Plant Species Biology (2004) 19, 1–12 © 2004 The Society for the Study of Species Biology Blackwell Science, LtdOxford, UKPSBPlant Species Biology1442-1984The Society for the Study of Species Biology, 2004April 2004191112Original Article FUNCTIONS OF STAMINATE FLOWERSM. A. SCHLESSMAN Et al. Correspondence: Mark A. Schlessman Email: schlessman@vassar.edu 1 Present address: Department of Biological Science, Florida State University, Tallahassee, FL 32306-1100, USA. 2 Present address: Department of Biology, Olin Rice 218, Macalester College, 1600 Grand Avenue, St Paul, MN 55105- 1899, USA. Functions of staminate flowers in andromonoecious Pseudocymopterus montanus (Apiaceae, Apioideae) MARK A. SCHLESSMAN, NORA UNDERWOOD, 1 TIM WATKINS, 2 LAURA M. GRACEFFA and DIANE CORDRAY Rocky Mountain Biological Laboratory, PO Box 512, Crested Butte, CO 81224, USA and Department of Biology, Box 187, Vassar College, 124 Raymond Avenue, Poughkeepsie, NY 12604-0187, USA Abstract The pollen donor and pollinator attractor hypotheses are non-exclusive alternative expla- nations for the functions of the staminate flowers of andromonoecious plants. We tested assumptions and predictions of both hypotheses using the andromonoecious perennial Pseudocymopterus montanus (Apiaceae, Apioideae), in which staminate flowers are approximately one-third as massive as perfect ones and sequentially blooming umbels bear varying proportions of staminate flowers. We confirmed that larger umbels receive more visits than smaller ones. We demonstrated that fluorescent dye is a suitable analog of P. montanus pollen and that larger umbels donate more dye to nearby recipient umbels than do smaller ones. We showed that experimental removal of staminate flowers lowers fruit set only in primary umbels. We used stepwise multiple regressions to examine the effects of staminate flowers on xenogamous pollination, overall (xenogamous and geito- nogamous) pollination and fruit set. The regressions revealed significant effects of stam- inate flowers only for cross pollination of primary umbels. We concluded that in P. montanus, staminate flowers function primarily as pollen donors. Keywords: andromonoecy, dichogamy, geitonogamy, protogyny, sex allocation, Umbelliferae. Received 13 December 2002; revision received 23 July 2003; accepted 15 August 2003 Introduction Andromonoecy is a sexual system of flowering plants in which each individual produces two morphologically dis- tinct kinds of flowers: perfect and staminate. The perfect flowers have both pistils and stamens, while the stami- nate flowers lack fully developed, functional pistils and cannot produce fruits. Andromonoecy is relatively uncommon, occurring in perhaps 2% or less of all angiosperms (Yampolsky & Yampolsky 1922; Schlessman et al. 2001). Several authors have hypothesized that andromonoecy evolved from hermaphroditism (all flow- ers perfect) by the conversion of perfect flowers into more numerous staminate flowers, producing larger inflores- cences and increasing visitation by pollinators (Bertin 1982; Willson 1983; Solomon 1987; Spalik 1991). Although details vary, most formulations of this hypothesis include the assumptions that the probability of setting fruit was low for at least some of the flowers of the hermaphroditic ancestor and that staminate flowers require less repro- ductive effort (allocation of resources) than perfect ones. Thus, conversion of non-fruiting hermaphroditic flowers to staminate flowers would neither decrease fitness by lowering maternal success nor necessarily increase over- all reproductive effort. Higher visitation rates might increase fitness in several ways including: (i) improve- ment of male reproductive success (seeds sired) by increasing the amount of pollen transferred to the stigmas of other plants; and (ii) improvement of female reproduc- tive success (seed set) by increasing the amount of pollen deposited on stigmas of the perfect flowers, the number of perfect flowers receiving pollen, or both. Following Solomon (1987), we call these the pollen donor and polli- nator attractor hypotheses. These hypotheses are not mutually exclusive: staminate flowers could enhance both donation and reception of pollen. As we apply them here specifically to andromonoecy, the two hypotheses are con-