guries and Lepine, however, there istiation among popula0.023). Populations offan order of magnitudetistics, reflecting strcamong localities. This ;by close examination <terns in allele frequencewhich differ sharply e^adjacent populations.

Pairwise genetic dist;populations of P. cilianUnited States are gener,observed for conspecificloid plants (e.g., mean =species tabulated by Amont populations of Fsimilar to each other t"lations from the AppaThis may be explainedPiedmont samples caifrom Georgia, while

TABLE 5. E

PopulalOCS

0.88140.83960.82160.88470.88160.85230.93090.9282

—0.06410.02880.07820.1349

TC

0.94650.92010.90310.94820.94540.93530.97380.93080.9379

0.09280.11080.0935

PD

0.86C0.8310.8i:0.86;0.85S0.84̂0.90C0.9410.9710.91)

0.09;0.14:

B)MDH-

GENETIC VARIABILITY IN A HAPLOID MOSS 1093

iloci of Plagiomnium ciliare: A) Me-1;irding the sample locations.

c identities (above the diagonal) for 13.bbreviations for populations are givenre 1.

ER

0.96950.99360.98920.96330.9651__

0.14190.0522

' 0.15990.06690.16960.05650.0152

cow

0.87490.85500.84090.87740.87340.8677

—0.11440.07160.02650.10510.17100.1640

PC

0.91180.94480.94390.90950.90700.94910.8919

—0.07460.07170.06030.03350.0320

Plagiomnium ciliare, despite its haploid-dominant life cycle, falls close to the meanvalues of dicots for the percentage of poly-morphic loci per population, the meannumber of alleles per locus, and gene di-versity (Hamrick et al., 1979). ComparingP. ciliare with diploid seed plants on thebasis of polymorphic loci only, this haploidmoss is above average for among-popula-tion diversity and below average for within-population diversity (see Loveless andHamrick, 1984). Overall levels of gene di-versity in P. ciliare are closely comparableto those measured in pitch pine (Pinus rig-idd) by Guries and Ledig (1982). In pitchpine, however, there is almost no differen-tiation among populations (mean GST =0.023). Populations of P. ciliare show valuesan order of magnitude larger for these sta-tistics, reflecting stronger dissimilarityamong localities. This pattern is reinforcedby close examination of geographical pat-terns in allele frequencies at particular loci,which differ sharply even between closelyadjacent populations.

Pairwise genetic distances among the 13populations of P. ciliare in the southeasternUnited States are generally within the rangeobserved for conspecific populations of dip-loid plants (e.g., mean = 0.0954 for Clarkiaspecies tabulated by Ayala [1975]). Pied-mont populations of P. ciliare were moresimilar to each other than were the popu-lations from the Appalachian Mountains.This may be explained by the fact that ourPiedmont samples came almost entirelyfrom Georgia, while the Appalachian

TABLE 5. Extended.

PopulationOCS

0.88140.83960.82160.88470.88160.85230.93090.9282_

0.06410.02880.07820.1349

TC

0.94650.92010.90310.94820.94540.93530.97380.93080.9379

0.09280.11080.0935

PD

0.86060.83140.81350.86350.85950.84400.90030.94150.97160.9114

—0.09320.1429

MS

0.92230.94320.93770.91720.91820.94510.84280.96700.92470.89510.9110

—0.0236

HLD

0.94840.98800.98450.94140.94350.98490.84870.96850.87380.91070.86680.9766—

Mountain populations were sampled fromfive states and were spread over a muchwider geographical area.

Plagiomnium ciliare is a dioecious mossthat reproduces regularly by sexual means.Sporophytes mature in late summer and re-lease approximately 1-2 x 105 wind-dis-persed spores. No specialized asexual prop-agules are produced, but like most mosses,P. ciliare is capable of regeneration from leafor stem fragments. This moss is a commonconstituent of the bryoflora of mesic decid-uous forests in eastern North America, witha continuous range across various physio-graphic provinces. Populations generallyconsist of millions of individual gameto-phores.

The discovery of considerable differen-tiation among populations of P. ciliare sug-gests that gene flow may be more restrictedthan one might expect from the large num-bers of wind-dispersed propagules pro-duced. Alternatively, it is possible that se-lection pressures for the loci we scored differstrongly among populations. Support for thisview comes from the observation of largedifferences in statistics of gene diversity fordifferent loci. It is also very likely, of course,that both selection and genetic drift in iso-lated populations act in concert to producethe observed pattern.

Intensive sampling of clumps of P. ciliarerevealed microscale genetic differentiation.Similarly, Cummins and Wyatt (1981) foundgenetic variation within small patches of themoss Atrichum angustatum. Given the lim-ited range of gene flow in this species andmost other bryophytes, such differentiationis to be expected (Wyatt, 1977, 1982, 1985;Wyatt and Anderson, 1984). Certainly, ifsuch differentiation exists within popula-tions, it is also to be expected among pop-ulations.

One of the most clear-cut differencesamong populations of P. ciliare is the sig-nificantly reduced genetic variation in thePiedmont. In the Appalachian Mountains,P. ciliare occurs in primary forests consist-ing of a highly diverse mixture of hardwoodtrees. Most of our sampling sites were inforests that had been minimally disturbed.On the other hand, populations in the Pied-mont occur mainly along streams in second-growth oak-hickory-pine forests. Most of

1094 R. WYATT ET AL.

these areas were cleared in the 1800's forcultivation of crops and have had only about100 years in which to recover. Therefore,although the present abundance of P. ciliarein Piedmont forests appears to be similarto that in the Appalachian Mountains, thegenetic diversity of Piedmont populationsis strikingly reduced. This impoverishmentof genetic stocks may have occurred becauseof the bottlenecks in population size to whichPiedmont populations were subjected.

To test this prediction, we sampled Pied-mont populations of P. ciliare from two sitesthat historical records suggested had neverbeen cleared or heavily logged: Gold MineCreek in the University of Georgia Arbore-tum and Eno River State Park in North Car-olina. These sites showed gene diversitiesmuch higher than other Piedmont sites. Infact, their values were more similar to thosefor sites from undisturbed forests in the Ap-palachian Mountains. It appears likely,therefore, that the reduction in genetic di-versity in the Piedmont of Georgia is dueto recent habitat destruction, which reducedpopulation sizes and forced colonies to re-establish from a limited number of surviv-ing sources. Piedmont populations are com-pletely or nearly fixed at all loci and totallylack rare or unique alleles, a pattern to beexpected in extreme cases of genetic drift.

The implications of our discovery of largeamounts of genetic variability in the haploidmoss P. ciliare are wide-ranging. Assumingthat the plants are truly haploid and thatthe genotype is therefore subjected directlyto natural selection, most models of geneticpopulation structure would predict reducedlevels of genetic variation (Ennos, 1983).Effects due to dominance or overdominancecannot be invoked to explain the mainte-nance of this variability in haploid organ-isms. Furthermore, models of temporallyvarying selection suggest that genetic vari-ation cannot be maintained in haploid pop-ulations by temporal heterogeneity (Ennos,1983). Rather, it appears most likely thatsome form of spatial heterogeneity, such asmultiple-niche selection, must be involvedif selection is indeed responsible for the ge-netic variation observed in P. ciliare. Ge-netic heterogeneity is most likely to bemaintained when there is restricted gene flowand when there are large differences in se-

lection coefficients between niches (Ennos,1983), a situation likely to be common inbryophyte populations (Wyatt, 1982; Wyattand Anderson, 1984).

On the other hand, Yamazaki (1981,1984) interpreted his discovery of extensivegenetic variation in Japanese populations ofthe liverwort Conocephalum conicum as aclear demonstration that allozyme poly-morphisms are selectively neutral. He ar-gued that "only under the model of selectiveneutrality of genetic variability do we expectthe equality of polymorphisms betweenhaploid and diploid organisms" (Yamazaki,1981 p. 374). Yamazaki's( 1981, 1984) data,however, are open to question. He chose touse a large number of nonspecific enzymesknown to be unusually variable in otherspecies, including five esterases and threeoxidases. This may have artifically elevatedhis estimates of genetic variability. Fur-thermore, his results conflict strongly withother studies of C. conicum (e.g., Szwey-kowski and Krzakowa, 1979; Szweykowskiet al., 198 la; Odrzykoski, unpubl.), in whichlittle polymorphism was found within dif-ferent races of this liverwort species. Allstudies do agree, however, that there is littledifferentiation among populations, a findingat odds with our results for P. ciliare.

We have, therefore, two contrasting pic-tures of genetic population structure withinbryophyte species: 1) the "Conocephalummodel," in which there are low levels ofvariation within races (which probably rep-resent separate biological species), weak in-terpopulation differentiation, and no mi-croscale heterogeneity; and 2) the"Plagiomnium model," in which there arehigh levels of genetic variation, strong in-terpopulation differentiation, and micro-scale heterogeneity. Wyatt (1985) has dis-cussed the ecological and evolutionaryimplications of these differing populationstructures. In any event, it is clear that atleast some bryophytes, despite their statusas "phylogenetic relicts," maintain signifi-cant stores of genetic variability. Perhapsfurther study will reveal that these organ-isms are as diverse in terms of genetic pop-ulation structure as are angiosperms or var-ious groups of animals. Certainly, theirgenetic systems appear to be similar in kindto those of diploid plants and do not ob-

viously constrain thtion and evolution.

ACKNOW

This research w;Grant BSR-840893the Department of Bof Georgia and the Jeemy of Natural Schelped to make I. Jthe U.S. possible. Wfield assistance and.ments on the manu;

LlTERAlANDERSON, L. E. 1963.

Mosses. Bryologist 66. 1980. Cytology

mosses, pp. 37-76. Initon (eds.), The MossiAssoc. Adv. Sci., San

AYALA, F.J. 1975. Gen<speciation process. Ev

CRUM, H. 1972. The geeof North America's iHattori Bot. Lab. 35:1

CUMMINS, H., AND R. Wability in natural popuangustatum. Bryologh

DANIELS, R. E. 1982. Ipopulations of SphaWarnst. J. Bryol. 12:6

. 1985. IsozymeSphagnum recurvum •*ain and Finland. J. Br

DE VRIES, A., B. O. VAN1983. Genetic variabiulations of two speciesBryopsida). Lindbergi;

DEWEY, R. 1989. GenetRiccia dictvospora (RioBot. 14:155-167.

ENNOS, R. A. 1983. Mailin plant populations. I

GEMMELL, A. R. 1950.The influence of sexuaduction and distributPhytol. 49:64-71.

GOTTLIEB, L. D. 1982. Cof isozymes in plants.

CURIES, R. P., AND F. T.versity and populationmis rigida Mill.). Evoli

HAMRICK, J. L., Y. B. Li1979. Relationships 1teristics and electrophvariation in plants. Ar200.

HARRIS, H., AND D. A. Hoof Enzyme Electroph<North-Holland. Amstc

HAUFLER, C. H. 1987. £

GENETIC VARIABILITY IN A HAPLOID MOSS 1095

fncients between niches (Ennos,tuation likely to be common inpopulations (Wyatt, 1982;Wyattson, 1984).other hand, Yamazaki (1981,

preted his discovery of extensiveiation in Japanese populations ofort Conocephalum conicum as aonstration that allozyme poly-s are selectively neutral. He ar-only under the model of selective3f genetic variability do we expectity of polymorphisms betweend diploid organisms" (Yamazaki,4).Yamazaki's(1981, 1984) data,are open to question. He chose to3 number of nonspecific enzymes

be unusually variable in othericluding five esterases and threeThis may have artifically elevatedates of genetic variability. Fur-his results conflict strongly with

lies of C. conicum (e.g., Szwey-d Krzakowa, 1979; Szweykowskila; Odrzykoski, unpubl.), in whichmorphism was found within dif-:es of this liverwort species. All> agree, however, that there is littleition among populations, a findingith our results for P. ciliare.rer therefore, two contrasting pic-enetic population structure within13 species: 1) the "ConocephalumI in which there are low levels ofi within races (which probably rep-karate biological species), weak in-tion differentiation, and no mi-

heterogeneity; and 2) the•mum model," in which there areIs of genetic variation, strong in-'.tion differentiation, and micro-;rogeneity. Wyatt (1985) has dis-he ecological and evolutionaryons of these differing population5. In any event, it is clear that atie bryophytes, despite their status^genetic relicts," maintain signifi-es of genetic variability. Perhaps|tudy will reveal that these organ-as diverse in terms of genetic pop-:ructure as are angiosperms or var-ups of animals. Certainly, their/stems appear to be similar in kindof diploid plants and do not ob-

viously constrain their potential for varia-tion and evolution.

ACKNOWLEDGMENTSThis research was supported by NSF

Grant BSR-8408931. The Palfrey Fund ofthe Department of Botany at the Universityof Georgia and the Jessup Fund of the Acad-emy of Natural Sciences in Philadelphiahelped to make I. J. Odrzykoski's visit tothe U.S. possible. We thank G. E. Wyatt forfield assistance and J. L. Hamrick for com-ments on the manuscript.

LITERATURE CITEDANDERSON, L. E. 1963. Modem species concepts:

Mosses. Bryologist 66:107-119.. 1980. Cytology and reproductive biology of

mosses, pp. 37-76. In R. J. Taylor and A. E. Lev-iton (eds.), The Mosses of North America. Amer.Assoc. Adv. Sci., San Francisco, CA.

AY ALA, F. J. 1975. Genetic differentiation during thespeciation process. Evol. Biol. 8:1-78.

CRUM, H. 1972. The geographic origins of the mossesof North America's eastern deciduous forest. J.Hattori Bot. Lab. 35:269-298.

CUMMINS, H., AND R. WYATT. 1981. Genetic vari-ability in natural populations of the moss Atrichwnangustatum. Bryologist 84:30-38.

DANIELS, R. E. 1982. Isozyme variation in Britishpopulations of Sphagnum pulchrum (Braithw.)Warnst. J. Bryol. 12:65-76.

. 1985. Isozyme variation in populations ofSphagnum recurvum var. mucronatum from Brit-ain and Finland. J. Bryol. 13:563-570.

DE VRIES, A., B. O. VAN ZANTEN, AND H. VAN DUK.1983. Genetic variability within and between pop-ulations of two species ofRacopilum (Racopilaceae,Bryopsida). Lindbergia 9:73-80.

DEWEY, R. 1989. Genetic variation in the liverwortRiccia dictyospora (Ricciaceae, Hepaticopsida). Syst.Bot. 14:155-167.

ENNOS, R. A. 1983. Maintenance of genetic variationin plant populations. Evol. Biol. 16:129-155.

GEMMELL, A. R. 1950. Studies in the Bryophyta. I.The influence of sexual mechanism in varietal pro-duction and distribution of British Musci. NewPhytol. 49:64-71.

GOTTLIEB, L. D. 1982. Conservation and duplicationof isozymes in plants. Science 216:373-380.

GURIES, R. P., AND F. T. LEDIG. 1982. Genetic di-versity and population structure in pitch pine (Pi-nus rigida Mill.). Evolution 36:387-402.

HAMRICK, J. L., Y. B. LINHART, AND J. B. MITTON.1979. Relationships between life history charac-teristics and electrophoretically detectable geneticvariation in plants. Ann. Rev. Ecol. Syst. 10:173-200.

HARRIS, H., AND D. A. HOPKINSON. 1976. Handbookof Enzyme Electrophoresis in Human Genetics.North-Holland, Amsterdam, Neth.

HAUFLER, C. H. 1987. Electrophoresis is modifying

our concepts of evolution in homosporous pteri-dophytes. Amer. J. Bot. 74:953-966.

HAUFLER, C. H., AND D. E. SOLTIS. 1986. Geneticevidence suggests that homosporous ferns with highchromosome numbers are diploid. Proc. Nat. AcadSci. USA 83:4389-4393.

KHANNA, K. R. 1964. Differential evolutionary ac-tivity in the bryophytes. Evolution 18:642-670.

KOPONEN, T. 1971. A monograph of PlagiomniumSect. Rosulata (Mniaceae). Ann. Bot. Fenn. 8-305-367.

KRZAKOWA, M., AND J. SZWEYKOWSKI. 1977a. Per-oxidases as taxonomic markers in two critical Pelliataxa (Hepaticae, Pelliaceae). Bull. Acad. Polon. Sci.Ser. Sci. Biol. 25:203-204.

. 19776. Peroxidases as taxonomic characters.II. Plagiochila asplenioides (L.) Dum. sensu Grolle(= P. maior S. Arnell) and Plagiochila porelloides(= P. asplenioides aucti non Grolle; Hepaticae, Pla-giochilaceae). Bull. Soc. Sci. Lett. Poznan Ser. D17:33-36.

1979. Isozyme polymorphism in naturalpopulations of a liverwort, Plagiochila asplenioides.Genetics 93:711-719.

LONGTON, R. E. 1976. Reproductive biology and evo-lutionary potential in bryophytes. J. Hattori Bot.Lab. 41:205-223.

LOVELESS, M. D., AND J. L. HAMRICK. 1984. Ecolog-ical determinants of genetic structure in plant pop-ulations. Ann. Rev. Ecol. Syst. 15:65-95.

NEI. M. 1972. Genetic distance between populations.Amer. Natur. 106:283-292.

. 1973. Analysis of gene diversity in subdivid-ed populations. Proc. Nat. Acad. Sci. USA 70:3321-3323.

. 1975. Molecular Population Genetics andEvolution. North-Holland, Amsterdam, Neth.

ODRZYKOSKI, I. J., M. A. BOBOWICZ, AND M.KRZAKOWA. 1981. Variation in Conocephalumconicum—The existence of two genetically differentforms in Europe, pp. 519-542. In J. Szweykowski(ed.), New Perspectives in Bryotaxonomy and Bry-ogeography. Adam Mickiewicz Univ., Poznan, Po-land.

ODRZYKOSKI, I. J., AND L. D. GOTTLIEB. 1984. Du-plications of genes coding 6-phosphogluconate de-hydrogenase in Clarkia (Onagraceae) and their phy-logenetic implications. Syst. Bot. 9:479^189.

ODRZYKOSKI, I. J., AND J. SZWEYKOWSKI. 1981. Aninteresting enzymatic polymorphism in some Eu-ropean populations of the liverwort Mannia fra-grans (Balbis) Frye and Clark, pp. 33-37. In J.Szweykowski (ed.), New Perspectives in Bryotax-onomy and Bryogeography. Adam MickiewiczUniv., Poznan, Poland.

ROGERS, J. S. 1972. Measures of genetic similarityand genetic distance. Univ. Texas Publ. Stud. Ge-net. 7213:145-153.

SCHUSTER, R. M. 1966. The Hepaticae and Antho-cerotae of North America, Vol. I. Columbia Univ.Press, N.Y.

SHAW, C. R., AND R. PRASAD. 1970. Starch gel elec-trophoresis of enzymes: A compilation of recipes.Biochem. Genet. 4:297-320.

SMITH, A. J. E. 1978. Cytogenetics, biosystematics

1096 R. WYATT ET AL.

and evolution in the Bryophyta. Adv. Bot. Res. 6:195-276.

SOLTIS, D. E., AND P. S. SOLTIS. 1989. Polyploidy,breeding systems, and genetic differentiation inhomosporous pteridophytes. In D. E. Soltis and P.S. Soltis (eds.), Plant Isozymes. Dioscorides Press,Portland, OR. In press.

STEERE, W. C. 1954. Bryophytes. Bot. Rev. (Lancas-ter) 20:425-450.

SWOFFORD, D. L., AND R. B. SELANDER. 1981. BIO-SYS-1: A Fortran program for the comprehensiveanalysis of electrophoretic data in population ge-netics and systematics. J. Hered. 72:281-283.

SZWEYKOWSKI, J., AND M. KRZAKOWA. 1979. Varia-tion of four enzyme systems in Polish populationsof Conocephalum conicum (L.) Dum. (Hepaticae,Marchantiales). Bull. Acad. Polon. Sci. Ser. Sci. Biol.27:37^1.

SZWEYKOWSKI, J., I. J. ODRZYKOSKI, AND R. ZIELINSKI.19 81 a. Further data on the geographic distributionof two genetically different forms of the liverwortConocephalum conicum (L.) Dum: The sympatricand allopatric regions. Bull. Acad. Polon. Sci. Ser.Sci. Biol. 28:437-449.

SZWEYKOWSKI, J., R. ZIELINSKI, AND M. MENDELAK.19816. Variation of peroxidase isoenzymes in cen-tral European taxa of the liverwort genus Pellia.Bull. Acad. Polon. Sci. Ser. Sci. Biol. 29:9-19.

WYATT, R. 1977. Spatial pattern and gamete dis-persal distances in Atrichum angustatum, a dioi-cous moss. Bryologist 80:284-291.

. 1982. Population ecology of bryophytes. J.Hattori Bot. Lab. 52:179-198.

1985. Species concepts in bryophytes: Inputfrom population biology. Bryologist 88:182-189.

WYATT, R., AND L. E. ANDERSON. 1984. Breedingsystems of bryophytes, pp. 39-64. In A. F, Dyerand J. G. Duckett (eds.), The Experimental Biologyof Bryophytes. Academic Press, London, U.K.

WYATT, R., I. J. ODRZYKOSKI, A. STONEBURNER, H. W.BASS, AND G. A. GALAU. 1988. Allopolyploidy inbryophytes: Multiple origins of Plagiomnium me-dium. Proc. Nat. Acad. Sci. USA 85:5601-5604.

WYATT, R., A. STONEBURNER, AND I. J. ODRZYKOSKI.1989. Bryophyte isozymes: Systematic and evo-lutionary implications. In D. E. Soltis and P. S.Soltis (eds.), Plant Isozymes. Dioscorides Press,Portland, OR. In press.

YAMAZAKI, T. 1981. Genie variabilities in naturalpopulation of haploid plant, Conocephalum coni-cum. I. The amount of heterozygosity. Jap. J. Ge-net. 56:373-383.

. 1984. The amount of polymorphism and ge-netic differentiation in natural populations of thehaploid liverwort Conocephalum conicum. Jap. J.Genet. 59:133-139.

Corresponding Editor: P. W. Hedrick

APPENDIXThe table below shows migration distances for electrophoretic variants of Plagiomnium ciliare under standardconditions (see text). All distances are expressed in mm from the origin. Enzyme codes are given in Table 1,and buffer compositions are described in the text. Code for phenotypes: 1 = single band; 2 = doublet of bandswhich do not segregate within populations; and 3 = triplet of bands which do not segregate within populations.Alleles encoding electrophoretic variants are represented by a-d.

Migration distanceEnzyme

GOT-lALD-1PGI-1EST-1MDH-1MDH-2PGM-1PGM-2TPI-1TPI-2ME-1PEP-1PEP-2PEP-3

Buffer

SSSSM(H)M(H)H(M)H(M)M(H)M(H)SSSS

Phenotypeiii22(3)311211111

a

45283338,4048,4638, 35, 32373145,483017473825

b

33,3049,4738, 29, 22343447,503515493523

C

44,4245,43

312744,4039

38

d

35