Biologicel C o ~ ~ ~ t'FiO a(1994)
n 25-31
0 1994 Etxvier E
d- L i m i
RimedinGrstBrilainAOdghurssprved
w3m7/94M7.00
GENETIC DIVERSITY ASSESSMENT IN A METAPOPULATION
OF THE BUTTERFLY Euphydryas gillettii
Diane M. Debinski*
Biology Deparhenl, Montana State University, Bozeman, Montana 59717, USA
(Received I2 March 1993;revised version received 9 November 1993;accepted 1 December 1993)
Abstract
Euphydryas gillettii is a Rocky Mountain endemic butterfy distributed in small, widely separated populations,
which appear to be declining. This paper presents the results of a genetic diversity assessment within and among
E. gdlettii colonies comprising the Glacier National Park
(GNP) metapopulation. Four colonies were assayed
within GNP during 1988 and 1989, and two outgroups,
colonies from Idaho and Wyoming, were also compared
with GNP populations for allelic uniqueness and heterogeneity. D%feences between the 1988 and 1989 data suggest that E. gillettii may exist as a biennial population.
Minor dfferences were found among GNP populatwnr
while larger dzrerences were found among M o n t m ,
Wyoming and Idaho populationr. Heterozygosity was
higher than expected, suggesting a potential selective
advantage for heterozygotes. Heterozygosity was also
found to be higher in populations that were distributed in
a linear space along stream corridors relative to those in
meadows. More research should be conducted with respect to dispersal along streams as it appears that these
corridors may be signifcant in the maintenance of the
me tapopulation.
or endangered, yet its population numbers are such
that there may be cause for concern in the future. The
goal of this project was to assess genetic diversity
within and among E. gillettii colonies composing the
Glacier National Park (GNP) metapopulation and to
determine the conservation implications of this analysis. Historically, five colonies were known to exist in
the GNP ecosystem (S. Kohler, pen. cornrn.).
A species whose range is composed of more or less
geographically isolated patches, interconnected through
patterns of gene flow, extinction, and recolonization, is
said to form a 'metapopulation', or 'population of populations' (Lande & Barrowclough, 1987). Each population is spatially and often times genetically distinct
from the others, but each is still a component of the
larger metapopulation, a term coined by Levins (1970).
Levins' (1970) mathematical model showed that metapopulation viability is dependent upon maintaining a
critical balance between patch colonization and extinction. Patch extinction is not necessarily detrimental to
the metapopulation unless there is a cumulative effect
of demographic instability andlor loss of genetic variation (Gipin & Soule, 1986).
The ecology and population genetics of several
species of checkerspot butterflies have been under intensive study in the western United States for the past
auarter centurv. In fact. one of the best documented
natural extinctions of butterfly populations is that of
Euphydryas editha bayensis (Ehrlich, 1983); thus ample
comparative data exist. Many of these data already
have made important contributions to the theory and
practice of conservation biology (e.g., Ehrlich & Murphy, 1987). Ehrlich (1983) found that when Euphydryar
populations fall below 3C-50 individuals, they appear
to go extinct promptly or to rebound to sues an order
of magnitude or more larger, and polymorphic loci
tend to maintain the same predominant alleles in a p
proximately the same frequencies through dramatic
changes in population sue.
Keywords: Glacier National Park, Euphydryas gillettii,
genetics, metapopulation.
Euphydryar gillettii, a northern Rocky Mountain
endemic checkerspot buttedy, is distributed in small,
widely separated patches, with apparently little migration between them (Holdren & Ehrlich, 1981). Further,
its populations appear to be declining. Williams (1988)
visited 29 localities in the Rocky Mountains where E.
gillettii had been collected in the past and found extant
populations in only 13. E. gillettii is also interesting
because it is the only North American member of the
Palaearctic species-group, the maturna. The taxonomic
&ties and geography both suggest that the animal is
an isolated relict (A. Shapiro, pers. wmm.).
E. gillettii thus offers an excellent opportunity to
study the genetics of a spatially fragmented relict popuIation. It is not so rare as to be considered threatened
Life history of Euphydryas giUenii
E. giUeftii is often found in populations of 30 adults or
less. Its range includes western Wyoming, northwestern
Montana, northeastern Idaho, and southwetern Alberta,
and it has a fragmented distribution, in some cam with
d i s t of
~ 3244
~
km between populations. E. gillettiiis
fragmented distribution can be partially explained by its
*Resent addrss: Department of A n i Ecology, Iowa State
University, Ames Iowa 50010, USA.
25
D. M. Debinski
larval foodplant specificity. It lays its eggs solely on the
black twinberry, Lonicera involucrata (Williams, 1988)
family. This perennial shrub must be located in a wet,
sunny area for the butterfly to find it suitable, often restricting E. gillettii to early successional habitats.
Disturbances are an important factor contributing to
E. gillettii habitat abundance. According to Williams
(1988), fire is the single most important factor in creating
its habitat, which may have been significantly reduced by
extensive fire suppression in the Rocky Mountains prior
to 1975. Encroachment by surrounding forests shades
the Lonicera, making them unattractive to females,
which always lay their eggs on sunny leaves. Other important disturbances include floods and the creation of
forest openings by beavers Cmtorfiber and logging.
Williams (1988) notes that some E. gillettii populations may fluctuate greatly in abundance from year to
year. Such fluctuations can have a major impact on the
genetics of the population because the effective population size (calculated as the harmonic mean) may be
smaller than the arithmetic mean population size.
There may also be certain high-elevation populations
that are biennial because the warm season is too short
for the larvae to mature in one summer.
A major factor limiting the recolonization of sites may
be the inability of dispersing individuals to reach a suitable habitat (Holdren & Ehrlich, 1981). Areas that appear perfectly habitable are never colonized by E. gillettii
due to the insect's poor dispersal ability. The purpose of
this study was to determine whether the observed inability to disperse actually translated into low heterozygosity
levels or limited gene flow between populations.
Questions related to genetic
diversity assessment include:
(1) What is the level of genetic diversity among gene
pools within the park relative to the diversity
among regions (Idaho, Wyoming, Montana)?
Given the short dispersal distance of the organism (Holdren & Ehrlich, 1981) one would expect
a high degree of differentiation between regions.
(2) Are the Glacier National Park populations components of a larger metapopulation, or are there
genetically distinct populations within the
Glacier ecosystem?
(3) What management decisions within or around
the park might result in the merging or destruction of gene pools?
METHODS
lar to that developed by May et al. (1979) and is described in Debinski (1991). In 1989, I conducted a brief
mark-recapture experiment to estimate population sizes
at four of the colonies (Debinski, 1991).
RESULTS
Populations
Five populations of E, gillettii (Christensen Meadow,
near Kiowa, Blacktail Hills, Swiftcurrent, and McGee
Meadow) were known to exist in the GNP ecosystem (S.
Kohler, pers. cornm.). During the course of this research,
previously known populations were located, and four new
colonies were identified at Red Meadow Creek Road,
Spot Mt, Elizabeth Lake, and Belly River. No E, gillettii
were observed at McGee Meadow during 1987-1989, and
only one eggmass was found at Swiftcurrent Lake in 1990
(surveyed 1989 and 1990). The Blacktail Hills site was not
surveyed due to gnzzly bear activity; thus, E. gillettii's
current status there is unknown.
A short mark-recapture study in 1989 revealed a
higher population size than Williams (1988) had estimated. My minimum population estimates were 142,
98, 56, and 28 individuals at Belly River, Marias Pass,
Red Meadow, and Christensen Meadow, respectively.
These data are based on two to four census dates per
site. The low number of recaptures suggests that the
true population size was much larger than that censused, or that the disturbance affected recapture rates.
Genetic analysis
Males of the species were collected from four sites in
GNP during 1988 and 1989 as well as central Idaho
and central Wyoming in 1989. The sample sizes were as
follows: Christensen Meadow (17 in 1988), Belly River
(23 in 1988, 9 in 1989), Red Meadow (25 in 1989),
Marias Pass (21 in 1989), W Central Idaho (20 in
1989), W Central Wyoming (22 in 1989).
1988 results
A total of 19 loci from two populations in GNP were
surveyed electrophoretically (AAT-1,2; AGP, DIA- 1,2;
GAPDH, GP, GPI, HBDH, IDH-1,2; LDH, MDH-1,2;
PEP, PGD, SOD-1,2; and PGM; Table 1) and the
computer program BIOSYS (Swofford & Selander,
Table 1. Polymorphic loci of E. gillettii, Glacier National
Park, Montana, 1988
Locus
Alleles present
-
Four colonies were assayed within GNP and two outgroups, colonies from Idaho and Wyoming, were also
compared with the GNP population for allelic uniqueness and heterozygosity. Because the entire metapopulation could not be surveyed, this subsample was chosen
to reflect the greater metapopulation. Electrophoretic
analysis was used to determine the level of heterozygosity at 24 presumptive enzyme loci (Debinski, 1991). A
minimum of 17 individuals was assayed from each site.
The methodology for starch gel electrophoresis is simi-
Christensen Meadow
IDH-2
PGD
GPI
MDH- 1
PGM
DIA-2
AGP
% polymorphic
-
Belly River
Genetic diversity asses:rment of a butterfly
Table 2. Contingency chi-square analysEoall loci 1988
Locus
No. alleles
2
d.f.
P
AGP- 1
IDH-2
PGD- 1
GPI- 1
MDH- 1
DIA-2
PGM-1
Totals
1981) was used to analyze the data. Seven loci were
found to be polymorphic in the Belly River population
(AGP, IDH-2, PGD, GPI, MDH-1, PGM, and DIA2), while four loci were found polymorphic in the
Christensen meadow population (PGD, MDH- 1, AGP
and DIA-2). All polymorphic loci in the Christensen
Meadow population were also polymorphic in the Belly
River population.
Belly River had a population mean heterozygosity of
0-084 with SE = 0.038 (unbiased estimate-Nei, 1978).
The mean number of alleles per locus was 1.5 (SE =
0.17). Using a 0.95 allele frequency as a minimum for
assuming heterozygosity (95% criterion), the percentage
of the loci polymorphic was 20%.
As compared to Belly River, Christensen Meadow
had a mean heterozygosity of 0.079, SE = 0.037 (unbiased estimate-Nei, 1978). The mean number of alleles
per locus was 1.25 (SE = 0.12), and the percentage of
loci polymorphic was 20% by the 0.95 criterion.
The summarized F statistics over all loci revealed inbreeding coefficients as follows: Fls = -0.125, FIT=
0.079, and FsT = 0.041. AGP, PGD, and MDH-1 exhibited significant differences in frequency between the
Christensen Meadow and Belly River populations in
1988 (Table 2). The total chi-square was 26.505, with a
corresponding probability of 0.00545, highly significant. Nei7s(1978) unbiased genetic identity between the
two populations was 0.997.
1989 Results
A total of 23 loci from Wyoming, Idaho, and Montana
Table 3. Polymorphic loci of E. giUettu-1989
Locus
IDH-2
PGD
GPI
MDH-1
PGM
AGP
MDH-2
MPI
GK-2
% polymorphic
Table 4. Observed heterozygosity at variable loci-1989
Alleles present
Wyoming Idaho
Red
Marias
n = 22 n = 20 Meadow Pass
n=25 n=21
were analyzed. The Montana population, however,
consisted of three separate subpopulations, Marias
Pass (n = 21), Red Meadow (n = 25), and Belly River
(n = 9), which were analyzed separately. The Idaho
and Wyoming analyses were conducted using samples
of 20 and 22 individuals respectively. Six of the 23
(26%) assayed loci were polymorphic in the Montana
and Idaho populations, while Wyoming had five of 23
(22%) loci polymorphic. Allelic composition is shown
in Table 3.
The diversity of alleles at each locus was similar
across the three populations. In several cases, a population could be distinguished by the presence or absence
of certain alleles. For example, Wyoming had the A
and B alleles for PGD, while Idaho had the B and C
alleles. Montana populations only had the B allele.
Similarly, Wyoming and Idaho had the C allele for
MDH-2 and GK-2, but Montana's Red Meadow population had B and C. At the MPI locus the Idaho and
Montana populations were similar, exhibiting the B
and C alleles, while the Wyoming had both C and D
alleles. PGM was polymorphic only in Idaho. At GPI,
Wyoming and Idaho were most similar, exhibiting the
B and C alleles, while Montana exhibited the C and D
alleles. The same alleles were present in all three populations at the AGP and MDH-I loci.
Genetic analyses revealed that the populations with
higher allelic diversity exhibited higher polymorphism.
Populations in Idaho, Wyoming, and Red Meadow
had six, five, and six polymorphic loci respectively,
compared to Marias Pass and Belly River's two. The
level of heterozygosity was near 0.05 in the Idaho,
Wyoming, and Red Meadow populations, while it was
0.030 and 0.035 for the Belly River and Marias Pass
populations (Table 4). Belly River's low polymorphism
estimate may not be as reliable as the others, however,
because of small sample size (n = 9).
The summarized F statistics over all loci revealed inbreeding coefficients as follows: FIs = 0-011, F, = 0.333,
and F, = 0.325. Three chi-square analyses were conducted, one which included all populations in 1989, a
second for the GNP 1989 populations only, and a third
that compared Belly River populations in 1988 and 1989.
Belly
River
n=9
Locus
(4
Wyoming
(22)
Idaho
(20)
Belly
River
(9)
Marias
Red
Pass Meadow
(20)
(25)
PGD
GPI
MDH- 1
PGM
AGP
MDH-2
MPI
GK-2
0.044
0.463
0.449
0400
0.172
0~000
0.139
0.000
0.095
0.049
0.180
0.145
0.180
0.000
0.480
0~000
n.a!
0.153
0.000 0~000
0~000 0.000
0.000 0.000
0.490
0.498
0,000 0.000
0.198
0.133
0.000 0.000
0.000
0.113
0.039
0.000
0.500
0.077
0.365
0.127
Mean H
SE
0.056
0.029
0.050
0.024
0.030
0.023
0.053
0.026
" Not applicable.
0.035
0.024
D. M. Debinski
Table 5. Contingency chi-square a n a l y L a l l loci, all populations 1989
Locus
No. alleles
,$
d.f.
4
3
204.450
90.588
11.717
4.432
61.085
36.401
5.660
9.977
424.307
12
8
4
8
8
4
AGP- 1
MPI- 1
PGM-I
PGD- I
GPI-1
MDH-1
MDH-2
GK-2
Totals
2
3
3
2
2
2
4
4
52
P
0~00000
0.00000
0.01958
0.81621
0~00000
0~00000
0.22606
0@4081
0~00000
Table 5 shows significant differences among all 1989 populations at the following loci: AGP, MPI, PGM, GPI,
MDH-1, and GK-2. The GNP populations are distinguished by significant variation at the MPI locus only
(Table 6). Belly River populations differ significantly
between years at loci AGP and IDH-2 (Table 7).
Similarity and distance estimates were calculated using
Rogers (1972) equations. Rogers' distance considers two
populations with fixation for different alleles farther
apart than ones where one or both are heteroallelic even
though they have no common allele (Wright, 1978).
Wyoming and Idaho populations cluster as the outliers
and the Montana populations grouped together, as one
would expect on the basis of geography (Fig. 1).
DISCUSSION
Colony numbers and distribution
The results of the brief mark-recapture project showed
higher population sizes than those reported by
Williams (1988). More rigorous censusing and a search
for new colony sites may prove the organism to be
more common than historically noted. Since many of
the historic colony sites are within easy access of roads,
there undoubtedly could be many more colonies in
remote areas that have not yet been discovered.
Genetic diversity
The electrophoretic analysis revealed that E. gillettii in
GNP, Montana, may consist of two separate biennial
populations. This hypothesis is supported by the major
genetic differences between the Belly River 1988 and
1989 analysis. Chi-square analysis showed significant
differences in allele frequencies at the AGP and IDH-2
Table 7. Contingency chi-square analysis-Belly River 1988
and 1989
Locus
,$
d.f.
P
75.539
10
0.00000
No. alleles
AGP- 1
IDH-2
PGM-1
PGD-1
GPI-1
MDH-I
Totals
loci between the two broods. The 1989 population had
only two polymorphic loci, while the 1988 population
had seven. Further, the C allele did not appear in the
1989 population at the AGP locus. This apparent difference may be a result of small sample sizes (n = 23 in
1988 while n = 9 in 1989); however, the results are so
strikingly different that they probably result from more
than simple sampling error. Calculations using 1988 allele frequencies to estimate those in 1989 revealed
major differences between observed and expected results. Several alleles present in 1988 were not observed
in 1989. The possibility of E. gillettics biennial nature
is suggested in Scott (1986).
A biennial life history could actually offer the species
a longer expected persistence time, as the two populations may experience the environment in different ways.
For example, during an extreme drought year, the population that is flying could become extinct. Presumably,
the hibernating population would also be affected, but
perhaps not as significantly. If the hibernating population is protected from a 'bad year', the species could
persist as one population that flies every other year.
This has been observed in Colorado populations of
Boloria acronema (Brussard & Britten, 1989).
Alternative explanations for the differences in gene
frequencies between years include multi-year diapausers
or the use of the Belly River habitat as a dispersal
corridor. Both Pieris napi pupae and larvae of other
Euphydryas, e.g. editha, are known to have multi-year
diapausers (A. Shapiro, pers. comm.). It is possible that
gillettii resembles many tundra species in having a flexible life history; rather than being strictly annual, bien.93
.95
.97
.99
Idaho
Table 6. Contingency chi-square analysis--all loci, GNP 1989
Locus
AGP- 1
MPI-1
PDG- 1
GPI-1
MDH-1
MDH-2
GK-2
Totals
No. alleles
,$
d.f.
Belly River, Montana
P
Marias Pass,Montana
Red Meadow, Montana
2
2
2
Wyoming
2
2
.95
2
Genetic Similarity
2
.97
.99
Fig. 1. cluster analysis of Euphydryas gillettii populations
using Rogers (1972) clustering algorithm for genetic analysis.
Genetic diversity assessment of a butterfly
nial, etc., it may be partly annual, biennial, triennial,
etc. Finally, if Belly River is a dispersal corridor, yearto-year differences could reflect variations in sources
rather than the coexistence of two populations.
Because of differences in allele frequency, the 1988
and 1989 results could not be combined. In 1988, Belly
River had almost twice the level of polymorphism
(35%) as Christensen Meadow (22%), but similar levels
of mean heterozygosity were found (Belly River =
0.084 and Christensen Meadow = 0.079). AGP, LDH-1,
and GPI significantly differentiated the populations
based upon chi-square results.
The 1989 survey allowed for comparisons among
three populations in GNP, as well as an inter-state comparison. Wyoming, Idaho, and Red Meadow, Montana
had comparable levels of heterozygosity (0.056, 0.050,
and 0.053 respectively). Montana's Marias Pass and
Belly River populations had the lowest heterozygosity
(0.035 and 0.030 respectively). The percentage of polymorphic loci was similarly distributed - Wyoming =
22%, Idaho = 26%, and Red Meadow, Montana = 26%,
while Marias Pass and Belly River, Montana were both
polymorphic at only 9% of the loci surveyed. However,
small sample size hindered thorough analysis of the
Belly River population.
The chi-square analysis of Montana, Wyoming, and
Idaho populations identified significant differences at
six out of eight polymorphic loci. This is not surprising,
given the fact that these populations are separated by
hundreds of kilometres. When GNP 1989 populations
were compared using the chi-square analysis, there
were differences only at the MPI locus, but this difference was highly significant ( p = 0.00000). The C allele
was predominant in Belly River and Marias Pass populations, but the B allele was predominant in the Red
Meadow population.
Within the park, spatial distribution of the population may be correlated with genetic diversity. For
example, the long, narrow populations distributed over
a few kilometers (e.g. Belly River and Red Meadow)
exhibited a higher level of heterozygosity than the
rectangular meadow populations that were less than
1 km on a side (e.g Christensen Meadow). The long,
narrow populations are much larger than the meadow
populations, but they may also be serving as dispersal
corridors between source populations.
F-statistics and genetic similarity
Wright's inbreeding coefficients showed different results
between years, but this may be expected due to the
different scales of analyses. The 1988 FIs of GNP populations was negative (indicating higher outcrossing than
random), and F, exhibited the highest value. In 1989,
the interstate FIS was small but not negative, F, was
highest, and F, was intermediate, but all values were
larger than in 1988.
F-statistics revealed that the two GNP subpopulations (1988 data) were more similar (F, = 0.041) than
the 1989 sample of Montana, Wyoming, and Idaho
populations (F, = 0.325). Using Wright's island model
29
(Hart1 & Clark, 1989), the actual number of migrants1
generation between populations is 5.80 for the 1988
GNP populations and 0.52 for the 1989 Montana,
Idaho and Wyoming populations. This implies, as one
would expect, that subpopulations between states are
more isolated, or more structured, than those within
GNP which appear panmictic given that the number of
migrants is greater than one (Crow, 1986).
Comparing populations in different states, Idaho and
Montana are much more similar than either is to
Wyoming. This is also to be expected given the fact
that the Idaho and Montana populations are closer
spatially. The Idaho site is closer to those in Montana
(453 km) than Wyoming is to the Montana sites
(485 km). There is also more continuous montane
habitat connecting the Idaho and Montana sites relative to those in Wyoming and Montana. Moving south
from GNP, there are breaks between Helena, Montana
and the Beartooth Mts of Wyoming, whereas there are
virtually no breaks from GNP to the Idaho site in the
Sawtooth Mountains.
Implications for management
The level of heterozygosity within the park is comparable to the diversity within other regions. Heterozygosity and percent polymorphic loci were the same
or slightly higher in Red Meadow, Montana relative to
the Idaho and Wyoming populations.
The PGD locus singled out Wyoming and Idaho by
characteristic alleles found nowhere else. GPI, MDH-1,
MDH-2, PGM, MPI, and GK-2 also were helpful in
distinguishing among state populations, as some sampled Wyoming and Idaho populations were fixed for
certain alleles at these loci.
Gene pools within GNP were somewhat different but
may be merged without adverse consequences. It appears that the genetic variation in some populations is
a subset of the variation expressed in the others. In
1988, Christensen Meadow and Belly River were
significantly different at AGP, MDH-1, and PGD
based upon chi-square results. In 1989, Belly River and
Marias Pass ranked high in similarity, but both were
different in allelic composition relative to the Red
Meadow population at GPI, MDH.1, MDH-2, MPI,
and GK-2.
Park management decisions are not likely to result in
the merging of gene pools, as geographical distance is
probably the largest separating factor given the low
vagility of this butterfly. However, management decisions could lead to the destruction of gene pools. The
butterfly's host-plant specificity, combined with the disturbance-loving nature of the host plant, suggests that
maintenance of successional-stage habitats is imperative
to the survival of E. gillettii.
Management policies that prevent natural changes
could deleteriously affect populations of this species. For
example, the fue suppression policies of the last few
decades may have been detrimental, as new openings
were not created. Moose Alces alces grazing on Lonicera
involucrata may also have an adverse effect on E. gillettii
D. M. Debinski
populations (E. H. Williams, pers. comm.). As successional changes occur, the species will face a continuous
loss of habitat. Corridors are of major importance in the
maintenance of the E. gillettii populations. E. gillettii
moves primarily up and down streams connecting
patches of good habitat; it does not use dry areas (E. H.
Williams, pers. comm.). As such, maintenance of the
larger metapopulation will be significantly influenced by
changes in the character of the stream corridors.
Given the results of the genetic analyses, it is unclear
whether the park is large enough to support more than
one viable gene pool over the long term. There are
minor differences in genetic composition of populations
within the park. However, dispersal distance is the critical factor in determining the true level of isolation between populations. Heterozygosity levels are higher than
one would expect for populations of 30 individuals. This
may indicate either a selective pressure favoring heterozygotes or dispersal among populations. Some of the
smaller populations may be sink populations relative to
the larger source populations (e.g. Pulliam, 1988). However, actual population sizes need to be estimated more
rigorously, and dispersal distances need to be determined
before this question can be answered.
Given the potential rarity of E, gillettii compounded
with its hypothesized biennial nature (Scott, 1986), continued population monitoring is critical. Genetic analysis is the only realistic way to distinguish between an
annual and biennial population structure, but responsible collecting is critical given E. gillettii's small population sizes. E. gillettii is known to exhibit major
fluctuations in population size, and its persistence in
low numbers is not easily explained. A biennial population structure could explain the observed allelic differences, population size fluctuations, and its continued
persistence. If this hypothesis is correct, the even-year
populations should be much more similar to each other
genetically than they are to the odd years.
The most interesting aspect of this research relates to
the species distribution as a whole. Given the apparently low dispersal ability of E. gillettii, the Montana,
Idaho, and Wyoming populations are actually more
similar than one would have expected. An historical
analysis of landscape patterns including fire frequencies, corridors, and habitat patch distribution would be
of interest in understanding the genetic and demographic structure of this population.
CONCLUSIONS
This research corroborates most of Williams' (1988)
ideas, who has studied E. gillettii most extensively.
Electrophoretic analyses revealed moderate levels of
heterogeneity both within and between the colonies
studied, and preliminary results suggest a potential
biennial nature of the organism. However, the level of
heterozygosity was higher than expected for population
sizes of 100 or more. Heterozygotes may have a selective advantage in these populations.
Using the Wyoming and Idaho populations for com-
parison, the GNP, Montana population has an equivalent genetic diversity relative to other areas. Montana
populations were more similar to Idaho than
Wyoming, which makes sense geographically. Highest
levels of polymorphism and mean heterozygosity were
both found in 1988 Belly River, Montana, a population
distributed along a riparian corridor. Two Montana
populations (Belly River and Marias Pass) were genetically indistinguishable, while Red Meadow, Montana
was slightly different. Dispersal remains the major unknown factor in determining genetic structure of E.
gillettii populations.
ACKNOWLEDGEMENTS
This research was supported by the National Park
Service at Glacier National Park, Montana. I would
like to thank Peter Brussard, Art Shapiro, and Richard
Gomulkiewicz for their comments on the manuscript.
REFERENCES
Brussard, P. F. & Britten, H. B. (1989). The status of the
Uncompahgre fritillary Boloria acronema: final report. U.S.
Forest Service (Cebolla District) Technical Report.
Crow, J . F . (1986). Basic concepts in population, quantitative,
and evolutionary genetics. W . H. Freeman & Co., New
York.
Debinski, D. M. (1991). Inventory and monitoring of biodiversity: an assessment of methods and a case study of
Glacier National Park, MT. PhD dissertation, Montana
State University, Bozeman, MT.
Ehrlich, P. R. (1983). Genetics and extinction of butterfly
populations. In Genetics and conservation--a reference for
managing wild animal populations, ed. C. M. SchonewaldCox, S. M. Chambers, B. MacBryde & L. Thomas.
Benjamin Cummings Publishing Co., Menlo Park, CA,
pp. 152-63.
Ehrlich, P. R. & Murphy, D. D. (1987). Conservation lessons
from long-term studies of checkerspot butterflies. Conserv.
Biol., 1, 122-31.
Gilpin, M. E. & Soult, M. E. (1986). Minimum viable populations: the processes of species extinction. In Conservation
biology: the science of scarcity and diversity, ed. M. E.
SoulC. Sinauer Associates, Sunderland, Massachusetts, pp.
13-34.
Hartl, D. L. & Clark, A. G. (1989). Principles of population
genetics. Sinauer Associates, Sunderland, Massachusetts.
Holdren, C. E. & Ehrlich, P. R. (1981). Long range dispersal
in checkerspot butterflies: transplant experiments with E.
gillettii. Oecologia, Berl., 50, 125-9.
Lande, R. & Barrowclough, G. R. (1987). Effective population size and genetic variation. In Viable populations for
conservation, ed. M. E. SoulC. Cambridge University Press,
Cambridge, pp. 87-123.
Levins, R. (1970). Extinction. In Some mathematical questions
in biology, ed. M. Gerstenhaber. American Mathematical
Society, Providence, Rhode Island, pp. 75-108.
May, B., Wright, J. E. & Stoneking, M. (1979). Joint segregation of biochemical loci in Salmonidae: results from experiments with Salvinus and a review of the literature of other
species. J. Fish. Res. Bd. Can., 36, 111428.
Nei, M. (1978). Estimation of average heterozygosity and
genetic distance from a small number of individuals.
Genetics, 89, 583-90.
Pulliam, R. (1988). Sources, sinks and population regulation.
Amer. Nut., 132, 652-61.
Rogers, J. S. (1972). Measurements of genetic similarity and
Genetic diversity assessment of a butterfly
genetic distance. Studies in Genetics. Univ. Texas Publs,
No. 7213, 14553.
Scott, James A. (1986). The butterflies of North America-+
natural hbtory and Jield guide. Stanford University Press,
Stanford, CA.
Swofford, D. L. & Selander, R. B. (1981). BIOSYS-1: A
fortran program for the comprehensive analysis of
electrophoretic data in population genetics and evolution.
J. Hered., 72, 281-3.
Williams, E. H. (1988). Habitat and range of E. gillettii
(Nymphalidae). J. Lepid. Soc., 42, 3745.
Wright, S. (1978). Evolution and the genetics of populations,
Vol. 4. Variability within and among natural populations.
University of Chicago Press, Chicago, Illinois.