Influence of plant reproductive systems on the evolution of hummingbird pollination

Abstract

Many hummingbird-pollinated plant species evolved from bee-pollinated ancestors independently in many different habitats in North and South America. The mechanisms leading to these transitions are not completely understood. We conducted pollination and germination experiments and analyzed additional reproductive traits in three sister species pairs of which one species is bee- and the other hummingbird-pollinated. All hummingbird-pollinated species showed higher seed set and germination rates in cross-pollinated than in self-pollinated flowers. In the self-compatible, bee-pollinated sister species this difference did not exist. As expected, seed set and germination rate were higher after cross-pollination in the largely self-incompatible genus Penstemon independently of the pollination syndrome. However, the bird-pollinated species produce only half of the amount of ovules and pollen grains per flower compared to the bee-pollinated sister species. This indicates that hummingbird pollination is much more efficient in self-incompatible populations because hummingbirds waste less pollen and provide higher outcrossing rates. Therefore, hummingbird pollination is less resource costly. Overall, we suggest that hummingbirds may increase the reproductive success compared to bees, influencing the evolution of hummingbird pollination in ecosystems with diverse bee assemblages.

Keywords: bee; germination rate; outcrossing; pollination efficiency; seed set; selfing.

FIGURE 1

The three investigated sister species pairs, on the left in the lateral view and on the right in the frontal view. The first mentioned species of each genus is bee‐pollinated and the second one hummingbird‐pollinated. (a, b) Penstemon neomexicanus; (c, d) Penstemon barbatus; (e, f) Lobelia siphilitica; (g, h) Lobelia cardinalis; (i, j) Mimulus lewisii; (k, l) Mimulus cardinalis

FIGURE 2

Differences between relative seed sets (a, b), seed weights (c, d), and germination rates (e, f) of hummingbird‐ and bee‐pollinated sister species pairs under self‐ and cross‐pollinations. The boxplots on the left side (a, c and e) show the groups used for the linear models while parts b, d and f show the data divided into species and pollination group with the results of the t‐tests between the outcrossing and the selfing treatment for the single species. Significance was indicated as ***p ≤ .001, **p ≤ .01, *p ≤ .05, ⁽*⁾ p ≤ .1, and n.s. = not significant. For the genera, “bee” indicates the bee‐pollinated species (blue colors), while “bird” indicates the hummingbird‐pollinated species (red colors). For the treatment, “out” represents the outcrossed group (darker colors), while “self” represents the self‐pollinated (brighter colors) group

FIGURE 3

Boxplots of (a) ovule numbers, (b) pollen grain numbers and (c) the ratio of ovules and pollen per flower. Differences between hummingbird‐ (red) and bee‐pollinated (blue) sister species are tested by t‐tests. Significance was indicated as *** p ≤ .001 and n.s. = not significant. On the x‐axes, “bee” indicates the bee‐pollinated species, while “bird” indicates the hummingbird‐pollinated species