Are trioecy and sexual liability in Atriplex canascens genetically based?: Evidence from clonal studies.
We address these alternate views by examining the extent to which the sexual phenotypes, fixed or labile, have a demonstrable genetic basis or may be attributed to mere sexual inconstancy. For this, we examine the occurrence and magnitude of sexual lability in clones of A. canescens grown in the controlled environmental conditions of common gardens and ask: 1) Is the ability to produce flowers of the opposite gender dependent on, or independent of genotype? 2) Is the proportion of alternate gender flowers produced dependent upon genotype?
Thus, we seek to determine whether the occurrence and proportion of alternate gender flowers in A. canescens is subject to, or independent of clonal sources. Where prior work in this area was based upon qualitative descriptions of sex change (McArthur, 1977; McArthur and Freeman, 1982; Freeman et al., 1984), we attempt here to examine the matter in quantitative fashion.
In one study, measurements of phenotypic sex expression were made on clones of A. canescens grown in a common garden with irrigation and control treatments. In both the treatment and the control environments, floral phenotypes of individuals were quantified repeatedly over time. This design allows us to distinguish the relative influence of genetics and environment on the ability to alter gender and on the magnitude of sex change. While the effects of genetics and environment are usually distinguished by the use of breeding experiments, such experiments would require more than a decade to elucidate the genetics of gender expression in this long lived species. A second study traces gender expression over a 20 year period in clonal and non-clonal plants of an A. canescens population not related to those of the first study.
Materials and Methods
Atriplex canescens (Chenopodiaceae) is a large wind pollinated shrub that inhabits the foothills, dry stream beds, and sandy habitats of the Great Basin and adjacent areas (Freeman and McArthur, 1989). The species forms a polyploid complex ([Chi] [bar] = 9) with tetraploid (4 [Chi]) populations being the most numerous and widespread (Stutz and Sanderson, 1979; Dunford, 1984). Diploid populations are strictly dioecous (Stutz et al., 1975; McArthur and Freeman, 1982) or nearly so (Barrow, 1987) whereas polyploid ([is greater than or equal to] 4 [Chi]) populations are subdioecious with substantitive numbers of hermaphroditic plants (McArthur and Freeman, 1982; Barrow, 1987). Flowers of A. canescens are small, only a few millimeters in diameter, and generally have reproductive organs of only one sex. Some individuals also bear a few perfect flowers. Fruit have been observed to develop from perfect flowers, but it is not known if the stamens are functional (Sanderson, unpubl. obs.). Individual plants were cloned using the methods of McArthur et al. (1984).
Rush Valley Common Garden Study
We cloned individuals of A. canescens from two tetraploid populations, one from Spanish Fork Canyon (SFC) in northern Utah and the other from Kingston Canyon (KC) in south central Utah. The resulting ramets were over-wintered in a greenhouse before being planted in a common garden located in Rush Valley, Utah, a mid-elevation cold salt desert shrub valley (1,585 m, 29 cm annual ppt.) of the Great Basin (Pendleton et al., 1992).
The SFC population is found on a steep slope; the sex of most individuals in this population was observed for three years prior to cloning. For the purposes of categorization, plants consistently displaying only the staminate phenotype were called "male" when cloned. "Females" were categorized in a similar manner, i.e., they had displayed only pistillate flowers. Plants showing any form of hermaphroditism (i.e., monoecy, sexual lability, or perfect flowers) were labeled as "hermaphrodites." Twelve male, 11 female, and 9 hermaphroditic plants were chosen at random from the 200 tagged plants of an earlier study (Freeman and McArthur, 1984b). A few hermaphroditic plants (those marked [Alpha]) came from plants not tagged in the earlier study (Freeman and McArthur, 1984b) because there were too few tagged plants of this category.
Individuals from the KC population had not been observed previously, and their gender was classified initially on the basis of the phenotype observed in 1981, when plants were cloned. Seventeen female, 10 male, and 4 hermaphroditic plants were cloned.
Several of the genets which had, prior to cloning, been preliminarily classified as either male or female produced ramets having flowers of the opposite gender. These plants were reclassified as hermaphroditic. This change in phenotype is not unexpected given the absence of a long flowering history on each individual. Ten clones from SFC and 6 from KC produced exclusively staminate flowers and are hereafter referred to as male clones. Similarly, 10 clones from SFC and 13 from KC produced exclusively pistillate flowers; these clones are referred to as female clones. Twelve clones each from SFC and KC produced both male and female flowers and are referred to as hermaphroditic clones.
Individual plants were cloned into 20 ramets each. Ten randomly chosen ramets of each plant were assigned to irrigation and control treatments. Ramets were transplanted into the common garden where all received supplemental water in 1982 to aid in their establishment. Thereafter, only the ramets in the irrigated treatment were watered receiving about five liters per plant per month in the growing period (May-August).
An early unexpected but low incidence of flowering occurred in 1982, the year of planting, and floral sex ratios were estimated that year from measurements of inflorescence lengths instead of direct flower counts. Because of differences in the method of identifying floral phenotype, observations and measurements from 1982 were not compared statistically with those taken later. In subsequent years, field estimation of gender was validated by laboratory examination (10 x dissecting microscope), of two inflorescences from all flowering ramets of hermaphroditic clones and from four flowering ramets of male and female clones. The gender of every flower on these inflorescences was recorded. In a preliminary survey we examined all of the flowers on five inflorescences of several clones identified as male or female based upon the field inspection. The results showed no variation in gender for male and female clones.
Chi square analyses and ANOVAs were used to determine if the production of flowers of the alternate gender differed among ramets of the same genotype within a treatment and year, and between both treatments and years (Pollard, 1977; SAS, 1985). ANOVAs were used to apportion environmental and genetic influences on flower production. The study was terminated in late 1984 by a plague of grasshoppers which consumed the garden (Hills and Davidson, 1984).
Long Term Clonal and Seedling Study
Atriplex canescens plants were established in common gardens at the Snow Field Station (SFS), Ephraim, Utah (1,700 m, 26.8 cm annual ppt., alluvial limestone derived clay), and the Upper Colorado Environmental Plant Center (UCEPC), Meeker, Colorado (1,980 m, 41.1 cm annual ppt., silt loam soil).
Snow Field Station Planting. -- This plantation, designated as U103p, a half-sib family originally from Rincon Blanco, New Mexico, was established in 1968 and consisted of 665 bushes (McArthur, 1977; McArthur et al., 1978; McArthur and Freeman, 1982). The plantation was removed in 1982 but clonal and seed materials were saved for future study. Plants derived from both clonal propagation and seed were established at UCEPC (below) and subsequently in 1985 reestablished at SFS as well (Appendix 1).
Upper Colorado Environmental Plant
Center Planting. -- This plantation consisted of a replicated block design with two blocks each of plants established from clones and seeds, respectively (Noller et al., 1984). The clones were generated from six constantly female and four constantly male bushes from the original SFS U103p plantation (see above). Panmictically produced seed from the initial SFS U103p plantation was used to produce the plants from seed. Each block included 253 ramets or seed derived plants for a total of 1,012 plants. Containerized rooted ramets and seedlings were established in their respective blocks in 1980. These plants began flowering in 1981. The 990 plants of Table 1 are those that flowered at least once (survival was over 90% for the plantation as a whole through the 1983 growing season). This planting was severely damaged by the cold temperatures of the 1983-1984 winter and was removed after the 1984 growing season. A new UCEPC planting has been established (Appendix 1). [TABULAR DATA 1 OMITTED]
Data Collection And Analysis.-- Data collection for SFS and UCEPC consisted of scoring gender, and measuring heights and crown diameters (McArthur et al., 1978; Freeman and McArthur, 1984b). Genders were scored by careful inspection of each plant during July or August when flower and fruit development is most evident. Plants that produced both male and female flowers were scored in three classes of hermaphroditism or monoecy (>60%, 40-60%, < 40% female flowers) for each year of observation. We thus scored six gender classes: 0 = no flowers; 1 = completely female; 2 = monoecious (hermaphroditic), >60% female; 3 = monoecious, 40-60% female; 4 = monoecious, <40% female; 5 = completely male. In addition, plant height and crown diameter were collected from three separate experiments (Appendix 2). Data were analyzed using ANOVA and t-test procedures (Pollard, 1977; SAS, 1985).
Rush Valley Common Garden Study
Sex Expression. -- The data on sex expression, survivorship, and flowering presented below are from Pendleton et al. (1992), and are based largely upon data collected in 1983. Before examining evidence of genetic control of gender expression, we present data on the patterns of gender expression, and gender variation, both within and among the clones. Specifically we ask the following:
Are different floral phenotypes manifest within a genotype?
For labile (or hermaphroditic) plants, the proportion of male and female flowers (floral phenotype) differed significantly among ramets of the same clone within a treatment. Of the clones that had at least 2 ramets in flower, 19 showed some manifestation of hermaphroditism. In 1983 14 of these 19 clones showed significant differences among the ramets in the proportion of male, female and perfect flowers (see Table 2 for examples). In 1984, 12 of the 19 clones flowered again, and 4 had ramets with statistically different floral sex ratios. [TABULAR DATA 2 OMITTED]
In the unirrigated (control) treatment, 17 clones showed some form of hermaphroditism. In 1983, the proportion of male, female, and perfect flowers differed significantly among ramets of the same clone in 10 of these 17 clones. In 1984, only 3 of the 17 clones from the unirrigated treatment survived and produced two or more flowering ramets. Of these clones, one had ramets with significantly different floral sex ratios. Taken all together, these observations indicate that "hermaphrodites" manifest variable sexual phenotypes in both treatments.
Does irrigation affect the sexual phenotype of hermaphroditic clones?
Chi square analyses show that significant variation exists in floral phenotypes of the 19 clones designated as hermaphroditic. These differences are manifest in total, year, treatment, and treatment x year analyses (Table 3). However, the effect of treatment is different among clones. For example, plants that increased their proportion of male flowers in the control treatment, averaged 27.4% more male flowers in that treatment than the irrigated treatment. Clones that had a higher proportion of male flowers in the irrigated treatment averaged 15.6% more male flowers in that treatment than in the unirrigated treatment. [TABULAR DATA 3 OMITTED]
Do individual ramets vary their gender from year to year?
Because of the infestation of grasshoppers, repeated observations of floral phenotype of individual ramets were severely limited (Table 4). In some cases, for example, ramet 1 of clone SFC45, the relative proportion of male and female flowers varied between years. In another example, ramets 2 and 3 of clone SFC75 changed from a bias favoring male (all male for ramet 2 and 94.4% male for ramet 3) flowers to all female flowers. Overall, of the 32 ramets for which such information was available, 6 showed clear differences in gender between years. The pooled data analyses of Table 3 also demonstrate year to year gender expression differences. [TABULAR DATA 4 OMITTED]
Does the floral sex ratio of a hermaphroditic clone vary between either treatments or years?
To answer this question the floral sex ratio (number of staminate flowers/number of pistillate flowers) for hermaphrodites was considered over a three year period (Table 5). The floral sex ratios of several clones differed by an order of magnitude. There are 17 instances where clones exhibited different floral sex ratios between treatments in the same year. In 11 of those, clones were more male biased in the control treatment, and in 6 cases they were more male biased in the irrigated treatment. There are 17 instances where the floral sex ratios of clones differed between years, for plants in the same treatment. In 1984, eight clones were male biased; nine clones were male biased in 1983. Twenty-three clones displayed female flowers alone, while 16 clones displayed exclusively male flowers during the period 1981-1984 (data not shown). Each clone had up to 20 ramets. [TABULAR DATA 5 OMITTED]
Is the ability to change sex genetically determined?
Now we return to the question of genetic control. If variation in sexual phenotype is solely environmental in origin, i.e., nongenetic phenotypic inconstancies, then the proportion of clones identified as being "all-male" should equate to that expected from a chance drawing of ramets from a population of all clones capable of producing male ramets. But if observed variations are genetically based then the observed frequency of clones of apparently fixed sex should exceed the expectation based on chance alone.
In doing this analysis, we ignore all prior gender designations, thus avoiding any circularity in logic. We use only the 1983 data and include only those clones with four flowering ramets and [is greater] 400 flowers. First, we identify all clones capable of producing at least one ramet with only male flowers. Then we examine these clones, and determine the total number of ramets from all such clones which produce only male flowers ([N.sub.1]), and the number of remaining ramets ([N.sub.2]). The probability of drawing an all-male clone (P.sub.m]) by chance from the total pool of ramets from all clones having one or more male ramets was determined by using binomial distribution (Pollard, 1977): [P.sub.m] = [([N.sub.1]/[N.sub.1] + [N.sub.2]).sup.4] = [(95/138).sup.4] = 0.225. We used the 4 power because that was the minimum number of ramets considered. This is a conservative treatment; there were up to 20 ramets per clone. The probability of forming an exclusively female clone ([P.sub.f]) was calculated similarly: [P.sub.f] = [(95/150).sup.4] 0.161.
Table 6 compares the number of "all-male" or "all-female" clones expected by chance with those observed. It shows clearly that the formation of an all-male or all-female clone occurs at a much greater frequency than expected by chance alone. The distribution of "sexually inconstant" ramets is not random among the clones, but is limited to certain clones, i.e., it occurs in significantly fewer clones than would be expected by chance. Thus, the ability to produce flowers of the alternate gender is confined to particular genotypes. [TABULAR DATA 6 OMITTED]
Is the magnitude of sexual lability under genetic control?
An ANOVA was used to determine the extent to which environmental versus genetic influences affect the production of alternate gender flowers (SAS, 1985; Table 7). For this, the dependent variables were the percentage of male and female flowers (arcsine transformed) of hermaphroditic clones in 1983. Because perfect flowers may constitute up to 10% of the flowers produced, the analysis was performed twice. First we included perfect flowers in the total flower production and in a second analysis they were excluded. In the ANOVA the ramet term was used as the error term in a model with unequal cell sizes allowing the use of data from all 19 hermaphroditic clones. [TABULAR DATA 7 OMITTED]
The mean squares computed for the various terms in the model are weighted sums of the variance components. In the model the mean square for treatment includes variance components for treatment, the treatment x clone interaction, and ramet (error). We have decomposed these mean squares into their variance components and then computed the percent of the variance accounted for by each component (Table 7).
The results of these analyses indicate the existence of strong genetic controls on the proportion and gender of flowers. Variance among clones accounted for between 44 and 62% of the total variance in the plants from the SFC population and between 90 and 92% of the variance for plants from the KC site. These, however, are undoubtedly overestimates as they are based upon data from only one year.
A significant interaction occurred between the treatment and clone terms for the SFC population. This difference may reflect differences in the sensitivity of various clones to changes in their environment. Differential sensitivities to environmental stress have been reported for different crop varieties and are genetically determined characteristics (Bulmer, 1980). Sensitivities of the various genotypes were calculated following Bulmer (1980). For SFC genets, the sensitivity coefficients for percent male flowers ranged from -0.01 to 4.19, and accounted for half of the interaction mean square (t = 22.61, P < 0.01). The sensitivity coefficients for the production of female flowers ranged from 0.1 to 4.75, again accounting for approximately half of the interaction variance (t = 6.97, P < 0.01).
Survivorship and Flowering. -- Ramets from male SFC clones in the irrigated treatment showed the lowest survivorship from 1981-1983 (90%) while ramets from SFC female clones had 98% survivorship; survivorship among SFC ramets from hermaphrodites was 96% in both the irrigated and nonirrigated treatment. Survivorship for ramets from KC ranged from 83% for hermaphrodites in the nonirrigated treatment to 94.6% for females in the nonirrigated treatment. The only genets with less than 90% ramet survivorship were those from clones of hermaphrodites from Kingston Canyon.
Few ramets survived the grasshoppers and flowered again in 1984. Where possible we report on these plants, but their numbers are not great. For example, in 1983, 17 clones in the nonirrigated treatment produced two or more flowering ramets. In 1984 only three of these clones had two or more ramets in flower.
Of the ramets which survived between 1982 and 1983, the majority also flowered in 1983. The percentages given below are based upon surviving ramets and not the total number of ramets planted. Male clones from SFC had the highest percentage of flowering (97%) in the irrigated treatment, and control treatment (95%). Female clones from SFC exhibited 93.8% flowering in the irrigated treatment and 76.3% in the control treatment. Clones of hermaphrodites from SFC had 85.7% flowering in both treatments. The percentage of flowering for clones from KC was less than those from SFC, ranging between 59% for females in the irrigated treatment to 89% for males in the irrigated treatment. Seventy-two percent of the ramets from female clones in the control treatment flowered, as did 73% and 81% of the ramets from hermaphroditic clones in the irrigated and control treatments, respectively. Eighty-seven percent of the ramets from male clones in the control treatment flowered.
Sex Expression in Long Term Clonal and Seedling Study
Genders of ramets from plants that had been scored as exclusively female or exclusively male over the life of the original U103p plantation were consistently and exclusively female or male in all common garden plantations (Table 1; Appendix 1). Table 1 presents the most complete data set (the initial UCEPC clonal plantation). The nearly 750 ramets derived from 11 female bushes (genets) produced only female and the more than 300 ramets derived from six male bushes produced only male flowers over 11 years of observation (1981-1991) in the three clonal plantations. On the other hand, both the UCEPC and SFS plantations derived from seed had plants that were 1) consistently and uniformly female, 2) consistently and uniformly male, and 3) monoecious or labile. Table 1 presents the data from the UCEPC plantation. There, of the 486 plants derived from seed that flowered (1,433 observations over 4 years), 208 (42.8%) displayed only the female state, 165 (34.0%) displayed only the male state, and the remaining 113 (23.2%) displayed states of monoecy and lability. Of the 113 monoecious seed derived plants, only 9 stayed in the same gender category for the four years of observation. All others varied among gender categories from year to year. Another nine plants varied from one extreme to the other with at least one representative in each of the monoecy gender states (2-4) were, during various years, exclusively male, exclusively female, and hermaphroditic.
These studies of A. canescens demonstrate that: 1) Clones transplanted from three different populations and grown in three common gardens reveal the existence of two distinctly different genetic controls that act to determine whether gender either varies or remains fixed as an all-male or all-female phenotype. In some clones gender expression is fixed as all-male or all-female floral phenotypes. In other clones, however, floral phenotype is seen to vary widely from those having all-male or all-female ramets to those with varying ratios of male and female flowers. Our studies (Table 1) suggest a single hermaphroditic genotype but do not exclude the possibility of multiple hermaphroditic states. The long term clonal and seedling study demonstrated that the all-male and all-female states are rigorously consistent in flowering characteristics. The flowering consistency has been observed in 11 all-female plants and 6 all-male plants continuously for up to 20 years. For five female and four male plants, these observations have been continuous from 1972-1991 and include the parental plants and two generations of clonal ramets, including hundreds of observations on millions of flowers (circa [10.sup.6] flowers per ramet per year) on individually marked plants. They collectively show no deviations from the pure male or pure female states. The data available suggest that the labile state also has distinctive characteristics. The three states differ in either individual characteristics or combinations thereof: flowers, flower production in young ramets, pollen grains per anther, seed production, juvenile growth rate, size and shape of mature plants, chlorophyll content, metabolite content, habitat of relative maximum fitness, and life expectancy (Appendix 2; McArthur, 1977; McArthur et al., 1978; Freeman et al., 1984; Tiedemann et al., 1987; Freeman and McArthur, 1989; Pendleton, 1990; Pendleton et al., 1992; Freeman, McArthur, Sanderson, and Tiedemann, unpubl. data; McArthur and Blauer, unpubl. data). 2) The ability to vary in floral phenotype is confined to specific genotypes. Variations in sexual phenotypes occur between the environmental treatments applied at the Rush Valley common garden, and especially within treatments and over time as a consequence of the combined effects of genotype plus environment.
We suggest that the labile or monoecious state became established in some populations of A. canescens because it has a relative adaptive fitness advantage in some habitats as also do both the staminate and pistillate genotypes (Freeman and McArthur, 1984a; Freeman, McArthur, Sanderson, and Tiedemann, unpubl. data). We believe these fitness differences in patchy environments coupled with polyploidy have allowed the species to exploit a wider range of environments. Other woody Atriplex species have variable frequencies of monoecy and dioecy in different populations in North America and Australia (Pope, 1976; Parr-Smith, 1977; Freeman and McArthur, 1984b). They may be playing variations of the same game. Other woody chenopods have demonstrated unusual breeding systems that may be adaptative (McArthur and Sanderson, 1984; Pendleton et al., 1988). The herbaceous chenopod, Spinacia oleracea, also has an unusual, probably adaptive breeding system (Smith, 1976; Vitale and Freeman, 1986). Other widely occurring and taxonomically disparate taxa demonstrate elements of the trioecious A. canescens system as we have described it. Examples are the polyploid Australian shrub Ptilotus obovatus, a polyploid complex including populations with distinctly different sex ratios (Kirby et al., 1987); the dioecious Iberian shrub Osyris quadripartita which exhibits dimorphism of morphological traits correlated with gender (Herrera, 1988); the Mediterranean Basin species Mercurialis annua with diploid dioecious populations and monoecious polyploid ones (Durand and Durand, 1985; Louis, 1989); and the Hawaiian endemic subshrub Schiedea globosa in which both environmental conditions and genetic constraints are involved in sexual lability (Sakai and Weller, 1991).
Sexual lability has been reported in plants for a century (Darwin, 1889; Atkinson, 1898; Schaffner, 1922, 1925, 1927; Heslop-Harrison, 1957; 1972; Dzhaparidize, 1967; McArthur, 1977; McArthur and Freeman, 1982; Freeman et al., 1980; Freeman and McArthur, 1984b; Lloyd and Bawa, 1984). The list of sexually labile species has grown and will likely continue to expand as the sexual phenotypes of marked individuals are scrutinized repeatedly. Such studies in the genus Atriplex have shown that up to 40% of the individuals alter their sexual phenotype McArthur and Freeman, 1982; Freeman and McArthur, 1984b).
The adaptive significance of gender modification has only been explored for a few species. Notable among these are Arisaema triphyllum and Atriplex canescens (Policansky, 1981; Freeman et al., 1984; Lovett-Doust et al., 1986 and references therein). Many seemingly dioecious species alter their sexual phenotype in response to a variety of conditions (see Freeman et al., 1980; Lloyd and Bawa, 1984), but, the importance and magnitude of such alterations have not been established, and remain, in most cases, a subject of discussion (Freeman et al., 1984; Lloyd and Bawa, 1984; Schlessman, 1986). The central issue of that polemic regards the nature of the effective controls of gender; are variations in floral phenotype independent of or dependent on genetic control? We posit the latter alternative, in which a genetic system for the determination of sex exists and the frequency of various genotypes therein may be determined by the action of selection.
Our argument would clearly be untenable if: (1) "all-male" and "all-female" individuals could not be established through breeding lines, or (2) sexual inconstancy occurred across subsequent generations in such lines, or (3) clones from hermaphrodites were shown to have fixed sex, and the frequency of hermaphrodites was unresponsive to selection. Thus, we posit an explanation for the control of plant gender that is consistent, in this instance, with both our data and current models common in classical genetics (McArthur, 1977; McArthur and Freeman, 1982; Barrow, 1987). In this regard, we note that pure breeding dioecious stocks of spinach (also a member of the Chenopodiaceae) have been developed, as has a pure breeding hermaphroditic stock (Bemis and Wilson, 1953; Thompson, 1956; Ramanna, 1976). Finally we concur with Barrow (1987) that all that is required to account for our observations, regarding these tetraploid plants, is for some variant "y" chromosomes (or linkage groups) to show incomplete dominance over their "x" chromosomes. This might occur either as a gene environment condition or as the consequence of a particular modifier.
To observe the dynamics of sexual phenotypes of A. canescens individuals, we use an approach here that measures the sexual phenotype of the same plant repeatedly. Had we not marked and repeatedly measured the same genotype, only TWO sexual morphs would have appeared to predominate with a few scattered individuals simultaneously showing appreciable numbers of flowers of both sexes. In the Rush Valley common garden in either year, up to 80% of the ramets from labile plants examined showed 5% or less apparent "alternate gender" flowers on a ramet. Although the actual number of individuals changing sexual phenotype was appreciable, in most instances the relative proportion of individuals simultaneously bearing large numbers of flowers of both sexes remained small. Clearly, the behavior of this sexual system is only evident from repeated observations of marked individuals. Furthermore, as evidenced by the continuity of sex expression in clones of males and females, and by the variance of that character in clones that are hermaphroditic, the ability to change sex is restricted to certain genotypes. The factor "genotype" (clone) accounted for from 41% to 93% of observable variation in sexual phenotype of hermaphrodites.
The use of clones allows us to address the issues of genetic control of sex expression in an expeditious manner, but the use of clones also has a potential limitation: If the initial environment experienced by a seedling predisposes it to have fixed or labile sex, then cuttings from that plant might also have fixed or labile sex. Chailakhyan (1979) has shown that gender in spinach, an annual member of the Chenopodiaceae, is determined at or about the time the plant has three leaves. Prior to that plants can be manipulated to produce either male or female flowers. However, there is no evidence that spinach plants can be environmentally induced to produce flowers of both genders or to exhibit labile expression. Nevertheless, in that species the initial expression of gender can be determined by the hormonal environment experienced early in life. We see several problems in extrapolating this argument from spinach to A. canescens. Spinach is an annual plant, and flowers for only a few weeks, after which the plants die. It is reasonable to suppose that the meristems of spinach plants do differentiate in advance of flowering, and that their pattern of differentiation is influenced by their hormonal environment as Chailakhyan has shown. Once the plant dies, so do the meristems. A. canescens is a perennial, and the meristems giving rise to the flower buds are produced annually. In order for the results from the spinach to falsify the results reported here for Atriplex, the spinach plants would have to flower a second season, and the gender expressed in the second season of the plant's life, must depend solely upon the environment experienced in the three leaf stage. No such evidence exists. Thus, we believe that the differences which occur between clones are due to genetic differences, while those occurring within a clone are due to environmental factors. We acknowledge, however, that our results do not entirely preclude the possibility that the environment experienced early in life affects the ability to change sex in all future years.
The extrapolation of these results to other systems is unwarranted without substantiating experimentation. The diverse methods and techniques by which studies on sexual lability have been conducted have engendered a spectrum of seemingly disparate observations in many such studies. While the hypothesis of sexual inconstancy may be relevant in some of those circumstances (Armstrong and Irvine, 1989), it is clearly in conflict with our results. Further studies on other species using repeated observations are essential for resolving the general extent to which sexual lability is under genetic control.
This research was supported, in part, by National Science Foundation Grant DEB 81-11010 (to D.C.F. and E.D.M.) and by the Pittman Robertson Wildlife Habitat Project W-82-R. Portions of the work were carried out at the Snow Field Station, the Upper Colorado Environmental Plant Center, and on an exclosure of the U.S. Department of Interior, Bureau of Land Management. These facilities are maintained, respectively, by the U.S. Department of Agriculture, Forest Service, Intermountain Research Station, the Utah Division of Wildlife Resources, Snow College, and Utah State University Agricultural Experiment Station; the White River and Douglas Creek Soil Conservation Districts and U.S. Department of Agriculture, Soil Conservation Service; and the Salt Lake District of the U.S. Department of the Interior, Bureau of Land Management. We thank A. C. Blauer, S. C. Garvin, G. L. Jorgensen, J. M. Massey, B. K. Pendleton, and R. L. Pendleton for assistance and synthesizing discussion and L. Floyd-Hanna, K. T. Harper, K. G. Wilson, and two anonymous reviewers for useful comments on earlier versions of the manuscript. H. G. Hilton provided statistical support.
Armstrong, J. E., and R. Irvine. 1989. Flowering, sex ratios, pollen ovule ratios, fruit set and reproductive effort of a dioecious tree, Myristica insipida (Myristicaceae) in two different rain forest communities. Am. J. Bot. 76:74-85. Atkison, G. G. 1898. Experiments on the morphology of Arisaema triphyllum. Bot. Gaz. 25:114. Barrow, J. R. 1987. The effects of chromosome number on sex expression in Atriplex canescens. Bot. Gaz. 148:379-385. Bemis, W. P., and G. B. Wilson. 1953. A new hypothesis explaining the genetics of sex determination in Spinacia oleracea L. J. Hered. 44:91-95. Bulmer, M. G. 1980. The Mathematical Theory of Quantitative Genetics. Claredon Press, Oxford, UK. Chailakhyan, M. Kh. 1979. Genetic and hormonal regulation of growth, flowering and sex expression in plants. Am. J. Bot. 66:717-736. Charnov, E. L. 1982. The Theory of sex Allocation. Princeton University Press, Princeton, NJ USA. Darwin, C. 1889. The Different Forms of Flowers on Plants of the Same Species. Murray, London, UK. Dunford, M. P. 1984. Cytotype distribution of Atriplex canescens (Chenopodiaceae) of southern New Mexico and adjacent Texas. Southwest. Natur. 29: 223-228. Durand, R., and B. Durand. 1985. Mercurialis, pp. 376-389. In A. H. Halevy (ed.), Handbook of Flowering, vol. 3. CRC Press, Boca Raton, FL USA. Dzhaparidize, L. I. 1967. Sex in plants. Part 1. Israel Programs for Scientific Translation, Jersualem, Israel. Freeman, D. C., K. T. Harper, and E. L. Charnov. 1980. Sex change in plants: Old and new observations and new hypotheses. Oecologia 47:222-232. Freeman, D. C., and E. D. McArthur. 1984a. Sex change in Atriplex canescens. Am. J. Bot. 71(5 pt. 2):65. --. 1984b. The relative influences of mortality, nonflowering, and sex change on the sex ratios of six Atriplex species. Bot. Gaz. 145:385-394. --. 1989. Atriplex canescens, pp. 75-86. In A. H. Halevy (ed.), Handbook of Flowering, vol 6. CRC Press, Boca Raton, FL USA. Freeman, D. C., E. D. McArthur, and K. T. Harper. 1984. The adaptive significance of sexual lability in plants using Atriplex canescens as a principal example. Ann. MO. Bot. Gard. 71:265-277. Hall, H. M., and F. E. Clements. 1923. The phylogenetic method in taxonomy, the North American species of Artemisia, Chrysothamnus, and Atriplex. Carnegie Inst. Washington Publ. 326:1-355. Herrera, C. M. 1988. Plant size, spacing patterns, and host-plant selection in Osyris quadripartita, a hemiparasitic dioecious shrub. J. Ecol. 76:995-1006. Heslop-Harrison, J. 1957. The experimental modification of sex expression in flowering plants. Biol. Rev. 32:38-90. --. 1972. Sexuality of the angiosperms,pp. 133-287. In F. C. Steward (ed.), Plant Physiology: A Treatise, Vol. 6C. Academic Press, N.Y., USA. Hills, B., and L. Davidson. 1984. Hefty, hungry hoppers gobble gardens, crops. Deseret News July 19:B1-B2. Kirby, G. C., B. A. Barlow, and S. Habel. 1987. Sex ratios in a gynodioecious plant Ptilotus obovatus (Gaudich.) F. Muell. (Amaranthaceae). Aust. J. Bot 35:679-687. Lloyd, D. G., and K. S. Bawa. 1984. Modification of the gender of seed plants in varying conditions. Evol. Biol. 17:255-338. Louis, J. P. 1989. Genes for the regulation of sex differentiation and male fertility in Mercurialis annua L. J. Hered. 80:104-111. Lovett-Doust, L., J. Lovett-Doust, and K. Turi. 1986. Fecundity and size relationships in Jack-in-the-Pulpit, Arisaema triphyllum (Araceac). Am. J. Bot. 73:489-494. McArthur, E. D. 1977. Environmentally induced changes of sex expression in Atriplex canescens. Heredity 38:97-103. McArthur, E. D., A. C. Blauber, and G. L. Noller. 1984. Propagation of fourwing saltbush (Atriplex canescens [Pursh] Nutt.) by stem cuttings, pp. 261-264. In A. R. Tiedemann, E. D. McArthur, H. C. Stutz, R. Stevens, and K. L. Johnson (eds.), Proceedings of the symposium on the biology of Atriplex and related chenopods. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, Ogden, UT USA. McArthur, E. D., and D. C. Freeman. 1982. Sex expression in Atriplex canescens: Genetics and environment. Bot. Gaz. 143:476-482. McArthur, E. D., A. P. Plummer, G. A. Van Epps, D. C. Freeman, and K. R. Jorgensen. 1978. Producing fourwing saltbush seed in seed orchards, pp. 406-410. In D. H. Hyder (ed.), Proceedings of the first international rangeland congress. Society for Range Management, Denver, CO USA. McArthur, E. D., and S. C. Sanderson. 1984. Distribution, systematics, and evolution of Chenopodiaceae: An over-view, pp. 14-24. In A. R. Tiedemann, E. D. McArthur, H. C. Stutz, R. Stevens, and K. L. Johnson (eds.), Proceedings of the symposium on the biology of Atriplex and related chenopods. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, Ogden, UT USA. Noller, G. L., S. E. Stranathan, and E. D. McArthur. 1984. Establishment and initial results from a |Rincon' fourwing saltbush [Atriplex canescens (Pursh) Nutt.] seed orchard, pp. 196-204. In A. R. Tiedemann, E. D. McArthur, H. C. Stutz, R. Stevens, and K. L. Johnson (eds.), Proceedings of the symposium on the biology of Atriplex and related chenopods. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station, Ogden, UT USA. Parr-Smith, G. A. 1977. Taxonomy of Atriplex vesicaria Heward ex Benth. (Chenopodiaceae) and related species. Ph.D. Diss. University Melbourne, Melbourne, Victoria Australia. Pendleton, B. K. 1990. Common garden studies of Atriplex canescens: Effect of genetics and environment on growth, reproduction and life history features. Ph.D. Diss. Wayne State University, Detroit, MI USA. Pendleton, B. K., D. C. Freeman, E. D. McArthur, and R. L. Pendleton. 1992. I. Life history features of three sexual morphs of Atriplex canescens clones grown in a common garden. Am. J. Bot. 79: 376-382. Pendleton, R. L., E. D. McArthur, D. C. Freeman, and A. C. Blauer. 1988. Heterodichogamy in Grayia brandegei (Chenopodiaceae): Report from a new family. Am. J. Bot. 75:267-274. Policansky, D. 1981. Sex choice and the size advantage model in Jack-in-the-Pulpit (Arisaema triphyllum). Proc. Natl. Acad. Sci. USA 78:1306-1308. Pollard, J. H. 1977. A Handbook of Numerical and Statistical Techniques. Cambridge University Press, Cambridge, UK. Pope, C. L. 1976. A phylogenetic study of the suffrutescent shrubs in the genus Atriplex. Ph.D. Diss. Brigham Young University, Provo, UT USA. Ramanna, M. S. 1976. Are there heteromorphic sex chromosomes in spinach (Spinacia oleracea L.)? Euphytica 25:277-284. Sakai, A. K., and S. G. Weller. 1991. Ecological aspects of sex expression in subdioccious Schiedea globosa (Caryophyllaceae). Am. J. Bot. 78:1280-1288. SAS. 1985. SAS/STAT Guide for Personal Computers, Version 6. SAS Institute, Cary, NC USA. Schaffner, J. H. 1922. Control of the sexual state in Arisaema triphyllum and A. dracontium. Am. J. Bot. 9:72-78. --. 1925. Experiments with various plants to produce change of sex in the individual. Bull. Torrey Bot. Club 52:35-47. --. 1927. Sex and sex determination in the light of observations and experiments on dioecious plants. Am. Nat. 61:319-332. Schlessman, M. 1986. Interpretation of evidence for gender choice in plants. Am. Nat. 128:416-420. --. 1988. Gender diphasy ("sex choice"), pp. 139-153. In J. Lovett-Doust and L. Lovett-Doust (eds.), Plant Reproductive Ecology. Oxford University Press, N.Y., USA. Smith, P. M. 1976. Minor crops, pp. 301-324. In N. W. Simmonds (ed.), Evolution of Crop Plants. Longman Group Limited, London, UK. Stutz, H. C., J. M. Melby, and G. K. Livingston. 1975. Evolutionary studies of Atriplex: A relic gigas diploid population of Atriplex canescens. Am. J. Bot. 62:236-245. Stutz, H. C., and S. C. Sanderson. 1979. The role of polyploidy in the evolution of Atriplex canescens, pp. 615-621. In J. R. Goodin and D. K. Northington (eds.), Arid Land Plant Resources. International Center for Arid and Semi-Arid Land Studies, Texas Tech University, Lubbock, TX USA. Thompson, A. E. 1956. Methods of producing first generation hybrid seed in spinach. Cornell Univ. Agric. Exp. Stn. Mem. 336:3-48. Tiedmann, A. R., E. D. McArthur, and D. C. Freeman. 1987. Variations in physiological metabolites and chlorophyll in sexual phenotypes of |Rincon' fourwing saltbush. J. Range Manage. 40:151-155. Vitale, J. J., and D. C. Freeman. 1986. Partial niche separation in Spinacia oleracea L.: An examination of reproductive allocation. Evolution 40:426-430.
Current Common Gardens
There are, currently, common gardens at both UCEPC and SFS. The clonal portion of UCEPC was replaced in 1984 by new rooted cuttings from the original but winter damaged UCEPC plantation. For this new plantation, ramets from five of the original female clones (the sixth clone performed poorly at UCEPC after 1982) and five consistently female plants from the seeded blocks at UCEPC were established. Male ramets came from the four original male clones and ramets from two consistently male plants from the seeded blocks. The new (current) clonal plantation consists of 264 ramets from 10 female and 6 male clones.
The new (established 1985) clonal plantation at SFS consists of 225 ramets derived from the same 10 female and 6 male genets that were used in the second clonal plantations at UCEPC. Nine of these genets had shown their respective constant floral phenotypes for the 11 years (1972-1982) of record in the original SFS plantation and the 4 years of record in the first UCEPC plantation. The other genets also displayed constant sexual expression but over a shorter time period. The seedling block at SFS, consisting of over 100 plants, was established (1985) from panmictically produced seed from the original clonal blocks at UCEPC.
Through 1991 all female and male ramets at the SFS and UCEPC have continued their absolute fidelity to "pure" male and female genotypes. These data are impressive when one considers that each A. canescens bush in these gardens a verages about 1,000,000 flowers (McArthur and Blauer, unpubl. data).
At SFS 103 plants produced from seed from the UCEPC clonal blocks were living in 1989. Of these, 30 were female, 36 were male, 30 were monoecious (14, monoecious class 2; 9, monoecious class 3; 7, monoecious class 4), and 7 did not flower.
Plant Height and Crown Diameters from [female], [male], and [[female, male]]S
Three experiments compare female, male, and labile plant growth characteristics. In the first experiment, 16 plants from the U 103p plantation were cloned. Twenty ramets each of 7 consistent female, 4 consistent male, and 5 labile plants were established in 15 cm pots and allowed to gorw in the greenhouse. The other two experiments were based on growth rates of the clonal ramets and seed plants used in the replicated block experiment at UCEPC. Data were collected from the first experiment (height only) prior to transplanting to field plots not otherwise reported here. Data from the other two experiments were collected each August after the plants were transplanted from 1980-1984.
Heights of the ramets from the potted plants in the growth comparison experiment were highly significantly different for the female genets (x [bar] = 13.2 [+ or -] 0.5 cm SE) in comparison to the male and labile (monoecious) genets (23.8 [+ or -] 0.9 and 22.6 [+ or -] 0.9, respectively) by analysis of variance and Student-Newman-Keuls multiple range comparison analyses (df = 319, F = 71.8, P < 0.001). Appendix Table 2A and 2B present growth rate data for the clonal ramets and seed-produced plants from 1980 UCEPC plantation. The female and male ramets of the clonal blocks were highly significantly different in height (1980), crown (1983), and significantly different in height to crown ratio (1983) at various times (Appendix Table 2A). Height and crown values for the seedlings reveal significant gender state differences (genders determined on a composite basis, 1981-1984, see Table 1) for height and crown in 1982 and 1983 (appendix Table 2B). When only the female and male plants are compared as was the case for the remets (Appendix Table 2A) then there are significant differences for height (1980, 1982) and height to crown ratio (1982), and highly siginificant differences for crowns (1982, 1983). [TABULAR DATA 2A OMITTED] [TABULAR DATA 2B OMITTED]
|Printer friendly Cite/link Email Feedback|
|Author:||McArthur, E. Durant; Freeman, D. Carl; Luckinbill, Leo S.; Sanderson, Stewart C.; Noller, Gary L.|
|Date:||Dec 1, 1992|
|Previous Article:||Temperature determination of male outcrossing ability in a simultaneous hermaphrodite.|
|Next Article:||Sex ratio selection and the evolution of environmental sex determination in laboratory populations of Menidia menidia.|