Population genetics of the rare alpine endemic Eutrema penlandii Rollins (Brassicaceae)

Material Information

Population genetics of the rare alpine endemic Eutrema penlandii Rollins (Brassicaceae) implications for conservation and management
Hardwick, Renea Chris
Publication Date:
Physical Description:
ix, 77 leaves : illustrations (some color) ; 29 cm


Subjects / Keywords:
Brassica -- Colorado ( lcsh )
Biodiversity conservation -- Colorado ( lcsh )
Biodiversity conservation ( fast )
Brassica ( fast )
Colorado ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references.
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Arts, Biology.
General Note:
Department of Integrative Biology
Statement of Responsibility:
by Renea Chris Hardwick.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
39694081 ( OCLC )
LD1190.L45 1997m .H37 ( lcc )

Full Text
Renea Chris Hardwick
B. S. University of Colorado at Denver, 1994
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts

This thesis for the Master of Arts
degree by
Renea Chris Hardwick
has been approved
Leo P. Bruederle
Linda K. Dixon

Hardwick, Renea Chris (M. A. Biology)
Population genetics of the rare alpine endemic Eutrema penlandii Rollins
(Brassicaceae): Implications for conservation and management
Thesis directed by Assistant Professor Leo P. Bruederle
Eutrema penlandii (Brassicaceae) is the sole representative of this genus in Colorado
and one of only two species of Eutrema occurring in North America. Endemic to the
alpine tundra of the Mosquito Range, E. penlandii grows at elevations in excess of
3659 meters near persistent snowfields that provide continually saturated soils and a
cold microclimate during the growing season. I evaluated population genetic
structure for seven populations of E. penlandii utilizing starch gel electrophoresis and
allozyme analysis. Low levels of genetic variability occurred at both the population
and species level. Only eight of the 28 loci (28.5.1%) exhibited variation (no
criterion) with a total of 38 alleles resolved. Mean population values for the
proportion of loci polymorphic, number of alleles per polymorphic locus, and
observed heterozygosity were only 11.2%, 2.22, and 0.029, respectively. These
values are less than those reported for other species with similar life history traits.
These data suggest two possible explanations for the paucity of genetic diversity

observed in E. penlandii: recent evolutionary origin and severe genetic bottleneck. At
the population level, the low levels of genetic differentiation have undoubtedly arisen
from genetic drift. A comparison of E. penlandii with its Alaskan congener E.
edwardsii indicates that the two are taxonomically distinct; furthermore, E. edwardsii
harbors more genetic diversity than E. penlandii both at the population and species
This abstract accurately represents the content of the candidates thesis,
its publication.
Leo P. Bruederle

I would like to dedicate this thesis to my husband for his moral, financial, and field
support for this work. Without his prodding and encouragement, I would never have
accomplished this goal.

I would like to thank Leo P. Bruederle for his help and understanding with this
project and for coming to the rescue whenever he would get a hysterical call from his
graduate student when she couldnt find the plant; my friends Sara Hill and Alicia
Esparza for their help in the field by putting up with arduous hiking and driving
without complaint; Jan Mckee at the U. S. Fish and Wildlife Service for information
and maps; Dave Murray and Alan Batten of the Herbarium of the University of
Alaska Museum for their assistance with Alaska locations; and The Colorado Natural
Areas Program, The Colorado Native Plant Society, and Sigma Xi for financial
support of this research.

List of Figures...........................................viii
List of Tables..............................................ix
1. INTRODUCTION........................................ 1
The Study Species.................................7
2. MATERIALS AND METHODS.............................. 22
3. RESULTS............................................ 32
4. DISCUSSION......................................... 44
Population Genetics............................. 44
Genetic Distinctness ofEutrema penlandii........ 57
5. CONSERVATION AND MANAGEMENT........................ 63
LITERATURE CITED............................................66

1.1 Colorado Distribution ofEutrema penlandii..................... 10
1.2 Northern Mosquito Range........................................12
1.3 Habitat ofEutrema penlandii....................................13
1.4 Illustration ofEutrema penlandii...............................14
1.5 Photograph of Eutrema penlandii................................15
1.6 North American Distribution of E. penlandii and E. edwardsii..17
1.7 Morphology Comparison of E. penlandii with E. edwardsii........19
2.1 Map of E. penlandii Locations..................................25
2.2 Alaskan Distribution of E. edwardsii...........................27
2.3 Location of E. edwardsii Populations Studied...................28
4.1 Dendrogram.....................................................53

1.1 Known E. penlandii Populations................................. 9
2.1 E. penlandii and E. edwardsii Populations Sampled............... 23
2.2 Site Characteristics of Those Eutrema Populations Sampled........24
3.1 Allele Frequencies for Polymorphic Loci in E. penlandii......... 33
3.2 Summary of Allozyme Data for E. penlandii........................34
3.3 Allele Frequencies for Polymorphic Loci in E. edwardsii......... 36
3.4 Summary of Allozyme Data for E. edwardsii....................... 37
3.5 Gene Diversity Statistics for E. penlandii and E. edwardsii..... 40
3.6 Matrix of Genetic Identities and Distances for Eutrema...........42
3.7 Mean Genetic Identities for Eutrema............................. 43
4.1 Population level comparison of E. penlandii with other plant
4.2 Species level comparison of E. penlandii with other plant
species......................................................... 46
4.3 Comparison of genetic diversity of E. penlandii with other plant

Many rare and endangered plants have experienced declines in population size and
number. A 1988 report from the Center for Plant Conservation identified 253 plants
in the United States that are susceptible to extinction within five years, and an
additional 427 that are susceptible to extinction within ten years. Globally, plants are
faring no better. Although fewer than 1000 of the 250,000 described species of
vascular plants have gone extinct in the past 100 years, nearly 60,000, mostly in the
tropics, are predicted to become extinct within the next 50 years (Raven, 1987). An
absence of knowledge about the biology of endangered plants is problematic since the
writing and implementation of recovery plans for threatened or endangered species is
dependent upon extensive biological research. This lack of understanding is reflected
in the fact that management and conservation efforts of endangered species are not
keeping pace with escalating losses. Between the 1973 enactment of the Endangered
Species Act and 1990, only 16 species were de-listed. Over the same time period, 26
species that were either listed or under review for listing have gone extinct (Schemske
et al., 1994).
Conservation biologists have directed their attention toward rare and endemic
species. However, in the face of an increasing human populations demand for land

and the resultant loss of habitat, protecting all remaining individuals of an endangered
species is frequently not possible, and decisions about which populations to protect
must be made (Lesica et al., 1988). One of the goals of conservation biology is to
preserve the evolutionary potential of species by maintaining natural levels of genetic
diversity. Thus, the long-term management and conservation of rare or threatened
species requires an understanding of both their genetic structure and population
biology (Godt et al., 1995). Although the short-term persistence of many species will
be determined by their ability to recover numerically from population declines and
extinctions, a species long-term viability is ultimately linked to the genetic variation
that it maintains. Species that lack genetic diversity overall, or those species lacking
variation for specific traits, may not have the ability to adapt to new and changing
environmental conditions and are thus more prone to extinction (Godt et al., 1995).
In the extreme, lack of genetic variation at one or more loci can lead to the loss of an
entire species. The extinction of natural populations of the American chestnut
ICastanea dentata Marshall) and stinking cedar (Torreva taxifolia Arnott) has been
directly linked to their inability to adapt to new pathogens and/or changing
environmental conditions (Godt et al., 1996). Therefore, genetic variation can
contribute to the long term fitness and survivorship of a species.
Population genetic theory predicts the loss of genetic diversity in populations that
remain small for several generations (genetic drift), in populations initiated from a
small number of colonists (founder effect), and in populations that suffer rapid

declines in size (population bottlenecks), particularly if recovery is slow or if size
fluctuations are frequent (Barrett and Kohn, 1991). Rare and/or locally distributed
plants might maintain lower levels of variation than common plants because of their
more restricted population sizes and, consequently, reduced opportunities for
outcrossing (Waller et al., 1987). In addition, directional selection within a narrow
range of environments, low population and individual variation coupled with allelic
loss from genetic drift may also lower levels of variation in rare plants (Young and
Brown, 1996).
Although relatively uncommon, population genetic investigations have been
conducted for rare plant taxa. However, few surveys have compared restricted, rare
species with their widespread congeners. Historical factors influencing species
distributions, population isolation, fluctuations in populations size, and migration
patterns will have a large impact on the patterns of genetic variation between closely
related taxa (Loveless and Hamrick, 1988).
Kruckeberg and Rabinowitz (1985) described several forms of rarity. Their
classification of rarity was based on local population size, habitat specificity, and
geographic range. For example, rare plants may be locally common, but occur in
only a few places (e.g., Lavia discoidea). They may be scarce where they occur, but
geographically widespread (e.g., Plantago cordata Lam.). They may be both scarce
and geographically restricted (e.g., Trifolium stoloniferum Muh.) (Holsinger and

Gottlieb, 1991). As a result, different types of rarity are likely to have different
genetic consequences.
Many rare plants are endangered in part because their populations are small and
geographically isolated (Holsinger and Gottlieb, 1989). However, populations may
be small for a variety of ecological reasons: 1) available sites are few and separated
by distances beyond a species normal dispersal ability, 2) carrying capacity of the
site is low, 3) habitability of the site is of short duration because of successional
replacement, or 4) colonization is in its early stages. Other factors may also be added
such as environmental catastrophes including fires, grazing, drought, floods, insect
and disease outbreaks (Barrett and Kohn, 1991).
When populations are small and geographically isolated from one another,
stochastic factors such as bottlenecks, founder events, genetic drift, and inbreeding
influence population genetic structure by reducing the amount of genetic variation
present. In fact, in small populations, these factors can be a stronger force than
natural selection in influencing population genetic structure (Primack, 1993).
A genetic bottleneck is a single event in time. It occurs when there is a sharp
reduction in the number of individuals comprising a population or species.
Bottlenecks are often associated with colonizing events; if one or a few individuals
establish populations in previously unoccupied territory, this is a founder event.
Bottleneck and founder events often involve sampling error small populations are
rarely genetically representative of the populations from which they are drawn. If

these populations remain small for any extended period of time, sampling effects
become cumulative (Karron et al., 1988). This gives rise to random changes in gene
frequency because of the sampling of gametes from generation to generation referred
to as genetic drift. In these situations, rare alleles tend to be lost; however, other
measures of genetic variation, such as heterozygosity, tend to remain at previous
levels, particularly if the population is not reduced to fewer than ten individuals,
population growth rate is high, and bottlenecks are not repeated. Therefore, large
populations (>100) should maintain higher levels of genetic variability than relatively
small populations (Barrett and Kohn, 1991). Though heterozygosity is not adversely
affected by bottlenecks, the stochastic loss of alleles that occurs at higher rates in
small populations is significant. The loss of alleles may affect the evolutionary
potential of a species (Schemske et al., 1994). These rare and unusual alleles may not
contribute to the fitness of the species at that point in time, but at some point in the
future these alleles may confer an increased fitness in the face of a changing
environment. The result of these events is that allele frequencies within populations
are changing not in response to natural selection, but due to random events.
Therefore, these populations may not be well adapted to their environment, resulting
in lack of fit between population phenotypes and the environment
Another genetic consequence of small population size is the occurrence of
inbreeding involving consanguineous mating. Inbreeding results in a reduction of
heterozygosity and an increase in the frequency of homozygous individuals within the

population in a proportion related to the inbreeding coefficient F. The level of
inbreeding in a population increases over time at a rate dependent upon the effective
population size (Ne) such that F = l/(2Ne) (Primack, 1993). Therefore, populations
become inbred more rapidly when they are of small size. The major effect of
inbreeding on fitness is inbreeding depression. The most commonly assumed genetic
basis for inbreeding depression is the presence in the gene pool of lethal or highly
deleterious recessive alleles (Primack, 1993).
Frankham (1995) investigated the relationship between inbreeding and extinction
utilizing data from previous studies of Drosophila melanoeaster. D. virilis. and Mus
musculus. Inbreeding is presumed to increase the risk of extinction because it
depresses components of reproductive fitness in naturally outbreeding species of
plants and animals. Using mathematical modeling, Frankham (1995) found large and
significant excess extinctions due to inbreeding. This work supports the fundamental
assumption of conservation genetics that inbreeding increases the risk of extinction in
naturally outbreeding species. Although, it has been argued that demographic
stochasticity, environmental stochasticity, and catastrophes are more important causes
of extinction in wildlife than inbreeding (Boyce, 1992; Schemske et al., 1994).
However, the response of populations to these events is affected by the level of
inbreeding and the resultant loss of genetic variation. For example, birth and death
rates are susceptible to inbreeding depression, and sex-ratio distortions are often
found in inbred populations. Furthermore, genetic differences in the ability to survive

climatic extremes and pollutants are also known; genetic variation for disease and
pest resistance is often critical for a populations survival. Therefore, the reduction in
genetic variation that accompanies inbreeding can have wide ranging effects
(Frankham, 1995).
The purpose of this study was to describe population genetic structure in Eutrema
penlandii Rollins, a rare plant that is endemic to the alpine tundra of the Mosquito
Range in the Colorado Rockies. The major goals were to (1) assess levels and
patterns of genetic variation in E. penlandii using allozyme analysis; (2) compare
levels of genetic variation in E. penlandii with other plant species, especially other
endemics; (3) and compare genetic variation and population structure in this species
with its close relative, the geographically widespread E. edwardsii R. Brown.
The Study Species
Eutrema penlandii (Penlands alpine fen mustard, Brassicaceae) is the sole
representative of this genus in Colorado and one of two species of Eutrema R. Brown
occurring in North America. Furthermore, Eutrema penlandii is the only
representative of this genus in the lower 48 states. Its nearest relative, Eutrema
edwardsii. is a circumboreal species. The other 15 species of Eutrema are found in
Asia (Naumann, 1988). First collected in 1935 at Hoosier Ridge in the Mosquito
Range (Park Co., Colorado) by C. W. T. Penland, E. penlandii is currently known
from only 13 distinct localities found along the 25 mile crest of this range in central

Colorado (Table 1.1, Figure 1.1). As such, it may well be one of the most rare plants
in Colorado (Naumann, 1988; U. S. Fish and Wildlife, 1993).
Eutrema penlandii exhibits a high degree of habitat specificity. A plant of the
Colorado alpine tundra, it grows in a macroclimate of long, cold, wet winters, and
cool windy summers. During the growing season, persistent snowfields provide
permanent water runoff, which produce the required microhabitat of sub-irrigated,
wet bryophyte bogs. Eutrema penlandii can be found growing in these sub-irrigated
bryophyte bogs along Streams or pools of water. It is restricted to flat or gently
sloping southerly to south-easterly facing limestone or dolomite substrate slopes
between 3688 m (12,150 ft) and 4054 m (13,300 ft) in elevation (Roy, et al., 1993).
Alpine winters in Colorado may last five months or more; furthermore, summer
temperatures are usually below 16 C with growing seasons only 0 to 70 days per
year. Thus, in its native habitat, this species grows at the adaptive limits of most
other plants (U. S. Fish and Wildlife, 1993).
The habitat specificity exhibited by E. penlandii rarely occurs in Colorado and
may account for the apparent restriction of E. penlandii to the Mosquito Range
(Naumann, 1988). Development of high elevation alpine fens requires a persistent
water source and flat to gently sloping topography. The northeast to southwest trend
of the Mosquito Range, unusual in the Colorado Rockies, permits the leeward
accumulation of heavy winter snowpack. The snowfields melt slowly and persist
throughout the summer providing a constant source of water (Figure 1.2). In

Table 1.1 History and demographics of known populations ofEutrema
penlandii Rollins (Brassicaceae) (Naumann, 1988; and Schwendinger
etal., 1991).
Population Date Discovered Size of Pop in Acres Estimated no. of Individuals
Hoosier Ridge 1935 40 2000
Cameron Amphitheater 1988 2 350+
Kite Lake 1991 3 200
Mount Buckskin 1988 3 500+
Cooney Lake 1988 2 1000+
Mosquito Pass 1967 3 1000+
North London Mine 1967 2 50+
Mount Evans 1991 5 1000+
Pennsylvania Creek 1985 1 200
Sacramento Creek 1991 2 500
Hilltop Mine 1967 3 300+
Dauntless Mine 1980 1 50
Horseshoe 1991 2 250

Figure 1.1 Range of Eutrema penlandii Rollins (Brassicaceae) in Park Co., Colorado

addition, the high elevation and gently sloping topography of the Mosquito Range
allow the aggregation and development of soils. A comparison of the Mosquito
Range with the nearby Ten Mile or Collegiate Range attests to the unusual
topography of the Mosquito Range (Figure 1.2). The Ten Mile Range has steep talus
slopes that extend to treeline, as well as a north- south trend. These factors preclude
the formation of high elevation alpine fens. In addition, the Mosquito Range is
underlain by an abundance of carbonate bedrock which is rare or absent elsewhere in
the Colorado mountains. Thus, there is a rare combination of environmental elements
found in the Mosquito Range that creates conditions suitable for development of an
alpine graminoid/forb wetland community with a bryophyte understory (Roy et al.,
1993) (Figure 1.3). In addition to E. penlandii. the Mosquito Range harbors other
disjunct, rare alpine species such as Saussurea weberi Hulten. Brava humilis C. A.
Meyer, Ipomopsis globularis Brand, and Eriophorum altaicum Meinshausen (Weber,
Eutrema penlandii is a small perennial herb that grows 3-8 cm in height. It has
shiny-green leaves and is glabrous. Basal leaves are long-petioled, 1.0 to 2.5 cm long
with a blade that is ovate to cordate and 5- 10mm long. Cauline leaves are sessile and
narrowly oblong. Clusters of small, white flowers crowd the inflorescence which is
borne atop a peduncle that grows from 1-15 cm long. The fruits, which are
cylindrical siliques with four prominent ribs, are diamond shaped in cross section
(Rollins, 1950) (Figures 1.4 and 1.5).

Figure 1.2 Mount Lincoln in the northern portion of the Mosquito Range (Park Co,
CO) as seen in July, 1996. The Ten Mile Range is seen to the north (right).

Figure 1. 3 Typical habitat of Eutrema penlandii Rollins (Brassicaceae) in the
Mosquito Range. Picture taken at Cameron Amphitheater September, 1995.

Figure 1.4 Depiction of Eutrema penlandii Rollins (Brassicaceae) illustrating
characters of taxonomic importance: silique and basal leaves. Drawing by Carolyn

Figure 1.5 Photograph of Eutrema penlandii Rollins (Brassicaceae) taken in
September, 1995 at Cameron Amphitheater

Eutrema penlandii was described in 1950 by Reed Rollins of the Gray Herbarium
at Harvard University. Until then, only a single species of Eutrema. E. edwardsii. had
been known from North America. The range of E. edwardsii is circumboreal in the
arctic and sub-arctic of North America and Siberia, extending southward into the
mountains of Asia. Rollins (1950) reported in his original description that he was
quite startled to receive a specimen from Colorado separated some 1000 miles from
its nearest relative (Figure 1.6).
As such, E. penlandii represents an extreme disjunction within a primarily Asiatic
genus. While such disjunctions are rare and provide a unique opportunity to
investigate biogeography, plant migration, and evolutionary relationships, there are
other instances of such disparate distributions within genera of the Brassicaceae.
Rollins (1982) reported Stroganowia. Smelowskia. Parrva. Brava. Halimolobos. and
Thellungiella as examples of taxa that reveal the strong affinity between the floras of
central Asia and western North America. Brava humilis provides a good example of
a distribution similar to that of E. penlandii. American species of Brava are restricted
to the arctic and sub-arctic. However, B. humilis extends southward to northern New
England and into the Rocky Mountains of Alberta. Interestingly, B. humilis subsp.
ventosa Rollins populations can be found in the Mosquito Range of Colorado within
the range and elevation of E. penlandii. some 1000 miles from its nearest populations
in Alberta (Rollins, 1982).

Figure 1.6 North American distribution of Eutrema (Brassicaceae).
Eutrema edwardsii
Eutrema penlandii

The extreme disjunction of E. penlandii has several possible explanations.
Eutrema penlandii may be a glacial relict from the Pleistocene. It has been suggested
that advancing glaciers pushed cryophilous species like Eutrema from the Arctic into
the Southern Rockies. As the Mosquito Range was glaciated during the Pleistocene,
it is unlikely that E. penlandii would have existed in situ during these glacial periods.
More than likely, E. penlandii would have been found in nearby South Park (Roy et
al., 1993). During the Holocene warming and glacial retreat, E. penlandii may have
found suitable refiigia by migrating up in elevation. The similarities between the
Brava and Eutrema disjunctions discussed previously seem to suggest that the
Colorado populations might be relictual with migration occurring southward during
glacial advance, as both disjunctions occur within the same general locality and at
approximately the same elevations. Alternatively, it has been postulated that E.
penlandii may be a relict of a more widespread Tertiary flora (Weber, 1965).
There is some controversy regarding the taxonomic status of E. penlandii due to
the morphological similarity of the two North American Eutrema species (Figure
1.7). In 1985, Weber relegated E. penlandii to subspecific status, i.e., Eutrema
edwardsii subsp. penlandii Rollins (Weber), suggesting that there are no
characteristics warranting recognition of E. penlandii at the species level (Weber,
pers. comm). However, Rollins (1993) emphatically stated that, while E. penlandii is
certainly closely related to E. edwardsii as noted in the original description of the
species, there is a clean set of characteristics that set it apart from E. edwardsii that

Figure 1.7 Photographs of a) Eutrema penlandii Rollins (Brassicaceae) taken at
Mosquito Pass Summit, Colorado in August 1996, and b) E. edwardsii Brown taken
at Eagle Summit 108, Alaska in July 1996.
a. b.

justify its recognition at the species level. Eutrema penlandii differs from E.
edwardsii in having more slender pedicels and petals which are narrowly ligulate
instead of broadly spatulate to obovate. Additional differences include E. penlandii
having fewer, more slender, and often somewhat decumbent stems; fewer and
narrower cauline leaves; smaller flowers; and siliques smaller than E. edwardsii
(Rollins, 1950).
One source of evidence that could contribute information to the taxonomic dispute
is karyotypic information; however this information is incomplete for both E.
penlandii and E. edwardsii. No counts have been reported for E. penlandii. although
varying counts have been reported for E. edwardsii including: 2n = 18 and 42
(Rollins, 1966), 2n = 28 (Johnson and Packer, 1968; Packer and McPherson, 1974),
and 2n = 56 (Knaben, 1968). Assuming a base chromosome number of x = 7, which
has strong support from the literature (Les et al.,1995), the aforementioned counts
excepting one, i.e., 2n= 18, undoubtedly represent tetraploid, hexaploid, and
octoploid individuals.
In August 1993, the U. S. Fish and Wildlife Service listed E. penlandii as a
threatened species. This protection was initiated in response to E. penlandiis small
population size, small number of populations, limited range, and extreme habitat
specificity (U. S. Fish and Wildlife, 1993). Despite years of searching by groups such
as the Colorado Native Plant Society, E. penlandiis documented distribution remains
restricted to the Mosquito Range (Naumann, 1988; Schwendinger et al.,1991).

Although nearly 80% of E. penlandii grows on public land, many of these populations
are found on patented mining claims (U. S. Fish and Wildlife, 1993). Increased
recreational use of public land, as well as mining activities, pose the greatest threats
to E. penlandii through destruction of their fragile alpine fen habitat, including
alteration of hydrologic conditions. Off road vehicles moving through E. penlandii
habitat create ruts that alter the surface flow of water to the extent that E. penlandiis
survival is both directly and indirectly affected (Naumann, 1988). The diversion of
runoff and acid mine drainage from mining activities also poses serious threats to E.
penlandiis existence if mining activity resumes in the Mosquito Range. In the face
of these threats, it is necessary that information about E. penlandii be obtained,
thereby permitting implementation of a conservation plan.

Eutrema penlandii sites were identified from previous fieldwork and reports in
the literature (Naumann, 1988; Schwendinger et al., 1991). A total of seven
populations was selected for study based upon habitat, location, and site accessibility
(Tables 2.1 and 2.2). These were distributed throughout the Mosquito Range from
Hoosier Ridge to the north to Hilltop Mine to the south. These seven sites
represented about 60% of all known populations (Figure 2.1). Vegetative tissue for
starch gel electrophoresis was collected in August and September of 1995 and 1996.
In each population, leaf tissue from between 14 and 25 randomly selected individuals
was collected at 1-m intervals to reduce the likelihood of re-sampling the same
genotype. Only cauline leaves were collected to ensure the identity of the tissue, as
the long-petioled basal leaves extend under the mossy substrate, emerging some
distance from the parent plant. Because of the rarity and threatened status of E.
penlandii. care was taken during collection to minimize impact to plants; therefore,
tissue collection was limited to two leaves per individual. Leaves from each
individual were placed in separate plastic bags with moist paper towels, and stored on
ice or under refrigeration for less than 48 hours until extraction of soluble enzymatic

Table 2.1 Locations for seven Eutrema penlandii Rollins and three
E. edwardsii R. Brown (Brassicaceae) populations sampled as part of this research.
Eutrema penlandii
Mount Buckskin Park Co. CO 8S/78W 39.19.05 106.09.43
Cameron Amphitheater Park Co. CO 8S/78W 39.20.45 106.06.30
Cooney Lake Park Co. CO 8S/79W 39.18.15 106.10.00
Hilltop Mine Park Co. CO 10S/79W 39.13.00 106.10.05
Hoosier Ridge Park Co. CO 8S/77W 39.21.50 106.01.40
Mosquito Pass Summit Park Co. CO 9S/79W 39.17.11 106.11.02
North London Mine Park Co. CO 9S/79W 39.17.05 106.10.20
Eutrema edwardsii
Eagle Summit Mile 108 NA AK NA 65.29.101 145.24.250
Eagle Summit Mile 106 NA AK NA 65.28.284 145.25.609
Twelve Mile Summit NA AK NA 65.24.020 145.58.510

Table 2.2 Site characteristics for those Eutrema R. Brown (Brassicaceae)
populations sampled as part of this research.
Eutrema oenlandii
Mount Buckskin 500+ 3 175 S 0-5 3872
Cameron Amphitheater 350+ 2 175 NW 0-2 3994
Cooney Lake 2000+ 2 1000 E 0-3 3826
Hilltop Mine 250+ 3 85 SW 0-3 3933
Hoosier Ridge 300+ 20 15 SSE 0-15 3872
Mosquito Pass Summit 1000+ 3 350 Sto SE 0-5 4009
North London Mine 50 2 25 E to S 0-15 3857
Eutrema edwardsii
Eagle Summit Mile 108 200+ 1 200 W . 0-5 1067
Eagle Summit Mile 106 100+ 1 100 Sto W 0-5 1006
Twelve Mile Summit 100+ 2 50 SE 0-5 973

Figure 2.1 Location of seven Eutrema penlandii Rollins (Brassicaceae) populations
sampled for this research, as well as all known populations of the taxon.

proteins. Approval for tissue collection was granted by the U.S. Fish and Wildlife
Service under the authority of permit PRT-704930, sub-permit 95-54.
Eutrema edwardsii populations were similarly sampled in July 1996 from three
sites in the White Mountains northeast of Fairbanks, Alaska (Figures 2.2 and 2.3;
Tables 2.1 and 2.2). These sites were selected based upon accession information
from the Herbarium at the University of Alaska Museum, as well as accessibility.
Due to time constraints and the inaccessibility of other known E. edwardsii
population sites, only three populations were collected from Alaska. Between 14 and
52 individuals in each population were sampled at a minimum of one meter intervals.
The entire stem was collected with fruits attached, placed in plastic bags with moist
paper towels, stored on ice, and expressed mailed to the University of Colorado at
Denvers plant systematics lab, where soluble enzymatic proteins were extracted
within 72 hrs from time of collection.
Proteins were extracted in approximately six drops of a cold Tris-HCl grinding
buffer modified by the addition of 4% w/v PVP 40,000 and 1% v/v B-
mercaptoethanol (Soltis et al., 1983). A small amount of sea sand facilitated
grinding. Extracts were adsorbed onto 1.5x11 mm sample wicks cut from No. 17
Whatman chromatography paper and stored at -76C until electrophoresis.
Eutrema penlandii samples were applied to each of three different starch gel
(10.5%) and electrode buffer systems (Nielsen and Johansen, 1986): lithium-borate,
gel buffer pH 8.3/electrode buffer pH 8.1; histidine HCl-tris, gel buffer pH

edwardsri R

Figure 2.3 Location of three Eutrema edwardsii R. Brown (Brassicaceae) sites in the
White Mountains of Alaska sampled as part of this study.
Sites sampled

7.5/electrode buffer 7.5; tris-citrate, gel buffer 7.8/electrode buffer 8.7. Gels were run
at 4 C until a bromophenol blue marker had migrated 12 cm. Electrophoretic
conditions varied with the system: 270 V constant for 10 hrs (lithium-borate system,
pH 8.3/8.1), 75 ma constant for 9 hrs (histidine-HCl, pH 7.5/7.5), and 50 ma constant
for 9 hrs (tris-citrate, pH 7.8/8.7).
Following electrophoresis, gels were sliced horizontally into six slices, each 1.5
mm thick. End slices were discarded and interior slices incubated in stains for
between one and 12 hrs at 22 C in the dark. Staining time varied with stain and
enzyme activity. Fifteen substrate stains, revealing 28 presumptive loci, were utilized
for data collection; enzyme nomenclature follows the International Union of
Biochemistry (1984). The lithium borate gel was stained for alcohol dehydrogenase,
ADH-2 (Gottlieb, 1973); malic enzyme, ME (Soltis et al., 1983);
phosphoglucoisomerase, PGI-1 and PGI-2 (Soltis et al., 1983), superoxide dismutase,
SOD-1 and SOD-2 (Soltis et al., 1983); and triose-phosphate isomerase, TPI-land
TPI-2 (Soltis, et al., 1983). The histidine-HCl gel was stained for isocitrate
dehydrogenase, IDH-2 (Soltis et al., 1983); malate dehydrogenase, MDH-1, MDH-2,
MDH-3, and MDH-4 (Soltis et al., 1983); menadione reductase, MNR-1, MNR-2,
and MNR-3 (modified from Conkle et al., 1983); 6-phosphogluconate dehydrogenase,
PGD-1, PDG-2, andPGD-3 (Gottlieb, 1973); phosphoglucomutase, PGM-2 and
PGM-3 (Soltis et al., 1983); and shikimate dehydrogenase, SDH-1 (Soltis et al.,
1983). The tris-citrate gel was stained for acid phosphatase, ACP-2, ACP-3, and

ACP-4 (Soltis et al., 1983); aspartate aminotransferase, AAT-2, and AAT-3 (Cardy et
al., 1981); and glyceraldehyde-3-phosphate dehydrogenase, G3PDH-2 (Soltis et al.,
1983). Four additional loci were examined in E. penlandii (ADH-1, EST, PGM-1,
SDH-2) but were not included in the data analysis due to staining inconsistency.
Eutrema edwardsii electrophoresis followed the procedures outlined above.
However, gel and electrode systems did not resolve as many loci; only 11 substrate
stains, revealing 19 loci, were found to be suitable. The lithium-borate gel was
stained for malic enzyme, ME; phosphoglucoisomerase, PGI-land PGI-2; superoxide
dismutase, SOD-2; and triose-phosphate isomerase, TPI-1 and TPI-2. The histidine-
HC1 gel was stained for isocitrate dehydrogenase, IDH-2; malate dehydrogenase,
MDH-1, MDH-2, MDH-3, and MDH-4; menadione reductase, MNR-land MNR-2;
6-phosphogluconate dehydrogenase, PGD-1 andPGD-3; phosphoglucomutase, PGM-
2 and PGM-3; and shikimate dehydrogenase, SDH-1. The tris-citrate gel was stained
for glyceraldehyde-3-phosphate dehydrogenase, G3PDH-2.
For each population, data were collected as individual genotypes. Designation of
presumptive loci and alleles was based upon relative mobility of the proteins. The
fastest locus was denoted 1, while the more slowly migrating loci were given
sequentially higher numbers. Alleles at a locus were assigned sequential letters in the
same manner with the fastest allele denoted a and subsequent alleles designated
b, etc. No formal genetic analyses were performed to document the pattern of
inheritance of putative alleles. Putative genetic loci and genotypes were inferred

from known substructure and intracellular compartmentalization of enzymes, and
electrophoretic patterns observed in individuals presumed heterozygous at these loci.
Data from 10 populations comprising 228 individuals were utilized for these
analyses with E. penlandii and E. edwardsii populations analyzed separately, as well
as together, due to the differing number of loci resolved for the two taxa. The
following statistics were calculated: allele frequencies, genotype frequencies, number
of alleles per locus (A), number of alleles per polymorphic locus (Ap), proportion of
loci polymorphic (Pp), effective number of alleles (A.p) calculated as Ae = 1/(1 He)
(Hamrick and Godt, 1990), observed and expected heterozygosity (Hobs and Hex), and
genetic identity (I), and genetic diversity statistics (Ht, Hs, DSt, Gst) (Nei, 1978).
Data were analyzed, in part, using BIOSYS-1 (Swofford and Selander, 1981) and the
GENESTAT program written by Whitkus and Lewis (Lewis, 1989). Statistics were
modified for polyploidy, as necessary.

The fourteen enzyme systems assayed for R penlandii are coded by 28 putative
loci. Twenty of the loci appeared monomorphic in all populations: AAT-2, ACP-2,
ACP-3, ACP-4, ADH-2, IDH-2, MDH-1, MDH-2, MDH-3, MDH-4, ME, MNR-2,
MNR-3, PGD-1, PGD-2, PGI-1, PGM-3, SOD-1, SOD-2, and TPI-2. Eight of the 28
loci were found to be variable in one or more populations (no criterion): AAT-3,
G3PDH-2, MNR-1, PGD-3, PGI-2, PGM-2, SDH-1, and TPI-1 (Table 3.1). The
proportion of polymorphic loci (Pp) averaged across the seven populations was
11.20% ranging from 3.60 % at Hilltop Mine to 21.40% at Hoosier Ridge (Table 3.2).
At the species level, Ps was 28.5%. Thirty-eight alleles were resolved (no criterion).
The mean number of alleles per locus (Ap) ranged from 1.03 for the Hilltop Mine
population to 1.25 for the Hoosier Ridge population with a mean of 1.13 (Table3.2).
The mean number of alleles per locus for the species (As) was 1.40 (Table 3.2). The
mean number of alleles per polymorphic locus averaged across nine populations
(App) was 2.22, ranging from 2.0 for the Mt. Buckskin and Hilltop Mine populations
to 2.50 for the North London Mine population. The mean number of alleles per
polymorphic locus for the species (Aps) was 2.25. The effective number of alleles
(Aep) averaged 1.03 ranging from 1.01 for the Hilltop Mine population to 1.04 for

Table 3.1 Allele frequencies for eight polymorphic gene loci (0.99 criterion)
observed in seven populations of Eutrema penlandii Rollins (Brassicaceae),
where N = number of individuals sampled.
AAT-3 N 25 25 25 24 26 25 14 164
a 1.00 1.00 1.00 1.00 1.00 0.92 1.00 0.99
b 0.00 0.00 0.00 0.00 0.00 0.08 0.00 0.01
G3PDH-2 N 25 25 25 22 26 25 14 162
a 1.00 0.96 1.00 1.00 1.00 1.00 1.00 0.99
b 0.00 0.04 0.00 0.00 0.00 0.00 0.00 0.01
MNR-1 N 25 25 25 24 26 25 14 164
a 0.70 0.54 0.74 1.00 0.69 0.65 0.66 0.71
b 0.30 0.39 0.22 0.00 0.16 0.35 0.32 0.25
c 0.00 0.07 0.04 0.00 0.15 0.00 0.02 0.04
PGD-3 N 25 25 25 24 20 25 14 158
a 0.02 0.07 0,03 0.00 0.01 0.00 0.00 0.02
b 0.98 0.93 0.97 1.00 0.99 1.00 1.00 0.98
PGI-2 N 25 25 25 24 20 25 14 158
a 0.00 0.00 0.00 0.00 0.06 0.00 0.00 0.01
b 1.00 1.00 1.00 1.00 0.94 1.00 1.00 0.99
PGM-2 N 25 25 25 24 20 25 14 158
a 1.00 1.00 1.00 1.00 0.91 1.00 1.00 0.99
b 0.00 0.00 0.00 0.00 0.09 0.00 0.00 0.01
SDH-1 N 25 25 25 22 26 25 14 156
a 0.00 0.00 0.00 0.00 0.06 0.00 0.00 0.01
b 1.00 1.00 1.00 1.00 0.94 1.00 1.00 0.99
TPI-1 N 25 25 25 24 26 25 14 164
b 0.60 0.00 0.00 0.92 0.00 0.01 0.21 0.25
c 0.00 0.36 0.01 0.00 0.01 0.01 0.00 0.056
d 0.40 0.64 0.99 0.08 0.99 0.98 0.79 0.70
Population 1 = Mt. Buckskin, 2 = Cameron Amphitheater, 3 = Cooney Lake
4 = Hilltop Mine, 5 = Hoosier Ridge, 6= Mosquito Pass, 7 = North
London Mine.

Table 3.2 Summary of allozyme variation over 28 loci for seven populations of
Eutrema penlandii Rollins (Brassicaceae), where N = number of individuals sampled,
P = proportion of loci polymorphic, A = number of alleles per locus, Ap = number of
alleles per polymorphic locus, Ae = effective number of alleles per locus,
Ho = observed heterozygosity, and He = expected heterozygosity.
Population N P (.99) P (.95) A Ap Ae Ho He
Mt Buckskin 25 10.70 7.10 1.10 2.00 1.03 0.032 0.033
Cameron Amphitheater 25 14.20 10.70 1.20 2.25 1.04 0.044 0.043
Cooney Lake 25 10.70 3.60 1.10 2.30 1.02 0.028 0.017
Hilltop Mine 24 3.60 0.00 1.03 2.00 1.01 0.000 0.005
Hoosier Ridge 26 21.40 14.20 1.25 2.20 1.03 0.036 0.032
Mosquito Pass 25 10.70 7.10 1.10 2.30 1.02 0.037 0.023
North London Mine 14 7.10 7.10 1.10 2.50 1.03 0.025 0.028
Population mean 23.4 11.20 7.11 1.13 2.22 1.03 0.029 0.026
(SE) 3.9 5.20 4.24 0.07 0.16 0.01 0.010 0.011
Species mean 28.50 7.14 1.40 2.25 1.06 0.030 0.040

the Cameron Amphitheater population. The effective number of alleles for the
species (Aes) was 1.06 (Table 3.2).
The 14 enzyme systems resolved for E. penlandii could not be resolved in their
entirety for E. edwardsii: however, a subset of 11 enzyme systems was resolved for E.
edwardsii. These are coded by 19 putative loci. Fifteen of the loci were
monomorphic in all of the E. edwardsii populations: G3PDH-2, MDH-1, MDH-2,
MDH-3, MDH-4, ME, MNR-2, PGD-1, PGD-3, PGI-1, PGM-2, PGM-3, SDH-1,
SOD-2, and TPI-2. Four of the loci were variable in one or more populations (no
criterion): IDH-2, MNR-1, PGI-2, and TPI-1 (Table 3.3). The proportion of loci
polymorphic averaged across populations (Pp) was 21.10% (Table 3.4). The
proportion of polymorphic loci for the species (Ps) was also 21.10%. A total of 26
alleles were resolved. The mean number of alleles averaged across the three
populations (Ap) was 1.32, ranging from 1.30 for Eagle Summit 106 and Twelve Mile
Summit populations to 1.37 for the Eagle Summit 108 population. The mean number
of alleles per locus for the species (As) was 1.40. The mean number of alleles per
polymorphic locus averaged across the three populations (App) was 2.50, ranging
from 2.25 for the Eagle Summit 106 population to 2.75 for the Eagle Summit 108
population, while the mean number of alleles per polymorphic locus for the species
(Aps) was 2.75. The effective number of alleles (Aep) also varied little, ranging from
1.10 for the Twelve Mile summit population to 1.11 for the Eagle Summit 106 and

Table 3.3 Allele frequencies for four polymorphic gene loci (0.99 criterion)
observed in three populations of Eutrema edwardsii R. Brown (Brassicaceae),
where N = number of individuals sampled.
IDH-2 N 24 25 14 63
a 0.85 0.89 0.89 0.88
b 0.15 0.11 0.11 0.12
MNR-1 N 24 25 14 63
a 0.53 0.24 0.58 0.45
b 0.47 0.51 0.42 0.47
c 0.00 0.25 0.00 0.08
PGI-2 N 24 26 14 64
a 0.32 0.35 0.35 0.34
b 0.68 0.65 0.65 0.66
TPI-1 N 24 25 14 63
a 0.36 0.39 0.46 0.40
b 0.42 0.27 0.23 0.31
c 0.00 0.05 0.21 0.09
d 0.22 0.29 0.10 0.20
Population 1 = Eagle Summit Mile 106, 2 = Eagle Summit Mile 108,
3 = Twelve Mile Summit.

Table 3.4. Summary of allozyme variation for 19 loci over three populations
of Eutrema edwardsii R. Brown (Brassicaceae), where N = number of individuals
sampled, P = proportion of loci polymorphic, A = number of alleles per locus,
Ap = number of alleles per polymorphic locus, Ae = effective number of alleles per
locus, Ho = observed heterozygosity, and He = expected heterozygosity.
Population N P (.99) P (.95) A Ap Ae Ho He
Eagle Summit Mile 106 24 21.10 21.10 1.30 2.25 1.11 0.180 0.100
Eagle Summit Mile 108 26 21.10 21.10 1.37 2.75 1.11 0.180 0.100
Twelve Mile Summit 14 21.10 21.10 1.30 2.50 1.10 0.170 0.100
Population mean 21.3 21.10 21.10 1.32 2.50 1.11 0.180 0.100
(SE) 5.3 0.00 0.00 0.03 0.20 0.01 0.010 0.000
Species mean 21.10 21.10 1.40 2.75 1.24 0.120 0.010

108 populations with a mean of 1.11. The effective number of alleles for the species
(Aes) was 1.24 (Table 3.4).
Of the 19 loci common to E. penlandii and E. edwardsii. a total of 31 alleles was
detected. Twenty-three of these alleles occurred at high frequencies in both the
endemic and the widespread species. Of the remaining eight alleles, five of these
were restricted to the endemic species (G3PDH-2b, MDH-2a, PGD-3a, PGM-2b, and
SDH-la) and three were found only in the widespread species (IDH-2b, MDH-2b,
and TPI-la). Most of the alleles that were restricted to one or the other species occur
at a low frequency or in only a few populations. However, at least two loci did vary
substantially between species. The MDH-2 locus was fixed for different alleles in the
two species, MDH-2a in the endemic and MDH-2b in the widespread species.
Additionally, the TPI-la allele occurred in frequencies ranging from 0.36 to 0.46 in
E. edwardsii. while it was absent from E. penlandii.
Observed heterozygosity (Hop) averaged across the seven populations of E.
penlandii ranged from 0.000 at Hilltop Mine to 0.044 at Cameron Amphitheater, with
a mean of 0.029 (Table 3.2). For the species, observed heterozygosity (Hos) was
0.030. Expected heterozygosity (Hep) averaged across the seven populations of E.
penlandii was 0.026, ranging from 0.005 at Hilltop Mine to 0.043 at Cameron
Amphitheater. Expected heterozygosity (Hes) for the species was 0.040. For three
populations ofE. edwardsii (Table 3.4), observed heterozygosity (Hop) averaged
0.180, ranging from 0.170 at Twelve Mile Summit to 0.180 at both Eagle Summit

Mile 106 and Mile 108. For the species, observed heterozygosity (Hos) was 0.120.
While expected heterozygosity (Hep) for each of the three populations of E. edwardsii
was 0.100, this was 0.100 for the species (Hes).
Neis (1978) gene diversity statistics were calculated and averaged over all loci
polymorphic for each species (Table 3.5). Gene diversity statistics are related by the
following sets of equations: Total gene diversity (Ht) is equal to gene diversity
within populations (Hs) plus gene diversity between populations (Dst). The
coefficient of gene diversity (Gst) is equal to Dst divided by Ht. Among the eight
polymorphic loci in E. penlandii. total genetic diversity (Ht) ranged from 0.011 to
0.452, with a mean of 0.127. Within population diversity (Hs) ranged from 0.011 to
0.404, with a mean of 0.093; while the among population diversity component (Dst)
averaged 0.033, ranging from 0.000 to 0.233. Mean Gst for E. penlandii was 0.262,
ranging from 0.011 to 0.515 indicating that most of the genetic diversity resides
within populations rather than among populations. A large proportion of the genetic
differentiation of E. penlandii is attributable to a single locus, TPI-1. This locus
exhibited high differentiation among populations with respect to the most common
allele. TPI-lb predominated in some populations, while TPI-ld predominated in
others. Without the contribution of the TPI-1 locus, Gst falls to 0.037. Among the
four polymorphic loci observed for E. edwardsii. total genetic diversity (Ht) ranged
from 0.216 to 0.706, with a mean of 0.487. The within population component (Hs)
ranged from 0.221 to 0.689, with a mean of 0.480. Between population diversity

Table 3.5 Gene diversity statistics (Nei 1978) calculated for Eutrema penlandii Rollins and E. edwardsii R. Brown
(Brassicaceae), where HT is total allelic diversity, Hs is that component found within populations, Dsx is that component
found among populations, and Gsx is population differentiation.
E. penlandii (8 loci) ________E. penlandii (7 loci) _______E. edwardsii (4 loci)
Locus Allele fq Hx Hs Dsx GSt Allele fq Hx Hs Dsx Gsx Allele fq Hx Hs Dsx Gsx
AAT-2 a = 0.990 0.020 0.022 0 0.051
b = 0.010
GPD-2 a = 0.990 0.011 0.011 0.000 0.015 a = 0.990 0.011 0.011 0.000 0.015 a = 1.000
b = 0.010 b = 0.010 b = 0.000
IDH-2 a = 1.000 a = 1.000 a = 0.880 0.216 0.221 0.000 0.000
b = 0.000 b = 0.000 b = 0.120
MNR-1 a = 0.710 0.432 0.404 0.028 0.064 a = 0.710 0.432 0.404 0.028 0.064 a = 0.430 0.591 0.549 0.042 0.071
b = 0.250 b = 0.250 b = 0.480
c = 0.040 c = 0.040 c = 0.090
PGD-3 a = 0.020 0.037 0.036 0.000 0.011 a = 0.020 0.037 0.036 0.000 0.011 a = 0.000
b = 0.980 b = 0.980 b = 1.000
PGI-2 a = 0.010 0.017 0.016 0.001 0.033 a = 0.010 0.017 0.016 0.001 0.033 a = 0.340 0.450 0.460 0.000 0,000
b = 0.990 b = 0.990 b = 0.660
PGM-2 a = 0.990 0.025 0.024 0.002 0.059 a = 0.990 0.025 0.024 0.002 0.059 a = 1.000
b = 0.010 b = 0.010 b = 0.000
SDH-1 a = 0.010 0.017 0.016 0.001 0,032 a = 0.010 0.017 0.016 0.001 0.032 a = 0.000
b = 0.990 b = 0.990 b = 1.000
TPI-1 a = 0.000 0.452 0.219 0.233 0.515 a = 0.000 0.452 0.219 0.233 0.515 a = 0.400 0.706 0.689 0.017 0.023
b = 0.250 b = 0.250 b = 0.310
c = 0.050 c = 0.050 c = 0.070
d = 0.700 d = 0.700 d = 0.220
mean 0.127 0.093 0.033 0.262 mean 0.142 0.104 0.038 0.267 mean 0.491 0.480 0.007 0.015

(Dst) ranged from 0.000 to 0.042, with a mean of 0.007. That proportion of genetic
diversity residing among populations (Gst) averaged 0.015, ranging from 0.000 to
0.071, indicating that most of the genetic diversity is maintained within populations
of E. edwardsii.
Neis genetic identities (I) and distances (D) obtained from inter-populational
pairwise comparisons reveal very high levels of similarity (Table 3.6). Populations
within species exhibited high genetic identities with values of 1.00 not unusual.
Genetic identities for populations of E. penlandii averaged 0.983, ranging from 0.950
to 1.000 indicating that the populations are closely related (Table 3.7). Eutrema
edwardsii populations are also very similar genetically; identities ranged from 0.996
to 1.000, with a mean of 0.997. Inter-taxon genetic identities ranged from 0.897 to
0.931 with a mean of 0.914 (Table 3.7), indicating that these two groups, while
closely related, are genetically distinct.

Table 3.6 Matrix ofNei's (1978) unbiased genetic identity (below diagonal) and genetic distance (above diagonal)
obtained from pairwise comparison of seven (1-7) populations of Eutrema penlandii Rollins (Brassicaceae) and
three (8-10) populations E. edwardsii R. Brown.
2 3 456789 10
1 Mt. Buckskin, CO
2 Cameron Amphitheater, CO .985
3 Cooney Lake, CO .981
4 Hhlltop Mine, CO .990
5 Hoosier Ridge, CO .979
6 Mosquito Pass, CO .982
7 North London Mine, CO .993
8 Twelve Mile, AK .924
9 Eagle Summit 106, AK .931
10 Eagle Summit 108. AK .920
.150 .190 .010
.008 .045
.992 .049
.956 .952
.991 1.000 .951
.994 1.000 .950
.995 .998 .968
.921 .906 .910
.919 .911 .916
.920 .907 .897
.021 .018 .007
.009 .006 .005
.000 .000 .002
.051 .051 .033
.002 .003
.998 .001
.997 .999
.906 .909 .918
.910 .914 .923
.909 .911 .919
.079 .072 .083
.083 .084 .084
.098 .094 .097
.095 .087 .109
.099 .095 .095
.096 .090 .093
.086 .080 .085
**** .000 .004
1.000 **** .003
.996 .997

Table 3.7 Matrix of mean interspecific and intraspecific genetic identities
(Nei, 1978) obtained from pairwise comparisons of seven populations of Eutrema
penlandii Rollins (Brassicaceae) and three populations of E. edwardsii. Ranges
are provided in parentheses.
Species N E. penlandii E. edwardsii
E. penlandii 7 0.983 (0.950 1.000)
E. edwardsii 3 0.914 (0.897- 0.931 0.997 (0.996- 1.000)

Population genetics
Eutrema penlandii exhibits notably low levels of genetic variation at both the
species and population level. Measures observed at the species level (P = 28.5, A =
1.40, Ae = 1.06, He = 0.040) are substantially lower than those reported by Hamrick
and Godt (1990) for all plant taxa (P = 50.5, A = 1.96, Ae = 1.21, He = 0.149). In
comparison to species with similar life history characteristics, i.e., dicots, short-lived
herbaceous perennials, and other endemics (Tables 4.1 and 4.2), E. penlandii exhibits
lower values for each of the aforementioned genetic measures.
Of particular interest, E. penlandii maintains less variation than that reported for
other endemic species (P = 40.0%, A = 1.80, Ae = 1.15, He = 0.096) a 50%
decrease in number of loci polymorphic, as well as the amount of expected
heterozygosity. The E. penlandii data support the trend seen in the majority of
studies in which populations of narrowly restricted taxa tend to have less genetic
variation than those of more widespread taxa (Karron, 1991). In a survey of the plant
allozyme literature, Hamrick and Godt (1990) reported that restricted endemic species
have fewer polymorphic loci, fewer alleles per polymorphic locus, and less than 50%
of the genetic variation of more widespread species. In addition, Hamrick and Godt

Table 4.1 Summary of allozyme variation at the population level for
Eutrema penlandii Rollins (Brassicaceae) and other plant species for which
these data are available (Hamrick and Godt, 1990), where N = number of
of populations sampled, Pp = % loci polymorphic ( 0.99 criterion), Ap =
number of alleles per locus, Aep = effective number of alleles per locus
(1/1-He), and Hep= expected heterozygosity.
Statistic N Pr Ap Aep Hep
E. oenlandii
MEAN 7.0 11.2 1.13 1.03 0.026
SD 5.2 0.07 0.01 0.011
Other plant species (473)
MEAN 12.7 34.2 1.53 1.15 0.113
SD 1.3 1.2 0.02 0.01 0.005
Other dicotyledonous species (338)
MEAN 11.9 29.0 1.44 1.13 0.096
SD 1.7 1.3 0.02 0.01 0.005
Other short-lived herbaceous species (159)
MEAN 8.8 28.0 1.40 1.12 0.096
SD 1.2 1.8 0.03 0.01 0.008
Other endemic species (100)
MEAN 6.5 26.3 1.39 1.09 0.063
SD 0.9 2.1 0.03 0.01 0.006

Table 4.2 Summary of allozyme variation at the species level for Eutrema
penlandii Rollins (Brassicaceae) and other plant species for which these data are
are available (Hamrick and Godt, 1990), where N = number of populations sampled,
Ps = % loci polymorphic (0.99 criterion), As = number of alleles per locus,
Aes = effective number of alleles per locus, and Hes = expected heterozygosity.
Statistic N Ps Aes Hes
E. nenlandii
MEAN 7.0 28.5 1.40 1.06 0.040
Other plant species (473)
MEAN 12.7 50.5 1.96 1.21 0.149
SD 1.3 1.4 0.05 0.01 0.006
Other dicotyledonous species (338)
MEAN 11.9 44.8 1.79 1.19 0.136
SD 1.7 1.5 0.04 0.01 0.007
Other short lived herbaceous species (152)
MEAN 8.8 41.3 1.70 1.15 0.116
SD 1.2 2.2 0.06 0.01 0.009
Other endemic species (81)
MEAN 6.5 40.0 1.80 1.15 0.096
SD 0.9 3.2 0.08 0.04 0.010

(1990) found that geographic range is one of the best predictors of genetic diversity
with endemic species tending to have the lowest levels; regionally distributed and
widespread species maintain the most. This reduction in genetic variation may result
from the erosive action of genetic drift in these smaller populations.
The power of the present study results from the comparison between the rare
endemic and its widespread congener, E. edwardsii. Eutrema penlandii exhibits far
less genetic diversity at the species and population level than its widespread congener
for three measures of genetic variation, i.e., P, Ap, and He, while the number of
alleles per locus is similar for the two taxa. Based on the 0.95 criterion at the species
level, there were only one third as many polymorphic loci observed in E. penlandii
(7.14 %) as in E. edwardsii (21.10 %). Eutrema penlandii also exhibited less genetic
diversity, i.e., Ht;; in fact, E. penlandii exhibited one quarter of the total genetic
diversity observed in E. edwardsii (0.127 and 0.491, respectively).
A review by Karron (1987) also found that geographically restricted species
exhibit lower levels of genetic polymorphisms than do widespread taxa, e.g., Clarkia
francisicana Lewis and Raven, Oenothera organensis Munz, Pinus torrevana. and
Stephanomeria malheurensis Gottlieb. At the extreme, several researchers have
found rare, endemic species that lacked any detectable variation: Pedicularis
furbishiae S. Watson (Waller et al., 1987), Howellia aquatilis Gray (Lesica et al,
1988), and Pinus torrevana Parry ex Carriere (Ledig and Conkle, 1983). However,
there have also been reports in which restricted or rare species exhibited moderate

amounts of variation or greater variation relative to their widespread congeners.
Karron (1987) noted moderate levels of genetic polymorphisms in the restricted
species Capsicum cardenasii, Gaura demareei Raven and Gregory, and Pinus
longaeva D. K. Bailey. Gottlieb et al. (1985) reported that Lavia discoidea Keck had
a high level of polymorphism. Lewis and Crawford (1995) found Pleistocene
refugium endemics within the genus Polygonella exhibited greater diversity than
widespread congeners. It is therefore difficult to state categorically that species with
restricted geographic range and small population size will have less genetic diversity
than widespread species.
There has been some speculation that, although endemics maintain less genetic
diversity, they partition it in a manner similar to other plant taxa (Hamrick and Godt,
1990). The Gst value for E. penlandii (0.262) (Table 4.3) is within the range reported
for other species with similar life history characteristics: dicots (0.0273), short-lived
herbaceous perennials (0.233), and other endemics (0.248). It should be reiterated
that the majority of the population differentiation is attributed to a single locus, the
TPI-1 locus (Gst= 0.515), which results from a major shift in allele frequencies
among populations (Table 3.5). In the Mt. Buckskin and Hilltop Mine populations,
the predominant allele was TPI-lb, while in the remaining populations the TPI-ld
allele predominated.
Eutrema penlandii has one third the total genetic diversity (Ht = 0.127) and
within population genetic diversity (Hs = 0.093) compared to that reported by

Table 4.3 Summary of gene diversity statistics (Nei, 1978) for Eutrema
penlandii Rollins (Brassicaceae) and other plant species for which these data are
are available (Hamrick and Godt, 1990), where Hx is total allelic diversity, Hs is
that component found within populations, and Gsx is genetic differentiation.
Hx Hs Gsx
E. uenlandii 0.127 0.093 0.262
Dicotyledonous 0.311 0.214 0.273
Narrowly distributed species 0.300 0.215 0.242
Widespread species 0.347 0.267 0.210
Endemic species 0.263 0.163 0.248

Hamrick and Godt (1990) for all plant taxa (Ht = 0.310; Hs = 0.230) (Table 4.3).
This trend is also observed when E. penlandii is compared to other dicots (0.311,
0.214), short lived herbaceous perennials (0.300, 0.222), and, most interestingly,
other endemic species (0.263, 0.163). Levels of total genetic diversity and within
population genetic diversity in E. penlandii are almost one-half of those reported for
other endemic species. Eutrema penlandii is a genetically depauperate species.
The reduced level of genetic diversity observed in E. penlandii is significant in
light of its increased ploidy level. Electrophoretic patterns, i.e., alleles maintained
and dosage effects, indicate that E. penlandii is most likely tetraploid, while the
sampled E. edwardsii populations are hexaploid. Species that are polyploid are
expected to harbor greater levels of genetic diversity than diploid organisms. The
duplication of entire sets of chromosomes increases the amount of variation possible
within polyploids. With the extra chromosomes present in a tetraploid species, any
mutations that arise may be tolerated as there are three other good copies of the
gene still present within the genome. These mutations may accumulate within the
genome without adverse effects on the organism, thereby increasing the genetic
variability of the population.
The majority of all genetic diversity maintained by E. penlandii. e.g., Ht= 0.127,
is attributable to differences among individuals within populations, i.e., Hs = 0.093.
Only 26.2 % of all genetic diversity is due to genetic differentiation of populations,
with 1.5 % of the genetic variation observed in E. edwardsii due to differences

among populations. While E. penlandii populations are only moderately
differentiated from one another, E. edwardsii populations are nearly identical with
little inter-populational variation whatsoever. There may be several reasons for this
lack of differentiation among E. edwardsii populations. 1) There may be strong gene
flow among these populations, as they are not distant from one another, nor are they
separated by high mountain ranges. 2) Populations across the entire range of this
species were not sampled, and there may be more variation than that seen from this
rather restricted sampling of E. edwardsii. 3) The continuity observed may also result
from recent historical factors. The E. edwardsii populations sampled were obtained
in an area that had not been glaciated during the Pleistocene; therefore, this area may
have provided refugia for E. edwardsii with the result that these populations may have
maintained similar population structure (Hulten, 1968). Endemic species, on the
other hand, might be expected to consist of smaller, more ecologically restricted
populations that have experienced population bottlenecks during their evolutionary
history resulting in greater inter-populational differences (Hamrick and Godt, 1990).
The observed reduction in overall genetic diversity seen in E. penlandii. specifically,
may be due to small population size (genetic drift), restricted distribution, and
evolutionary history (founding events) (Karron, 1991).
Mean genetic identity among E. penlandii populations (I = 0. 983) was higher
than that reported by Gottlieb (1977) for conspecific populations (I = 0.95 0.02).
Genetic identities varied little among pairwise comparisons of populations revealing

that E. penlandii populations are genetically very similar to one another (Figure 4.1).
Furthermore, they are clearly conspecific. There is no apparent relationship between
genetic identity and geographic distance among the seven populations.
Eutrema penlandiis small population sizes makes this species susceptible to
factors that influence levels and apportionment of genetic variation. Founding events
or genetic bottleneck, genetic drift, and inbreeding all could contribute to the low
genetic variation seen in E. penlandii.
Genetic bottlenecks resulting from founding events undoubtedly have played a
role in the evolutionary history of E. penlandii. During glacial and interglacial
periods, E. penlandii would have experienced both range expansion and contraction.
Its penchant for cold microclimates and the fact that its nearest relative, E. edwardsii.
is a circumboreal species, suggest that E. penlandii may have arisen in the high
northern latitudes, and migrated southward during glacial advances, finding suitable
habitat at the glaciers terminus. The Mosquito Range was glaciated and, while E.
penlandii could not have existed in situ, it may have found suitable habitat in South
Park (Roy et al., 1993). With Holocene warming and subsequent glacial retreat, R
penlandii may have migrated elevationally, finding suitable refiigia within the
Mosquito Range due to its unusual geology and topography. During this period of
range contraction, a small number of individuals may have founded the current sites,
resulting in another genetic bottleneck and possible loss of genetic variation. Severe
reduction in genetic variability has been reported for other species, e.g., Sullivantia

Figure 4.1 Dendrogram obtained using Neis (1978) unbiased genetic identities and UPGMA analysis.
.90 ,91 ,92 ,93 ,94 ,95 ,96 ,97 ,98 , 99 1, 00
0o o o 006-0 o oQo-e-e oo o o ooe-o oo o o oo-o o oQo oooQooo o o o o o
Mt Buckskin, Co
Hilltop Mine, Co
Cameron Amph, Co
Cooney Lake, Co
Mosquito Pass, Co
Hoosier Ridge, Co
North London, Co
Twelve Mile, Ak
Eagle Summit 106, Ak
Eagle Summit 108, Ak
o o o oooooe-e-o oo-o o oo o o oo o o oo o o oo o o oo o o oo o o o
,90 ,91 ,92 ,93 ,94 ,95 ,96 ,97 ,98 ,99 1,00

oregana S. Watson (Solits, 1982), Chrvsosplenium iowensis Rydberg (Schwartz,
1985), Pedicularis furbishiae (Waller et al., 1987), and Howelia aquatica (Lesica et
al., 1988). These species all share a requirement for cool, moist conditions, and
probably experienced range contraction and genetic bottlenecks during the warming
period of the Holocene.
Eutrema penlandii populations are generally small in number and area. While
they range in size from one acre with 50 individuals to 20 acres with several
thousand, most populations average two to three acres in size with 200 to 500
individuals. However, population size can be influenced by fluctuations in the alpine
environment. For example, some summers are preceded by a cold, wet winter with
heavy snowpack, resulting in late season exposure of the alpine fens with few
individuals able to flower and set seed. Thus populations already small in size may
be susceptible to further genetic drift. The restricted area and distribution of alpine
fens precludes any increase in range or distribution of E. penlandii. It is unlikely that
populations of this species will increase in number or size; therefore, E. penlandii will
remain susceptible to the effects of genetic drift. Variation that has arisen due to
mutations, favorable or neutral, runs a higher risk of being lost through genetic drift.
The reproductive biology of E. penlandii has been poorly documented and
therefore, one can only speculate as to the degree of inbreeding that may have
occurred from selfing. In two field seasons, no pollinators were observed visiting E.
penlandii flowers (Hardwick, pers. obs.). However, while genetic variation in the

species is low, there does not appear to be any evidence of reduction in the number of
heterozygous individuals as would be expected with inbreeding, especially selfing.
Because this species is polyploid, Hardy-Weinberg equilibrium was not calculated for
this study. Hardy-Weinberg analysis on polyploids is possible only with the
knowledge of whether inheritance is double disomic ortetrasomic (Vrijenhoek, pers.
comm.), which is unknown for E. penlandii.
Eutrema penlandii may be able to tolerate higher levels of inbreeding because
species with low genetic diversity will have lower genetic load, and thus, be less
constrained to evolve self-compatibility. Secondly, some investigators have
suggested that the evolution of selfing is based on reproductive assurance. For
example, populations that occupy marginal areas may need to set seed before
conditions become unfavorable marginal areas may be less favorable to pollinators
thereby reducing pollinator abundance, rare species may have to compete with more
abundant species for pollinator visits, and periodic environmental fluctuations may
reduce population sizes of pollinators (Wyatt, 1988). Limnanthes floccosa provides
an interesting example of the evolution of autogamy. Limnanthes floccosa occupies
areas that suffer from periodic droughts. Therefore, autogamy is advantageous for
insuring reproduction; plants germinate and flower earlier in the spring, avoiding
drought and setting seed without waiting for the emergence of pollinators. Thus,
self-fertilizing individuals may be favored by natural selection (Wyatt, 1988).
Eutrema penlandii emerges and flowers as soon as the sites are free from snow.

Perhaps at this time, these areas are not hospitable to pollinators, and self-fertilization
may be a method that ensures early fertilization and seed set during the short and
often unpredictable alpine summers.
Narrow endemics have previously been shown to be genetically depauperate, e.g.,
Silene diclinis (Prentice, 1984), Pinus torrevana (Ledig and Conkle, 1983),
Pedicularis furbishiae (Waller et al., 1987). Soltis (1982) found low levels of genetic
variation in three species of Sullivantia which is characterized by small, isolated
populations. This lack of variation was attributed to founder effects resulting from
colonization at the margins of the Pleistocene glacial boundary, as well as genetic
drift and inbreeding within these small geographically and ecologically restricted
populations. Similar mechanisms were suggested by Schwaergerle and Schall (1979)
to account for reduced variability and heterozygosity in populations of Saracenia
purperea L. from northerly populations in previously glaciated portions of the
species distribution. A similar explanation seems to have been used to explain
reduced genetic variability in species whose present ranges largely occupy previously
glaciated regions. Loveless and Hamrick (1988) found large differences in genetic
variability between Cirsium pitcheri Torr. ex Eat., a rare regional endemic of the
western Great Lakes shoreline, and Cirsium canescens Petrak, its widespread
congener. The endemic exhibited much less genetic variability than its widespread
congener. While the widespread congener exists in unglaciated areas, the rare
endemic had recently colonized the Great Lakes shoreline following Wisconsin

glaciation. The authors suggested that current geographic distribution may be less
important in determining the genetic organization of a species than the historical and
evolutionary processes that have effected that distribution. While ecological features
such as breeding system, seed dispersal, and seed longevity are important in
organizing genetic variation in plant species, other factors, particularly historical
events such as fluctuations in population size, can intervene to produce divergent
patterns of genetic variation.
Genetic distinctness of Eutrema penlandii
In his description of E. penlandii. Rollins (1950) acknowledged that while E.
penlandii was certainly closely related to its nearest congener, E. edwardsii. there
was a clean set of characteristics that set it apart from E. edwardsii [and justifies] its
recognition at the species level. Nevertheless, Weber (1985) treated E. penlandii as
a subspecies of E. edwardsii (E. edwardsii ssp. penlandii! indicating that he felt there
were no significant differences between the two taxa warranting specific status.
Weber attributes the marked size difference between the two taxa to variation in
alpine and arctic summers (personal comm.). To date, Webers (1985) treatment has
not been widely accepted; currently the U. S. Fish and Wildlife Service, Colorado
Native Plant Society, and the Colorado Natural Areas Program recognize E. penlandii
at the species level (U.S. Fish and Wildlife Service, 1993; Colorado Native Plant
Society, 1997; Naumann, 1988).

There have been other allozyme studies of endemic and/or rare species and their
widespread congeners, e.g., Purdy and Bayer, 1995; Karron et al., 1988; Smith and
Pham, 1996; Linhart and Premoli, 1993;Young and Brown, 1996. Through
examination of allozyme variation in endemics and their congeners, the divergence
across many loci can be estimated. This divergence is referred to as genetic distance
and, conversely, similarity is referred to as genetic identity. The relationship between
the two is not strictly linear. Genetic identity (I), estimates the proportion of genes
that are identical in structure in two populations, and may range in value from zero
(no alleles in common) to one (the same alleles at the same frequencies are found in
both populations). Genetic distance (D) estimates the number of allelic substitutions
per locus that have occurred in the separate evolution of the two populations, and may
range in value from zero ( no allelic change at all) to infinity.
In 1977, Gottlieb reported that mean genetic identities were very high in a
comparison of conspecific plant populations, averaging 0.95. Since Gottliebs early
reports, mean genetic identities reported for conspecific plant populations have not
changed significantly, averaging nearly 0.90 or above. However, mean genetic
identities obtained from pairwise populational comparison of congeneric plant species
do vary considerably, ranging from 0.48 to 0.97 (Crawford, 1983). Summarizing
the data reported for 11 progenitor-derivative species pairs for which speciation was
thought to be relatively recent, many following glacial retreat, Pleasants and Wendel
(1989) reported mean genetic identities ranging from 0.78 to 0.99. Whereas some

authors have demonstrated primary speciation, evidenced by morphological
differentiation, with little or no genetic divergence, e.g., Tetramolopium (Lowrey and
Crawford, 1985), others have reported genetic differentiation without the expected
morphological divergence, e.g., Carex (Bruederle andManos, 1986; Bruederle, in
review). Therefore, while a genetic identity of 0.90 may be an average lower limit
for populations of conspecific taxa, each species must be interpreted within the
context of several factors including morphology, life history traits, and biogeography,
and not solely upon electrophoretic data.
Allozyme data indicate that E. penlandii and E. edwardsii. while structurally
similar, are, in fact, genetically distinct. To a large extent, the alleles present in E.
penlandii are also found in E. edwardsii. A total of 31 alleles was detected in the two
species. Twenty-three of these were shared by both species and were, in general, the
most common. Of those alleles remaining, five were unique to E. penlandii and three
found only in E. edwardsii. Mean genetic identity obtained from interspecific
comparisons (0.914) was far lower than that obtained from infraspecific comparisons;
however, it was not below the 0.90 genetic identity reported for congeneric
populations (Crawford, 1983). In addition, electrophoretic patterns indicate thatE.
penlandii is most likely tetraploid, while the sampled E. edwardsii populations are
hexaploid. Furthermore, while polyploids maintain the same alleles found in their
diploid progenitors, E. penlandii and E. edwardsii do not maintain the same alleles at
the same frequencies.

While genetic identity indicates that these two taxa are conspecifics and may be
recognized as subspecies, genetic identity alone should not be used as a measure of
speciation. Infraspecific comparisons revealed extremely high genetic identities
among populations: E. penlandii with a mean genetic identity of 0.983, ranging from
0.950 to 1.000; and E. edwardsii with a mean genetic identity of 0.997, ranging from
0.996 to 1.000. Clearly, populations within each of these two taxa are very closely
related and exhibit little genetic differentiation.
Additional support for the recognition of E. penlandii at the specific level derives
from the presence of a fixed difference at the MDH-2 locus. Eutrema penlandii
populations were monomorphic for the a allele, while E. edwardsii was
monomorphic for the b allele. Taking into account polyploidy, the presence of a
fixed difference is surprising. In addition to MDH-2, another locus appeared to
exhibit a fixed difference between the two species. PGD-2 locus was resolved in E.
penlandii as monomorphic for the a allele; however, this locus could be resolved in
only two of the three populations of E. edwardsii. which were monomorphic for the
b allele. Due to the inability to resolve the PGD-2 locus for all populations, it was
omitted from the study. If PGD-2 had been included in the analysis, then two loci
would have exhibited fixed differences, substantially lowering the genetic identity
substantially below 0.90.
In addition, there are strong interspecific morphological differences evident in the
field. Eutrema penlandii is quite diminutive, growing only 5-7 cm tall; E. edwardsii

grows 10 40 cm tall. It is unlikely that this difference is due solely to the arctic
growing season with its greater daylength, as suggested by Weber (pers. comm.).
Syntopic species such as those present in Colorado and Alaska did not exhibit any
obvious difference in habit and appeared nearly identical. In addition to size
differences, the shape of the basal leaves is quite different between the two taxa;
Eutrema penlandiis leaves are cordate or heart shaped, while E. edwardsiis leaves
are ovate to elliptical.
Combining allozyme data, including the fixed genetic difference, ploidy, and
morphology allows a more synthetic evaluation of the taxonomy of the two taxa. It is
my opinion that taken together, these factors provide evidence that E. penlandii is
taxonomically distinct from E. edwardsii and should be recognized as a distinct
Not surprisingly, the results invoke new questions. While the allozyme data, when
coupled with other data, suggest a clear and separate identity for E. penlandii. further
examination of allozymes utilizing additional populations and more loci should be
undertaken. Increasing the number of loci and alleles resolved would allow the
collection of a more comprehensive, and possibly, definitive data set. For example, if
PGD-2 could be resolved and included in the analyses, a significant decrease in
genetic identity would probably result, further supporting the distinctiveness of the
two taxa. Also, increasing the number of loci examined and the number of
individuals sampled would increase the probability of identifying rare alleles. To

further increase our understanding of genetic variation, DNA variation should be
examined using novel molecular techniques, e.g., RAPDs; such data could provide
greater genetic detail, as well as further insight into the relatedness and taxonomic
relationships of E. penlandii and its congeners. Because E. penlandii is rare, and its
habitat sensitive to mining operations, greater knowledge of its population genetic
structure, as well as the identification of rare alleles within populations, is crucial to
maximizing the amount of genetic variation conserved should some populations need
to be sacrificed.

One of the goals of conservation programs, in addition to habitat preservation, is
the maintenance of existing levels of genetic variation in species that are rare or
threatened (Barrett and Kohn, 1991). The long term objective is to maintain the
evolutionary viability of the taxon, maximizing its chances for persistence in the face
of changing environments (Huenneke, 1991). With this in mind, the first step toward
the conservation of E. penlandii is to maintain naturally occurring levels of genetic
variation. Given the low genetic diversity among E. penlandii populations, it would
appear that it is not critical that all populations be preserved in order to preserve a
large portion of the genetic variability present within the species. However,
conservation priority should be given to populations large in number and area,
populations that are more genetically variable, and populations that harbor rare
alleles. Rare alleles, while they may not currently confer any evolutionary fitness to
the individuals possessing them, may confer evolutionary fitness in the future.
It should be noted that populations that appear uniform or similar allozymically
may actually possess adaptations to their specific environments that are not apparent
through starch-gel electrophoresis of a limited number enzymes (Richter et al., 1994).
Whenever possible, conservation efforts should attempt to conserve genetic variation
from as many populations as possible. If the goal is to preserve overall genetic

variation, then large population sizes must also be maintained. Small populations
may suffer from erosion of genetic variation due to inbreeding and genetic drift
(Hamrick et al., 1991). Sadly, several biological features of E. penlandii place this
species at high risk of extinction, including small population size, restricted
ecological niche, and low genetic diversity (Primack, 1993).
The existence of infraspecific variation has profound implications for the
conservation of a species. At each level, variation can affect fitness and viability. At
the level of the population, the increase in homozygosity that can result from loss of
variation can lead to inbreeding depression, which is often expressed as a severe
decline in population health. Furthermore, variation is often correlated with
adaptation to the local environment. Reductions in variation or other disruptions of
the local gene pool may decrease the chance that a population can adapt to changing
conditions. At the species level, loss of diverse populations reduces the potential of
the species to respond to environmental changes at regional and global scales (Millar
and Libby, 1991). Thus, a species continued existence depends to a large part on the
genetic health at each of the infraspecific levels.
While population genetic structure of E. penlandii is better understood as a result of
this study, this information needs to be placed within the context of this species life
history to be truly valuable. Unfortunately, very little is known about E. penlandiis
life history. Therefore, information regarding E. penlandiis reproductive biology is
crucial for the development of a sound management and recovery plan. Future work

should include: 1) demographic studies to determine if populations are increasing,
decreasing or remaining stable in number of individuals; 2) pollinator studies to
determine the degree to which E. penlandii outcrosses, as well as the mode of
pollination; 3) reproductive studies quantifying seed set, seed viability, and seed
dormancy requirements; and 4) observations of the effects of increased recreational
use and mining activity on the alpine fen habitats upon which E. penlandii depends.
There is a great deal of information that remains unknown both to the scientific
community and the management community; without this information, the
development of a successful conservation plan is almost certainly destined to fail.
While it is not apparent that E. penlandii provides any major functional support to
the alpine fen community, E. penlandii does add complexity to the local specialized
food web and the significance of this protected species should not be diminished.
The alpine fens in which E. penlandii occurs are rare and unique habitats. These fens
provide a rich source of food to the herbivores of the alpine tundra including marmot,
pica, elk, deer, and bighorn sheep. Unfortunately, alpine fens occur in an area that is
also desirable for recreational use, including off-road four wheeling, as well as
mining of metal ores. At this time, it is not certain what effects these activities may
have on the alpine fen, but it is highly probable that any change in hydrology or
chemistry of these areas will result in their reduction or possible elimination, thereby
removing a very unique and beautiful habitat from the Colorado Rockies.

Barrett, Spencer C. H. and Joshua Kohn. 1991. Genetic and evolutionary
consequences of small populations size in plants: implications for
conservation. Pp. 3-30 in Genetics and Conservation of Rare Plants, eds.
Donald A. Falk and Kent E. Holsinger. Oxford: Oxford University Press.
Boyce, Mark S. 1992. Population viability analysis. Annual Review of Ecological
Systematics 23: 481-506.
Bruederle, L. P. and P. S. Manos. 1986. Conservation genetics of the rare sedge,
Carex mitchelliana M.A. Curtis (Cyperaceae). American Journal of Botany
76: 663 (abstract).
Cardy, B. J., C. W. Stuber, and M. M. Goodman. 1981. Techniques for starch gel
electrophoresis of enzymes from maize (Zea mays L.). Institute of Statistics
Mimeograph Series No. 1317. Raleigh: North Carolina.
Colorado Native Plant Society. 1997. Rare Plants of Colorado. Helena, Montana:
Falcon Press Publishing Co., Inc.

Conkle, M. Thompson, Paul D. Hodgskiss, Lucy B. Nunnally and Serena C. Hunter.
1983. Starch gel electrophoresis of conifer seeds: a laboratory manual
United States Department of Agriculture. Berkeley Ca: Pacific Southwest
Forest and Range Experiment Station.
Crawford, Daniel J. 1983. Phylogenetic and systematic inferences from
electrophoretic studies. Pp. 257-287 in Isozymes in Plant Genetics and
Breeding, eds. S. D. Tanksley and T J. Orton. Amsterdam: Elsevier.
Frankham, Richard. 1995. Inbreeding and extinction: a threshold effect.
Conservation Biology 9: 792-799.
Godt, Mary J. W., J. L. Hamrick, and Susan Bratton. 1995. Genetic diversity in a
threatened wetland species, Helonias bullata (Liliaceae). Conservation
Biology 9: 596-604.
----------, Bart R. Johnson, and J. L. Hamrick. 1996. Genetic diversity and
populations size in four rare Southern Appalachian plant species.
Conservation Biology 10: 796-805.
Gottlieb, L. D. 1973. Genetic differentiation, sympatric speciation, and the origin of
a diploid species of Stephanomeria. American Journal of Botany 69: 545-553.

---------- 1977. Electrophoretic evidence and plant systematics. Annual of
Missouri Botanical Gardens 64: 161-180.
----------^ S. I. Warwick, and V. S. Ford. 1985. Morphological and electrophoretic
divergence between Lavia discoidea and L. glandulosa. Systematic Botany 10:
Hamrick, J. L. and M. J. W. Godt. 1990. Allozyme diversity in plant species. Pp. 43-
63 in Plant Population Genetics, Breeding, and Genetic Resources, eds. H. D.
Brown, Michael T. Clegg, Alex L. Kahlerr, and Bruce S. Wier. Sunderland,
Massachusetts: Sinauer Associates Inc.
----------, M. J. W. Godt, D. A. Murawski, and M. D. Loveless. 1991. Correlations
between species traits and allozyme diversity: implications for conservation
biology. Pp. 75-86 in Genetics and Conservation of Rare Plants, eds. Donald
A. Falk and Kent E. Holsinger. Oxford: Oxford University Press.
Holsinger, K. E. and L. D. Gottlieb. 1989. The conservation of rare and endangered
plants. Trends in Ecological Evolution 4: 193-194.

---------- 1991. Conservation of rare and endangered plants: principles and
prospects. Pp. 195-208 in Genetics and Conservation of Rare Plants, eds.
Donald A. Falk and Kent E. Holsinger. Oxford: Oxford University Press.
Huenneke, Laura F. 1991. Ecological implications of genetic variation in plant
populations. Pp. 31-44 in Genetics and Conservation of Rare Plants, eds.
Donald A. Falk and Kent E. Holsinger. Oxford: Oxford University Press.
Hulten, E. 1968. Flora of Alaska and Neighboring Territories. A Manual of the
Vascular Plants. Stanford, Ca: Stanford University Press.
International Union of Biochemistry. 1984. Enzyme Nomenclature 1984. New York:
Academic Press.
Johnson, A. W. and J. G. Packer. 1968. Chromosome numbers in the flora of
Ogotoruk Creek. N.W. Alaska. Bot. Not 121:403-456.
Karron, Jeffrey D. 1987. A comparison of levels of genetic polymorphism and self-
compatibility in geographically restricted and widespread plant congeners.
Evolutionary Ecology 1: 47-58.

--------- 1991. Patterns of genetic variation and breeding systems in rare plant
species. Pp. 87-98 in Genetics and Conservation of Rare Plants, eds. Donald
A. Falk and Kent E. Holsinger. Oxford: Oxford University Press.
---------, Yan B. Linhart, Christine A. Chaulk, and Cliff A. Robertson. 1988.
Genetic structure of populations of geographically restricted and widespread
species of Astragalus (Fabaceae). American Journal of Botany 75: 1114-
Knaben, G. 1968. Chromosome numbers of flowering plants from central Alaska.
Nytt Mag. Bot. 15: 240-254.
Kruckeberg, A. R. and D. Rabinowitz. 1985. Biological aspects of endemism in
higher plants. Annual Review of Ecological Systematics 16: 447-479.
Ledig, F. T. and M. T. Conkle. 1983. Gene diversity and genetic structure in a
narrow endemic, Torrey Pine (Pinus torrevana Parry Ex. Carr.). Evolution 37:
Les, D. H., G. J. Anderson, and M. A. Cleland. 1995. Sterility in the North
American lake cress Neobeckia aquatica (Brassicaceae). Rhodora97: 185-

Lesica, P., Robb F. Leary, Fred W. Allendorf, and David E. Bilderback. 1988. Lack
of genic diversity within and among populations of an endangered plant,
Howellia aquatilis. Conservation Biology 2: 275-282.
Lewis, Paul O. and Daniel J. Crawford. 1995. Pleistocene refugium endemics
exhibit greater allozymic diversity than widespread congeners in the genus
Polygonella (Polygonaceae). American Journal of Botany 82: 141-149.
Lewis, Paul O. 1989. Genestat for microcomputers. American Society of Plant
Taxonomists Newsletter 2: 15-16
Linhart, Yan B. and Andrea C. Premoli. 1993. Genetic variation in Aletes acaulis
and its relative, the narrow endemic A. humilis (Apiaceae). American Journal
of Botany 80: 598-605.
Loveless, M. D. and J. L. Hamrick. 1988. Genetic organization and evolutionary
history in two North American species of Cirsium. Evolution 42: 254-265.
Lowrey, T. K. and D. J. Crawford. 1985. Allozyme divergence and evolution in
Tetramolopium (Compositae: Astereae) on the Hawaiian Islands. Systematic
Botany 10: 64-72.

Miller, Constance I. and William J. Libby. 1991. Strategies for conserving clinal,
ecoptypic, and disjunct population diversity in widespread species. Pp. 149-
170 in Genetics and Conservation of Rare Plants, eds. Donald A. Flak and
Kent E. Holsinger. Oxford: Oxford University Press.
Naumann, T. S. 1988. Status report for Eutrema penlandii. Colorado Natural Areas
Program. Unpublished manuscript, available from Colorado Natural Areas
Program, 1313 Sherman St., Denver, Co 80203.
Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a
small number of individuals. Genetics 89: 583-590.
Nielsen, G. and H. B. Johansen. 1986. Proposal for the identification of barley
varieties based on the genotypes for 2 hordien and 39 isoenzyme loci of 47
reference varieties. Euphytica35: 717-728.
Packer, J. G. and G. D. McPherson. 1974. Chromosome numbers in some vascular
plants from northern Alaska. Canadian Journal of Botany 52: 1095-1099.
Pleasants, John M. and Jonathan F. Wendel. 1989. Genetic diversity in a clonal
narrow endemic, Erythronium propullans. and its widespread progenitor,
Erythronium albidum. American Journal of Botany 76: 1136-1151,

Prentice, H. C. 1984. Enzyme polymorphism, morphometric variation and
population structure in a restricted endemic, Silene diclinis (Caryophyllaceae).
Biological Journal of the Linnaen Society 22: 125-144.
Primack, Richard B. 1993. Essentials of Conservation Biology. Sunderland,
Massachusetts: Sinauer Associates, Inc.
Purdy, Brett G. and Randall J. Bayer. 1995. Genetic diversity in the tetraploid sand
dune endemic Deschampsia mackenzieana and its widespread diploid
progenitor D. cespitosa (Poaceae). American Journal of Botany 82: 121-130.
Raven, P. H. 1987. The scope of the plant conservation problem world-wide. Pp.
19-29 in Botanic Gardens and the World Conservation Strategy, eds. D.
Bramwell, O. Hamann, V. Heywood, and H. Synge. London: Academic Press.
Richter, Tracy S., Pamela S. Soltis, and Douglas E. Soltis. 1994. Genetic variation
within and among populations of the narrow endemic, Delphinium viridescens
(Ranunculaceae). American Journal of Botany 81: 1070-1076.
Rollins, Reed C. 1950. Studies on some North American Cruciferae. Pp. 42-53 in
Contributions From the Gray Herbarium of Harvard University. Cambridge,
Massachusetts: The Gray Herbarium of Harvard University.

. 1966. Chromosome numbers of Cruciferae. Contributions of the Gray
Herbarium 198: 3-8.
. 1982. A new species of the Asiatic genus Stroganowia (Cruciferae) from
North America and its biogeographic implications. Systematic Botany 7: 214-
--------- 1993. The Cruciferae of Continental North America. Systematics of the
Mustard Family from the Arctic to Panama. Stanford, Ca: Stanford
University Press.
Roy, Graham, Sylvia Kelso, and Alex Tonnesen. 1993. Habitat characteristics of
Eutrema penlandii (Brassicaceae) in the Colorado Rockies: a study of alpine
endemism. Madrono 40: 236-245.
Schemske, Douglas W., Brian C. Husband, Mary H. Ruckelshaus, Carol Goodwillie,
Ingrid M. Parker, and John G. Bishop. 1994. Evaluating approaches to the
conservation of rare and endangered plants. Ecology 75: 584-606.
Schwaegerle, Kent E. and Barbara A. Schall. 1979. Genetic variability and founder
effect in the pitcher plant Sarracenia purpurea L. Evolution 33: 1210-1218.

Schwartz, O. A. 1985. Lack of protein polymorphism in the endemic relict
Chrvsosplenium iowensis (Saxifraga). Canadian Journal of Botany 63: 2031-
Schwendinger, R. B., G. K. Carlson, and C. O. Spielman Jr. 1991. Status report on
Eutrema penlandii Rollins as a result of field investigations in Park, Summit,
Gunnison, Chaffee, and Clear Creek Counties, Colorado in July and August,
1991. Unpublished report submitted to the U.S. Fish and Wildlife Service: 1-
Smith, James F. and T. V. Pham. 1996. Genetic diversity of the narrow endemic
Allium aaseae (Alliaceae). American Journal of Botany 83: 717-726.
Soltis, Douglas E. 1982. Allozymic variability in SulUvantia (Saxifragaceae).
Systematic Botany 7: 26-34.
----------, C. H. Haufler, D. C. Darrow, and G. J. Gastony. 1983. Starch gel
electrophoresis of ferns: a compilation of grinding buffers, gel and electrode
buffers, and staining schedules. American Fern Journal 73: 9-27.

Swofford, D. L. and R. B. Selander. 1981. Biosys-1: A Computer Program for the
Analysis of Allelic Variation in Genetics. User's Manual. Urbana, Illinois:
University of Illinois.
U. S. Fish and Wildlife Service, Department of Interior. 1993. Endangered and
threatened wildlife and plants; the plant Eutrema penlandii (Penland Alpine
Fen Mustard) determined to be a threatened species. Federal Register (CFR)
58(143): 40539-40547.
Waller, D. M., D. M. O'Malley, and S. C. Gawler. 1987. Genetic variation in the
extreme endemic Pedicularis furbishiae (Scrophulariaceae). Conservation
Biology 1: 335-340.
Weber, W. A. 1965. Plant geography in the Southern Rocky Mountains. Pp. 453-
468 in The Quaternary of the United States, eds. H. E. Wright Jr. and David
G. Frey. Princeton, New Jersey: Princeton University Press.
--------- 1985. New names and combinations, principally in the Rocky Mountain
Flora. Phytologia 58: 382-384.
--------- 1990. Colorado Flora Eastern Slope. Niwot, Colorado: University Press
of Colorado.

Wyatt, Robert. 1988. Phylogenetic aspects of the evolution of self-pollination. Pp.
109-131 in Plant Evolutionary Biology, eds. Leslie D. Gottlieb and Subodh
K. Jain. New York: Chapman and Hall.
Young, A. G. and A. H. D. Brown. 1996. Comparative population genetic structure
of the rare woodland shrub Daviesia suaveolens and its common congener D.
mimosoides. Conservation Biology 10: 1220-1228.