OIKOS 89: 461 – 470. Copenhagen 2000 Pollinator abundance and behavior on Aconitum lycoctonum (Ranunculaceae): an analysis of the quantity and quality components of pollination Anna-Barbara Utelli and Barbara A. Roy Utelli, A.-B. and Roy, B. A. 2000. Pollinator abundance and behavior on Aconitum lycoctonum (Ranunculaceae): an analysis of the quantity and quality components of pollination. – Oikos 89: 461 – 470. The yellow-flowered monkshood Aconitum lycoctonum has variable degrees of in- breeding among populations, despite being primarily bee-pollinated. Here we exam- ined the pollination ecology of A. lycoctonum. We investigated pollinator community and frequency in four populations at two altitudes over two years. We found that flowers were more often visited at low elevation than at high elevation. However, because flowers lived longer at high elevation than at low elevation, and plants at high elevation had more flowers than at low elevation, the overall chance of a plant being visited may be greater at high elevation. Breeding system experiments showed that at least some populations of A. lycoctonum were self-compatible and also were not completely protandrous. Thus selfing, especially by geitonogamous pollen, will be common if pollinators often visit several flowers within a plant, as we found to be true for the main pollinator (bumblebees) of A. lycoctonum. Although the stereotyp- ical behavior of bees is to move upwards, 17% of the within-plant movements of the main pollinator (Bombus gerstaeckeri ) were downwards, i.e. from a male-phased to a female-phased flower because the flowers open from the bottom to the top of the plant. Other pollinators of A. lycoctonum moved less often within plants. We conclude that in addition to differences in self-compatibility, differences in pollinator abundance and behavior could have led to variation in the realized mating system in different A. lycoctonum populations. A.-B. Utelli and B. A. Roy, Geobotanical Inst. ETH, Zu ¨richbergstr. 38, CH-8044 Zu ¨rich, Switzerland (roy@geobot.umnw.ethz.ch). The mating system of plants is influenced by genetic, demographic, and environmental factors (Barrett and Eckert 1990). For example, outcrossing rates may vary as a result of differences in the degree of protandry (Schoen et al. 1997), population density (Norman et al. 1997) and population size (Raijmann et al. 1994, Rout- ley et al. 1999). However, the mating system of plants can also be influenced by several factors in concert (e.g. Barrett and Husband 1990). An important aspect of pollination in insect-polli- nated plants is variation in pollinator species and in their activity (Schemske and Lande 1985). Successful pollination of an insect-pollinated plant depends upon its attractiveness to pollinators, and on the ability of the pollinator to deposit enough compatible pollen on the flower’s stigma. Thus, pollination can be separated into quantity and quality components: the abundance and visitation rates of pollinators, versus the amount of pollen deposited on stigmas, selection of floral sexual stage by pollinators, and movement patterns of pollina- tors (Herrera 1987). Abundance and visitation rates of pollinators can vary markedly among and within years and populations (e.g. Herrera 1989, Fishbein and Ven- able 1996, Roy 1996). Quality of visits also varies among pollinating species (e.g. Herrera 1987, Fishbein and Venable 1996, Thøstesen and Olesen 1996), but is Accepted 23 November 1999 Copyright © OIKOS 2000 ISSN 0030-1299 Printed in Ireland – all rights reserved OIKOS 89:3 (2000) 461