Factors Determining Fruit Set in Heterostylous Bluets, Houstonia caerulea (Rubiaceae)

Subpopulations of heterostylous Houstonia caerulea consist of both pin and thrum plants (heterogeneous) or of plants of a single morph (homogeneous). Crossing experiments demonstrate that pollination between morphs is essential for maximal fruit set. In heterogeneous subpopulations percent fruit set increases as the ratio of pins to thrums approaches unity. Fruit set also is positively correlated with subpopulation size but shows little relationship with distance to the nearest compatible subpopulation. In homogeneous subpopulations, however, fruit set decreases with increasing subpopulation size, and distance to the nearest compatible subpopulation is more important in determining percent fruit set. Substructuring, therefore, may profoundly affect the reproductive biology of plant populations. Heterostylous plants present population biologists with a unique tool for examining problems concerned with plant reproductive biology and spatial patterns in natural populations. Individuals of such species can easily be segregated morphologically into two genetic incompatibility groups. In Houstonia caerulea members of the long-styled group, commonly known as "pins," have anthers attached low in the corolla tube, small pollen grains, and large stigmatic papillae (Rothrock, 1868; Darwin, 1877; Meehan, 1880; Lewis, 1962, 1976; Ornduff, 1977). "Thrums" have short styles, anthers attached high in the corolla tube, large pollen grains, and small stigmatic papillae. Thrums usually represent the "Ss" heterozygote and pins, the "ss" homozygote, in terms of the genetics of the incompatibility system (Vuilleumier, 1967; Yeo, 1975). Such a system of enforced outcrossing should maintain an equal ratio of pins to thrums. The comparative ease in working witha visible genetic polymorphism such as heterostyly has allowed researchers to gather much information on disassortative pollination and pollination efficiencies in natural systems (Levin, 1968; Mulcahy & Caporello, 1970; Ornduff, 1970a, 1970b, 1971, 1975a, 1975b, 1976; Ganders, 1974, 1975a, 1975b, 1976; Dulberger, 1975). It also has made possible the demonstration of spatial patterning in plant populations (e.g., Levin, 1974; Ornduff & Weller, 1975). One question which has not been approached in heterostylous plants however, is the identification and quantification of extrinsic factors determining levels of fruit set. In non-heterostylous plant populations such factors as plant spacing (Levin & Kerster, 1968; Levin, 1972), plant density (Levin & Kerster, 1969a, 1969b), inflorescence size and form (Willson & Rathcke, 1974; Willson & Price, 1977; Wyatt, 1978, 1980) and ' Botany, Duke University, Durham, NC 27706; Present address of RW: Botany, University of Georgia, Athens, GA 30602.