Luizzi, Victoria; Friberg, Magne; Petrén, Hampus
Fragrance compounds produced by flowers are important in attracting pollinators, deterring nectar thieves (organisms that consume nectar without pollinating), and controlling potentially harmful nectar microbes. Floral scent is often highly variable among and even within populations of plants, but this variation is only sometimes associated with variation in pollinator communities. An alternative, but little-researched, explanation for floral scent variation is phenotypic plasticity in response to environmental factors. This would mean that plants that experience different local conditions—for instance, different levels of soil moisture or nutrients—might produce different amounts or kinds of floral scent compounds, which could be relevant to pollinators without any genetic differences between plants. We tested this idea using alpine rock-cress (Arabis alpina), an herbaceous alpine perennial in the mustard family. Since some alpine rock cress populations are self-incompatible (they must be fertilized with pollen from another plant to set seed) and some are self-compatible (they can be fertilized with their own pollen), we also tested for differences in phenotypic plasticity for plants from the two categories. Finally, we asked whether individuals that invest more in floral scent had fewer resources for other functions like growth or reproduction.
We grew A. alpina plants originating from both self-incompatible and self-compatible populations under different water and nutrient conditions and measured the amount and kind of floral scent they produced. We also hand-pollinated plants in the low nutrient treatment to test whether more strongly-scented plants had less resources available for seed production.
Plants that received fewer nutrients produced less fragrance on a per-flower basis, regardless of whether they were self-compatible or self-incompatible, but the differences were relatively small compared to differences due to the plants’ population of origin. This indicates that genetic differences were more important than environmental differences in determining floral scent. Plants receiving fewer nutrients produced far fewer flowers than plants with sufficient nutrients, resulting in a dramatic decrease in floral scent at the level of the whole plant. Water availability had no effect on scent emission, and neither nutrients nor water affected scent composition (i.e., identity and amount of compounds present). Despite the lower floral scent emission in nutrient-limited plants, we found no evidence that producing scent is costly for A. alpina, because more scented individuals did not produce fewer seeds. Our findings suggest that floral scent in A. alpina may not be affected by changes in precipitation patterns caused by climate change. However, differences in soil nutrients may affect the fitness of A. alpina, but this will depend on how scent relates to pollinator attraction, and whether it functions to attract pollinators at long or short distances.