Printer Friendly

The effect of inbreeding in diploid and tetraploid population of Epilobium angustifolium (Onagraceau): implications for the genetic basis of inbreeding depression.

Inbreeding depression is the reduced fitness of inbred offspring relative to outbred offspring, where inbred individuals are a product of matings between relatives. The phenomenon has attracted the attention of biologists for over a century (Darwin 1876). Currently, interest in inbreeding depression is widespread because of its role in the evolution of reproductive systems in natural populations (Schemske 1983; Schoen 1983; Kalisz 1989; Dudash 1990; Holtsford and Ellstrand 1990; Johnston 1992; Agren and Schemske 1993; Dole and Ritland 1993; Eckert and Barrett 1994; Latta and Ritland 1994; Husband and Schemske 1995; Norman et al. 1995; Sakai et al. 1997; Husband and Schemske 1996 and references therein) and its potential importance in conservation biology (Allendorf and Leary 1986; Barrett and Kohn 1991). Despite the ecological and evolutionary implications, the genetic basis and the causes of variation in the magnitude of inbreeding depression in natural populations are not completely understood.

There are two prevailing genetic models of inbreeding depression: partial dominance and overdominance (Charlesworth and Charlesworth 1987). Under the partial dominance model, individuals possess a number of deleterious mutations that are recessive or partially recessive; inbreeding depression is caused by the increased frequency of recessive homozygotes due to inbreeding. In the overdominance model, heterozygotes are superior to homozygotes at loci affecting fitness, and inbreeding depression is caused by a loss in allelic interactions due to reduced heterozygosity. While experimental studies of individual genes have provided some support for the partial dominance model (Wright et al. 1942; Klekowski 1976; Willis 1992), distinguishing the two genetic models is often difficult because of the large number of loci controlling inbreeding depression. Additional information regarding mutation rates as well as their degree of dominance and magnitude of effects is required.

One issue that has received little attention from empiricists is the effect of chromosome doubling on inbreeding depression (but see Johnston and Schoen 1996). In the partial dominance model, the prediction is quite clear. When recessive alleles are lethal or sublethal, the equilibrium inbreeding depression of a tetraploid should be nearly half that of the diploid progenitor (Lande and Schemske 1985). Hedrick (1987) made a similar prediction for homosporous ferns. This expectation can best be understood in terms of the equilibrium frequency of recessive deleterious mutations in diploids and polyploids and the rate at which homozygosity rises after one generation of selfing. For a given mating system, the frequency of deleterious mutations will be higher in tetraploids than diploids. This would appear to enhance inbreeding depression in the tetraploid. However, after one generation of selfing, full homozygosity at a diploid heterozygous locus increases by 50%, while in tetraploids with tetrasomic inheritance, full homozygosity (aaaa, AAAA) increases more slowly (17-21%, depending on the pattern of segregation) (Haldane 1930; Wright 1938; Parsons 1959). The rise in homozygosity is slowed further and inbreeding depression thereby reduced in tetraploids because the frequency of double heterozygotes (aaAA) carrying mutations, from which most full homozygotes segregate, will be less common than heterozygotes in diploid populations (Lande and Schemkse 1985). Lande and Schemske (1985) showed that the difference in inbreeding depression between diploids and tetraploids should hold for both autopolyploids and allopolyploids when mutations are recessive and lethal or sublethal. However, for alleles that are partially recessive, they predicted no differences between diploids and tetraploids in inbreeding depression.

The theoretical effects of chromosome doubling in the overdominance model are less clear, primarily because of the uncertainty regarding the allelic interactions in polyploids. Nevertheless, theoretical studies by Busbice and Wilsie (1966) and Bennett (1976) suggest that inbreeding depression in tetraploids may actually be larger than their diploid progenitors. This is possible because tetraploid genotypes may have higher-order allele interactions in addition to two-allele interactions, and the rate at which the number of allelic interactions decreases upon selfing may be higher in tetraploids. Therefore, if inbreeding depression is caused by a decrease in heterotic interactions among alleles, fitness may actually fall more rapidly upon selfing a tetraploid than a diploid (Bever and Felber 1992).

The effects of chromosome doubling on the magnitude of inbreeding depression will have important implications for the evolution of polyploidy, which is extremely prevalent (estimated 30-47% of species) in flowering plants (Stebbins 1971; Grant 1981). Current models of the evolutionary dynamics of polyploids indicate that the circumstances favoring a polyploid in a diploid population, or favoring mixed cytotypes, are quite restrictive but will depend, in part, on the relative differences in fertility and viability between cytotypes (Felber 1991; Rodriguez 1996). Thus, differences in inbreeding depression among cytotypes may influence opportunities for the evolution of polyploidy, assuming both cytotypes have similar selfing rates. Furthermore, differences in inbreeding depression may alter the outcome of selection on self-fertilizing genotypes and thereby affect the direction and rate of mating system evolution in related cytotypes.

In this study we tested the specific predictions of the partial dominance model by comparing the magnitude of inbreeding depression in diploid and tetraploid populations of Epilobium angustifolium (Onagraceae). E. angustifolium is a perennial herbaceous plant distributed throughout the northern hemisphere. In North America it exists as diploid and tetraploid cytotypes, which coexist in the Beartooth range of the Rocky Mountains in Wyoming and Montana (Mosquin 1967). Because equilibrium inbreeding depression will depend on the mating system, we first estimated the selfing rate of a tetraploid population from the Beartooth Mountains, to compare it to an earlier estimate obtained from a diploid population in this region (Husband and Schemske 1995). Then, in a greenhouse experiment, we compared the magnitude of inbreeding depression in two diploid populations to that in three tetraploid populations of E. angustifolium. We were also interested in knowing whether the differences in inbreeding depression between the cytotypes were consistent across life stages. A previous survey of the literature (Husband and Schemske 1996) found that inbreeding depression in predominantly selfing taxa was less than that in outcrossed taxa for early life stages but not for later ones. One hypothesis is that inbreeding depression expressed during early life stages involves mutations of major effect, which can be purged through inbreeding, while inbreeding depression late in life is a product of mutations of small effect that are difficult to purge. If this is true, differences in inbreeding depression among cytotypes should be largest during early life stages and weak or non-existent in later stages.



Epilobium angustifolium seed was collected from two diploid and three tetraploid populations along the Beartooth Pass highway (Hwy 212), which crosses the Beartooth Mountains from southern Montana into northern Wyoming. Flint (1980) showed that both diploid and tetraploid cytotypes of E. angustifolium occurred in this region, with diploids predominantly at higher altitudes. Of the sites we sampled, one tetraploid population (T26, site numbers are from Flint 1980) occurred on the southern end of the mountain pass, in Wyoming, and two tetraploid (T2, T4) and two diploid populations (D2, D6) occurred on the northern end, in Montana. The ploidy for each population was initially classified according to Flint (1980). We later confirmed the distribution of cytotypes with a survey of DNA content using flow cytometry (Husband and Schemske, unpubl. data).

Mating System

The mating system for one of the diploid populations of E. angustifolium, D2, was reported previously (Husband and Schemske 1995) as a measured selfing rate of [r.sub.m] = 0.06 and a primary selfing rate of r = 0.45 (see below). Since the mating system for tetraploid populations in the mixed cytotype region has not been characterized, the rate of self-fertilization ([r.sub.m]) was estimated for one population, T2, using allozyme markers and the mixed-mating model MLTET developed by Ritland (1989).

Before estimating the mating system, we conducted several controlled pollinations between known genotypes and examined the segregation of genotypes in their progeny to determine the inheritance of allozymes in tetraploid E. angustifolium. This step was necessary since the model for estimating mating systems assumes that allozymes exhibit tetrasomic inheritance without double reduction (i.e., chromosomes in a gamete can originate from any random combination of homologous chromosomes but it is impossible for sister chromatids to be in the same gamete). Offspring from three different mating combinations (mating pairs by genotype: ssmm x ssmm; ssss x ssmm; ssss x smmm) that involved five different parents were collected and their genotypes scored using electrophoresis. Each offspring was scored at the seedling stage at a single polymorphic locus using horizontal starch gel electrophoresis. Approximately 0.5 [cm.sup.2] of leaf material was ground in a 0.1M Tris buffer (pH 7.5) containing 0.02M sodium bisulfite and 0.5M sucrose; the homogenate was absorbed onto 3 x 6 mm paper chromatography wicks and inserted into a 12% starch gel and a current was run at 50 mA and 325 V for six hours (Husband and Schemske 1995). The enzyme phosphoglucose isomerase (PGI) was resolved on a Tris-borate (pH = 8) buffer system (Soltis et al. 1983). Genotypes of the offspring from the three crosses were compared using a goodness of fit G test to genotype ratios expected under three different genetic models: tetrasomic inheritance with no double reduction; tetrasomic inheritance with double reduction; and two-gene disomic inheritance. The first two models correspond to patterns of inheritance when there are four homologous chromosomes and gametes originate from random chomosome and random chromatid segregation, respectively. The two-gene model depicts segregation when the four homologous chromosomes behave like two nonhomologous pairs.

The selfing rate was estimated using the mixed-mating model MLTET of Ritland (1989). This estimation model is based on the assumption that progeny genotypes are a product of either self-fertilization or random outcrossing, that mating system is uniform among maternal parents, and that alleles exhibit tetrasomic inheritance without double reduction. Open-pollinated seed was collected from 46 different plants in population T2 in September 1991. Seeds were germinated and seedlings were planted into individual pots and grown to the four to 10 leaf stage. At least eight seedlings from each seed family were scored at the Pgi-2 locus using horizontal starch gel electrophoresis as described above. The selfing rate estimate is based on a maximum-likelihood procedure which infers the genotype of the maternal parents, the frequency of alleles in the pollen pool (p) and the proportion of progeny per family that were derived from outcrossing (t). Standard errors for each selfing rate estimate were generated by bootstrapping 100 times among the 46 progeny arrays.

As with the analysis of selfing rate in a diploid population (Husband and Schemske 1995) the primary selfing rate (r), defined as the selfing rate at the time of fertilization rather than at the seedling stage, was estimated by adjusting the measured selfing rate ([r.sub.m] = 1 - t) for any inbreeding depression that may have occurred before progeny were analyzed, i.e., at seed-set and germination using the formula (Maki 1993),

r = [r.sub.m]/[1 - [Delta] + [r.sub.m] [Delta]] (1)

where [Delta] is the inbreeding depression measured at seed maturation and germination (see below).

Inbreeding Depression

Population- and family-level inbreeding depression for diploid and tetraploid populations was estimated by generating and comparing selfed and outcrossed offspring in the greenhouse. Fifteen plants, one offspring from each of 15 families of open-pollinated seed, were grown and used as maternal plants for each of the five populations (in total, n = 75 maternal plants). For each plant, four flowers were self-pollinated (selfed flowers, hereafter), four flowers were cross-pollinated (outcrossed flowers, hereafter) and two were left unpollinated to test for any inadvertent pollen transfer (control). All flowers including controls were emasculated prior to pollination and the equivalent of one anther's worth of pollen was applied evenly across each of the four stigma lobes. For outcrossed flowers, one pollen donor was selected at random from the remaining plants in that population for each stigma lobe. Four anthers per stigma were sufficient to ensure that all ovules could be pollinated. Only two flowers per plant were ever pollinated on a single day. Since anthesis within an inflorescence as well as on different plants is not completely synchronous, pollinations were conducted over three weeks by pollinating a selfed and outcrossed pair of flowers on each plant every second day, when possible. If the number of potential pollen donors dropped below eight on any given day, no flowers were pollinated. Fruit were collected as they began to dehisce, about two to three weeks after pollination.

The effect of inbreeding was estimated at four life stages: seed maturation, seed germination, survival to nine weeks, and dry mass at nine weeks. Mean fitness for selfed and outcrossed offspring in each population at each life stage was estimated as the mean of all families. The number of seeds was estimated for each of the eight fruit per maternal plant (control fruit had no seed) as the number of filled seeds, expressed as a proportion of the total number of fertilized ovules. Filled seeds were plump, oval shaped, and with light brown colon Untilled seeds were classified into two groups, those that had expanded but were collapsed and those that had not expanded. While the last category may consist of some unfertilized ovules, the non-expanded seeds were relatively uncommon ([less than] 10%) in all populations examined and their frequencies did not differ between the two cytotypes.

Germination was estimated by placing 20 seeds from each of four selfed and four outcrossed fruit per plant into a separate petri dish. In some cases less than 20 seeds were available. When one fruit was missing, seed was collected at random from the remaining fruit on that plant and placed in a separate petri dish. After two weeks at 4 [degrees] C, dishes were placed in an incubator (12 h day/night; 25 [degrees] C day/21 [degrees] C night). The number of germinating seeds were counted and then removed weekly, starting one week after sowing, until germination ceased.

Seedling survival and dry mass were assessed in a greenhouse experiment. Nine days after sowing the seeds in petri dishes, five outcrossed and five selfed seedlings from each family in each of the five populations were each transplanted into separate half-gallon tubs. To allow for damage and mortality as a result of transplanting, any seedlings that died within the first week of transplanting (59 out of 710 seedlings) were replaced. While diploids experienced higher mortality in the first week than tetraploids, there were no differences in early mortality between selfed and outcrossed progeny ([[Chi].sup.2] = 0.83, df = 1, P [greater than] 0.35, n = 59) and no difference between cytotypes in the effects of selfing on early mortality ([[Chi].sup.2] = 0.64, df = 2, P [greater than] 0.40, n = 59). Therefore, early replacement likely had no influence on the final results. Selfed and outcrossed plants from all populations were then randomly arranged in an array. Extra seedlings, drawn randomly from dishes of bulk self and outcrossed seeds from both ploidy levels, were placed as a border around the outside of the array to reduce edge effect. The plants were then allowed to grow for nine weeks, at which time at least 80% of all plants in each cytotype had flowered. Plants were harvested, dried to a constant weight at 80 [degrees] C, and weighed.

The effects of chromosome number and pollination treatment, and differences among populations and families within ploidy, with respect to seed-set, germination, survival, dry mass at nine weeks, and cumulative fitness were examined using a two-factor split-plot ANOVA with nesting. The sources of variation for analyses of seed-set, germination, and drymass were Ploidy, Population[ploidy] {random}, Parent[population,ploidy]{random}, Pollination, Pollination x Ploidy, Pollination x Population[ploidy]{random}, and Pollination x Parent[population, ploidy]{random}. Since survival and cumulative fitness were estimated at the family level, the Pollination x Parent[population,ploidy] could not be estimated for these variables. An arcsine-square root transformation of seed-set, germination, and survival data, and a log transformation of cumulative fitness was used to improve normality of residuals and reduce heteroscedasticity. Denominator mean squares were synthesized and degrees of freedom for significance testing were determined using Satterthwaite approximations (SAS Institute 1994). In the ANOVA, it is the Pollination x Ploidy interaction that is most important for indicating whether the effects of selfing differ among cytotypes. The significance of this interaction term was determined using log-tranformed data (applied after any original transformation), since this is necessary to assess the relative rather than absolute differences between selfing and outcrossing (Johnston and Schoen 1994). All statistical analyses were conducted on a Power Macintosh 6100/60 computer, using JMP 3.0.1 software (SAS Institute 1994).

Inbreeding depression, the reduced fitness of progeny derived from selfing relative to outcrossing, was expressed as 8, which is equal to 1 - [w.sub.s]/[w.sub.o], where [w.sub.s] and [w.sub.o] are the mean fitness of selfed and outcrossed progeny, respectively. To estimate inbreeding depression at each life stage for a population, fitness measures were family means. For lifetime measures of inbreeding depression, [w.sub.s] and [w.sub.o] were the product of family mean fitnesses at each life-stage. The measure of inbreeding depression, [Delta], ranges from [infinity], when the relative fitness of outcross progeny approaches zero, to one, when the relative fitness of selfs is zero.


Mating System of Tetraploids

Inheritance of the Pgi-2 isozyme locus was based on electrophoretic analyses of progeny from three kinds of crosses: ssmm x ssmm (n = 105), ssss x ssmm (n = 61) and ssss x smmm (n = 35) (Table 1). Balanced and unbalanced heterozygotes were detected in each cross and the number of progeny with heterozygous and homozygous genotypes was reported for each (Table 1). The segregation ratios uniformly preclude monogenic-disomic inheritance and were compared to expectations under tetrasomic inheritance and digenic-disomic inheritance. Segregation frequencies deviated most from the expectations based on disomic inheritance; these deviations were significant for one of the three types of crosses. The allozyme segregation patterns were consistent with tetrasomic inheritance, although sample sizes were not sufficient to distinguish the two modes of tetrasomic segregation: random chromosomal (no double reduction) and random chromatid (complete double reduction).

The estimate of the population selfing rate ([r.sub.m]) in the tetraploid population T2, based on one polymorphic locus (Pgi-2 allele frequencies: 0.075/0.202/0.730) and a sample of 499 offspring from 46 maternal plants, was 0.28 (SE = 0.06). [TABULAR DATA FOR TABLE 1 OMITTED] This selfing rate was significantly different from zero (t = 4.6, df = 44, P [less than] 0.001). The primary selfing rate (r), calculated by correcting for differential seed production and germination of selfed and outcrossed offspring (see Methods), was 0.43. The parental two gene inbreeding coefficient (F) based on the single polymorphic locus was -0.301. Such a low value is unusual and may be caused by an exceptionally high number of heterozygotes in the sample, errors during scoring, and biases in the way the MLTET program infers the genotype of maternal parents (K. Ritland, pers. comm.). The latter two explanations were explored (#3, using data from controlled crosses) but could not account for the unusually low value of F. The pollen two-gene fixation index was 0.327 (SE = 0.083).

Inbreeding Depression

Of the 75 plants pollinated (15 per population), two each from tetraploid populations T2 and T26, did not set seed after either self- or cross-pollination. Because this apparent sterility may have occurred for reasons other than mutational load, these individuals were excluded from all analyses. Seed counts were not available for one of the 15 plants pollinated in population D6 because of mishandling, but it was represented in the germination trial and greenhouse comparison. This family was excluded for all analyses of cumulative inbreeding depression.

Of the 70 maternal plants for which seed-set was estimated, the mean percent filled seeds for selfed and outcrossed flowers ranged from 0% to 94% and only three maternal plants (1 in population T4, 2 in population T26) exhibited higher mean seed-set upon self-fertilization than cross-fertilization. On average, outcrossed flowers produced more seed than selfed flowers (Tables 2, 3); however, the effects of pollination treatment were heterogeneous among ploidy levels as well as populations within ploidy levels and maternal parents [TABULAR DATA FOR TABLE 2 OMITTED] within populations (Table 2). In diploids, selfed flowers produced 84% less seed than outcrossed flowers, while in tetraploids the selfed flowers produced 41% less seed than outcrossed flowers. Mean inbreeding depression for seed-set in diploids was twice that in tetraploids (diploid [Delta] = 0.84, tetraploid [Delta] = 0.41) [ILLUSTRATION FOR FIGURE 1 OMITTED].

Germination occurred in both selfed and outcrossed offspring from all 71 families, although the proportion germinating varied from 0.15 to 1.0. Overall, the effect of pollination treatment on germination was highly significant (Table 2). In all but six families (1 in D2, 1 in D6, 1 in T4, 3 in T26), germination frequencies of outcrossed offspring were greater than selfed offspring. A nonsignificant Ploidy x Pollination interaction indicated that the negative impact of self-fertilization in diploids was not different from that in tetraploids. Mean inbreeding depression at germination was [Delta] = 0.22 and [Delta] = 0.11 for diploids and tetraploids, respectively [ILLUSTRATION FOR FIGURE 1 OMITTED]. The effects of self-fertilization did vary significantly among families within populations but not among populations (Table 2).

In the greenhouse experiment, 78 of the original 710 plants died before the harvest at nine weeks. Overall, mortality of selfed progeny exceeded that of outcrossed progeny in 63 of 71 families (outcrosses exceeded selfs in 8 families: 1 in D2, 1 in T2, 4 in T4, 2 in T26), a significant difference between self and outcross treatments. The effects of pollination treatment on mortality, however, did not differ between ploidy levels, between populations within ploidy levels, nor between maternal parents within populations (Tables 2,3). Mean inbreeding depression for survival to nine weeks was [Delta] = 0.19 for diploids and [Delta] = 0.05 for tetraploids [ILLUSTRATION FOR FIGURE 1 OMITTED].

After nine weeks, 82% of all diploids and 91% of all tetraploids transplanted had flowered. Flowering first occurred as early as 48 days after the germination trials began and, [TABULAR DATA FOR TABLE 3 OMITTED] on average, diploids started flowering earlier than tetraploids, regardless of pollination treatment. Diploids flowered an average of 58 days after the germination trials, which was six days earlier than the mean for tetraploids [ILLUSTRATION FOR FIGURE 2 OMITTED]. For both cytotypes, selfed progeny flowered later and were significantly less likely to flower than outcrossed progeny (outcrossed and selfed plants flowering, n = 370 and 247, respectively; [[Chi].sup.2] = 24.5, P [less than] 0.0001). In diploid populations, 61% of plants that flowered were outcrossed, while 59% of flowering tetraploid plants were outcrossed. The proportion of outcrossed versus selfed plants flowering in diploids was not significantly different from that in tetraploids (G = 0.25, P [greater than] 0.50; [ILLUSTRATION FOR FIGURE 2 OMITTED]). In addition, it appeared that selfed offspring were more likely to exhibit flower abnormalities and pollen sterility, although this was not quantified.

Mean dry mass in 71 selfed and 71 outcrossed families was 15.6 g and ranged from 1.7 to 32.6 g. Overall, tetraploids had a higher mass than diploids (x tetraploid mass = 19.6 and diploid mass = 9.5 g; Tables 2,3). Outcrossed progeny had a significantly larger mass than selfed progeny; selfs had a greater mean mass than outcrossed progeny in only 5 of 71 families examined (1 in D2, 2 in D6, 1 in T4, 1 in T26). In contrast to seed-set, the effect of pollination treatment on mass did not differ between diploids and tetraploids (mean [Delta], tetraploids = 0.35, diploids = 0.47, [ILLUSTRATION FOR FIGURE 1 OMITTED]), nor did the effect of pollination differ among populations within ploidy levels or among families within populations.

Cumulative fitness, the product of fitness at seed-set, germination, survival, and dry mass after nine weeks, ranged from 0.02 to 25.15 among selfed and outcrossed maternal families (self-fertilization [Mathematical Expression Omitted] vs. outcross [Mathematical Expression Omitted], F = 130.7, P [less than] 0.01) [ILLUSTRATION FOR FIGURE 3 OMITTED]. Mean cumulative fitness differed significantly among ploidy levels and pollination treatments. For the 70 maternal families examined, mean cumulative fitness of outcrossed progeny was always greater than in selfed progeny and the ratio of cumulative fitness in selfed and outcrossed offspring ranged from 0.002 to 0.980. The effect of pollination treatment, however, differed between diploids and tetraploids (Pollination x Ploidy interaction, F = 18.4 P [less than] 0.05). Mean inbreeding depression in diploid populations was [Delta] = 0.95, which is 1.4 times higher than the mean for tetraploid populations ([Delta] = 0.68). Significant heterogeneity in the effects of selfing was also observed among populations within ploidy levels.


This study represents the first comparison of mating system and inbreeding depression between sympatric natural populations of diploid and tetraploid cytotypes. We found that tetraploid E. angustifolium had a primary selfing rate of r = 0.43, which was similar to an estimate from a geographically separate tetraploid population (r = 0.53; Parker et al. 1995), and not significantly different from the estimate for a diploid population in the Beartooth Mountains (Husband and Schemske 1995). However, based on the measured selfing rate, adult populations are essentially randomly outcrossing due to the fact that most inbred offspring die before reproducing. Self-fertilization had a significant and negative effect on mean fitness at seed-set, germination, survival, and dry mass at nine weeks as well as on cumulative fitness in all populations (Table 2). The magnitude of the pollination effect, however, differed widely between diploid and tetraploid populations [ILLUSTRATION FOR FIGURE 1 OMITTED]. All three tetraploid populations had less cumulative inbreeding depression (29% less, on average) than the diploid populations. This difference between diploid and tetraploid cytotypes in population inbreeding depression is consistent with the predictions of the partial dominance model of inbreeding depression (Lande and Schemske 1985).

Lower inbreeding depression in tetraploids may be explained not only by partial dominance in tetraploids but also by historical events which reduced genetic load in tetraploids. For example, if tetraploids are the result of a single chromosome doubling event, inbreeding depression may be lower in tetraploids because most load was eliminated during the bottleneck. We cannot exclude this possibility entirely, although the present population sizes of tetraploids are so large that it is very likely they have reached an evolutionary equilibrium for inbreeding depression. In addition, isozyme surveys in other tetraploid species indicate that variability is higher in autopolyploids than their diploid progenitors, which supports the idea that tetraploids have persisted long enough to acquire variability or that bottlenecks associated with the origin of tetraploids are not that severe (Soltis and Rieseberg 1986; Wolf et al. 1990).

Other empirical studies providing data on the relationship between inbreeding depression and ploidy give mixed results; some show a decrease in inbreeding depression with ploidy, while others show the converse. For example, Hedrick (1987) showed that variation in inbreeding depression among ferns was consistent with the prediction that inbreeding depression decreases with ploidy, and Belaoussoff and Shore (1995) suggested that relatively low inbreeding depression in nine populations of Turnera ulmifolia may be attributed to it being an allohexaploid. Similarly, inbreeding depression in synthesized tetraploids of wheatgrass (Dewey 1969), maize (Alexander 1960) and clover (Davies 1961; Townsend and Remmenga 1968) were lower than their respective diploid progenitors. In contrast, Johnston and Schoen (1996) found that inbreeding depression in the highly self-fertilizing tetraploid Amsinckia gloriosa was higher than that in selfing populations of the diploid A. spectabilis. Also, studies on crop plants that are naturally polyploid show larger effects of inbreeding in polyploids. Tysdal et al. (1942) found that with repeated generations of selfing in alfalfa, plant fertility and yield declined faster than one would predict based on theoretical changes in homozygosity (Bingham 1980). Dewey (1966) found higher levels of inbreeding depression in polyploid wheatgrass relative to diploids, as did Kalton et al. (1952) for orchardgrass. Unfortunately, it is extremely difficult to interpret the data on polyploid crops because these studies often involved small numbers of maternal plants and measurements at only a small number of life stages (usually yield, measured as biomass or seed production after open pollination). Nevertheless, Busbice and Wilsie (1966) argued that such elevated levels of inbreeding depression in natural polyploids may be caused by the loss of allelic interactions, or overdominance, rather than partial dominance.

The theoretical effects of selfing in diploids and tetraploids under the overdominance model have been investigated, but these models do not include differential selection among genotypes and hence may not be directly comparable to natural populations. The results suggest, however, that inbreeding depression should be greater in tetraploids (Bever and Felber 1992). With overdominance, inbreeding depression is affected by the rate at which heterozygosity declines with inbreeding; however, there are four classes of heterozygous genotypes possible in tetraploids (aaaA, aaAA, aaAB, aABC), compared to only one in diploids (Aa). Busbice and Wilsie (1966) have shown that the consequences of multiple heterozygotes in tetraploids are that heterozygosity can involve higher order interactions as well as two-allele interactions, the number of allele combinations falls rapidly as partial heterozygotes are formed during inbreeding, and allele combinations will fall as rapidly or more rapidly than full homozygosity rises. Therefore, if inbreeding depression is caused by a reduction in heterotic interactions among alleles, fitness may actually fall more rapidly upon selfing a tetraploid than a diploid. Bennett (1976) showed that this rapid loss of heterozygosity applies not only to loci with four distinct alleles but also to those with only two alleles. While these patterns suggest a distinct difference between diploids and tetraploids, further information about the effect of selection against genotypes and about the relationship between allele dosages and fitness in tetraploids is necessary to predict inbreeding depression in natural populations.

The partial dominance theory of inbreeding depression predicts that diploids and tetraploids should differ with respect to the magnitude of inbreeding depression when mutations are completely recessive and lethal or sublethal (Lande and Schemske 1985). No differences should exist, however, when mutations are partially recessive or of small effect. One would predict, therefore, that the differences in inbreeding depression between diploids and tetraploids would be constant across all life stages if the genetic basis of inbreeding depression is comparable at different stages. In this study, differences in inbreeding depression between diploid and tetraploid E. angustifolium were greatest for seed production and there were no significant differences for germination, survival, and dry mass at nine weeks. This variation in the relative magnitude of diploid and tetraploid inbreeding depression may have two possible explanations. First, variation in the timing of inbreeding depression between cytotypes may arise if diploid and tetraploid populations of E. angustifolium experience substantially different stage-specific selection and, therefore, differ in the rate and extent to which mutations are purged. We attempted to minimize these differences by sampling populations of each cytotype from throughout the mixed cytotype zone (Beartooth Pass, Montana/Wyoming); however, tetraploid populations did occur at lower elevations and thus may have been associated with different environments and selection regimes. The second explanation is that inbreeding depression in later stages may have a different genetic basis than in early stages. The specific pattern of variation observed in this study would occur if inbreeding depression expressed during reproduction was the result of partially recessive mutations, each of small effect. This explanation was suggested independently in a survey of the magnitude and timing of inbreeding depression in plants (Husband and Schemske 1996), which showed that inbreeding depression in outcrossed and selfed species differed most in early life stages. These results, along with the findings from this study, suggest that mutations expressed during late life history stages may have a different genetic basis and selective effect, and are thus affected differently by self-fertilization.

Inbreeding depression is an important selective force acting against the evolution of breeding systems that promote inbreeding (Lande and Schemske 1985). Therefore, if inbreeding depression is reduced by chromosome doubling, as the partial dominance theory predicts, selfing should evolve more readily in tetraploids than diploids. Higher rates of self-fertilization in polyploids are also expected because selfing would increase the likelihood of establishment in a diploid population (Stebbins 1971; Rodriguez 1996). A polyploid that is self-fertilizing is more likely to spread in a diploid population because most matings will result in fertile polyploid offspring; outcrossing polyploids are more likely to produce inviable or sterile triploid offspring from 2x-4x matings. Few mating system estimates are available for polyploids (Murawski et al. 1994; Parker et al. 1995). The prediction of higher rates of self-fertilization in tetrapioids conflicts with our result that tetraploid E. angustifolium has a selfing rate similar to diploids, and also runs counter to the observation that all known naturally occurring autopolyploid angiosperms are outcrossing (Stebbins 1957; MacKey 1970) and that there are no examples of successful, selfing autopolyploid crop plants (Bingham 1980). However, mating system estimates from homosporous ferns indicate that polyploids are often highly setting relative to diploid races (Masuyama and Watano 1990; Watano and Masuyama 1991). More information regarding the joint evolution of mating system and inbreeding depression in polyploids is required to understand differences in selfing rates associated with ploidy.

The observed differences in inbreeding depression between diploid and tetraploid E. angustifolium may have important implications for the evolution of polyploidy in natural plant populations. Theoretical studies by Levin (1975), Felber (1991), and Rodriguez (1996) indicate that the establishment of a polyploid in a diploid population is unlikely, but its probability would be enhanced if the fitness of polyploids greatly exceeds that of diploids. Characteristics of tetraploid E. angustifolium that may confer a fitness advantage over diploids include lower inbreeding depression, large plant mass, and high ovule production/flower (Table 3). While the latter two attributes were not our primary focus, our study suggests that lower inbreeding depression in tetraploids may be particularly significant for the evolution of polyploidy, and would be adaptively significant in E. angustifolium, given the high rates of self-fertilization. Given that at least 47% of all plants species are polyploid (Grant 1981), additional comparisons of inbreeding depression in diploid and tetraploid populations would contribute greatly to our understanding of the significance of inbreeding depression for the evolution of polyploidy and mating systems in plants.


We thank M. Conrath, S. Curtis, D. Ewing, D. Lello, K. Ritland, J. Ross, and A. Sutton for technical assistance; and T Burton, D. Charlesworth, R. Lande, S. Schultz, and J. Willis for comments on the manuscript. This study was supported by a Royalty Research Fund grant to DWS, and a postdoctoral fellowship and an operating grant from the Natural Sciences and Engineering Research Council of Canada to BCH.


AGREN, J., AND D. W. SCHEMSKE. 1993. Outcrossing rate and inbreeding depression in two annual monoecious herbs, Begonia hirsuta and B. semiovata. Evolution 47:125-135.

ALEXANDER, D. E. 1960. Performance of genetically induced corn tetraploids. Am. Seed Trade Assoc. 15:68-74.

ALLENDORF, F. W., AND R. F. LEARY. 1986. Heterozygosity and fitness in natural populations of animals. Pp. 57-76 in M. E. Soule, ed. Conservation biology: the science of scarcity and diversity. Sinauer, Sunderland, MA.

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. Oxford Univ. Press, New York.

BELAOUSSOFF, S., AND J. S. SHORE. 1995. Floral correlates and fitness consequences of mating-system variation in Turnera ulmifolia. Evolution 49:545-556.

BENNETT, J. H. 1976. Expectations for inbreeding depression on self-fertilization of tetraploids. Biometrics 32:449-452.

BEVER, J. D., AND E FELBER. 1992. The theoretical population genetics of autopolyploidy. Pp. 185-217 in J. Antonovics and D. Futuyma, eds. Oxford surveys in evolutionary biology. Vol. 8. Oxford Univ. Press, New York.

BINGHAM, E. T 1980. Maximizing heterozygosity in autopolyploids. Pp. 269-273 in W. H. Lewis, ed. Polyploid, biological relevance. Plenum Press, New York.

BUSBICE, T. H., AND C. P. WILSIE. 1966. Inbreeding depression and heterosis in autotetraploids with application to Medicago sativa. Euphytica 15:52-67.

CHARLESWORTH, D., AND B. CHARLESWORTH. 1987. Inbreeding depression and its evolutionary consequences. Annu. Rev. Ecology Syst. 18:237-268.

DARWIN, C. R. 1876. The effects of cross and self-fertilization in the vegetable kingdom. John Murray, London.

DAVIES, W. E. 1961. Herbage legume breeding: investigations on polyploid clovers. Report for the Welsh Plant Breeding Station, pp. 40-44.

DEWEY, D. R. 1966. Inbreeding depression in diploid, tetraploid, and hexaploid crested wheatgrass. Crop Sci. 6:144-147.

-----. 1969. Inbreeding depression in diploid and induced-autotetraploid crested wheatgrass. Crop Sci. 9:592-595.

DOLE, J., AND K. RITLAND. 1993. Inbreeding depression in two Mimulus taxa measured by multigenerational changes in the inbreeding coefficient. Evolution 47:361-373.

DUDASH, M. R. 1990. Relative fitness of selfed and outcrossed progeny in a self-compatible, protandrous species, Sabatia angularis L. (Gentianaceae): a comparison in three environments. Evolution 44:1129-1139.

ECKERT, C. G., AND S.C. H. BARRETT. 1994. Inbreeding depression in partially self-fertilizing Decodon verticillatus (Lythraceae): population-genetic and experimental analyses. Evolution 48: 952-964

FELBER, E 1991. Establishment of a tetraploid cytotype in a diploid population: effect of relative fitness of the cytotypes. J. Evol. Biol. 4:195-207.

FLINT, E. P. 1980. Ecology and distribution of diploid and tetraploid Epilobium angustifolium (fireweed) in the Beartooth Mountains of Wyoming and Montana. Ph.D. diss. Duke Univ., Durham, NC.

GRANT, V. 1981. Plant speciation. 2d ed. Columbia Univ. Press, New York.

HALDANE, J. B. S. 1930. Theoretical genetics of autopolyploids. J. Genet. 22:359-372.

HEDRICK, P. W. 1987. Genetic load and the mating system in homosporous ferns. Evolution 41:1282-1289.

HOLTSFORD, T. P., AND N. C. ELLSTRAND. 1990. Inbreeding effects in Clarkia tembloriensis (Onagraceae) populations with different natural outcrossing rates. Evolution 44:2301-2046.

HUSBAND, B.C., AND D. W. SCHEMSKE. 1995. Magnitude and timing of inbreeding depression in a diploid population of Epilobium angustifolium (Onagraceae). Heredity 75:206-215.

-----. 1996. Evolution of the magnitude and timing of inbreeding depression in plants. Evolution 50:54-70

JOHNSTON, M. O. 1992. Effects of cross and self-fertilization on progeny fitness in Lobelia cardinalis and L. siphilitica. Evolution 46:688-702.

JOHNSTON, M. O., AND D. J. SCHOEN. 1994. On the measurement of inbreeding depression. Evolution 48:1735-1741.

-----. 1996. Correlated evolution of self-fertilization and inbreeding depression: an experimental study of nine populations of Amsinckia (Boraginaceae) Evolution 50:1478-1491.

KALISZ, S. 1989. Fitness consequences of mating system, seed weight, and emergence date in a winter annual, Collinsia verna. Evolution 43:1263-1272.

KALTON, R. R., A. G. SMIT, AND R. C. LEFFEL. 1952. Parent-inbred progeny relationships of selected orchardgrass clones. Agron. J. 44:481-486.

KLEKOWSKI, E. J. 1976. Genetics of recessive lethality in the fern, Osmunda regalis. J. Hered. 67:146-148.

LANDE, R., AND D. W. SCHEMSKE. 1985. The evolution of self-fertilization and inbreeding depression in plants. I. Genetic models. Evolution 39:24-40.

LANDE, R., SCHEMSKE, D. W., AND S. T SCHULTZ 1994. Maintenance of high inbreeding depression under partial self-fertilization, and the threshold selfing rate for purging recessive lethal mutations. Evolution 48:965-978.

LATTA, R., AND K. RITLAND. 1994. The relationship between inbreeding depression and prior inbreeding among populations of four Mimulus taxa. Evolution 48:806-817.

LEVIN, D. A. 1975. Minority cytotype exclusion in local plant populations. Taxon 24:35-43.

MACKEY, J. 1970. Significance of mating systems for chromosomes and gametes in polyploids. Hereditas 66:165-176.

MAKI, M. 1993. Outcrossing and fecundity advantage of females in gynodioecious Chionographis japonica var. kirohimensis (Liliaceae). Am. J. Bot. 80:629-634.

MASUYAMA, S., AND Y. WATANO. 1990. Trends for inbreeding in polyploid pteridophytes. Plant Species Biol. 5:13-17.

MOSQUIN, T 1967. Evidence for autopolyploidy in Epilobium angustifolium (Onagraceae). Evolution 21:713-719.

MURAWSKI, D. A., T. H. FLEMING, K. RITLAND, AND J. L. HAMRICK. 1994. Mating system of Pachycereus pringlei: an autotetraploid cactus. Heredity 72:86-94.

NORMAN, J. K., A. K. SAKAI, S. G. WELLER, AND T. E. DAWSON. 1995. Inbreeding depression in morphological and physiological traits of Schiedea lydgatei (Caryophyllaceae) in two environments. Evolution 49:297-306.

PARKER, I. M., R. NAKAMURA, AND D. W. SCHEMSKE. 1995. Reproductive allocation and the fitness consequences of selfing in two sympatric species of Epilobium (Onagraceae) with contrasting mating systems. Am. J. Bot. 82:1007-1016.

PARSONS, PA. 1959. Some problems in inbreeding and random mating in tetrasomics. Agron. J. 51:465-467

RITLAND, K. 1989. A series of FORTRAN computer programs for estimating plant mating systems. J. Hered. 81:235-237

RODRIGUEZ, D. J. 1996. A model for the establishment of polyploidy in plants. Am. Nat. 147:33-46.

SAKAI, A. K., S. G. WELLER, M.-L. CHEN, S. -Y. CHOU, AND C. TASANONT. 1997. Evolution of gynodioecy and maintenance of females: the role of inbreeding depression, outcrossing rates, and resource allocation in Schiedea adamantis (Caryophyllaceae). Evolution. In press.

SAS INSTITUTE. 1994. Statistical software for the Macintosh. JMP, statistics and graphics guide. Vers. 3. SAS Institute, Inc., Cary, NC.

SCHEMSKE, D. W. 1983. Breeding system and habitat effects on fitness components in three Neotropical Costus (Zingiberaceae). Evolution 37:523-539.

SCHOEN, D. J. 1983. Relative fitness of selfed and outcrossed progeny in Gilia achillieifolia (Polemoniaceae). Evolution 37:292-301.

SOLTIS, D. E., AND L. H. RIESEBERG. 1986. Autopolyploidy in Tolmiea menziesii (Saxifragaceae): genetic insights from enzyme electrophoresis. Am. J. Bot. 73:310-318.

SOLTIS, D. E., C. H. HAUFLER, D.C. DARROW, AND G. J. GASTONY. 1983. Starch gel electrophoresis of ferns: a compilation of grinding buffers, gel and electrode buffers and staining schedules. Am. Fern J. 73:9-27.

STEBBINS, G. L. 1957. Self-fertilization and population variability in the higher plants. Am. Nat. 91:337-354.

-----. 1971. Chromosomal evolution in higher plants. Edward Arnold, London.

TOWNSEND, C. E., AND E. E. REMMENGA. 1968. Inbreeding in tetraploid alsike clover, Trifolium hybridum L. Crop Sci. 8:213217.

TYSDAL, H. M., J. A. KIESSELBACH, AND H. L. WESTOVER. 1942. Alfalfa breeding. Nebr. Agric. Exp. Stat. Bull. 124:46.

WATANO, Y., AND S. MASUYAMA. 1991. Inbreeding in natural populations of the annual polyploid fern Ceratopteris thalictroides (Parkeriaceae). Syst. Bot. 16:705-714.

WILLIS, J. H. 1992. Genetic analysis of inbreeding depression caused by chlorophyll-deficient lethals in Mimulus guttatus. Heredity 69:562-572.

WOLF, P. G., D. E. SOLTIS, AND P.S. SOLTIS. 1990. Chloroplast DNA and allozymic variation in diploid and autotetraploid Heuchera grossulariifolia (Saxifragaceae). Am. J. Bot. 77:232-244.

WRIGHT, S. 1938. The distribution of gene frequencies in populations of polyploids. Proc. Nat. Acad. Sci. USA 24:372-377.

WRIGHT, S., T. DOBZHANSKY, AND W. HOVANITZ. 1942. Genetics of natural populations. VII. The allelism of lethals in the third chromosome of Drosophila pseudoobscura. Genetics 27:363394.
COPYRIGHT 1997 Society for the Study of Evolution
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1997 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Husband, Brian C.; Schemske, Douglas W.
Date:Jun 1, 1997
Previous Article:Evolution of gynodioecy and maintenance of females: the role of inbreeding deprssion, outcrossing rates, and resource allocation in Schiedea...
Next Article:How discrete are oak species? Insights from a hybrid zone between Quercus grisea and Quercus gambelii.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters