A quantitative description of the gender of monoecious plants in the plots was provided by a functional gender analysis, a technique developed by Lloyd (1980a, b).
There is a different E value for each plot, calculated as the total number of female flowers produced by all monoecious plants in the plot divided by the total number of male flowers.
Half of the monoecious plants at A produced only male flowers, compared to 18% at M and none at P.
The rank plots of gender [ILLUSTRATION FOR FIGURE 6 OMITTED] show the pattern of distribution of gender for the monoecious plants in each site.
I examined the relationship between gender and plant size in the monoecious plants in the three sites and found that in A, the most stressful environment, the smallest plants tended to be males.
dioicum produced more seeds per fruit and smaller seeds than monoecious plants. These differences were maintained at all sites ([ILLUSTRATION FOR FIGURE 7 OMITTED]; Table 6).
There was a 99% increase in seed production for dioecious plants in P2 compared to that in P1, while monoecious plants showed only a 76% increase.
One of the most striking results of this experiment is the gender pattern observed in the monoecious plants in the experimental plots: monoecious plants tended to fall into two gender classes (male or cosexual) in the drier sites (in areas where dioecious populations occur naturally), exhibiting a trend toward dimorphism, while in the wet site (in an area where monoecious plants occur naturally), they showed much less variation, appearing monomorphic with respect to gender.
If the present parapatric distributions of the monoecious and dioecious subspecies of Ecballium could be explained by environmentally induced fertility differences between them, then one would expect monoecious plants: (1) to outperform plants of the dioecious subspecies in the northern wet site (Pamplona), (2) to do equally as well on the subspecies' boundary (Madrid), and (3) to be outperformed in the southern dry site (Almeria).
Another important issue to consider is that potential differences in the colonizing abilities of the subspecies may underlie their present geographic distributions: given that monoecious plants are self-compatible, one seed can start a population.
Dioecious Ecballium in the experimental plots (where matings are ensured) were strong performers, even in novel environments, and persisted longer than monoecious plants; however, the lack of suitable mates in a natural colonizing population is a critical problem for unisexual plants that monoecious plants can circumvent through selfing.