Biology Reference
In-Depth Information
is particularly significant that the shorter bud development time reported here
for selfing annuals is predicted explicitly by the time-limitation hypothesis but
not by selection associated with pollinator/mate limitation. Although, we can-
not of course rule out the possibility that shorter bud development time may be
a pleiotropic consequence of the evolution of autonomous selfing through other
mechanisms.
Designing empirical studies that clearly distinguish between mechanisms in-
volving time-limitation versus pollinator/mate limitation remain a challenge but
we anticipate that our results and our discussion of these issues may help to in-
spire further research along these lines. Future studies may be designed to test
more directly the role of limited pollination time (vis-à-vis Figure 1b) by compar-
ing the time required for effective pollination under selfing versus outcrossing for
closely related species or ecotypes within natural habitats, taking care of course to
control for other aspects of the pollination environment (such as mate and pol-
linator availability) that might affect time-to-effective pollination.
Methods
Data Base Analyses
The literature was surveyed to obtain breeding information (i.e. selfing versus
outcrossing) for as many annuals species as possible. For each species, data on
plant height, flower size, and seed size were obtained where possible from stan-
dard floras and other published literature. A complete database was assembled for
118 species from both Europe and North America, involving 14 families (Table
1). For each species, the maximum published value for each trait was used. Plant
height was the maximum recorded vertical extent of the plant. The measure used
for flower size depended on the usual convention specific for each family, e.g.
maximum petal length, corolla width, lemma length (in the Poaceae). Seed size
was measured as the length of the longest axis. Within each family, for each trait,
the median value across selfing species and the median value across outcrossing
species was calculated and used in the phylogenetically-independent contrasts.
The contrasts were based on 14 phylogenetically-independent pairs, where
each pair consisted of median values of the selfing and outcrossing species within
one family, which by definition are species that are more closely related to each
other than to any other species in the data set [19]. For plant height, only 13 pairs
were used due to missing information. The data were analyzed using a Wilcoxon
matched pairs test.
