RELATIVE CONTRIBUTION OF BREEDING SYSTEM AND ENDEMISM TO GENOTYPIC DIVERSITY: THE OUTCROSSING ENDEMIC TARAXACUM CALIFORNICUM VS. THE WIDESPREAD APOMICT T. OFFICINALE (SENSU LATO)

7 Pages·1998·4.9 MB·English

by  Jennifer C Lyman

Madrono, Vol. 45, No. 4, pp. 283-289, 1998
RELATIVE CONTRIBUTION OF BREEDING SYSTEM AND ENDEMISM TO
GENOTYPIC DIVERSITY: THE OUTCROSSING ENDEMIC
TARAXACUM CALIFORNICUM VS. THE WIDESPREAD APOMICT
OFFICINALE (SENSU LATO)
T.
Jennifer C. Lyman1-2 and Norman C. Ellstrand1
'Department of Botany and Plant Sciences, University of California,
Riverside, CA 92521-0124
and
2Department of Biology, Rocky Mountain College, 1511 Poly Drive,
MT
Billings, 59102-1796
Abstract
We used allozymes to compare the population genetic structure of Taraxacum californicum Munz &
I. M. Johnston, an outcrossing endemic of the San Bernardino Mountains of California, to that of its
widespread weedy relative, T. officinale Wigg. (sensu lato), an obligate apomict. The average number of
allozyme phenotypes per population of the endemic was six times that of the widespread species. The
endemic had abouttwice as muchgenotypicdiversity (estimatedbyD)perpopulationthanthe widespread
species, and that diversity was distributed much more evenly per population than in the widespread
species. Both species had about the same level ofaverage interpopulation differentiation. Forthis pairof
related species, breeding system apparently plays a more important role than endemism in determining
population genetic structure.
Restricted geographical distribution and small whereas clonal or selfing populations will tend to
population size are characteristics ofendemic plant start with relatively low variation.
species. Population genetic theory predicts that Population genetic comparisons of predominant-
these conditions should influence population genet- ly outcrossing endemic species and related selling
ic structure, and, as aconsequence, endemic species or asexual widespread species would reveal valu-
are expected to lack or to have reduced genetic able information about the relative contributions of
polymorphism (reviewed by Ellstrand and Elam breeding system and endemism to population struc-
1993). Empirical evidence that supports those the- ture. Ifendemism is the more important factor, then
oretical expectations is accumulating; and genetic genetic analyses should reveal a paucity of genetic
diversity tends to increase with species range size polymorphism within populations of the endemic
(e.g., Hamrick and Godt 1990; Karron 1991). species compared to a widespread selfing or asex-
Breeding systems are also known to influence ual congener. Alternatively, if breeding system in-
population genetic structure. Theory predicts that fluences population structure more than endemism,
predominantly selfing and asexual species should then within-population genotypic diversity of an
exhibit lower variation within populations and outcrossing species with a restricted distribution
greater interpopulation differentiation than out- should be greater than that of a selfing or asexual
crossers (Jain 1976; Baker 1959; Levin and Kerster widespread species. And because interpopulation
1971). Again, empirical data support these expected gene flow is typically higher in outcrossing species
trends (e.g., Ellstrand and Roose 1987; Hamrick than a selfing orasexual species (HamrickandGodt
and Godt 1990). 1990), then we would expect that an outcrossing
It is not clear which factor, breeding system or endemic should exhibit less interpopulation differ-
endemism, should have a greater influence on pop- entiation than a widespread selfing or asexual spe-
ulation genetic structure. The answer to this ques- cies.
tion will be an important one forplant conservation A genus well suited for comparative study ofthe
managers who are rarely able to assess the genetic effects of endemism and breeding system on pop-
diversity of every species put in their charge. Such ulation genetic structure is Taraxacum. The great
information will help managers judge under what majority of species in this large genus are asexual,
circumstances loss of genetic variation might be producing seeds that are genetically identical to the
most severe. For example, all other things being maternal parent through apomixis (agamospermy);
equal, if breeding system is more important than and most of the remainder are predominantly or
endemism, then outcrossing species will have more obligately outcrossing species (Grant 1981). The
genetic variation to lose as their populations be- species we chose for study are the T californicum
come increasingly fragmented under disturbance Munz & I. M. Johnston (California dandelion) and
.
MADRONO
284 [Vol. 45
T. officinale Wigg. (common dandelion). Although wind. The heads were transported to Riverside for
T californicum is in the section Ceratophora and germination as described below.
T. officinale is in Ruderalia, the two species are so In collecting both species, plants sampled were
closely related that they occasionally spontaneously separated spatially (:»1 m) to insure that rosettes
hybridize when they come in contact (Skinner and were not attached to the same taproot (Naylor
Pavlik 1994). 1941). All collection sites were isolated from each
Taraxacum californicum is a predominantly out- other by at least 4 kilometers.
crossing perennial herb endemic to moist, subalpine
(1950-2400 m) meadows of the eastern San Ber- Germination. Achenes were germinated in4-inch
nardino Mountains of southern California (Munz styrofoam cups in a greenhouse at the University
1974; Krantz 1980; Hickman 1993). The species of California at Riverside and later transferred to
was recently listed as endangered by the U.S. Fish clay pots in the lathhouse. Germination occurred
readily within 3-7 days. One seedling per maternal
and Wildlife Service (Federal Register 1998).
Taraxacum officinale, also a perennial herb na- parent was grown to maturity for electrophoretic
tive to Europe, is a pantemperate weed of lawns, amnoaslyosmise. cAodudnittsi.onal seedlings were used for chro-
meadows, and disturbed places (in California, from
0-3300 m); and, it occurs throughout North Amer- Chromosome counts. Chromosome numbers
ica (Munz 1974; Hickman 1993). The taxon is ob- were counted for the offspring of five individuals
ligately apomictic. While treated as a single species from the Bluff Lake population of T californicum
by most North American researchers, T. officinale and from two individuals from each population of
probably represents what many European experts T. officinale. We germinated seeds from these pop-
judge to be a complex of agamospecies (Grant ulations on moist filter paper in petri dishes under
1981; Richards personal communication). Like oth- lab conditions. Four- to six-day-old root tips were
er North American researchers (Taylor 1987; King placed in vials with 0.2% colchicine solution for 2
1993), we were unable to identify any character h and then fixed in 1:3 aceto-alcohol (Richards
that allowed for the easy assignment of separate 1972a). The root tips were hydrolyzed in 0.1 N HC1
taxa. Therefore, we consider the species tobe "sen- for 11 minutes at 60°C and then stained in Feulgen
su lato" solution for at least 1 h before the squashes were
The purpose of this study was to survey the ge- prepared (Love and Love 1975). We squashed 2-
notypic diverity within and among populations of mm lengths ofroot tips on microscope slides in 1%
T. californicum, an outcrossing endemic, and T. of- aceto-propionic acid before viewing root tip mei-
ficinale, its widespread apomictic congener. We otic cells under the microscope. A minimum of
compared their patterns of genetic diversity to dis- three metaphase plates per plant were examined
tinguish the relative contributions of breeding sys- and counted.
tem and endemism on their population genetic
structure. Additionally, we add some baseline in- TestingfoWrapomixis andautogamy in California
formation on the genetic biology of T. californicum dandelion. re bagged flowers ofT. californicum to
with bagging experiments and chromosome counts. test whether it can set seed without the assistance
of animal vectors (i.e., by apomixis or autogamy).
In June 1982 at Wildhorse Meadows, the least dis-
Methods and Materials
turbed site, small-mesh net bags were placed over
Collection ofmaterial. We used unopened flower one unopened capitulum on each of five plants of
buds for our genetic analysis. In the case of T. cal- T californicum. The bags were secured tightly to
ifornicum, in June 1982 we collected one unopened the ground with metal spikes to prevent insect
flower bud from each of 30 randomly selected movement in and out of the bags. In July the bags
plants in each of 5 representative populations from were removed, and each head was placed in a sep-
its range in the eastern San Bernardino Mountains arate plastic bag so that seed set, as judged by
ofCalifornia (Fig. 1). The buds were stored inplas- whether achenes were fully developed and filled,
could be determined in the laboratory.
tic bags and kept cool until they could be extracted
for allozyme study in Riverside. We also collected Electrophoretic analysis. Unopened flower buds
achenes to be germinated forchromosomal analysis from up to 30 plants per population of T. califor-
from 3 individuals located in the population near nicum collected from the field were subjected to
Bluff Lake. isozyme analysis by starch gel electrophoresis. For
In the case of T. officinale, flower buds were ob- T officinale, we analyzed unopened buds from up
tained from plants grown from seed. During June to 20 plants per population of plants grown from
and July 1980 we collected a mature infructescence field-collected seed. Individuals buds were homog-
M
(head) from each of 30 randomly selected plants in enized in two drops of extraction buffer (0.01
each of22 T. officinale populations across the Unit- DTT buffered with 0.1 M Tris-HCl, pH 7.0). Ho-
ed States, detailed in Lyman and Ellstrand (1984). mogenates were adsorbed to paper wicks which
Heads were considered to be mature when they were inserted vertically into 12% electrostarchgels.
were fully opened, with achenes exposed to the We used the Tris-EDTA-borate continuous gel and
1998] LYMAN AND ELLSTRAND: GENETIC DIVERSITY IN TARAXACUM 285
Fig. 1. Collection sites of 5 populations of Taraxacum californicum, San Bernardino Mountains, CA (inset area
enlarged).
electrode buffer system of Heywood (1980). Elec- number ofchromosomes was 31 (2n = 31). Similar
trophoresis was conducted for 4 h at 50 milliamps. levels of unusual aneuploid tetraploidy have been
Plastic containers of ice were placed on the gels reported in some sexual European species of Tar-
during the run to prevent overheating. Internal stan- axacum (Richards 1972a, b, 1973). Taraxacum of-
dards were run on each gel to determine the elec- ficinale, however, has been shown to be a triploid,
trophoretic equivalence of bands from di—fferent X = 8, 2n = 24 (Munz 1974). Our analysis oftwo
populations. We assayed for three enzymes alco- individuals from each of22 populations ofthis spe-
hol dehydrogenase (ADH), phosphogluc—oisomerase cies showed the same chromosome number and
(PGI), and phosphoglucomutase (PGM) using the ploidy level. Chromosome size was distinctly dif-
staining procedures described by Heywood (1980). ferent in the two species. Taraxacum californicum
Genetic analysis ofisozyme patterns in related spe- chromosomes were observed to be more than twice
cies (e.g., Roose and Gottlieb 1976) suggest that the size of those of T. officinale.
electrophoresis resolves five loci (one for ADH, Three of five bagged capitula produced no seed
two for PGI and PGM). Despite the polyploidy of at all. The number of unfertilized ovules per head
these species (see Results), banding patterns were was 53, 73, and 76. Two remaining capitula pro-
simple, and alleles were easily assigned. duced a few filled achenes, 8 of 78 in one, and 2
of 67 in another. Overall, 2.9% of the ovules pro-
Results duced seed when foreign pollen was excluded. Un-
bagged inflorescences on the same plants had full
Chromosome studies revealed a tetraploid chro- seed set (100%). These data support the hypothesis
mosome complement in the nucleus of all individ- that T. californicum is a self-incompatible outcross-
uals of T. californicum surveyed. In all counts, the ing species.
MADRONO
286 [Vol. 45
Table 1. Alleles in T. californicumand T. officinale.
Taxon
T.
Locus Allelle californicum T. officinale
PGI-1 a X X
b X X
c X X
d X X
e X
f X Mini
PGM-1 a X X ill
b X i
c X X 1 2 3 4 5 6 7 8 G9ENO1T0YPE1S1 12 13 14 15 16 17 18 19
d X X
Fig. 2. Number of allozyme phenotypes per population
e X
for T. californicum and T. officinale.
f X
ADH
a X X
b X X
c X X
d X others, the clones were equitably distributed among
e X the individuals of the population. All five T. cali-
fornicum populations, however, had evenness val-
ues close to 1.0 (Fig. 3). Both evenness and diver-
sity differed significantly between these species (P
Genotypic variation was present within and < 0.01; Mann-Whitney U test).
oafmfoicnignalteheinpvoesptuilgaattieodn.sWoef Tr.epocratligfeonrontiycpuimcarnadtheT.r cieIsntwearspompuelaastuironeadludsiifnfgereHnetdiraitciko'nswi(t1h9i7n1)eafcohrmsuple-a
than allele frequency data throughout this study be- for genotypic similarity. This index is as follows:
cause dosage effects due to polyploidy in both spe-
cies were not clearly discernible on the starch gels, y
p p
making it impossible to determine allele frequen-
cies. Levels of variation were different for the two Ixy = tl .
species. Of the five loci considered, two were
monomorphic for both T. officinale and T. califor-
nicum. The polymorphic loci were shared by both
species. Some alleles werecommontoboth species, The values for this formula range from 0 (no sim-
wfhoeurneda5s6otuhneirqsuewegreenostpyepcieess-(saplelciofziycme(Tpahbelneot1)y.peWse) iildaernittiytyamaomnogngpoppouplualtaitoinosn)s)t.o T1h(ecovmapllueetseogfentohetyfpiivce
Tahmeonrgan1g4e7oifndgievniodtuyalpsesofpeT.rcpaolpiufloartniiocnumwasusrv1e7y-e1d9. phiogphu,laftiroonms0.o7f0-T0..9c6al(ixfo-rni0c.u83m).wTehree vuanliufeosrmfloyr
t(xhan=s1ix8.2i;ndFiivgi.du2a)l.sNoof tphoepuslaamteiognencootnytpaei.neIdnmcoorne- t0h.e35-212.0po(pxul=at0i.o8n1s).of T. officinale ranged from
trast, only 21 genotypes were found among the 518
individuals surveyed, ranging from 1-9 (x = 3.2)
genotypes per population (Fig. 2).
Pielou's correction version ofthe Gini Index (D)
describes the degree of diversity within a popula-
tion (cf. Ellstrand and Roose 1987). This index em-
phasizes changes in common rather than rare class-
D
es. can be as little as 0 (a uniform sample) or as
large as 1 (every individual different). The correct-
ed values forthe five populations ofT. californicum
ranged from 0.94-0.97 (x = 0.96). Values for T.
officinale ranged from 0.00-0.89 (x = 0.50). Even-
ness values (£), which reflect how evenly the ge-
notypes within a population are distributed among
individuals, were also calculated (Fig. 3) (cf. Ells-
trand and Roose 1987). This value also ranges from
zero (extreme skewness) to 1.0 (complete equita-
bility). The striking bimodal distribution ofEfor T. 01 0.2 0.3 0.4 0.E5VENNESS0.6 0.7 0.8 0.9 1
officinale showed that in some populations one or Fig. 3. Evenness values per populations for T. officinale
a few dominant clones predominated but that in the and T. californicum.
1998] LYMAN AND ELLSTRAND: GENETIC DIVERSITY IN TARAXACUM 287
Discussion perimentation, but, presently, all of the evidence
supports outcrossing as T. californicum's breeding
In our comparison of two congeners having dif- system.
ferent recombination systems and range sizes, we We recognize that using a small numberofmark-
found that the predominantly outcrossing endemic er loci can underestimate genotypic diversity. Add-
species T. californicum had substantially more ge- ing more characters may identify more genotypes
notypic diversity than the apomictic and wide- (Ellstrand and Roose 1987). Nonetheless, it is clear
spread species T. officinale {sensu lato). The aver- that the two species have different population ge-
age number ofallozyme phenotypes per population netic structures for the same set of genetically
ofthe endemic was six times that ofthe widespread based markers. Genotypic diversity in T. officinale
species. Likewise, the endemic had about twice as is low, but virtually every individual of T. califor-
much genotypic diversity per population as the nicum was found to be genotypically distinct. In-
widespread species, and that diversity was distrib- deed, we are confident that if we were able to add
uted much more evenly per population than in the a few more markers, we would be able to distin-
widespread species. guish among those few individuals of T. californi-
Both species had about the same level ofaverage cum that shared a genotype in this study.
interpopulation differentiation as estimatedby Hed- For this pair of species, the difference in breed-
rick's /. The similarity is curious, given thatthe two ing system is apparently more important in deter-
species were sampled over such different scales (T. mining population genetic structure than the differ-
californicum over a scale of tens of kilometers; T. ence in range size. The relatively high levels of
officinale over hundreds of kilometers). We would genotypic diversity persist in the apparently out-
expect much more interpopulation differentiationin crossing T. californicum despite its limited geo-
T. officinale fortwo reasons. First, we wouldexpect graphic range. We are aware of only one other
more isolation (and consequently more differentia- study that compares an endemic, outcrossing spe-
tion) among distantly sampled populations com- cies with a widespread species that has an alterna-
pared to those sampled at a finer scale. Second, as tive breeding system. The self-incompatibleendem-
noted above, all otherthings being equal, we would ic of New Mexico's Organ Mountains, Oenothera
expect an apomictic species to have reduced gene organensis Munz, is nearly monomorphic at several
flow relative to a sexual species (and consequently allozyme loci despite high polymorphism at self-
more differentiation). Gene flow by seed is the only incompatibility locus (Levin et al. 1979). Most of
means available to those species that reproduce its congeners that have been studied have more ge-
without fertilization, whereas sexual species can netic diversity (Levin et al. 1979), and almost all
disperse their genes by both seed and pollen. The ofthese have anessentially clonal reproductive sys-
best interpretation we can make is that the level of tem (permanent translocation heterozygosity; Grant
isolation among the meadows that make the home 1981). Thus, the trend in Oenothera is opposite that
ofthe California dandelion are roughly the same as found here for Taraxacum.
the level of isolation among our widely sampled Taraxacum californicum is believed to be arelict
common dandelion populations. of the section Ceratophora (Jepson 1925) that is
Results from our bagging experiment are com- thought to have spread through the northern hemi-
patible with an outcrossing breeding system for T. sphere during an interglacial period (Richards
californicum based on self-incompatibility. Bagged 1973). Fossil records ofthe section date to 1 X 105
capitula set few or no seed; unbagged capitula on years B.P. (Chearney and Mason 1936). Subse-
the same plants had full seed set. Although this quently T. californicum is presumed to have be-
species may be pseudogamous or semigamous, re- come isolated from the many other species of the
quiring pollination to stimulate apomictic seed pro- section, which occur in a circumpolar distribution
duction (Richards 1986), such syndromes are un- (Richard 1973). Given its long isolation, the fact
known for Taraxacum (Grant 1981). Also, net bags that its total population size may be less than 6 X
might have raised the temperature of the capitula 103 plants, and its restricted habitat, the polymor-
to the point that apomictic seed production was dis- phism of T. californicum is surprisingly high.
rupted (Stebbins personal communication). How- The remarkable genetic diversity that T. califor-
ever, as noted above, self-incompatible, endemic nicum exhibits despite its endemism suggests that
montane Taraxacum are known in Europe (Rich- breeding systemplays a greaterrole in maintenance
ards 1973), and apomixis requiring pollination has of genetic variation than the constraints of endem-
never been reported for the genus (Grant 1981). ism do to limit it. But there is another possibility
Finally, the high genotypic diversity we discovered to account for the origin and maintenance of ge-
in T. californicum argues against high levels ofap- netic variation. The ploidy difference between T.
omictic seed production. Prior genetic surveys of californicum and T. officinale may explain at least
agamospermous species typically show much lower some of the greater genotypic variation displayed
genotypic diversity (e.g., Ellstrand andRoose 1987; in T. californicum. If this species does, in fact, be-
Diggle et al. 1998). We acknowledge that demon- long to the section Ceratophora, then, like the tet-
strating sexuality conclusively requires further ex- raploid European members of this section, it is
MADRONO
288 [Vol. 45
probably an allotetraploid derived from two diploid parum. International Journal of Plant Sciences 159:
sexual species (Richard 1973). The resulting sexual 606-615.
amphiploid will initially breed true for ahighly het- Ellstrand, N. C. and D. R. Elam. 1993. Population ge-
erozygous genotype but will release that variation netic consequences of small population size: impli-
cations for plant conservation. Annual Review of
slowly over generations through rare tetrasomic re-
Ecology and Systematics 24:217-242.
combination (Grant 1981; Roose and Gottlieb Ellstrand, N. C. and M. L. Roose. 1987. Patterns of
1976). genotypic diversity in clonal plant species. American
Another possibility is that T. californicum is not Journal ofBotany 74:123-131.
yet in evolutionary equilibrium. If fragmentation Federal Register. 1998. Endangered and threatened
and endemism are relatively recent in terms of the wildlife and plants; Final rule to determine endan-
mean generation time of the species, the patterns gered or threatened status for six plants from the
we observed better represent historic inertia than mountains of southern California. Federal Register
63:49006-49022.
the factors that are currently molding patterns of
diversity (Ellstrand and Elam 1993). Because the Granvte,rsVi.ty1P9r8e1s.s,PlNaentwSpYeocrika,tiNoYn,.2nd ed. Columbia Uni-
kspneocwieslitwtlaesofointlsyhidsetsorcyr.ibIefdT.incatlhiifsorcneinctuumryp,opwue- Hamrviecrks,ityJ.iLn. palnadntM.speJc.ieWs.. GPopd.t.431-96930.. AinllAo.zymH.e dDi.-
lation genetic structure is not in equilibrium, we Brown, M. T. Clegg, A. L. Kahler, and B. S. Weir
might expect to see the effects of isolation and en- (eds.) Plant Population Genetics, Breeding, and Ge-
demism eventually work to erode the current levels netic Resources. Sinauer, Sunderland, MA.
of genetic diversity, but such changes could take Hedrick, P. W. 1971. A new approach to measuring ge-
decades. netic similarity. Evolution 25:276-280.
A puzzling feature ofthe chromosome studies is Heywood, J. S. 1980. Genetic correlates of edaphic dif-
the aneuploid condition found in the individuals of ferentiation and endemism in Gaillardia. Ph.D. dis-
sertation. University ofTexas, Austin, TX.
T. californicum at Bluff Lake. Investigators have Hickman, J. C. 1993. The Jepson Manual: Higher Plants
noted the same phenomenon in a number of Euro- of California, University of California Press, Berke-
pean Taraxacum species (Malecka 1962, 1967a, b, ley, CA.
1969; Sorensen and Gudjonsson 1946; Richards Jain, S. K. 1976. The evolution of inbreeding in plants.
1970, 1972a, b, 1973), including many sexual spe- Annual Review of Ecology and Systematics 7:469-
cies (e.g., Sorensen and Gudjonsson 1946). The ex- 495.
tent of the aneuploid condition and its role in re- Jepson, W. L. 1925. A manual offlowering plants ofCal-
productive events in T. californicum are presently ifornia. University ofCalifornia Press, Berkeley, CA.
unknown. Karron, J. D. 1991. Patterns of genetic variation and
breeding systems in rare plant species. Pp. 87-98. in
This research has uncovered relatively high lev- D. A. Falk and K. E. Holsinger (eds.) Genetics and
els of genotypic variation in an endemic species in Conservation of Rare Plants. Oxford Press, New
comparison with its widespread apomictic conge- York, NY
ner. We conclude that endemics need not necessar- King, L. M. 1993. Origins ofgenotypic variation inNorth
ily be genotypically depauperate species relative to American dandelions inferred from ribosomal DNA
widespread congeners. Instead, it is clear that the and chloroplast DNA restriction enzyme analysis.
organization of genetic variation may be subject to Evolution 47:136-151.
other constraints such as breeding system and his- Krantz, T. P. 1980. Rare plant status report. California
Native Plant Society, Sacramento, CA.
tory. Further descriptive and experimental popula- Levin, D. A. and H. W. Kerster. 1971. Neighborhood
tion genetics studies are needed for this and other structure in plants under diverse reproductive meth-
species pairs to determine the nature of these con- ods. American Naturalist 105:345-354.
straints. , K. Ritter, and N. C. Ellstrand. 1979. Protein
polymorphism in the endemic Oenothera organensis.
Acknowledgments Evolution 33:534-542.
Love, A. and D. Love. 1975. Plantchromosomes. Strauss
WewishtothankJanetClegg, AdrienneEdwards,Betty and Cramer, Vaduz. Liechtenstein.
Lord, Maile Neel, Andy Schnabel, Nancy Smith-Huerta, Lyman,J. C. andN. C. Ellstrand. 1984.Clonaldiversity
Bob Soost, and Frank Vasek for their helpful comments in Taraxacum officinale (Compositae), an apomict.
on an earlier draft of this paper. Special thanks to Laura Heredity 53:1-10.
Heraty for her expert editing and work on the figures. Malecka, J. 1962. Cytological studies in the genus Tar-
axacum. Acta Biologica Cracovienska 4:117-136.
Literature Cited . 1967a. Chromosome numbers in five Taraxacum
species from Mongolia. Acta BiologicaCracovienska
Baker, H. G. 1959. Reproductive methods as factors in 10:73-83.
speciation in flowering plants. Cold Springs Harbor . 1967b. Cytoembryological studies in Taraxacum
Symposium. Quantitative Biology 24:177-191. scanicum. Acta Biologica Cracovienska 12:195-206.
Chearney, R. W. and H. L. Mason. 1936. A Pleistocene . 1969. Further cyto-taxonomic studies in the ge-
flora from Fairbanks, Alaska. American Museum nus Taraxacum section Erythrospermum. Acta Biol-
Novitates 887:17. ogica Cracovienska 12:57-70.
Diggle, P. K., S. Lower and T. A. Ranker. 1998. Clonal Munz, P. A. 1974. A flora of southern California. Uni-
diversity in Alpine populations of Polygonium vivi- versity ofCalifornia Press, Berkeley, CA.
1998] LYMAN AND ELLSTRAND: GENETIC DIVERSITY IN TARAXACUM 289
Naylor, E. 1941. The proliferation of dandelions from Roose, M. L. and L. D. Gottlieb. 1976. Genetic and
the roots. Bulletin of the Torrey Botanical Club 68: biochemical consequences of polyploidy in Trago-
351-358. pogon. Evolution 30:818-830.
Richards, A. J. 1970. Hybridization in Taraxacum. New Skinner, M. W. and B. M. Pavlik. 1994. California Na-
Phytologist 69:1103-1121. tive Plant Society's inventory ofrare and endangered
. 1972a. The karyology of some Taraxacum spe- vascular plants of California. CNPS Special Publi-
cies from alpine regions ofEurope. Botanical Journal cation No. 1, 5th ed., Sacramento, CA.
boefrt.hr1ee9p7oLr2itbns.n.aITenuaAsx.oSnLooc2vi1ee:t,1y615.6-501:.64P76.-.B5.7.Chromosome num- Sorencsherno,moTs.omaendabeCr.raGnutsdjionnsaspoomni.cti1c946T.araSxpaocnat.anKeoonu-s
1973. The origin of Taraxacum agamospecies. geligo danske Videnskabernes Selskab, Skrifter Biol-
Bot.anical Journal of the Linnean Society 66:189- ogiske 4:1-48.
211. Taylor, R. J. 1987. Population variation and biosyste-
1986. Plant breeding systems. Allen & Unwin, matic interpretations in weedy dandelions. Bulletinof
.
London, England. the Torrey Botanical Club 114:109-120.