Sex Allocation in Ragweed (Ambrosia artemisifolia)
|
by: Sebastian Molnar
Abstract Sex allocation has important implications for the evolution of plant breeding systems. One prediction from sex allocation was tested through studying the effect of plant size on the sex ratio in natural populations of Ambrosia artemisifolia (common ragweed). Larger plants are expected to allocate more resources to female functions than to male functions. Therefore, larger plants should favor a female-biased sex ratio, whereas smaller plants should favor a male-biased sex ratio. ANOVA regression analyses were carried out on two measures of plant size (i.e. plant height and branch number) as the independent variables, with the sex ratio as the dependent variable. In the seven ragweed populations sampled, only one showed strong support for the expected predictions, while two other populations had weaker support. No relationship between plant size and sex ratio was found in the remaining populations. Overall, the above-mentioned prediction from sex allocation is weakly supported in Ambrosia artemisifolia. Introduction Sex allocation is a general theory that includes three areas of study: sex ratio evolution in dioecious species, sex change in sequential hermaphrodites, and resource allocation to male versus female functions in simultaneous hermaphrodites (Charnov, 1986; Tuljipurkar, 1990). Ultimately, the study of sex allocation deals with the partitioning of resources to particular male and female structures (or functions), and how that partitioning influences reproductive strategies. The formation of structures that are involved in reproduction require some level of resource allocation (Charlesworth and Morgan, 1991), and this has typically been measured as biomass or as nutrient content (Doust and Cavers, 1982; Campbell, 1992). The majority of flowering plants are hermaphrodites (Dellaporta and Calderon-Urrea, 1993). Hermaphrodites (including monoecious plants) must allocate resources to both male and female functions in the same plant -- whether the same resources are limiting (or not) to both male and female functions in a species needs to be determined empirically (e.g. Doust and Cavers, 1982). In dioecious species, since male and female functions exist in separate plants, resources are allocated to one particular sex function. Sex determination mechanisms may be associated with resource allocation. Certain genes (and in a few plant species, sex chromosomes) may function to control sex determination and therefore provide heritable variation on which selection may act (Leigh Jr. et al., 1985). Thus, studying sex allocation is important in understanding the evolution of plant breeding systems. Several mathematical models have been developed for sex allocation theory (e.g. Charnov, 1985, 1986; Leigh Jr. et al., 1985; Lloyd, 1985; Proulx, 2000; Taylor, 1985; Tulijapurkar, 1990) and have been reviewed by Charlesworth and Morgan (1991). Certain assumptions present in theoretical models, such as fixed limited resources and ‘trade-offs’ or ‘compensation’ between male and female functions, may have little direct supporting evidence (Savolainen et al., 1993). However, some investigations do suggest that such processes can take place (Korpelainen, 1992; Wolfe and Shmida, 1997). One difficulty in certain plant species -- especially those with perfect flowers -- is in assigning gender to particular structures (such as the corolla and the calyx; Ritland and Ritland, 1989) and therefore quantitative analyses of total resource allocation to a particular sex function would be inaccurate and incomplete. Dioecious and monoecious species are probably easier to work with in these types of analyses and may be used as model systems with which to compare to other systems. The focus of this study is on one prediction derived from sex allocation theory, which was tested in common ragweed, a monoecious plant species (or a ‘simultaneous hermaphrodite’). The effects of two measures of plant size on the male-female sex ratio were investigated. In this paper, larger relative sex ratio values are ‘male-biased’, while smaller relative sex ratios are ‘female-biased’. Smaller plants are expected to have a relatively more male-biased sex ratio than larger plants. Thus, smaller plants will also have larger values of the relative sex ratio. Conversely, large plants are expected to have a female-biased sex ratio (or a smaller relative sex ratio). This is a preliminary step in studying sex allocation in ragweed. Materials and Methods Ambrosia artemisifolia (Asteraceae) is an annual, wind-pollinated monoecious species. Clusters of male flowers develop at the terminal regions of branches (i.e. raceme-type inflorescences), while female flowers are located below the male flower clusters, usually near the basal portions of branches. Seven populations of ragweed found adajcent to parking lots and found on roadsides around the York University campus (4700 Keele St., Toronto, ON.) were sampled late in September and early October of the year 2000. Whole plants were uprooted and brought to the laboratory for the following measurements. Height was measured in centimeters for the above ground portion of each plant, from the stem-root junction (or the base of the primary stem) to the apex of the highest branch. The criterion for ‘branch’ was a ‘stem that had a terminal cluster of male flowers.’ The total number of branches was counted for small plants. For larger plants, total branch number was estimated by averaging the number of tertiary branches on at least three secondary branches, and then multiplying the average by the total number of secondary branches. Similarly, male and female flowers were either counted completely for small plants, or were estimated per branch (as above) for larger plants. Height and total branch number were used as indicators for resources available to the plants. Sex ratio (total male flowers divided by total female flowers) was calculated per plant and used as the dependent variable in regression ANOVAs, with height and total branch number as independent variables. Three regression ANOVAs were carried out for each sample, differing only in the following independent variables (i.e. ‘sex ratio’ was the dependent variable used in all cases): 1) height separate, 2) branch number separate, and 3) both height and branch number together. All statistical analyses were performed using SPSS 10.0. Results Descriptive statistics for samples taken from seven ragweed populations are provided in Table-1. Results from regression ANOVAs are shown in Table-2. Descriptions of each sample and of the statistical analyses used are given below.
TABLE-1: Sample Sizes and Means of Ragweed Populations. Means include their associated standard errors.
TABLE-2: Regression ANOVA F-values for Ragweed Samples. All ANOVA analyses used the male/female sex ratio as the dependent variable. Independent variables used are indicated below. Numbers in parentheses show the significance of the F-values.
Sample 1 was taken from a roadside population located on the west side of York University campus on The Pond Road. This was an area that had probably been mowed during the growing season -- most plants taken from this area were rather small in size (height ranged from 11.5 to 23 cm; branch numbers ranged from 2 to 18) and several had cut stems. Sample 1 appeared to be the only sample in which the population had been mowed before the data were collected. It is conceivable that mowed plants would confound analyses dealing the effects of plant size on sex ratio. For example, although the above ground plant size had been reduced, the below ground plant size probably remained the same after mowing. The root system (left intact after mowing) could contain resources accumulated over the growing season. Presumably, this would favor a female-biased sex ratio. Comparing the population means of sex ratios indicated that this was not the case (Table-1). The sample mean from population 1 had the second highest male-biased sex ratio at 4.83. Also, most plants in this sample had a sex ratio above 2.00 (ranging between 1.42 to 12.93). Scatterplots of height vs. sex ratio and of branch number vs. sex ratio are shown in Fig-1A and –1B. ANOVA regression analyses did not reveal any dependency relationships of plant size on the sex ratio for this sample (Table-2). Sample 2 was taken from a small population located on a plot of land adjacent to the Steacie Library parking lot. These plants were much larger in size compared to plants from population 1 (Table-1). Height ranged from 23 to 67 cm, and branch number ranged from 6 to 28 cm, in sample 2. The sex ratio ranged from 1.01 to 2.82. The mean sex ratio for this sample was the smallest, equaling 1.82 (Table-1). Scatterplots of height vs. sex ratio and of branch number vs. sex ratio are shown in Fig-2A and –2B. Whereas regression ANOVAs with branch number alone, and with branch number and height together as independent variables did not reveal any relationship, the null hypothesis rejected when height was used as the independent variable. That is, the slope of the linear regression did not equal zero, and height appears to have an effect on the sex ratio for this sample (Table-2). Therefore, this appears to support the notion that larger plants will exhibit relatively more “femaleness” than smaller plants. On the other hand, this sample had the smallest size (N=5) and therefore had relatively weaker statistical power than the other samples. At the very least, this sample provides weak support for the above-mentioned notion. Samples 3, 4 and 5 were all taken from separate populations (isolated by several meters) located on “grassy islands” in a York University parking lot. Many plants from this site were apparently dying off at the time samples were taken (indicating either poor growing conditions or that the plants were old). Heights ranged from 13 to 29 cm, 11 to 31 cm, and 19 to 44 cm for samples 3, 4, and 5, respectively. Branch number ranged from 1 to 19, 2 to 12, and 3 to 13, for samples 3, 4, and 5, respectively. Sex ratios ranged from 0.76 to 4.67, 1.12 to 4.27, and 1.40 to 6.50 for samples 3, 4, and 5 respectively. The mean values for each of these parameters are given in Table-1. Scatterplots of height vs. sex ratio and of branch number vs. sex ratio are shown in Fig-3A and –3B. Regression ANOVAs for samples 3 and 4 did not reveal any effect of plant size on the sex ratio (Table-2). All three regression ANOVAs for sample 5, however, indicates that plant size did have an effect on the sex ratio (Table-2). Sample 6 was taken from a population located adjacent to the newly developed Seneca@York parking lot on the south-east side of the York University campus. Plants from this site were much larger with respect to spread (and were healthier in appearance) than plants from other sites, with branch number ranging from 4 to 67 branches and with mean branch number at 28.90 (Table-1). Inter- and intra-specific competition was probably low at this site, as plants were well spaced-apart and vegetation was sparse. Plant height was modest, ranging from 18 to 53 cm and averaging 29.04 cm (Table-1). The sex ratio in this sample ranged from 0.37 to 8.05, and the mean sex ratio was 2.41 (Table-1). Scatterplots of height vs. sex ratio and of branch number vs. sex ratio are shown in Fig-4A and –4B. Only the regression ANOVA with branch number as the independent variable rejected the null hypothesis for this sample, suggesting that plant size has an effect on sex ratio (Table-2). Sample 7 was taken from a roadside population located directly across from the York University information center near The Commons. Plant height averaged at 32.36 cm and ranged between 16.0 and 51.0 cm. Mean branch number was 13.26, and ranged from 3 to 38 branches. The sex ratio in this sample was from 1.75 to 29.35, and averaged 6.43. Scatterplots of height vs. sex ratio and of branch number vs. sex ratio are shown in Fig-4A and –4B. ANOVA regressions did not reveal any effect of plant size on the sex ratio (Table-2). Discussion The number of male flowers versus female flowers was used as an indicator of sex functioning in individual plants. This method is potentially an inaccurate account of the true sex ratio. For example, some flowers might develop into a male flower, but not produce any pollen. Also, the amount of pollen produced far exceeds the amount of ovules that are produced in Ambrosia artemisifolia (Crane, 1986). Therefore, pollen and ovule abundance estimates are probably more accurate indicators of sex allocation (Ritland and Ritland, 1989). In the current study however -- since Ambrosia is a monoecious species, therefore counting male and female flowers is a relatively simple task -- the presence of a male or female flower was assumed to count towards the sex ratio. In the majority of individual plants, the sex ratio (male/female) was found to be greater than 1.0 across all Ambrosia artemisifolia populations studied. The lowest mean sex ratio was 1.82 in population 2 (Table-1). Ambrosia is considered to be an outcrossing wind-pollinated species (Goldman and Willson, 1986), and outcrossers are expected to exhibit a higher relative investment towards male function (Doust and Cavers, 1982). The results obtained for A. artemisifolia appear to support this notion. Male function may be favored in wind-pollinated species since large amounts of pollen is probably lost and therefore they need to produce large amounts of pollen in order increase their chances of fertilizing ovules (Doust and Cavers, 1982). The main focus of this study was to test the hypothesis that larger plants have relatively higher female functioning than smaller plants. Regression ANOVAs were used to analyse values of size and sex ratio. With height as the independent variable and sex ratio as the dependent variable, populations 2 and 5 revealed a relationship between size and sex ratio, whereas the other populations showed no relationship (i.e. the slope was not statistically different from zero in those cases). When branch number was used as the independent variable, populations 5 and 6 had a relationship between size and sex ratio, while no relationship was found in the other populations. When both height and branch number were used together as independent variables, only population 5 had a dependent relationship of sex ratio on size. Branch number is probably a better measure for size (and therefore resource availability) than plant height, and several different measurements of size should be used collectively in this sort of analysis. That being the case, only population 5 shows statistical support for the effect of plant size on sex ratio. Overall, the effect of large size to enhance female functioning in Ambrosia artemisifolia is weakly supported. Several factors may influence plant size. For example, population 2 was located in an area with shrubs and slight shade -- the majority of these plants were taller than plants from other samples (therefore these plants may be growing tall due to competition for sunlight). Soil quality was probably also higher quality than at the other sites, since this was a landscaped site with some maintenance during the year. In contrast, population 6 was located in a well-lit area adjacent to a newly developed parking-lot -- vegetation was sparse and the soil quality was clay-like. These plants could therefore spread out more (i.e. develop more branches) than at the other sites where plants tended to be very close together (e.g. populations 1 and 7). Thus, intra- and inter-specific competition plays a role in plant size. There may be other factors (e.g. size-independent effects) that are involved in generating the observed patterns of sex ratios in ragweed. In dioecious species of Rumex, Korpelainen (1992) found that size-independent effects can have a greater influence on reproductive effort than size. Further research into sex allocation in ragweed would include measurements of biomass allocation and of nutrient content to particular sex functions at particular life stages (pollen or ovule production versus seed production). Genetic controls of sex determination and of developmental patterns would also need to be identified and compared to biomass allocation and sex ratios for a more thorough understanding of the processes involved in sex allocation in ragweed.
References:
|
Get your own Free Home Page