Population Density and Mating Pattern in
Western Larch
Y. A. El-Kassaby and B. Jaquish
Variation in the mating systems of western larch as affected by different stand
density was investigated in two adjacent populations (natural and seed tree) in two
locations in the interior of British Columbia. Population multilocus outcrossing
rates, individual tree multilocus outcrossing rates, and the correlation of outcrossed paternity were estimated from analysis of allozyme variation in embryonic
and megagametophytic tissue of wind-pollinated seeds. All population outcrossing
rate estimates were significantly different from t = 1. Positive association between
population density and outcrossing rate was detected for one location (0.903 versus 0.876) and inverse association was obtained for the second (0.793 versus
0.912). In three populations, the distributions of individual tree outcrossing rates
were skewed toward high outcrossing. In the fourth population, which produced a
low multilocus estimate of outcrossing, individual tree outcrossing rates were
widely distributed. High and significant correlated matings (6% and 10%) were detected in the populations with high densities, while low (0.1% and 2.0%) estimates
were detected in populations with low density. It appears that the effects of population density observed in this study and reported in previous studies can be overridden by the presence of barriers that limit pollen movement.
From Pacific Forest Products Ltd., Saanich Forestry
Centre, 8067 East Saanich Road, Saanichton, British Columbia V8M 1K1, Canada, the Forest Sciences Department, Faculty of Forestry, University of British Columbia (El-Kassaby), and the British Columbia Ministry of
Forests, Research Branch, Kalamalka Research Station
and Seed Orchard, 3401 Reservoir Road, Vernon, British Columbia V1B 2C7, Canada (Jaquish). We thank K.
Ritland for providing the computer program used in
this study, M. Carlson, R. G. Latta, M. D. Meagher, J. B.
Mitton, and A. D. Yanchuk for reviewing an earlier version of the manuscript, and V. Ashley and S. A. Barnes
for technical assistance. This study is supported in
part by the British Columbia Ministry of Forests, Research and Silviculture Branches, Forest Renewal British Columbia grant no. TIP-002, and Pacific Forest Products Ltd.
Journal of Heredity 1996;87:438-443; 0022-1503/96/S5.00
438
Detailed knowledge and understanding of
the mating system in natural and artificial
plant populations are fundamental prerequisites for understanding their genetic architecture and evolutionary potential. Traditionally the mating system has been assessed by the probability of self-fertilization versus random outcrossing via the
"mixed-mating" model (Fyfe and Bailey
1951). Recently, several areas of research
into plant mating systems, such as the extent of consanguineous mating (Ritland
and El-Kassaby 1985) and the similarity of
paternity among siblings (Ellstrand and
Marshall 1986; Meagher 1986) or correlated matings (Ritland 1989), have received
increased attention as other facets of the
mating system (Ritland and El-Kassaby
1995). While most of the mating systems
studies have focused on the population's
mating system estimate [see Schemske
and Lande (1985) for a review], variation
among individuals' selfing rate and the extent of correlated mating (i.e., sibs' genetic relatedness) in an individual progeny array have proven to be revealing and more
informative. For example, constant outcrossing rate estimates over 3 years were
observed for a western white pine (Pinus
monticola) population. However, individu-
als' outcrossing rates varied within (among
trees) and among years (for the same
trees), revealing the existence of consistent inbreeders and outcrossers (El-Kassaby et al. 1987, 1993).
The mating system in a population is influenced by genetic and environmental
factors (Clegg 1980). Population density
represents one of the environmental factors potentially influencing the mating pattern in forest tree populations. Positive relationships between population density
and outcrossing rate were reported for
ponderosa pine (Pinus ponderosa; Farris
and Mitton 1984), tamarack (Larix laricina;
Knowles et al. 1987), and Engelmann
spruce (Picea engelmannii; Shea 1987), but
not for coastal Douglas-fir (Pseudotsuga
menziesii var. menziesii; Neale and Adams
1985), Jeffrey pine (Pinus jeffreyi; Furnier
and Adams 1986), subalpine fir (Abies lasiocarpa; Shea 1987), Norway spruce (Picea abies; Morgante et al. 1991), and Bombacaceous tree (Cauanillesia plantanifolia;
Murawski and Hamrick 1991). It should be
emphasized, however, that density should
be considered in concert with other factors, such as reproductive ecology (ElKassaby and Davidson 1991), reproductive
phenology (El-Kassaby and Ritland 1986;
Table 1. Log-likelihood G test on segregation ratios of seven polymorphic loci in western larch seeds
Locus
Genotype
IDH
PGM
G6P
G6P
1:2'
MDH1
MDH4
PGI
6PG
1:2
1:2
1:3
1:2
1:2
1:2
1:2
Number of
trees
Observed ratio
Pooled G°
Heterogeneity C
41
25
26
7
54
16
3
57
851:794
522:447
592:448
158:147
1,019:1,114
315:325
63:67
1,177:1,106
0.160"
0.016"
0.000"
0.629"
0.088"
0.883"
0.684"
0.137"
59.318'
42.296'
75.101'
11.719"
0.498"
0.061"
2.406"
57.218"
• Pooled G values indicate the overall deviation from 1:1 ratio.
6
Heterogeneity G values indicate the amount of heterogeneity in the segregation ratio among trees.
'Allozymes were numbered starting with " 1 " for most common alleles; faster and slower alleles were given even
and odd numbers, respectively.
" Not significant.
• Significant at 5%.
'Significant at 1%.
El-Kassaby et al. 1988), fecundity (Denti
and Schoen 1988; Schoen and Stewart
1986, 1987), and the distance between
mates (Adams et al. 1992; Erickson and
Adams 1989).
Western larch [Larix occidentalis Nutt.
(Pinaceae)] is a monoecious, deciduous
conifer that occurs in the upper Columbia
River Basin of North America (British Columbia, Montana, Idaho, Washington, and
Oregon). Its seed and pollen cones commonly occur in the upper and lower parts
of the crown, respectively, but considerable overlap exists. It is characterized by
infrequent cone production; however, in
good cone years it is a prolific seed producer. The relatively large seed wings on
its small seed enable long-distance seed
dispersal by wind, although it is reported
that only about 5% of the total sound seed
produced is likely to reach beyond about
130 m of the seed donor (Boe 1953). Fire
is essential to the maintenance of western
larch in natural populations. The species
is one of the most shade intolerant conifers in western North America and thus
grows in even-aged stands, often after wild
fires (Schmidt and Reymond 1990). Regeneration systems that promote the establishment of even-aged stands, such as
seed tree and clear-cut, are best suited for
western larch ecological requirements
(Smith 1962).
The seed tree regeneration system is
the most suitable silvicultural system for
the natural regeneration of shade intolerant conifers. In this system, all trees in the
area are removed with the exception of
regularly distributed seed trees, which are
left either individually or in clusters in a
uniform fashion throughout the area to
serve as a seed source. This regeneration
system is expected to affect the mating
system pattern and the level of gene flow
among the remaining individuals (i.e.,
higher inbreeding and correlated matings)
(El-Kassaby and Namkoong 1994).
Genetic evaluation of the seed tree regeneration system, specifically assessment of the rate of selfing, mating among
relatives, and the extent of correlated mating, is important due to the impact of
these mating system features on the regeneration potential of the seed tree system. Moreover, this evaluation will help in
understanding the genetic architecture
and evolutionary potential of the new populations produced by this regeneration
system (El-Kassaby and Namkoong 1994).
This study investigates the importance
of stand density in western larch on both
population and individual-tree outcrossing
rate estimates, and the level of correlated
mating in paired populations of different
densities in two locations in the interior of
British Columbia.
mately 21%). The average height of the
western larch dominant/codominant trees
was 25-30 m. The adjacent seed tree stand
is part of a 36 ha forest that was harvested
in 1991-1992, leaving about 60 regularly
distributed western larch trees per hectare. Prior to harvest, the seed tree stand
was part of the natural stand.
The Flathead population (49°04'N,
114°26'W; elevation 1380 m) is also an
even-aged stand that originated from a fire
approximately 100 years ago. The natural
stand consists of approximately 75% western larch, with the remaining 25% being
lodgepole pine and black cottonwood
(Populus trichocarpd). The stand density is
about 1,250 trees/ha. Average western
larch dominant/codominant height was
similar to that of the Becker Lake location
(about 25 m). The adjacent seed tree
stand is located in a 25 ha block that was
harvested in 1987, leaving about 30 regularly distributed western larch trees per
hectare. Again, both stands (seed tree and
natural) were parts of the same stand prior to harvest.
Collections of seed cones were conducted in the two locations during early September 1993 (a very good seed and pollen
cone producing year). From each stand,
34-35 trees were selected, based on their
fecundity (i.e., trees bearing a sufficient
number of seed cones) and being 50-75 m
apart. Maternal identity of the seeds was
maintained throughout seed extraction
and electrophoresis.
Starch-gel electrophoresis of megagametophytic (In) and corresponding embryonic (2n) tissues were assayed following El-Kassaby et al. (1982). Seven enzyme
loci (JDH, PGM, G6P, MDH1, MDH4, PGI,
Materials and Methods
The western larch trees used in this study
were sampled from two "populations"
(natural and seed tree) at two locations
(Becker Lake and Flathead) in the interior
of British Columbia, Canada.
The Becker Lake population (50°15'N,
119°10'W; elevation 1200 m) is an evenaged stand that originated from a fire approximately 120 years ago. The natural
stand is a diverse mix of western larch,
interior Douglas fir (Pseudotsuga menziesii
var. glauca), lodgepole pine (Pinus contortd), white spruce (Picea glauca), paper
birch (Betula papyrifera), and trembling
aspen (Populus tremuloides) in the overstory with a developing understory of
western red cedar (Thuja plicata). Overall
stand density was about 1,150 trees/ha, of
which 240 were western larch (approxi-
and 6PG) were assayed for estimation of
mating system parameters. Maternal tree
genotypes were inferred from the segregation of allozymes in megagametophytic
tissue of about 40 seeds per heterozygous
tree used in the mating system estimation.
The number of trees included in the segregation analysis are listed in Table 1. A
tandem assay of the haploid megagametophyte and its corresponding diploid embryo tissue revealed the pollen contribution. Differences between allele frequencies in the maternal tree populations and
their corresponding outcrossing pollenpool allele frequencies were checked for
significance (P < .05) by comparing
bounds of confidence intervals (Table 2).
To confirm that the seed tree populations
were parts of the natural populations, the
level of genetic variation in the four populations and their respective seed crops
El-Kassaby and Jaquish • Western Larch Population Density 4 3 9
Table 2. Allelic frequencies and their 95% confidence intervals (1.96 SE) for seven polymorphic loci in the maternal and outcrossing pollen pools of western
larch populations
Flathead
Becker Lake
Natural
Seed Tree
Locus
Allele
Ovule
Pollen
Ovule
IDH
1
2
3
1
2
1
2
3
1
2
1
2
1
2
3
1
2
0.957 ± 0.010
0.043 ±0.010
0.962 ± 0.002
0.038 ± 0.002
0.957
0.043
0.700
0.229
0.071
0.686
0.314
0.971
0.029
0.958
0.014
0.028
0.314
0.686
0.983
0.017
0.746
0.204
0.050
0.762
0.238
0.963
0.037
0.981
0.001
0.018
0.452
0.548
0.972
0.014
0.014
0.986
0.014
0.800
0.171
0.029
0.671
0.329
0.929
0.071
0.972
0.014
0.014
0.471
0.529
PGM
G6P
MDHl
MDH4
PGI
6PG
±
±
±
±
±
±
±
±
±
±
±
±
±
±
0.010
0.010
0.049
0.041
0.015
0.050
0.050
0.007
0.007
0.009
0.003
0.006
0.050
0.050
± 0.001
+ 0.001
± 0.010
± 0.009
± 0.003
± 0.010
± 0.010
± 0.002
i 0.002
± 0.001
± 0.000°
± 0.001
± 0.013
±0.013
Natural
Pollen
± 0.006
± 0.003
± 0.003
± 0.003
± 0.003
± 0.037
±0.033
± 0.007
± 0.052
± 0.052
±0.015
±0.015
± 0.006
± 0.003
± 0.003
± 0.058
± 0.058
0.959
0.040
0.001
0.980
0.020
0.791
0.158
0.051
0.749
0.251
0.963
0.037
0.991
0.001
0.008
0.451
0.549
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
±
0.002
0.002
0.001
0.001
0.001
0.009
0.007
0.003
0.010
0.010
0.002
0.002
0.000
0.000
0.000
0.013
0.013
Seed Tree
Ovule
Pollen
Ovule
Pollen
0.456 ± 0.059
0.544 ± 0.059
0.465 ± 0.013
0.535 ±0.013
0.557 ± 0.058
0.443 ± 0.058
0.481 ± 0.013
0.519 ± 0.013
0.897
0.103
0.957
0.029
0.014
0.838
0.162
0.897
0.103
0.971
0.014
0.014
0.765
0.235
0.909
0.091
0.959
0.037
0.004
0.813
0.187
0.902
0.098
0.995
0.001
0.004
0.758
0.242
0.829
0.171
0.958
0.028
0.014
0.743
0.257
0.957
0.043
0.986
0.014
0.874
0.126
0.979
0.015
0.007
0.750
0.250
0.951
0.049
0.999
0.001
± 0.022
±0.022
± 0.010
± 0.007
± 0.003
± 0.032
±0.032
± 0.022
±0.022
± 0.007
± 0.003
± 0.003
± 0.043
± 0.043
± 0.004
± 0.004
± 0.002
± 0.002
± 0.002
± 0.000
±0.008
± 0.005
± 0.005
± 0.000
± 0.000
± 0.000
± 0.010
±0.010
± 0.033
±0.033
± 0.009
± 0.009
± 0.003
± 0.045
± 0.045
± 0.010
± 0.010
± 0.003
± 0.003
0.671 ± 0.053
0.329 ± 0.053
± 0.006
±0.006
± 0.001
± 0.001
± 0.000
± 0.010
± 0.010
i 0.002
± 0.002
* 0.000
± 0.000
0.730 ±0.012
0.270 ± 0.012
Cl < 0.0004.
was assessed by Nei's (1978) genetic distance method.
Single-locus (/s) and multilocus (t^) estimates of outcrossing rate, and outcrossing pollen allelic frequencies (p) were estimated using the maximum likelihood
procedure of Ritland and El-Kassaby
(1985). The correlated matings (rp) were
estimated following the method of Ritland
(1989) after modification to account for
the added information obtained from the
megagametophytic tissue of conifer seeds
(Ritland, in preparation). The sample sizes
(number of trees and seeds per stand)
used in estimating the mating system pa-
lie combinations were not significantly different from the expected 1:1 ratio (Table
1), indicating that these allozymes exhibited distinct, codominant expression and
simple Mendelian segregation in their
mode of inheritance. The heterogeneity G
Results
test among trees' segregation ratios was
The seven loci were polymorphic in all the
significant for three allelic combinations
four populations studied. Four loci (PGM, (IDH, PGM, and G6P) (Table 1); however,
MDHl, MDH4, and 6PG) were diallelic,
close examination of these allelic combiwhile IDH (Flathead seed tree population), nations on an individual-tree level indicatPGI (both of the Flathead populations and ed a lack of systematic bias toward any
the Becker Lake natural population), and
specific allele (data not given).
G6P were triallelic (Table 2). Segregation
Estimates of allelic frequencies and
analyses for all the observed pooled alle- their 95% confidence intervals for the
ovule (maternal trees) and outcrossing
pollen pools for the seven loci studied are
Table 3. Estimates of single-locus ( f j , multilocus (O> and correlated paternity ( r j for the four
listed in Table 2 for each of the four poppopulations
ulations. Ovule and pollen pool allelic frequencies differed at the 95% level (Table
Flathead
Becker Lake
2). In general, the natural populations proLocus
Seed Tree
Natural
Natural
Seed Tree
duced fewer significant differences (25%
IDH
0.847 (0.096)
0.923(0.113)
0.910 (0.044)
0.775 (0.056)
and 63% of the allelic comparisons bePGM
0.859 (0.099)
0.936 (0.202)
0.837 (0.042)
0.748 (0.040)
G6P
tween the ovule and the outcrossing pol0.827 (0.043)
0.711 (0.049)
1.004(0.445)
0.538(0.115)
MDHl
0.946 (0.035)
0.867 (0.050)
0.987 (0.042)
0.836 (0.063)
len pools were significant for Becker Lake
MDH4
0.890(0.119)
0.905 (0.029)
0.907(0.175)
0.673 (0.088)
and Flathead populations, respectively)
a
PGI
0.823(0.158)
0.515(0.234)
6PG
0.841 (0.046)
0.920 (0.038)
0.887 (0.035)
0.852 (0.058)
than observed for the seed tree populations (73% and 77% for Becker Lake and
0.864 (0.018)
0.914 (0.018)
0.838 (0.028)
0.771 (0.028)
/
0.903 (0.015)
0.876 (0.022)
0.793 (0.026)
0.912 (0.018)
Flathead populations, respectively) (Table
0.062 (0.009)
0.104(0.034)
rp
0.020 (0.008)
0.001 (0.018)
2). This suggests that the maternal trees
M
35
34
35
35
were not representative of the local pollen
1,380
1,400
1,360
1,343
pools, especially for the seed tree popuNo. of trees/ha
1,250
30
1,150
60
lations.
No. of larch trees/ha
940
30
240
60
Estimates of Wright's inbreeding coeffi% larch
75
100
21
100
cient (F; Wright 1922), which measures the
SDs are given in parentheses.
excess of homozygosity above Hardy° Single-locus outcrossing was not calculated due to the lack of information in the pollen gene pool (see Table 2,
Weinberg expectations, were significantly
for pollen pool extreme allele frequencies).
different from zero (P < .05) for maternal
* Single-locus minimum variance mean.
trees. Negative F values were observed for
c
Number of trees sampled.
three of the maternal populations [Becker
" Number of seeds censused.
4 4 0 The Journal of Heredity 1996:87(6)
rameters [single-locus (tj and multilocus
(?„) estimates of outcrossing rate, outcrossing pollen allelic frequencies (p), and
correlated matings (rp)] are listed in Table 3.
B- Becker Lake (Seed Tree)
A- Becker Lake (Natural Stand)
10
8
| a
J'
2
0
.40
.50
.60
.70
.90
.80
.40
.50
.60
.70
.80
.90
1.0
.90
1.0
Outcrossing Rate
Outcrossing Rate
D- Flathead (Seed Tree)
C- Flathead (Natural Stand)
10
10
8
o
6
2
0
.40
.50
.60
.70
.80
.90
1.0
.50
.60
.70
.80
Outcrossing Rate
Outcrossing Rate
Figure 1. Frequency distribution ol single-tree outcrossing rate by population: (A) Becker Lake natural; (B) Becker Lake seed tree; (Q Flathead natural; and (D) Flathead
seed tree.
Lake natural (F = -0.110) and seed tree
(F = -0.066) and Flathead seed tree (F =
-0.208)], indicating an excess of heterozygotes. A positive value (F = 0.064) was
recorded for the Flathead maternal natural
population, indicating a deficiency of heterozygotes. In contrast, the means of Fva\ues over progeny loci were positive [Flathead natural (0.053) and seed tree (0.116)
and Becker Lake natural (0.078) and seed
tree (0.034)] for the four populations, indicating that some inbreeding had taken
place in the seed crops (see below).
Single-locus outcrossing rate estimates
ranged between 0.515 (PGl, Flathead seed
tree population) and 1.004 (G6P, Becker
Lake seed tree population) and were significantly heterogeneous (P < .05) among
loci for all populations studied (Table 3).
Multilocus outcrossing rate estimates
ranged between 0.793 and 0.912 for Becker
Lake natural and seed tree populations, respectively (Table 3). All multilocus outcrossing rates were significantly different
from t = 1.0 (Table 3). In addition, significant differences were detected between
the natural and seed tree populations' tm
estimates for the two locations; however,
this difference was not consistent in its direction (Table 3). In the Flathead location,
the natural population tm estimate was
higher than the seed tree estimate, while
at Becker Lake the seed tree population tm
estimate was higher than the natural
stand (Table 3). With the exception of the
Becker Lake seed tree population, differences between the mean single locus and
the multilocus estimates were detected
(Table 3), indicating the existence of mating among relatives in the remaining three
populations.
Estimates of outcrossing rates varied
among trees within populations (Figure
la-d). With the exception of the Becker
Lake natural population, the individual
tree estimates in the three remaining populations produced skewed distributions,
with the majority of the trees yielding a
high outcrossing rate (Figure lb-d). The
Becker Lake natural population produced
a wide distribution of single-tree outcrossing rate estimates (Figure la).
The correlation of outcrossed paternity
varied between the natural and seed tree
populations (Table 3). In general, higher
estimates of correlated matings were obtained for natural populations than from
their seed tree counterparts. Natural populations produced significant 6% (Flathead) and 10% (Becker Lake) correlated
matings, while seed tree populations yielded only 2% (Flathead) and 0.1% (Becker
Lake) (Table 3).
To assess the genetic similarity between
the seed tree populations and their respective natural populations, the level of
genetic variation present in the two locations (between the Flathead and Becker
Lake), between the natural and seed tree
populations within each location, and between the seed crops and their respective
maternal populations was assessed using
Nei's (1978) genetic distance (Figure 2).
The dendrogram confirmed the genetic
similarity between each seed tree stand
and its natural population within each location, as well as the presence of genetic
differences between the Flathead and
Becker Lake locations. In addition, as expected, the genetic variation present in
the four seed crops was similar to that of
their respective maternal tree populations
(Figure 2).
Discussion
The mating system pattern of a plant species is expected to show variation that is
due to either genetic or environmental factors (Clegg 1980). The mixed mating sys-
El-Kassaby and Jaquish • Western Larch Population Density 4 4 1
FHNE
FHNM
FHSE
FHSM
€
€
.40
.33
.27
.20
. 13
.07
BLNE
BLNM
BLSE
BLSM
.00
Distance
Figure 2. Phenogram, based on Nei's genetic distance, of the four western larch populations and their respective
seed crops (FH = Flathead; BL = Becker Lake; N = natural population; S = seed tree population; M = maternal
trees; E = seed crop).
tern of coniferous tree species is not an
exception. While the majority of conifers
are almost completely outcrossed (Mitton
1992), some species have produced low
estimates; for example, in western red cedar the outcrossing rate is t = 0.32 (ElKassaby et al. 1994). In conifers, variation
in outcrossing rate estimates have been
reported among populations, individuals,
seasons, and tree crown strata, and are
caused by variation in reproductive ecology, phenology, and output (i.e., fecundity), distance between mates, and density
as well as variation among trees in propensity to selfing [see Mitton (1992) for a
review].
In species that are vector-dependent for
their pollination, plant density represents
one of the most important environmental
factors affecting their mating system pattern, extent of gene flow, and level of genetic differentiation among and within
populations (Ellstrand 1992; Mitton 1992).
Higher plant densities are expected to affect pollinator flight distance. Thus, higher
inbreeding, caused by mating among nearby relatives, restricted gene flow, and a
pronounced level of genetic differentiation, is common. In wind-pollinated species, lower density is also expected to produce similar effects to high-density vectordependent plants due to the patchy nature of their pollen dispersal (Farris and
Mitton 1984). Therefore, in coniferous tree
species, plant density is expected to represent an important factor affecting mating pattern. However, plant density should
be considered in concert with other factors, such as reproductive phenology and
presence of pollen barriers, that might
4 4 2 The Journal of Heredity 1996:87(6)
work as mechanisms counteracting the
limitation of unrelated pollen in low plant
density.
If population density represents the
main factor affecting the mating system of
western larch, then high and low estimates of outcrossing rates should have
been produced for both natural (high density) and seed tree (low density) populations, respectively. This trend (i.e., positive relation between tree density and outcrossing rate estimates) was observed
only for the Flathead location (0.903 versus 0.876) and not for Becker Lake (0.793
versus 0.912) (Table 3). Population density, however, could provide an explanation
for the observed anomaly. Western larch
stocking in the Flathead natural and seed
tree populations are 940 and 30 western
larch trees per hectare, respectively. However, the western larch proportion in the
Flathead natural population represented
75% of the total number of trees, with almost half of the remaining trees being the
deciduous black cottonwood, which had
not leafed out during western larch pollination. The observed densities and outcrossing rate estimates for the Flathead location meet our expectations. In the Becker Lake location, western larch stocking is
240 and 60 trees/ha for the natural and
seed tree populations, respectively. However, the 240 western larch trees in the
natural population represent only 21% of
the stocking in the canopy. In this case, it
is likely that the evergreen associates (i.e.,
portion of the 79%) of the western larch
might have created physical barriers preventing the free movement of western
larch pollen within the stand, thus result-
ing in a lower population outcrossing rate
estimate (tm = 0.793). These results
seemed to violate our expectation of association between plant density and outcrossing. In the seed tree populations, all
other species were removed, so barriers
to pollen movement were minimal.
If the above interpretation is correct,
then we should expect individual-tree outcrossing rate estimates of the four populations to produce patterns that are similar to the multilocus estimates. The distributions of the individual-tree outcrossing
rate estimates for the Flathead natural,
Flathead seed tree, and Becker Lake seed
tree populations were skewed and the majority of trees had high single-tree outcrossing rate estimates (Figure lb-d).
However, in the Becker Lake natural population, individual trees had a wide distribution of single-tree outcrossing rate estimates, thus supporting our earlier observation on the impact of physical barriers
to pollen movement (10 trees produced a
ts < 0.6; Figure la).
Several factors could explain why density alone cannot account for the mating
system pattern detected in the present
study. While western larch is known to be
a poor pollen and seed cone producer,
this study was conducted in a mast year
when lack of pollen production likely was
not a factor. Western larch releases its pollen well before vegetative bud burst (Owens et al. 1994), thus further reducing the
impact of physical barriers. The species
also has a very compact pollination season that is characteristic of the interior
continental climate (Worrall 1993) and a
very short receptive period (Owens et al.
1994). Moreover, the species has an "active" pollination mechanism (i.e., the anatomical structures of the female strobila
participate in directing the pollen toward
the nucellus) that does not discriminate
among self, related, and unrelated pollen.
This active pollination mechanism would
increase the chance of inbreeding caused
by first-pollination primacy, especially
when family structure is common (i.e., no
pollen competition). Since conifers are
known to display strong inbreeding depression (Franklin 1970), it should be expected that inbred seed would be at a selective disadvantage and that the amount
of viable inbred seed available for the detection of inbreeding would be low. In the
present study, the relatively low estimates
of population (Table 3) and individual-tree
(Figure 1) outcrossing rates indicate that
western larch does not exhibit this strong
inbreeding depression, even with the ex-
istence of polyembryony in its reproductive biology (Owens et al. 1994) and its
known impact on postzygotic selection
(Sorensen 1982). The large sample size
used in this study (Table 3) may have contributed to the increased efficiency of inbreeding detection; however, lower outcrossing rate estimates have been detected for related Larix species (Knowles et al.
1987) using a sample size half that in our
study. These factors (pollination ecology,
reproductive phenology and biology, and
lack of physical barriers) act collectively
to promote free pollen movement in a
very short period, nullifying the effect of
density. In addition, density differences
(fewer trees per unit area) between the
natural and seed tree populations also
permit greater air movement, thus allowing less restricted gene flow among mates.
This scenario is further supported by the
observed differences between (1) maternal trees and outcrossing pollen gene
pools [the natural populations produced
fewer significant combinations (25% and
63%) than did the seed tree populations
(73% and 77%)], indicating more pollen
movement in the seed tree populations,
and (2) correlated matings for natural
(high density) and seed tree (low density)
populations.
The genetic similarity observed between the seed tree populations and their
respective natural populations, as well as
the genetic differences between the two
locations, confirmed that the pairs of populations in each location originated from
two different founder populations. This is
expected in species that rely on fire for
their regeneration, further suggesting that
genetic drift might have a major role in the
genetic history of the species (Fins and
Sib 1986) as well as in the existence of
family structure after natural regeneration. The results obtained from this study
indicate that the population structure after regeneration in the natural and seed
tree populations may be different. Lower
levels of inbreeding and correlated matings are expected in the seed tree regeneration system than under natural regeneration, however, the contrasting results
obtained from the two studied locations
call for further investigations to verify the
genetic effects of the seed tree silvicultural system.
In conclusion, the present study demonstrated that generalization regarding
the relation between the mating system
and environmental factors, such as popu-
lation density, could be misleading and requires an in-depth understanding of the biological and ecological peculiarity of the
species in question.
References
Adams WT, Griffin AR, and Moran GF, 1992. Using paternity analysis to measure effective pollen dispersal in
plant populations. Am Nat 140:762-780.
Boe KN, 1953. Western larch and Douglas-fir seed dispersal into clearcuttings. USDA Forest Service Research Note 129, Northern Rocky Mountain Forest and
Range Experiment Station. Washington, D.C.:U.S. Government Printing Office.
Clegg MT, 1980. Measuring plant mating systems.
BioScience 30:814-818.
Denti D and Schoen DJ, 1988. Self-fertilization rates in
white spruce: effect of pollen and seed production. J
Hered 79:284-288.
El-Kassaby YA and Davidson R, 1991. Impact of pollination environment manipulation on the apparent outcrossing rate in a Douglas fir seed orchard. Heredity
66:55-59.
El-Kassaby YA, Meagher MD, and Davidson R, 1993.
Temporal variation in the outcrossing rate in a natural
stand of western white pine. Silvae Genet 42:131-135.
El-Kassaby YA, Meagher MD, Parkinson J, and Portlock
FT, 1987. Allozyme inheritance, heterozygosity and outcrossing rate among Pinus monticola near Ladysmith,
British Columbia. Heredity 58:173-181.
El-Kassaby YA and Namkoong G, 1994. Impact of forest
management practices on genetic diversity and its conservation. In: Proceedings of the International Symposium on Genetic Conservation and Production of Tropical Forest Tree Seed (Drysdale RM, John SET, and
Yapa AC, eds). Saraburi, Thailand: ASEAN-CANADA Forest Tree Seed Centre Project, Muak-Lek: 205-213.
El-Kassaby YA and Ritland K, 1986. The relation of outcrossing and contamination to reproductive phenology
and supplemental mass pollination in a Douglas fir
seed orchard. Silvae Genet 35:240-244.
El-Kassaby YA, Ritland K, Fashler AMK, and Devitt WJB,
1988. The role of reproductive phenology upon the
mating structure of a Douglas fir seed orchard. Silvae
Genet 37:76-82.
El-Kassaby YA, Russell J, and Ritland K, 1994. Mixedmating in an experimental population of western redcedar, Thuja plicala. J Hered 85:227-231.
El-Kassaby YA, Yeh FC, and Sziklai 0, 1982. Inheritance
of allozyme variants in coastal Douglasfir[Pseudotsuga
menziesii var. menziesii (Mirb.) Franco]. Can J Genet
Cytol 24:325-335.
Furnier GR and Adams WT, 1986. Mating system in natural populations of Jeffrey pine. Am J Bot 73:1022-1008.
Fyfe JL and Bailey NTJ, 1951. Plant breeding studies in
leguminous forage crops. I. Natural cross-breeding in
winter beans. J Agric Sci 41:371-378.
Knowles P, Furnier GR, Aleksiuk MA, and Perry DJ,
1987. Significant levels of self-fertilization in natural
populations of tamarack. Can J Bot 65:1087-1091.
Meagher TR, 1986. Analysis of paternity within a natural population of Chamaetirium luieum. I. Identification
of most-likely male parents. Am Nat 128:199-215.
Mitton JB, 1992. The dynamic mating systems of conifers. New For 6:197-216.
Morgante M, Vendrain GG, and Rossi P, 1991. Effects of
stand density on outcrossing rate in two Norway
spruce (Picea abies) populations. Can J Bot 69:27042708.
Murawski DA and Hamrick JL, 1991. The effect of the
density offloweringindividuals on the mating systems
of nine tropical tree species. Heredity 67:167-174.
Neale DB and Adams WT, 1985. The mating system in
natural and shelterwood stands of Douglas fir. Theor
Appl Genet 71:201-207.
Nei M, 1978. Estimation of average heterozygosity and
genetic distance from a small number of individuals.
Genetics 89:583-590.
Owens JN, Morris SJ, and Catalano GL, 1994. How the
pollination mechanism and prezygotic and postzygotic
events affect seed production in Larix occidentalis. Can
J For Res 24:917-927.
Ritland K, 1989. Correlated matings in the partial seller
Mimulus guttatus. Evolution 43:848-859.
Ritland K and El-Kassaby YA, 1985. The nature of inbreeding in a seed orchard of Douglas fir as shown by
an efficient multilocus model. Theor Appl Genet 71:
375-384.
Ritland K and El-Kassaby YA, 1995. Evolutionary and
ecological significance of plant mating systems. In: International Symposium on New Genetical Approaches
to Crop Improvement (Siddiqui KA, ed). Tandojam
Sind, Pakistan: Atomic Energy Agricultural Research
Centre.
Schemske DW and Lande R, 1985. The evolution of selffertilization and inbreeding depression in plants. II. Empirical observations. Evolution 39:41-52.
Schmidt WC and Reymond RC, 1990. Larix occidentalis.,
western larch. In: Silvics of North America: 1. Conifers
(Burns RM and Honkala BH, tech coords). USDA Forest
Service, Agriculture Handbook 654. Washington, D.C.:
U.S. Government Printing Office; 160-172.
Schoen DJ and Stewart SC, 1986. Variation in male reproductive investment and male reproductive success
in white spruce. Evolution 40:1109-1120.
Ellstrand NC, 1992. Gene flow among seed plant populations. New For 6:241-256.
Schoen DJ and Stewart SC, 1987. Variation in male fertilities and pairwise mating probabilities in Picea glauca. Genetics 116:141-152.
Ellstrand NC and Marshall DL, 1986. Patterns of multiple paternity in populations of Raphanus salivus. Evolution 40:837-842.
Shea KL, 1987. Effects of population structure and cone
production on outcrossing rates in Engelmann spruce
and subalpine fir. Evolution 41:124-136.
Erickson VJ and Adams WT, 1989. Mating success in a
coastal Douglas fir seed orchard as affected by distance and floral phenology. Can J For Res 19:1248-1255.
Sorensen FC, 1982. The roles of polyembryony and embryo viability in the genetic system of conifers. Evolution 36:725-733.
Farris MA and Mitton JB, 1984. Population density, outcrossing rate, and heterozygote superiority in ponderosa pine. Evolution 38:1151-1154.
Fins L and Sib LW, 1986. Genetic variation in allozymes
of western larch. Can J For Res 16:1013-1018.
Smith DM, 1962. The practice of silviculture. New York:
Wiley.
Franklin EC, 1970. Survey of mutant forms and inbreeding depression in species of the family Pinaceae. USDA
Forest Service Research Paper SE-61, Southeastern Forest Experiment Station. Washington, D.C.: U.S. Government Printing Office.
Worrall J, 1993. Temperature effects on bud-burst and
leaf-fall in subalpine larch. J Sustain For 1(2):1-18.
Wright S, 1922. Coefficients of inbreeding and relationship. Am Nat 56:330-338.
Received June 20, 1995
Accepted March 11, 1996
Corresponding Editor: Thomas Mitchell-Olds
El-Kassaby and Jaquish • Western Larch Population Density 4 4 3