Jd. Ackerman et al., VARIATION IN POLLINATOR ABUNDANCE AND SELECTION ON FRAGRANCE PHENOTYPES IN AN EPIPHYTIC ORCHID, American journal of botany, 84(10), 1997, pp. 1383-1390
Extraordinary floral variation is common among some orchids that emplo
y deception to attract pollinators. This variation may be maintained b
y frequency-dependent selection where rare phenotypes are preferred. O
ver a 2-yr period, 1993-1994, we monitored the reproductive success of
Tolumnia variegata, an obligately outcrossing epiphytic orchid, at th
ree localities in Puerto Rico that differed in pollinator service. Pla
nts varied in floral morphology and fragrance characteristics. Artific
ial arrays of varying frequencies of scentless and fragrant phenotypes
were established to test for frequency-dependent selection. Where pol
linators were rare (Cambalache, range of census average = 0-0.2 bees/h
), 0.9-1.2% of the flowers were effectively visited (pollinarium remov
als and pollinations). At Tortuguero where 0.4-1.1 bees/h were observe
d, 4-9.2% of the flowers were visited. At Pi (n) over tilde ones where
bees were the most abundant (1.4-5.2 bees/h), 20.9-25.0% of the flowe
rs were visited. A significant portion of the variance in all measures
of reproductive success (male, female, and combined) was explained by
differences among populations, which we attribute mostly to variation
in pollinator abundance. Neither the fragrance phenotype nor its freq
uency had a significant effect on success as revealed by a split-plot
ANOVA. There was a significant interaction between population and phen
otypic frequencies in all our measures of reproductive success, but on
ly for the 1994 flowering season. Thus, variation in floral fragrance
phenotypes is not likely maintained by frequency-dependent selection.
High levels of variation remain unexplained.