T. Pailler et Jd. Thompson, DISTYLY AND VARIATION IN HETEROMORPHIC INCOMPATIBILITY IN GAERTNERA VAGINATA (RUBIACEAE) ENDEMIC TO LA REUNION ISLAND, American journal of botany, 84(3), 1997, pp. 315-327
Documenting the floral biology of species throughout the Rubiaceae fam
ily is of particular interest since heterostyly and dioecy may have ev
olved more than once in this large family. Unfortunately many species
in several tropical regions remain unstudied. The purpose of this pape
r is to describe the floral biology, the nature of self-incompatibilit
y, morph ratios, and fecundity in natural populations of Gaertnera vag
inata, a small tree endemic to the island of La Reunion in the Indian
Ocean. Measurements of floral characters in populations across the ent
ire distribution of this species showed that G. vaginata exhibits a re
ciprocal stigma height and anther height dimorphism characteristic of
a distylous species. Pollen grain size and corolla tube length are con
sistently greater in short-styled plants and long-styled plants produc
e more pollen per flower. Controlled pollinations in a natural populat
ion showed that 25% of the short-styled plants gave at least one fruit
on intra-morph (illegitimate) pollination, whereas no long-styled pla
nts set fruit on illegitimate pollination. In total, 19.4% of illegiti
mate pollinations produced fruit on short-styled plants. No self-polli
nation gave fruit on either morph and between-morph pollinations produ
ced 92.2 and 92.8% for short and long-styled plants, respectively. Ove
rall, short-styled plants were significantly more abundant than long-s
tyled plants. Short-styled plants outnumbered long-styled plants in 16
of the 19 populations. In three of these populations the morph ratio
was significantly different from 1:1. In two natural populations, frui
t set was significantly higher on long-styled plants, although the num
ber of seeds per fruit was not significantly different between the two
morphs. The possible effect of variation in the strength of heteromor
phic incompatibility on observed variation in morph abundance and the
possible causes for the variation in fruit set are discussed.