E. Melendezackerman et Dr. Campbell, ADAPTIVE SIGNIFICANCE OF FLOWER COLOR AND INTER-TRAIT CORRELATIONS INAN IPOMOPSIS HYBRID ZONE, Evolution, 52(5), 1998, pp. 1293-1303
Flower color is often viewed as a trait that signals rewards to pollin
ators, such that the relationship between flower color and plant fitne
ss might result from its association with another trait. We used exper
imental manipulations of flower color and nectar reward to dissociate
the natural character correlations present in a hybrid zone between Ip
omopsis aggregata and Ipomopsis tenuituba. Isozyme markers were used t
o follow the male and female reproductive success of these engineered
phenotypes. One field experiment compared fitnesses of I. aggregata pl
ants that varied only in flower color. Plants with flowers painted red
received more hummingbird visits and sired more seeds than did plants
with flowers painted pink or white to match those of hybrids and I. t
enuituba. Our second held experiment compared fitnesses of I. aggregat
a, I. tenuituba, and hybrid plants in an unmanipulated array and in a
second array where all flowers were painted red. In the unmanipulated
array, I. aggregata received more hummingbird visits, set more seeds p
er flower, and sired more seeds per flower. These fitness differences
largely disappeared when the color differences were eliminated. The hi
gher male fitness of I. aggregata was due to its very high success at
siring seeds on conspecific recipients. On both I. tenuituba and hybri
d recipients, hybrid plants sired the most seeds, despite showing lowe
r pollen fertility than I. aggregata in mixed donor pollinations in th
e greenhouse. Ipomopsis tenuituba had a fitness of only 13% relative t
o I. aggregata when traits varied naturally, compared to a fitness of
36% for white relative to red flowers when other traits were held cons
tant.