| A plant's ability to attract animal pollinators should be influenced by two important factors: the amount of available resources that is invested to the production of rewards such as nectar, and the individual's size or resource state. Using Chamerion (= Epilobium) angustifolium (Onagraceae) as a model, I investigate: (1) selection pressures that determine the evolutionarily stable investment to nectar production, and (2) allocation tactics that vary with an individual's current resource state. A conceptual model is presented, in which a plant's attractiveness to pollinators increases with relative investment to nectar production ('social competition'). The evolutionarily stable investment strategy occurs where the expense of further nectar production outweighs the potential disadvantage of being slightly less attractive than competitors. The model may explain why C. angustifolium inflorescences produce, on average, such large amounts of nectar. Within populations, however, individuals varied greatly in the amount of nectar offered to pollinators, mainly due to variation in floral display size (number of open flowers---a correlate of resource state). Given that large, attractive inflorescences maximize pollen export by limiting the amount of pollen removed by each visitor, I predicted a size-dependent distribution of nectar within the vertical inflorescences of C. angustifolium. As predicted, small inflorescences distributed nectar nearly evenly among (lower) female- and (upper) male-phase flowers, whereas larger inflorescences allocated extra nectar to female-phase flowers. In experimental inflorescences, I distributed the same volume of nectar to mimic the 'large' and 'small' allocation types. Nectar-foraging bumblebees visited a mean of 3.2 fewer male-phase flowers on the 'large', relative to 'small' type, as expected if the nectar distribution of large displays functions to limit pollen removal during individual visits. I propose that the nectar gradient may adaptively mediate the schedule of pollen removal by manipulating the patch departure behaviour of pollinators. |
