Measurements of selection on floral traits in black mustard, Brassica n&a

The large diversity of floral displays in angiosperms has been attributed largely to natural selection exerted by animal pollinators (Darwin, 1859; Grant and Grant, 1965; Waser, 1983). Quantitative measurements of selection on floral traits have only been made recently, however, and only on a handful of plant species (Campbell, 1989; Galen, 1989; Schemske and Horvitz, 1989; Campbell et al., 1991; Johnston, 1991; Herrera, 1993; Conner et al., 1996a, b). Even fewer studies have used lifetime fitness data for selection measurements (Campbell, 1991; Conner et al., 1996a, b), in spite of the fact that selection measurements based on only part of an organism’s reproductive life can be misleading (Lande and Arnold, 1983; Arnold and Wade, 1984; Endler, 1986; Travis and Heinrich, 1986; Clutton-Brock, 1991). Here we report measurements of selection on floral traits based on lifetime female fitness measurements in black mustard, BI-crssicc~ nigm (Brassicaceae). Several studies have reported directional selection for increased flower size and/or number (Galen, 1989; Campbell et al., 1991; Johnston, 1991; Herrera, 1993; Conner et al., 1996a, b); these traits are important in pollinator attraction (e.g., Clements and Long, 1923; Willson and Rathcke, 1974; Willson and Bertin, 1979; Thomson et al., 1982; Bell, 1985; Geber, 1985; Schmid-Hempel and Speiser, 1988; Thomson, 1988; Eckhart, 1991; Conner and Rush, 1996), and may also be important determinants of pollen and ovule production. Flower number is particularly important in determining both male and female fitness of plants (e.g., Schoen and Stewart, 1986; Broyles and Wyatt, 1990; Devlin and Ellstrand, 1990; Herrera, 1991, 1993; John-