From Plant Press Vol. 11, No. 1 from January 2008.
By W. John Kress and Ethan J. Temeles
A major goal of systematic, evolutionary, and ecological studies is to understand the processes that shape adaptations of organisms to their environment. One approach towards this goal is to use patterns of morphological variation among closely related species to make inferences about the kinds of ecological processes, such as competition, predation, parasitism, or mutualism, through which these characteristics have evolved. This type of study traces back to Charles Darwin and his classic observations in the Galapagos Islands of beak variation and food types among species of ground finches in the genus Geospiza.
Since Darwin’s time, interactions between flowering plants and their pollinators have provided model examples of hypothesized feeding adaptations, species specialization, and coevolution. However, some investigations suggest that such interspecific interactions are seldom constant throughout a species’ range and that populations differ in the traits shaped by the interaction. These observations have resulted in the John Thompson’s proposed “Geographic Mosaic Theory of Coevolution.” Accordingly, studies of co-adaptation between species require an analysis of populations across the broad geographic distribution of the species. Some populations across this geographic landscape are hotspots of reciprocal selection and others are coldspots with little coevolution.
We have recently described an interaction between a plant and its hummingbird pollinator found in the Eastern Caribbean that provides an ideal tropical system to test Thompson’s theory. Both the pollinators and the plants exhibit character polymorphisms, which can be readily linked to fitness measures, throughout their ranges. Over the last several years and probably for a number of years to come we plan to study co-adaptations in this plant-hummingbird association through extensive observations of both floral and sexual dimorphisms across the island archipelago in the Eastern Caribbean.
The focal system of this research involves the purple-throated carib hummingbird, Eulampis jugularis, and its Heliconia food plants, considered one of the strongest examples of ecological causation of sexual dimorphism to date and a well-documented plant-pollinator mutualism. Many biology textbooks note that flowers pollinated by hummingbirds have long, tubular corollas matching the size and shape of the birds’ beaks. Darwin himself stated that bills of hummingbirds are specially adapted to the various kinds of flowers they visit.
Our work has provided evidence for co-adaptation by demonstrating that the purple-throated carib hummingbird is the primary pollinator of H. caribaea and H. bihai, with flowers of the former corresponding to the short, straight bills of males (the larger sex) and flowers of the latter to the longer, curved bills of females. Further evidence for co-adaptation comes from a reversal in the floral dimorphism of the Heliconia food plants on two islands: on St. Lucia, the female-pollinated H. bihai develops a second morph with shorter, straighter flowers matching the bills of males, whereas on Dominica the male-pollinated H. caribaea develops a second morph with longer, curved flowers matching the bills of females. The nectar rewards of all Heliconia morphs on both islands are consistent with each sex’s choice for the morph corresponding to its bill morphology and energy requirements, supporting the hypotheses of ecological causation of sexual dimorphism and co-adaptation.
Our work started nearly a decade ago when Ethan Temeles initiated studies of the purple-throated carib hummingbird, Eulampis jugularis, on the island of St. Lucia. The males and females are identical in plumage, but display some of the most extreme sexual dimorphisms of any hummingbird species. First, bills of males average 24.8 mm in length, whereas bills of females average 29 mm in length, a difference of 15 percent. Moreover, bills of males are relatively straight, curving downwards at a 15 degree angle, whereas bills of females are highly curved, curving downwards at a 30 degree angle.
These differences in bill morphology were associated with differences in the use of the two Heliconia species on the island. Males were associated with a red-bracted morph of H. caribaea, which they defended against other males and females. In contrast, females intruded onto male territories to feed and also fed at undefended patches of H. caribaea, but more importantly they were the predominant visitor to the other Heliconia, a green-bracted morph of H. bihai. The morphology of the flowers of the two heliconias reflected this sexual preference by the hummingbirds. Flowers of H. caribaea were short and straight, measuring 38 mm in length and 20 degrees in curvature, which approximates the short, straight bills of male purple-throats, their primary pollinator.
In contrast, flowers of the green-bracted H. bihai were long and curved, measuring nearly 44 mm in length and 31 degree in curvature, which is almost identical to the 30 degree curvature of female bills. The close correspondence between bills of male and female purple-throats and the sizes and shapes of the Heliconia flowers they visit argues strongly for reciprocal adaptation. This hypothesis receives additional support from our discovery of a second, red-green morph of H. bihai on the island. This red-green morph acts as a geographic replacement of H. caribaea and is defended and visited primarily by male purple-throated caribs. Its flowers are both significantly shorter and straighter than flowers of the green-bracted morph visited by females, which is what you would expect given its visitation by males.
We then extended our investigations of this plant-pollinator system to the island of Dominica, located 100 miles north of St. Lucia. Bills of both sexes of purple-throated caribs on this island are slightly longer and slightly less curved than on St. Lucia. As on St. Lucia, however, the primary food plants of these birds are H. caribaea and H. bihai. But in contrast to St. Lucia which has two morphs of H. bihai, green and red-green, on Dominica there is only one morph of H. bihai which is red with a yellow stripe. Similarly, whereas there was only a single, red color morph of H. caribaea on St. Lucia, on Dominica there are now two: a red morph and a yellow morph.
As on St. Lucia, males associate primarily with H. caribaea. Females also visit H. caribaea and are the sole pollinator of H. bihai. Once again, flowers of the two heliconias correspond closely to the bills of their primary hummingbird visitor. Flowers of H. caribaea are short and straight, measuring 36 mm in length and 20 degrees in curvature, which roughly corresponds to the 15-degree bill curvature of males. In contrast, flowers of H. bihai measure 48 mm in length and 30 degrees in curvature, a perfect match for the 30 degree bill curvature of females. The flowers of H. bihai are 4 mm longer than on St. Lucia which we interpret as a coevolutionary response to the exclusive visitation by female purple-throats with their long bills.
Vinita Gowda, a graduate student at George Washington University working with us on this project, has provided an important temporal component to our observations. As part of her dissertation, she monitored the flowering times of the heliconias and the visitation rates of the hummingbirds on three islands: Dominica, St. Kitts to the north, and St. Vincent to the south. On Dominica flowering patterns of H. bihai and H. caribaea were significantly different: H. bihai flowers throughout the year but has a distinct peak flowering season in April, whereas H. caribaea has a distinct flowering season from March until September (with a peak in May) and does not produce any flowers the rest of the year. The peak flowering season of the two species did not overlap.
With respect to pollinator visitation patterns H. bihai shows a clear female dominated interaction throughout the flowering season while H. caribaea inflorescences are always strictly defended as territories by males with females often allowed to enter the territories to feed on the flowers during the mating season. Other pollinators that were rarely observed on H. caribaea are Green throated Carib Hummingbirds (Eulampis holosericeus) and Antillean Crested Hummingbirds (Orthorhyncus cristatus). Both of these visitors are aggressively chased away by territorial males.
As our studies and observations have spread throughout the archipelago we have tracked the long-distance transition between generalized and specialized plant-pollinator interactions in this heliconia-hummingbird system. What we have found is that one species of plant, H. bihai, changes from a generalist strategy in the southern range of the island chain (Trinidad) where its flowers are pollinated by many hummingbird species to a highly specialized pollination strategy in the core of the archipelago (St. Lucia and Dominica) where its flowers are pollinated by one sex of one hummingbird species.
We are only now beginning to understand the floral and pollinator traits associated with the transition from generalization to specialization. Where H. bihai is pollinated by many hummingbird species with bills of many different lengths, the flowers are 30% shorter in length than on the islands where the flowers are only visited by female purple throated Carib hummingbirds with long-curved bills. We hypothesize that this transition across 300 miles in the Caribbean archipelago results from character release associated with a decrease in the number of species on islands as one moves north farther and farther from the mainland. The number of endemic Heliconia species decreases from five on Trinidad to two on St. Lucia and Dominica, whereas the number of hummingbird species decreases from twelve on Trinidad to three on St. Lucia and four on Dominica. This reduction in species diversity may permit H. bihai and H. caribaea to undergo character displacement where the two heliconias overlap. Similarly, a reduction in competing hummingbird species combined with two abundant and rewarding Heliconia food plants may drive the evolution of sexual dimorphism within purple-throated caribs. Our available evidence supports this hypothesis.
These results indicate that a geographic approach can identify populations of plant-pollinator specialization and generalization. Highly-specialized populations may represent hotspots of reciprocal evolution. In this regard, it is significant that the longest flowers of H. bihai and the longest bills of female purple-throated caribs occur where it is pollinated exclusively by females. Whether such associations represent coevolutionary arms races requires future study.
We have much more to understand about the mosaic of interactions between the species of heliconia and their hummingbirds pollinators in the Lesser Antilles. Through a grant from the National Science Foundation we have funding support to continue our investigations in the Eastern Caribbean for several more years. Like the Galapagos Islands off the coast of Ecuador, the Eastern Caribbean archipelago is an ideal natural laboratory for understanding the evolutionary processes of adaptation and coevolution.