Visitation of plants by different pollinators depends on individual plant traits, spatial context, and other factors. A neglected aspect of small-scale variation of plant-pollinator interactions is the role of vertical position of flowers. We conducted a series of experiments to study vertical stratification of plant-pollinator interactions in a dry grassland. We observed flower visitors on cut inflorescences of Centaurea scabiosa and Inula salicina placed at different heights above ground in two types of surrounding vegetation: short and tall. Even at such a small-scale, we detected significant shift in total visitation rate of inflorescences in response to their vertical position. In short vegetation, inflorescences close to the ground were visited more frequently, while in tall vegetation, inflorescences placed higher received more visits. Moreover, we found major differences in the composition of the pollinator community on flowers at different heights. In a second experiment, we measured flower visitation rate in inflorescences of Salvia verticillata of variable height. Total flower visitation rate increased markedly with inflorescence height in this case. Data on seed set of individual plants provide evidence for a corresponding positive pollinator-mediated selection on increased inflorescence height. Overall, our results demonstrate strong vertical stratification of plant-pollinator interactions at the scale of mere decimetres. This may have important ecological as well as evolutionary implications.

• Keywords
• Bees, Entomology, Foraging biology, Plant reproduction, Plant–pollinator interactions, Pollination,
• Publication type
• Journal Article MeSH

Previous studies demonstrated the effects of polyploidy on various aspects of plant life. It is, however, difficult to determine which plant characteristics are responsible for fitness differences between cytotypes. We assessed the relationship between polyploidy and seed production. To separate the effects of flowering phenology, flower head size and herbivores from other possible causes, we collected data on these characteristics in single flower heads of diploid and tetraploid Centaurea phrygia in an experimental garden. We used structural equation modelling to identify the main pathways determining seed production. The results showed that the relationship between polyploidy and seed production is mediated by most of the studied factors. The different factors acted in opposing directions. Wider flower heads displayed higher above the ground suggested higher seed production in diploids. In contrast, earlier flowering and a lower abundance of herbivores suggested higher seed production in tetraploids. However, because phenology was the strongest driver of seed production in this system, the sum of all the pathways suggested greater seed production in tetraploids than in diploids. The pathway linking ploidy level directly to seed production, representing unstudied factors, was not significant. This suggests that the factors studied likely are drivers of the between-cytotype differences. Overall, this study demonstrated that tetraploids possess overall higher fitness estimated as seed production. Regardless of the patterns observed here, strong between year fluctuations in the composition and diversity of insect communities have been observed. The direction of the selection may thus vary between years. Consequently, understanding the structure of the interactions is more important for understanding the system than the overall effects of cytotype on a fitness trait in a specific year. Such knowledge can be used to model the evolution of species traits and plant-herbivore and plant-pollinator interactions in diploid-polyploid systems.

• Keywords
• AMOS, Asteraceae, Carduoideae, pollinators, polyploid, pre-dispersal seed predation, seed set, structural equation modelling,
• Publication type
• Journal Article MeSH