FLORAL DISPLAY IN PHLOX AND GERANIUM: ADAPTIVE ASPECTS.

Abstract

Floral displays are subject to selection pressures from several different factors, including pollinator attraction, competition for mates, time available for flowering and seed maturation, seed dispersal and predation (e.g., Janzen, 1969, 1977; Mulligan and Kevan, 1973; Stebbins, 1974; Willson, 1978). Although ways in which individual flowers are morphologically adapted to pollinators are well-documented (e.g., Grant and Grant, 1965; Faegri and van der Pijl, 1966; Beattie, 1974; Gentry, 1974) and flower or inflorescence size has attracted some attention (von Frisch, 1967; Leppik, 1970; Mulligan and Kevan, 1973; Willson and Rathcke, 1974; Cruden, 1976; Smith and Snow, 1976), few published data concern the adaptive nature of the entire floral display (size and number of inflorescences). Perhaps the adaptive value of flower-arranging has seemed self-evident. However, the variety of selection pressures resulting in the adaptive compromises we call organisms, the difficulty of demonstrating adaptive value for numerous traits that has led to arguments for selective neutrality, and the current flurry of interest in plant/animal coevolutionary interactions have all contributed to our opinion that an attempt to demonstrate a relationship between floral display and some measure of fitness is worthwhile. We embarked on this study in hopes of comparing seed prodution as a function of certain aspects of the floral display (number of flowers/stem, the clustering of flowering stems, and the distance between neighboring clusters [often genetic individuals]) for the two perennial woodland herbs, Phlox divaricata L. (Polemoniaceae) and Geranium maculatum L. (Geraniaceae). These two species were chosen because of their abundance, their differing floral displays, and their availability at the time we began field work in the spring. In particular, we tested the following hypotheses concerning the floral displays: 1) successful pollinator visits and total seed production per inflorescence and per individual increase with increasing size of the presented to the pollinators. Furthermore, the effectiveness of each flower in producing seeds might also increase with increasing target size, unless pollination among flowers of the same individual (geitonogamy) results in reduced seed set. Target size was indexed by the total number of flowers per stem, the number of simultaneously blooming flowers per stem, and the size of the cluster of flowering stems (see below). 2) the most successful floral display should be approximately the most frequent in the population (Lack, 1954; Johnson and Cook, 1968). Success may be measured both in terms of total seed output and of efficiency of seed production, i.e., number of seeds per flower. 3) successful pollinator visits and seed production should increase with decreasing distance between individuals (when cross-pollination prevails). At this point, we cannot assess the relative importance of all selective factors impinging on floral displays; rather, we direct our attention primarily to some of the consequences of such factors. Although the possible importance of pollen donation and of