Abstract Detail

Reproductive Processes

Yomai, Viann Marie [1], Williams, Joseph [1].

Evidence for relaxed selection on style length on an isolated, oceanic island.

Plants on oceanic islands should evolve smaller flowers than their mainland relatives due to the repeated evolution of selfing syndromes. Among animal-pollinated species, shifts to more selfed mating systems should relax selection on pollinator attraction traits such as petal size, the usual metric for floral size. Changes in petal size should incur allometric changes in other floral organs, if genetic or developmental correlations are strong. But style length is often proposed to be a sexually selected organ because longer styles magnify beneficial effects of pollen competition on offspring vigor. If style length is maintained by sexual selection in more outcrossed mainland populations, then we hypothesized that shifts to more selfed mating systems on islands should relax selection on style length, which should evolve independently of petal length among colonizing populations.   A previous study documented high levels of pollen limitation and shifts to more autonomous breeding systems on Pohnpei, Micronesia, a small, isolated volcanic island. In this study, we asked three questions: 1) Do plants on Pohnpei have shorter style and petal lengths than mainland intra- or inter-specific sister taxa? 2) Are floral organ sizes of mainland sister taxa unusual relative to their mainland genus? 3) Do island-mainland differences in style length scale with those of flower size (petal length), as predicted by the genetic correlations hypothesis; or are they uncorrelated or negatively allometric, as predicted by sexual selection theory? We collected petal and style length data for 73 plant species in 61 genera on Pohnpei, and from 15 online flora databases and the primary literature for data on mainland taxa. We used phylogenetic generalized least squares to test for petal length and style length correlations between island and mainland relatives. Our results showed that flowers on Pohnpei were not smaller relative to their mainland sister taxa (N = 50 pairs, P = 0.255), whereas styles were about 40% shorter on Pohnpei (N = 43 pairs, P = 0.005). Reduced style length was especially evident in conspecific sister pairs (P = 0.009, N = 16). The log ratios (LRs) of island-mainland petal lengths were not correlated with the LRs of style lengths (P = 0.407, N = 28 pairs). Mainland sister taxa had similar petal and style lengths to the mean of their mainland genus.   Our results reflect other studies that have found mixed evidence for flower size reduction on islands. On the other hand, there was strong evidence that species have evolved shorter style lengths during colonization of Pohnpei. Pollen competition intensity is lowest when stigmas receive few pollen grains and relatedness among pollen grains is high, consistent with high levels of pollen limitation and long-term selfing on Pohnpei. Thus, our results support a major prediction of sexual selection theory - that if pollen competition maintains longer styles in more outcrossed populations, then shifts to higher levels of self-fertilization should relax selection on costly styles, leading to shorter style lengths.

1 - University of Tennessee, Ecology and Evolutionary Biology, Knoxville, TN, 37996, USA

pollen competition
Floral traits
Petal size
Style length
Island Ecology
Pacific islands
breeding system

Presentation Type: Oral Paper
Number: RP5002
Abstract ID:513
Candidate for Awards:None

Copyright © 2000-2022, Botanical Society of America. All rights reserved