|
|||||||
|
|
|
|||||
|
|
|||||||
Evolution, August 1993 v47 n4 p1094(11)
The effects of population size and
plant density on outcrossing rates in locally endangered Salvia pratensis. R.
van Treuren; R. Bijlsma; N.J. Ouborg; W. van Delden.
Abstract: Multilocus outcrossing rates were studied in natural and
experimental populations of Salvia pratensis
and related to population size, density and spatial distribution. Multilocus
estimation procedures were used to determine outcrossing rates. Hermaphroditic
plants in natural populations had outcrossing rates ranging from 38.2%-81.8%
and those in experimental groups had a 71.5%-95.5% rate. However, no
correlations were noted between outcrossing rate and population size.
Full Text: COPYRIGHT 1993 Allen Press
In population genetic studies, knowledge of the mating system is of major
importance, because patterns of mating influence the amount and distribution of
genetic variability and hence evolutionary processes in populations (Hedrick
1990). Populations may diverge genetically when gene flow is restricted (Schaal
1975; Schoen 1982a; Knight and Waller 1987; Holtsford and Ellstrand 1989). In
general, populations of out-breeding plants show higher levels of genetic
diversity and less differentiation among each other than populations of
inbreeding plants (Allard et al. 1968; Brown 1979; Hamrick et al. 1979; Clegg
1980; Loveless and Hamrick 1984; Van Delden 1985; Van Dijk et al. 1988).
Outbreeding populations also are expected to show a higher frequency of recessive
deleterious alleles, hidden in heterozygous individuals, and carry a higher
genetic load. Therefore, fitness in outbreeding populations should decrease
more with selfing than fitness in habitually inbreeding populations that have
purged their genetic load during many generations of inbreeding (Charlesworth
and Charlesworth 1987; Barrett and Charlesworth 1991). Darwin (quoted in
Richards 1986) already showed that normally outbreeding species are more
susceptible to inbreeding depression than regularly inbreeding species. Lande
and Schemske (1985) modeled the interaction of inbreeding depression and the
degree of self-fertilization and predicted a bimodal distribution of
outcrossing rates, resulting in primarily outcrossing and primarily
self-fertilizing species. They argued that inbreeding depression decreases the
genetic load and subsequently decreases the inbreeding depression in later
generations. If the load caused by inbreeding depression is decreased below 0.5
(because the contribution of self-fertilization to the next generation is two
times the contribution of cross-fertilization), the model predicts the
evolution of selfing, otherwise the evolution of outcrossing is favored. A
survey of the outcrossing rates of 55 plant species fits this bimodal distribution
(Schemske and Lande 1985). However, Aide (1986) demonstrated that this bimodal
distribution applies primarily to wind-pollinated species and that
animal-pollinated species show a random distribution of outcrossing rates with
high interpopulational variation.
Factors influencing the mating system can be roughly divided into intrinsic
characteristics of the plant and characteristics of its environment. Dioecy,
androdioecy, gynodioecy (Van Damme 1983; Wolff et al. 1988), protogyny (Bos et
al. 1985), or protandry (Schoen 1982b; Kesseli and Jain 1985) and the presence
of self-incompatibility systems (Ennos 1985) are obvious mechanisms of the
plant that promote outcrossing. Additionally, flower color (Ennos and Clegg
1983; Epperson and Clegg 1987a) and morphology (Marshall and Abbott 1982;
Epperson and Clegg 1987b) can influence outcrossing rates. Characteristics of
the environment that influence patterns of pollen flow and hence outcrossing
rates, include the foraging behavior of the pollinators (Levin and Kerster
1969; Campbell 1985; Devlin and Stephenson 1985) and the density of flowering
plants (Schmitt 1983; Farris and Milton 1984; Vaquero el al. 1989; Murawski et
al. 1990; Watkins and Levin 1990; Murawski and Hamrick 1991).
Studies of the mating system are of particular interest to conservation
biology (Karron 1991). Rare species often occur in small and isolated
populations in which they face an increased probability of extinction through
random demographic and environmental forces (Goodman 1987; Shaffer 1987).
Additionally, increased levels of inbreeding and random genetic drift may be
promoted in small populations (Franklin 1980; Frankel and Soule 1981).
Inbreeding depression and the loss of alleles may further reduce the fitness of
individuals and hence increase the probability of extinction for the population
(Frankel and Soule 1981; Schonewald-Cox et al. 1983). The rate at which
populations become inbred depends on the effective population size, increasing
when effective population size becomes smaller (Barrett and Kohn 1991). Because
individuals rarely contribute gametes equally to the next generation, the
effective population size is rarely equivalent to the number of individuals in
a population. Mating patterns are important factors that influence the
effective size of populations, decreasing when plants experience higher levels
of self-fertilization (Barrett and Kohn 1991). Consequently, the rate of
outcrossing influences the rate of loss of genetic variation.
To assess the importance of genetic factors for population extinction, we
have started a research project to investigate these factors in relation to
demographic parameters in two plant species, Salvia
pratensis and Scabiosa columbaria. The number of populations of both species
has decreased by more than 50% in the
Because the rate of loss of genetic variation is influenced by mating
patterns, our next objective was to examine the outcrossing rate in both
species. In this paper, we present data on the rate of outcrossing in relation
to population size, density, and spatial distribution in S. pratensis. By
quantifying mating systems in a number of natural and experimental populations,
our study also addresses Aide's (1986) hypothesis of random interpopulational
variation in outcrossing rate.
MATERIALS AND METHODS
Study System
S. pratensis is a gynodioecious, protandrous,
perennial that is pollinated primarily by bumblebees. Flowers are arranged in
whorls on flower stalks. Bagging experiments have shown that pratensis is
dependent on an insect vector for pollination. Experimental pollinations
indicate self-compatibility in hermaphroditic plants.
Estimation of Outcrossing Rates
Outcrossing rates were calculated using the multilocus estimation procedure
of Shaw et al. (1981). This model compares mother-offspring combinations such
that the offspring can be assigned to one of two classes: discernible
outcrosses (progeny that have received an allele not represented in the
maternal genotype and hence must be the result of an outcross) and ambiguous
matings (progeny that carry alleles similar to those of the maternal genotype
and thus may be the result of either self-fertilization or outcrossing). By
calculating the expected proportion of ambiguous outcrosses, the rate of
outcrossing is estimated by the sum of discernible and ambiguous outcrosses
relative to the total sample size. Obviously, the probability of
nonidentification of an outcross, given that an outcross has occurred (denoted
by |Alpha~) has to be determined. The expected |Alpha~ depends on the maternal
genotype, the allele frequencies of the maternal alleles in the pollen pool, and
the number of loci examined. Generally speaking |Alpha~ can be decreased, and
the probability of direct observation of outcrosses increased, by examining
more loci, and by focusing on maternal genotypes homozygous for alleles that
have a low frequency in the population. Because |Alpha~ decreases with the
number of loci examined, multilocus estimates of the outcrossing rate are more
resistant to bias than single-locus estimates.
The estimation procedure of outcrossing rates depends on the correct estimation
of |Alpha~ and consequently on the correct estimation of allele frequencies in
the pollen pool. Assuming that male steriles sample the same pollen pool allele
frequencies TABULAR DATA OMITTED on average as hermaphrodites, the male
steriles can be used to check the correctness of the estimation procedures for
outcrossing rates of hermaphrodites. Because the progeny of male steriles are
obligatory outcrossed, the allele frequencies sampled by male steriles
(provided that they are homozygous) from the pollen pool can be determined
unambiguously. In the present study, differences between these frequencies and
the estimated frequencies in the pollen pool seen by hermaphrodites were tested
for significance using the standardized variable Z, transformed from the
binomial distribution (
Natural Populations
In 1989, we examined four populations, with characteristics given in table
1. Additional information about the populations and electrophoretic procedures
are described in Van Treuren et al. (1991). Allozyme variation for nine loci
was determined for all plants (both flowering and nonflowering) at Neerijnen
and Lexmond. At Bijland-2 and Koekoekswaard, plants were sampled at about 1 -
to 2-m intervals along three linear transects in different parts of the
populations |respectively, distances between the transects (1-3: |is similar
to~ 100 m; 2-3: |is similar to~ 250 m; 1-2: |is similar to~ 300 m), (1-2: |is
similar to~ 7 m; 2-3: |is similar to~ 15 m; 1-3: |is similar to~ 30 m)~. This
enabled us to check for genetic differentiation within populations, using an
analysis of gene diversity (Nei 1987). The sex phenotype of flowering plants
was recorded during the flowering season. In addition to hermaphrodites and
male steriles, six partial male steriles in Bijland-2 and four in Koekoekswaard
were observed. In the small populations, allele frequencies in the pollen pool
were estimated from genotypes of hermaphrodites only (table 1; 22 and 27 hermaphrodites
in Neerijnen and Lexmond, respectively). Because the allozyme genotypes of only
a fraction of the plants present in the large populations were determined, all
these plants were used to estimate allele frequencies in large populations,
regardless of flowering and sex phenotype (table 1: 92 and 129 plants in
Bijland-2 and Koekoekswaard, respectively). In both populations, the allele
frequencies in hermaphrodites did not significantly deviate from those in male
sterile and nonflowering plants (P |is greater than~ 0.30 in all cases), using
the standardized variable Z. Respectively 5, 11, 11, and 19 sampled
hermaphrodites in the four populations were selected for progeny analysis based
on the probability of direct observation of outcrosses. These plants were
homozygous for the N-allele for most of the loci. Because these alleles have
high frequencies in populations the selected plants represent a large fraction
of the population. Furthermore, these maternal plants were not clustered but
were randomly distributed within populations. Progeny of these plants were
grown in the greenhouse and used for electrophoretic determination of the
number of discernible outcrosses. The allele frequencies in the pollen pool and
the genotypes of the selected maternal plants were used to estimate |Alpha~ in
each population. Subsequently, the analyzed progeny of these hermaphrodites and
the estimated |Alpha~ were used to determine the multilocus outcrossing rate
together with the variance according to the procedures of Shaw et al. (1981).
Based on the observed (|H.sub.o~) and the expected (|H.sub.e~) frequency of
heterozygotes, fixation indices (|F.sub.IS~ = 1 - |H.sub.o~/|H.sub.e~) were
estimated for the adult populations (Wright 1965). From the multilocus
outcrossing rate (|t.sub.m~), the expected value of the equilibrium fixation
index was calculated using the equation: |F.sub.eq~ = (1 - |t.sub.m~)/(1 + |t.sub.m~). |F.sub.IS~ and
|F.sub.eq~. are expected to be equal if all
Hardy-Weinberg assumptions are met, and if the mating system is the only factor
determining genotype frequencies (Brown 1979).
Experimental Populations
Compared to natural populations, estimation of outcrossing rates in
experimental populations has the following advantages: (1) Provided that many
suitable genotypes are available, the probability of direct observation of
outcrosses can be increased and the bias in the estimates of outcrossing rates
reduced; (2) The genetic structure of the populations can be designed; (3) A
regular density pattern can be created.
In 1989, four experimental populations with different numbers of individuals
and densities were created in an experimental garden. The number of individuals
was either 27 (small populations: 1 and 3) or 54 (large populations: 2 and 4),
planted at either 20-cm distance (dense populations: 2 and 3) or 100-cm
distance from each other (sparse populations: 1 and 4). The closest populations
were separated by 8 m (populations 3 and 4), whereas the most isolated
populations were separated by 25 m (populations 1 and 3). The first individuals
flowered 29 days after planting and the experiment was terminated at day 86.
Twice weekly up to and including peak flowering and once weekly thereafter,
data were collected about the floral characteristics of each plant. These data
included the sex phenotype (hermaphroditic, male sterile,
or partial male sterile), the number of stalks flowering simultaneously, and
the flowering period of each stalk. To avoid a disproportionate contribution to
the pollen pool and to reduce the probability of larger plants experiencing
more self-pollination, flower stalks were removed if the number of
simultaneously flowering stalks per plant exceeded five. All seeds of each
individual flower stalk were sampled at the end of the experiment.
TABLE 2. Frequencies of the alleles of nine variable loci, computed from the
genotypes of all sampled plants in the four natural populations.
Koekoeks- Locus Allele Neerijnen Lexmond Bijland-2
waard
Me-1 S -- -- 0.098 0.047
|I.sub.1~ -- -- -- 0.059
N 1.000 1.000 0.902 0.894
Pgd-1 S -- -- 0.033 --
N 1.000 1.000 0.957 0.899
F -- -- 0.011 0.101
Pgd-2 S -- 0.059 0.049 0.248
N 0.848 0.941 0.934 0.752
F 0.152 -- 0.016 --
Got-1 |I.sub.1~ 0.065 0.137 0.054 0.050
N 0.935 0.863 0.946 0.950
Got-2 S 0.087 -- -- 0.052
N 0.891 0.971 0.929 0.948
F 0.022 0.029 0.071 --
Pgm-1 N 0.978 1.000 0.967 1.000
F 0.022 -- 0.033 --
Pgm-2 |I.sub.1~ 0.043 -- -- --
N 0.935 1.000 0.989 1.000
F 0.022 -- 0.011 --
Tpi-1 S -- -- 0.065 0.074
|I.sub.1~ -- -- 0.005 0.070
N 1.000 1.000 0.929 0.857
Tpi-2 N 1.000 0.951 0.995 0.981
F -- 0.049 0.005 0.019
The plants used in this experiment originated from six crosses in which the
Mendelian inheritance of the allozymes of seven loci (Pgd-2, Per-ox, Got-1,
Got-2, Pgm-1, Pgm-2, and Tpi-1) was tested (Van Treuren et al. 1991).
Consequently, plant genotypes for the seven loci used as maternal plants for
the experiment were known. Individuals were planted such that the probability
of direct observation of outcrosses was more or less equal in all populations.
The plants derived from the different crosses varied with respect to allelic
variation and were distributed over the populations so that gene flow between
populations could be inferred. As a consequence, the progeny of the six crosses
were not equally represented in the four populations. Allele frequencies in the
pollen pool were estimated in each population on all days of observation.
Hermaphrodites were assumed to contribute equally and the partial male steriles
to contribute half to the pollen pool, irrespective of the number of flower
stalks. The progeny of 17 to 28 hermaphroditic flower stalks per population
were selected based on the probability of direct observation of outcrosses and
the flowering period of the flower stalk (so that the combined sample in each
population covered all stages of the flowering season of the population). These
progenies were grown in the greenhouse and used for electrophoretic analyses.
Variation in the flowering periods of the selected stalks may have led to
sampling of different pollen pool allele frequencies. Therefore, values of
|Alpha~ (based on the maternal genotype and the allele frequencies in the
pollen pool during the flowering period of the stalk analyzed) and the
multilocus outcrossing rate (using the procedures of Shaw et al. 1981) were
estimated on a per stalk basis.
If the estimate of |Alpha~ for a flower stalk is based on the allele
frequencies of only its own population, and flowers on the stalk also received
pollen from the other populations, then outcrossing rates will be
overestimated. Therefore, we devised an estimation procedure for flower stalks
that also takes the allele frequencies of the other populations into account.
First, the allele frequencies in the pollen pool of each population separately
were estimated during the flowering period of the stalk analyzed. Next, we
computed the mean allele frequencies of the four populations, weighted by the
fraction that each population contributed to the total progeny of the stalk
analyzed, to estimate |Alpha~. In all populations with male sterility, flower
stalks of three individuals, located in different parts of the population, were
used to check the correctness of the estimation procedure for |Alpha~, as
described above. These plants differed with respect to their allozyme genotypes
and the flowering period of the stalk analyzed.
Because on many plants more than one flower stalk was sampled, not all
estimates of the outcrossing rate were independent. Therefore, multiple
observations from one individual plant were averaged before the examination of
the effects of population size and density on outcrossing rates in a two-way
ANOVA. Outcrossing rates were normally distributed but sample sizes and
variances were not equal in all populations (Cochran's C = 0.417, P = 0.097).
RESULTS
Natural Populations
Little genetic differentiation was found between the two adjacent clusters
(|is similar to~ 10 m distance) of the population Lexmond (|G.sub.ST~ = 0.028),
between the two subunits of the population Neerijnen (arbitrarily divided in
two subunits of equal size, |G.sub.ST~ = 0.033), and among the three transects
of the population Koekoekswaard (|G.sub.ST~ = 0.026; table 3). Genetic
differentiation within the population Bijland-2 (|G.sub.ST~ = 0.049) was almost
equal to that among the four populations (|G.sub.ST~ = 0.053). Altogether,
|G.sub.ST~-values were relatively low.
In the population Koekoekswaard, the estimated allele frequencies in the
pollen pool did not significantly deviate from those observed in the offspring
of two male steriles, located on different transects (P |is greater than~ 0.05
in all cases). Also the differences between the expected and observed number of
ambiguous and discernible outcrosses in male steriles were not significant (P
|is greater than~ 0.40 in both cases). Estimates of the outcrossing rate ranged
from 0.382 |+ or -~ 0.092 in Bijland-2 to 0.818 |+ or -~ 0.087 in Lexmond,
indicating departure from |t.sub.m~ = 1 in all populations. No relation existed
between population size and the rate of outcrossing. Estimates of the
outcrossing rate were higher in the populations with a relatively higher
density of flowering plants (Lexmond and Koekoekswaard), than those with a
relatively low density of flowering plants (Neerijnen and Bijland-2; table 1).
However, this relation is by no means unequivocal. For example the outcrossing
rate of Bijland-2 (0.075 flowering individuals/|m.sup.2~, |t.sub.m~ = 0.382 |+
or -~ 0.092) was lower than that of Neerijnen (0.01 flowering
individuals/|m.sup.2~, |t.sub.m~ = 0.669 |+ or -~ 0.079).
No significant deviation was found between the observed and the expected
heterozygosity in the adult plants, chi-square tested for each locus separately
in each population (results not shown). Consequently, the mean fixation indices
(|F.sub.IS~) of the populations were close to zero. |F.sub.IS~ values were
considerably smaller than the equilibrium fixation indices (|F.sub.eq~) in all
populations.
Experimental Populations
The number of plants that did not flower during the experiment ranged from
zero in population 3 to three in population 1. However, substantial differences
in the percentage male sterility (range 0-48.1%) and the percentage immigration
(range 1.6-35.9%) were found between the populations). In the populations 2, 3,
and 4, respectively 13 out of 18, 15 out of 17, and 16 out of 17 tests to
assess accuracy of our estimations for |Alpha~ showed no significant
differences between the estimated allele frequencies in the pollen pool and
those observed in the offspring of the male steriles. The expected number of
ambiguous and discernible outcrosses, based on the estimated values of |Alpha~,
did not significantly deviate from those observed when corrections for gene flow
among populations were made (P |is greater than~ 0.10 in all cases). Without
the correction procedure, two out of nine tests showed significant differences.
Therefore, the estimation procedure of |Alpha~, corrected for gene flow among
populations, seems to be sound and was used to determine outcrossing rates of
hermaphroditic flower stalks.
The mean outcrossing rate of hermaphroditic flower stalks, weighted by the
number of offspring analyzed, ranged from 0.715 |+ or -~ 0.058 in population 4
to 0.955 |+ or -~ 0.054 in population 3. As in natural populations, no
relationship existed between population size and rate of outcrossing (small
populations 1 and 3, respectively |t.sub.m~ = 0.882 |+ or -~ 0.057 and 0.955 |+
or -~ 0.054; large populations 2 and 4, respectively |t.sub.m~ = 0.940 |+ or -~
0.026 and 0.715 |+ or -~ 0.058). In the populations with low plant density
(100-cm distance between neighboring plants), |t.sub.m~ values were lower
(respectively 0.882 |+ or -~ 0.057 and 0.715 |+ or -~ 0.058 in populations 1
and 4) than in the populations with high plant density (20-cm distance between
neighboring plants: respectively 0.940 |+ or -~ 0.026 and 0.955 |+ or -~ 0.054
in populations 2 and 3). The effect of plant density on the rate of outcrossing
was statistically significant (ANOVA: |F.sub.1,67~ = 5.57, 0.025 |is less than~
P |is less than~ 0.01), whereas the effect of population size and the
interaction between the two TABULAR DATA OMITTED parameters were not
significant (ANOVA: respectively, |F.sub.1,67~ = 3.76, 0.05 |is less than~ P
|is less than~ 0.1 and |F.sub.1,67~ = 0.47, 0.4 |is less than~ P |is less than~
0.5, N = 71).
The flowering periods of the stalks used to estimate outcrossing rates
varied over the flowering season. However, the number of plants simultaneously
flowering and the plant density in the populations were not constant over the
flowering season. In addition, the differences in percentage of male sterility
affected both the number and density of the pollen producing plants in the populations.
Furthermore, the number of stalks simultaneously flowering per plant ranged
from one to five over the flowering season. Therefore, for each flowering
period of the stalks used to estimate outcrossing rates, we computed the number
of pollen producing plants flowering in the population, the mean distance
between the maternal plant of the stalk analyzed, and the other pollen
producing plants flowering in the population, and the number of stalks
flowering simultaneously on the plant of the stalk analyzed. Outcrossing rates
were negatively correlated with the mean distance separating the plants
analyzed from other pollen producing plants, whereas no correlation existed
between the outcrossing rates and the number of pollen producing plants in the
population. The number of stalks per plant flowering simultaneously was
positively correlated with the mean distance between plants. However, this
could not account for the significant negative correlation between the mean
plant distance and the rate of outcrossing because the partial correlation
coefficient between the rate of outcrossing and the number of flowering stalks
was not significant.
DISCUSSION
The results of this study showed that S. pratensis is predominantly
outcrossing in both natural and experimental populations. Furthermore,
differences in outcrossing rates in hermaphroditic plants were found both
between natural and between experimental populations. Results from the
experimental populations suggested that the density of flowering plants and, particularly,
the density of pollen producing plants contributed to the differences between
the experimental populations.
TABLE 5. Matrix of partial correlation coefficients (Sokal and Rohlf 1981)
between the outcrossing rate of flower stalks, the number of pollen producing
plants in the population, the mean distance separating the plants analyzed
from other pollen producing plants, and the number of stalks flowering
simultaneously on the plants analyzed. The values represent partial
correlation coefficients between two parameters, holding the others constant.
(N = 92; ** P |is less than~ 0.005; NS, not significant).
Number of pollen- Mean plant Number of
producing plants distance flower
stalks
Outcrossing rate -0.024 (NS) -0.435(**) 0.160 (NS)
Number of pollen-
producing plants -0.156 (NS) 0.152 (NS)
Mean plant distance 0.382(**)
Outcrossing rates were estimated using the procedure of Shaw et al. (1981).
The validity of this model rests on four major assumptions. First, it is
assumed that there is no linkage among the loci. Violation of this assumption
seems neglectable because the alleles of the loci of S. pratensis assort
independently (Van Treuren et al. 1991), and deviations from linkage
equilibrium were never found.
Second, it is assumed that allele frequencies in the pollen pool are
distributed uniformly over the population of maternal plants. In this study,
the frequencies of the alleles were estimated irrespective of the number of
flowering stalks per plant. Furthermore, no data were collected about the
number of whorls per flower stalk, the number of flowers per whorl, and the
pollen production per flower. Despite the possibility that variation in these
traits may have contributed to unequal contributions of plants to the pollen
pool, the analyses of male steriles suggested that the estimation procedure for
pollen pool allele frequencies was sound and that the assumption of uniform
allele frequencies within the population of maternal plants was generally
fulfilled. However, in the natural populations, only two male steriles were
examined in Koekoekswaard. In maize, temporal variation of allele frequencies
in the pollen pool may influence the estimation of outcrossing rates (Bijlsma
et al. 1986). This variation may have also affected the estimation of
outcrossing rates in natural populations of S. pratensis as estimates of pollen
pool allele frequencies did not take into account the flowering times of the
plants sampled. In contrast, in the experimental populations, each estimate of
the outcrossing rate was based on the estimated allele frequencies during the
flowering period of the stalk analyzed.
Third, it is assumed that the probability of an outcross is independent of
the maternal genotype. It seems unlikely that this assumption was violated in the
natural populations because the maternal plants sampled represented a large
fraction of the population and were randomly distributed within populations.
However, in the experimental garden, the progeny of the six crosses used as
maternal plants were not distributed equally in the four populations.
Consequently, outcrossing rates in these populations were estimated using
different numbers of offspring from each cross. Nonetheless, differences in
outcrossing rate between the maternal genotypes were not significant
(Kruskal-Wallis one-way nonparametric ANOVA: P |is greater than~ 0.15 in all
four populations).
Fourth, it is assumed that no selection affecting the marker loci intervenes
between fertilization and the determination of progeny genotypes. If fertilization
by alien pollen is favored and selection takes place between mating and scoring
of progeny genotypes, outcrossing rates will be biased upwards. In S. pratensis
significant inbreeding depression was found for ovule abortion rate and
germination percentage (Ouborg et al. unpubl. data). Because progeny in the
present study were analyzed at the juvenile stage and the fraction of seeds
surviving to be genotyped ranged from 0.179 to 0.738, outcrossing rates might
have been upwardly biased by postfertilization selection.
The smaller |F.sub.IS~ values relative to the |F.sub.eq~ values might
suggest that there has been selection against homozygotes for the marker loci
or genes linked to them. However, for selection alone to account for these
disparities, particularly for Bijland-2, would require very large selection
coefficients. In addition to the possible temporal or spatial structuring of
the pollen pool, fluctuations in the breeding system between years may have
influenced the large differences between |F.sub.IS~ and |F.sub.eq~.
Outcrossing rates were correlated with plant density rather than population
size. Positive correlations between plant density and outcrossing rates were
also found for thyme (Valdeyron et al. 1977), ponderosa pine (Farris and Mitton
1984), Plantago coronopus (Wolff et al. 1988), rye (Vaquero et al. 1989), and
several species of tropical trees (Murawski et al. 1990; Murawski and Hamrick
1991). Differences in foraging behavior of the pollinators in populations with
different plant densities may underlie the effect of plant density on
outcrossing rates in S. pratensis. Bumblebees are known to switch among
individuals more often in dense populations than in sparse populations
(Heinrich 1979). In Viola populations, the frequency of interplant flights by
insects is also proportional to plant density (Beattie 1976). Unfortunately,
the distances between the experimental populations were inadequate to prevent
interpopulational flights by pollinators. Accordingly, it cannot be ruled out that
the behavior of the pollinators in separate populations was influenced by
earlier experience in neighboring populations. This constraint may limit how
applicable our findings are to the amount and distribution of genetic variation
in natural populations of S. pratensis.
Substantial differences in the percentage of male sterility existed both
between natural and experimental populations. In the experimental garden, a
high percentage of male sterility appeared to influence the rate of outcrossing
negatively in the sparse populations (table 4: |t.sub.m~ = 0.715 |+ or -~ 0.058
in population 4 and |t.sub.m~ = 0.882 |+ or -~ 0.057 in population 1), but had
little effect on outcrossing rates in the dense populations (Table 4: |t.sub.m~
= 0.955 |+ or -~ 0.054 in population 3 and |t.sub.m~ = 0.940 |+ or -~ 0.026 in
population 2). The outcrossing rates of hermaphrodites correlated negatively
with the mean distance separating the plants analyzed from other pollen
producing plants. The combined effects of a low overall plant density and a
relatively high percentage of male sterility may also have contributed to the
low outcrossing rate of the natural population Bijland-2 (|t.sub.m~ = 0.382 |+
or -~ 0.092). Higher selling rates in hermaphroditic plants also accompany higher
frequencies of male steriles in natural populations of thyme (Valdeyron et al.
1977) and Plantago coronopus (Wolff et al. 1988).
Plant densities in the experimental populations were considerably higher
than those of the natural populations. Density of pollen producing plants in
the densest natural population (Koekoekswaard) resembled the sparsest
experimental population (population 4), and their rates of outcrossing were
nearly equal. This pattern suggests that the high outcrossing rates seen in the
densest experimental populations (2 and 3) are probably never observed in
natural populations of S. pratensis in the
The results of this study corroborate the results of Aide (1986), who
demonstrated that animal-pollinated species show a random distribution of
outcrossing rates with high interpopulational variation. Results of the present
study suggest that plant density and sex polymorphism are among the
characteristics that contribute to this variation.
Implications for Conservation Biology
The rate at which populations become inbred depends on the effective
population size. To preserve genetic variation in populations it is often
recommended to maintain numerous individuals. Because higher rates of
self-fertilization influence the effective size of plant populations
negatively, the rate of outcrossing also determines the rate of loss of genetic
variation in plant populations. Therefore, effective management to preserve
genetic variation in plant populations should provide for the maintenance of
high outcrossing rates. Our experimental study suggests that in S. pratensis,
the rate of outcrossing depends on densities rather than on numbers. This means
that the maintenance of high plant densities in S. pratensis populations will
slow the rate of loss of genetic variation and, if loss of genetic variability
affects the illness of individuals negatively, the rate of reduction in
fitness.
ACKNOWLEDGMENTS
We would like to thank C. Galen, J. Heywood, and two anonymous reviewers for
their helpful comments on earlier drafts of the manuscript, A. Rumahloine, L.
Hoeksema-du Pui, J. Haeck, and K. Reinink for support in carrying out the
experiments, and L. Bruinzeel, C. van Hoogmoed, R. Hut, and R. van der Wal for
their part in the electrophoretic analyses. This research project has been
partly subsidized by the Ministry of Agriculture, Nature Management and
Fisheries.
LITERATURE CITED
Aide, T. M. 1986. The influence
of wind and animal pollination on variation in outcrossing rates.
Evolution 40:434-435.
Allard, R. W., S. K. Jain, and P. L. Workman. 1968.
The genetics of inbreeding populations. Advances in Genetics 14:55-131.
Barrett, S.C.H., and D. Charlesworth. 1991. Effects
of a change in the level of inbreeding on the genetic load. Nature 352:522-524.
Barrett, S.C.H., and J. R. Kohn. 1991. Genetic and
evolutionary consequences of small population size in plants: implications for
conservation. Pp. 3-30 in D. A. Falk, and K. E. Holsinger, eds. Genetics and
conservation of rare plants.
Beattie, A. J. 1976. Plant dispersion, pollination and
gene flow in Viola. Oecologia 25:291-300.
Bijlsma, R., R. W. Allard, and A. L. Kahler. 1986.
Nonrandom mating in an open-pollinated maize population. Genetics 112:669-680.
Bijlsma, R., N. J. Ouborg, and R. Van Treuren.
1991. Genetic and phenotypic variation in relation to population size in two
plant species: Salvia pratensis and Scabiosa
columbaria. Pp. 89-101 in A. Seitz and V. Loeschcke, eds. Species conservation:
a population-biological approach. Birkhauser,
Bos, M., R. Steen, and H. Harmens. 1985. Protogyny
in Plantago lanceolata populations: an adaptation to pollination by wind? Pp. 327-338 in P. Jacquard, ed. Genetic differentiation and
dispersal in plants. Springer,
Brown, A.H.D. 1979. Enzyme polymorphism in plant
populations. Theoretical Population Biology 15: 142.
Campbell, D. R. 1985. Pollen and gene dispersal: the influences of
competition for pollination. Evolution 39:418-431.
Charlesworth, D., and B. Charlesworth. 1987.
Inbreeding depression and its evolutionary consequences. Annual Review of
Ecology and Systematics 18:237-268.
Clegg, M. T. 1980. Measuring plant mating systems.
Bioscience 30:814-818.
Devlin, B., and A. G. Stephenson. 1985. Sex
differential floral longevity, nectar secretion, and pollinator foraging in a
protandrous species. American Journal of Botany 72:303-310.
Ennos, R. A. 1985. The mating system and genetic structure in a perennial
grass, Cynosurus cristatus L. Heredity 55:121-126.
Ennos, R. A., and M. T. Clegg. 1983. Flower color
variation in the morning glory, Ipomoea purpurea. Journal of Heredity
74:247-250.
Epperson, B. K., and M. T. Clegg. 1987a.
Frequency-dependent variation for outcrossing rate among flower-color morphs of
Ipomoea purpurea. Evolution 41:1302-1311.
-----. 1987b. First-pollination primacy and pollen
selection in the morning glory, Ipomoea purpurea. Heredity 58:5-14.
Farris, M. A., and J. B. Mitton. 1984. Population
density, outcrossing rate, and heterozygote superiority in ponderosa pine.
Evolution 38:1151-1154.
Frankel, O. H., and M. E. Soule. 1981. Conservation
and evolution.
Franklin,
Goodman, D. 1987. The demography of chance extinction.
Pp. 11-34 in M. E. Soule, ed. Viable populations for
conservation.
Hamrick, J. L., Y. B. Linhart, and J. B. Mitton.
1979. Relationship between life-history characteristics and electrophoretically
detectable genetic variation in plants. Annual Review of Ecology and
Systematics 10:173-200.
Hedrick, P. W. 1990. Mating systems and evolutionary
genetics. Pp. 83-114 in K. Wohrmann and S. Jain, eds. Population
biology: ecological and evolutionary viewpoints. Springer,
Heinrich, B. 1979. Resource
heterogeneity and patterns of movement in foraging bumblebees. Oecologia
40:235-245.
Holtsford, T. P., and N. C. Ellstrand. 1989.
Variation in outcrossing rate and population genetic structure of Clarkia
tembloriensis (Onagraceae). Theoretical and Applied Genetics 78:480-488.
Karron, J. D. 1991. Patterns of genetic variation and
breeding systems in rare plant species. Pp. 87-98 in
D. A. Falk and K. E. Holsinger, eds. Genetics and conservation of rare plants.
Kesseli, R. V., and S. K. Jain. 1985. Breeding
systems and population structure in Limnanthes. Theoretical and Applied
Genetics 71:292-299.
Knight, S. E., and D. M. Waller. 1987. Genetic
consequences of outcrossing in the cleistogamous annual, Impatiens capensis. I.
Population-genetic structure. Evolution 41:969-978.
Lande, R., and D. W. Schemske. 1985. The evolution
of self-fertilization and inbreeding depression in plants. I. Genetic models.
Evolution 39:24-40.
Levin, D. A., and H. W. Kerster. 1969.
Density-dependent gene dispersal in Liatris. American Naturalist 103:61-74.
Loveless, M. D., and J. L. Hamrick. 1984.
Ecological determinants of genetic structure in plant populations. Annual
Review of Ecology and Systematics 15:65-95.
Marshall, D. F., and R. J. Abbott. 1982.
Polymorphism for outcrossing frequency at the ray floret locus in Senecio
vulgaris L. I. Evidence. Heredity 48:227-235.
Murawski, D. A., and J. L. Hamrick, 1991. The effect of the density of flowering individuals on the mating
system of nine tropical tree species. Heredity 67: 167-174.
Murawski, D. A., J. L. Hamrick, S. P. Hubbell, and R. B.
Foster. 1990. Mating systems of two Bombacaceous trees of a neotropical moist forest. Oecologia 82:501-506.
Nei, M. 1987. Molecular evolutionary genetics.
Ouborg, N. J., R. Van Treuren, and J.M.M. van Damme.
1991. The significance of genetic erosion in the process of extinction. II.
Morphological variation and fitness components in populations of varying size
of Salvia pratensis L. and Scabiosa
columbaria L. Oecologia 86:359-367.
Richards, A. J. 1986. Plant breeding systems. Allen
and Unwin,
Schaal, B. A. 1975. Population structure and local
differentiation in Liatris cyllindracea. American Naturalist
109:511-528.
Schemske, D. W., and R. Lande. 1985. The evolution
of self-fertilization and inbreeding depression in plants. II. Empirical
observations. Evolution 39:41-52.
Schmitt, J. 1983. Density-dependent pollinator foraging, flowering phenology, and temporal pollen dispersal patterns in
Linanthus bicolor. Evolution 37:1247-1257.
Schoen, D. J. 1982a. Genetic variation and the breeding
system of Gilia achilleifolia. Evolution 36:361-370.
-----. 1982b. The breeding system of Gilia
achilleifolia: variation in floral characteristics and outcrossing rate. Evolution
36:352-360.
Schonewald-Cox, C. S., S. M. Chambers, B. Macbryde, and L.
Thomas. 1983. Genetics and conservation: a reference for managing wild
animal and plant populations. Benjamin-Cummings,
Shaffer, M. 1987. Minimum viable populations: coping with uncertainty. Pp.
69-86 in M. E. Soule, ed. Viable populations for
conservation.
Shaw, D. V., A. L. Kahler, and R. W. Allard. 1981. A multilocus estimator of
mating system parameters in plant populations. Proceedings of the National
Sokal, R. R., and F. J. Rohlf. 1981. Biometry, 2d
ed. W. H. Freeman,
Valdeyron, G., B. Dommee, and P. Vernet, 1977.
Self-fertilization in male-fertile plants of a gynodioecious species: Thymus
vulgaris L. Heredity 39:243-249.
Van Damme, J.M.M. 1983. On gynodioecy in Plantago
lanceolata L. Ph.D. dissertation.
Van Delden, W. 1985. The
significance of genetic variation in plants as illustrated by Plantago
populations. Pp. 219-239 in J. Haeck and J. W. Woldendorp, eds.
Structure and functioning of plant populations 2.
Van Dijk, H., K. Wolff, and A. De Vries. 1988.
Genetic variability in Plantago species in relation to their ecology. 3.
Genetic structure of populations of P. major, P. lanceolata and P. coronopus.
Theoretical and Applied Genetics 75:518-528.
Van Treuren, R., R. Bijlsma, W. Van Delden, and N. J.
Ouborg. 1991. The significance of genetic erosion in the process of
extinction. I. Genetic differentiation in Salvia
pratensis and Scabiosa columbaria in relation to
population size. Heredity 66: 181-189.
Vaquero, F., F. J. Vences, P. Garcia, L. Ramirez, and M.
Perez de la Vega. 1989. Mating system in rye: variability in relation to
the population and plant density. Heredity 62:17-26.
Walpole, R. E. 1982. Introduction to statistics. 3d ed.
Watkins, L., and D. A. Levin. 1990. Outcrossing
rates as related to plant density in Phlox drummondii. Heredity 65:81-89.
Wolff, K., B. Friso, and J.M.M. van Damme. 1988.
Outcrossing rates and male sterility in natural populations of Plantago
coronopus. Theoretical and Applied Genetics 76:190-196.
Wright, S. 1965. The interpretation of population
structure by F-statistics with special regard to systems of mating.
Evolution 19:395-420.