Sexual Dimorphism and Floral Polymorphism in a Hummingbird-Plant Association
Sexual dimorphism in size and morphology is widespread in the animal kingdom. Charles Darwin drew attention to these differences, and offered three explanations for their evolution based on mechanisms of sexual selection, fecundity selection, and ecological causation, e.g., resource- partitioning. Although empirical studies demonstrate that the first two mechanisms operate in natural populations, unambiguous examples of ecological causation of sexual dimorphism have been hard to find, because in many animal species, sexual differences in the size of feeding structures scale positively with body size. The positive relationship between body size and feeding morphology makes it unclear whether ecological differences between the sexes are the cause or the consequence of sexual dimorphism.
Beginning in 1999, I have been studying sexual dimorphism in the purple-throated carib (Anthracothorax jugularis), initially on the island of St. Lucia, West Indies. Although the wings and body masses of males average 8.6% and 25% larger than those of females, the bills of females average 20% longer and 40% than those more curve than those of males, one of the most extreme bill dimorphisms of any hummingbird (Fig. 1A & B).
Fig. 1. Polymorphisms in bills of E. jugularis (A and B) and in flowers (C and D) and inflorescences (E to J) of Heliconia species on St. Lucia (E to G) and Dominica (C to D and H to J), West Indies. (A) E. jugularis, female bill. (B) E. jugularis, male bill. (C) H. bihai, flower. (D) H. caribaea, flower. (E) H. bihai, green inflorescence morph, St. Lucia. (F) H. bihai, red-green inflorescence morph, St. Lucia. (G) H. caribaea, inflorescence, St. Lucia. (H) H. bihai, red-and-yellow–striped inflorescence, Dominica. (I) H. caribaea, red inflorescence morph, Dominica. (J) H. caribaea, yellow inflorescence morph, Dominica. From Temeles & Kress 2003 Science.
In studies with Amherst students Irvin Pan (1999), Jill Brennan (2001), and Jed Horwitt (2001), I found that the purple-throated carib is the sole pollinator of two species of Heliconia, a red-bracted morph of H. caribaea (Fig. 1G), and a green-bracted morph of H. bihai (Fig. 1E) endemic to St. Lucia. Flowers of H. caribaea (Fig. 1D) correspond in size and shape to the bills of male caribs, whereas flowers of the green-bracted H. bihai (Fig. 1C) correspond in size and shape to the bills of female caribs. Further evidence for ecological causation of sexual dimorphism (and plant-pollinator coevolution) was provided by a geographic replacement of H. caribaea by H. bihai where the former is rare, and the subsequent development of a floral dimorphism in H. bihai, with a red-green bracted morph (Fig. 1F) with flowers matching the bills of males, and a green-bracted morph with flowers matching the bills of females (Temeles et al. 2000 Science).
Subsequently, John Kress, Curator of Botany at the Smithsonian Institution and an authority on Heliconia taxonomy and systematics, and I found additional support for ecological causation of sexual dimorphism in the purple-throated carib from studies on the island of Dominica, 120 kilometers to the north. As on St. Lucia, Dominica has both H. caribaea and H. bihai, with flowers of the former matching bills of males and flowers of the latter matching bills of females. But on Dominica, H. bihai has only one morph, which is red with a yellow stripe (Fig. 1H). Heliconia caribaea is more common than H. bihai, and it is the Heliconia species with two color morphs: a yellow-bracted morph (Fig. 1J) with shorter, straighter flowers matching the bills of males, and a red-bracted morph (Fig. 1I) with longer, more curved flowers matching the bills of females. Not only that, but the replacement morphs on Dominica or St. Lucia either reduce (Dominica) or increase (St. Lucia) their nectar reward through decreases or increases in their numbers of bracts and flowers in correspondence to the size and energy requirements of their female or male pollinator. This latter finding suggests that sexual differences in body size of purple-throated caribs, and not just their bill morphology, are maintained by feeding specialization on the different Heliconia food plants (Temeles & Kress 2003 Science).
The Roles of Food and Sex in Heliconia use by Purple-throated Caribs
Both intersexual food competition and sexual selection may result in male-male competition for territories, either as food sources during the non-breeding period, or as feeding-and-mating territories during the breeding period. Competition between males and females for flowers during the non-breeding season is an unambiguous indicator of intersexual food competition. Because many females that intrude onto male territories during breeding periods copulate with males, however, it is unclear whether male territorial defense during the breeding period is mostly for food, and driven by intersexual food competition, or is mostly for sex, and driven by sexual selection.
To discriminate between these two alternatives for territory maintenance, my students and I experimentally reduced flower numbers on territories of four different males by covering Heliconia flowers with clear plastic bags on male territories in 2000, 2001, and 2002 on the island of St. Lucia. If breeding territories are defended primarily for sex, then experimental reductions of flower numbers on male territories should have little effect on the amount of time a male spends on the territory. This was why we covered bracts with clear bags: if breeding territories are for sex, then clear plastic bags should not reduce territory attractiveness if inflorescences (flower stalks) are a signal of male quality. Alternatively, if breeding territories are defended primarily for food, we expected males to spend more time off of territories feeding following experimental reductions of flower numbers on their territories, as well as lose weight, which we monitored by weighing males with electronic balances fitted to Heliconia inflorescences (Fig. 2).
Fig. 2. Ohaus electronic balance with weighing pan removed and replaced with an inflorescence of H. bihai attached to a wooden dowel.
Following inflorescence reductions, males reduced the time they spent feeding on their territories and lost a significant amount of weight, but increased their time feeding and regained weight to pre-manipulation levels following restoration of flower abundance. Territorial males also reduced the amount of time they spent in defense and in mating chases following inflorescence reductions. Both results indicate that males derive an energy benefit from defense of Heliconia territories during the breeding season, in contrast to studies of some other hummingbird species. In all three years, significantly more females intruded onto the more rewarding H. caribaea than onto H. bihai territories, suggesting that the quality of a male’s territory may be a cue for female choice of mates in this hummingbird-flower system (this research was published as Temeles et al. 2004 Condor with honors students Amanda B. Muir, Elon B. Slutsky, and Maren N. Vitousek).
Sexual Dimorphism and Foraging and Territorial Economics at Three Heliconias
Additional support for the role of food in driving this hummingbird Heliconia system comes from a study of foraging and territorial economics at the three heliconias on St. Lucia. Across three years of study, H. caribaea territories defended by males were significantly smaller in area and had higher densities of flowers than red-green H. bihaiterritories, and both kinds of male Heliconia territories were smaller and had higher densities of flowers than a green H. bihai territory maintained by a female (females were rarely observed on Heliconia territories and the single female post-breeding territory was the only female territory found in three years of studies). My undergraduate students and I used measured metabolic costs of sleeping, resting, and hovering for purple-throated caribs obtained by Wolf and Hainsworth (1971 Ecology), and Gill’s and Wolf’s (1975 Ecology) equations for territory and foraging economics of sunbirds (modified for the hummingbirds), to estimate the relative costs of foraging and defense for male and female purple-throated caribs at the three heliconias. The smaller territory areas and higher flower densities of H. caribaeaterritories lowered males’ foraging time and energy costs per flower relative to red-green H. bihai territories, which theoretically allowed them to meet their energy demands in less time and at lower cost. Males’ estimated foraging time and energy costs were greatest at the green morph of H. bihai, and compared to females, they would save a higher proportion of time and energy by foraging at H. caribaea and the red-green morph of H. bihai. (Fig. 3). This asymmetry in relative gains from foraging at each of the three heliconias for males and females may further reinforce resource partitioning between them, in addition to differences in size and fighting abilities (this research was published as Temeles et al. 2005 Auk with honors students Robin S. Goldman and Alexi U. Kudla).
Fig. 3. Comparisons of foraging time and energy cost budgets for male (circles) and female (triangles) purple-throated caribs at flowers of green H. bihai (A and B), red-green H. bihai (C and D), and H. caribaea (E and F). Foraging times are presented as percentages of 7.5 h during which birds were on territories during daytime (0630–1400 hours) and were computed using the equation for balanced costs and benefits from Gill and Wolf (1975). For these calculations, males were assumed to weigh 10 g and females were assumed to weigh 8 g, and both males and females were assumed to spend 23.87 s per flower at green H. bihai, 15 s per flower at red-green H. bihai, and 10.8 s per flower at H. caribaea. Arrow indicates that calculated foraging time exceeds the time available. Foraging time and energy cost budgets for males and females at flowers of green H. bihai at nectar volumes of 5 μL greatly exceeded time available and are not illustrated in graphs (A) and (B). From Temeles et al. 2005 Auk.
Feeding Performance and Bill Morphology of the Sexes
As noted above, sexes of purple-throated carib hummingbirds exhibit extreme sexual dimorphism in bill morphology, with males having short, straight bills and females having bills 20% longer and 40% more curved. We analyzed feeding performance of males and females in order to understand how these differences in bill morphology between the sexes contribute to niche partitioning. These experiments involved the use of natural Heliconia flowers on St. Lucia and 34 artificial flowers differing in length, curvature, and diameter. The use of artificial flowers allows us to examine birds’ feeding performance at floral phenotypes not found in nature, and gives us insight into how their bills would respond to selection were there environment to change.
Our most exciting finding was evidence for a trade-off in the relative advantages of short, straight bills and long, curved bills. Specifically, the longer, more highly curved bills of female purple-throated caribs allow them to feed from longer, more highly curved flowers than males, and to do so more quickly, when the birds perch on Heliconia bracts to feed (Fig. 4). Males had great difficulty extracting nectar from flowers of the green morph of H. bihai, and in many cases could not insert their bills into flowers of the green morph due to mismatches between bill and flower shape.
Fig. 4. Handling times (HT, log-transformed) of male (M) and female (F) Anthracothorax jugularis at (a) three St. Lucian (West Indies) heliconias for 10 and 20 µl nectar volumes (mean ± SE) and (b-d) in relation to total bill length (BL, only results for 10 µl are shown). For flowers of green Heliconia bihai, handling times of females were significantly shorter than handling times of males at both nectar volumes. Both sexes had significantly shorter handling times at H. caribaea than at either morph of H. bihai (P = 0.05 overall, t-tests and paired t-tests with Bonferroni corrections). (b) Relationship between log(handling time) and total bill length of both sexes at green H. bihai. Regression statistics are: M10HT = 3.32 – 0.103BL, r2 = 70.9%, F1,4 = 9.8, P = 0.04, n = 6; F10HT = 2.41 – 0.063BL, r2 = 93.6%, F1,3 = 43.7, P = 0.01, n = 5. (c) Relationship between log(handling time) and total bill length of both sexes at red-green H. bihai. Regression statistics are: M10HT = 2.81 – 0.094BL, r2 = 75.2%, F1,4 = 12.1, P = 0.025, n = 6; F10HT = -0.23 + 0.025BL, r2 = 7.2%, F1,3 = 0.23, P = 0.66, n = 5. (d) Relationship between log(handling time) and total bill length of both sexes at H. caribaea. Regression statistics are: M10HT = 1.10 – 0.029BL, r2= 16.5%, F1,4 = 0.8, P = 0.42, n = 6; F10HT = 0.28 + 0.002BL, r2 = 0.3%, F1,3 = 0.01, P = 0.94, n = 5. Handling time was measured in seconds. From Temeles et al. 2009 Ecology.
Experiments with artificial flowers supported those with real Heliconia flowers. Only 1 of 6 males was able to extract nectar from an artificial flower of dimensions 44 mm long x 3 mm wide x 0.04 1/mm curvature, replicating the dimensions of the green morph of H. bihai on St. Lucia visited primarily by females. In contrast, even though the bills of males are much shorter than the bills of females, they could extract nectar from the same maximum length of straight flowers while hovering Moreover, males had significantly faster handling times and better feeding performance than females when hovering to feed from perfectly straight flowers (Fig. 5). (This work was published as Temeles et al. 2009 Ecology with honors students Carolyn R. Koulouris and Sarah E. Sander).
Fig. 5. Handling times (mean ± SE, inverse-transformed) of male (M) and female (F) Anthracothorax jugularis at straight artificial flowers (curvature, K = 0 mm-1) measuring 35, 40, 45, and 50 mm in length (L), with 10 µl of nectar. For both sexes, handling times were significantly longer at 50-mm flowers than at shorter lengths (P = 0.05 overall, paired t-tests with Bonferroni adjustments). Handling times of males were significantly shorter than handling times of females at 35-mm and 40-mm flowers (P = 0.05 overall, t-tests with Bonferroni adjustments). From Temeles et al. 2009 Ecology.
Mate Choice and Mate competition in a Tropical Hummingbird
In addition to examining feeding performance of male and female purple-throats in the context of natural selection, we examined sexual selection acting on males. Our earlier studies demonstrated a close correspondence between energy rewards of the heliconias and the body size and energy needs of sexes of purple-throats and suggested that natural selection may have had a role in the evolution of body sizes of the sexes (Temeles et al. 2000, 2005; Temeles & Kress 2003). Because male purple-throats also defend heliconias for breeding, however, we noted that sexual selection may have had a role in the evolution of their larger size. In a study of a population of 12 males defending territories of H. caribaea on the island of Dominica, we found that males defended territories greatly in excess of their energy needs, and that females preferred to mate with males having the highest standing crops of nectar on their territories. Males specifically maintained plants on their territories for use by females, a behavior we have termed “flower farming” (see Fig. 6). A male’s ability to accumulate high standing crops of nectar on his territory depended on his ability to evict intruders, as well as on his size, with larger males being the most successful. Thus, sexual selection contributes to the larger size of male purple-throated caribs (this work was published as Temeles & Kress 2010 Proc. R. Soc. B).
Fig. 6. Flower farming by male purple-throated caribs (Anthracothorax jugularis, formerly Eulampis jugularis). When breeding, male caribs defend an excess of plants which they use to attract females as mates. Males feed from specific plants on their territories, and allow females to feed from a different set of plants. Females prefer to mate with males that have the most nectar in their flowers, which is a function of male size and fighting ability.
Sexual Dimorphism in Hermit Hummingbirds
The purple-throated carib hummingbird and its Heliconia food plants provide some of the best evidence to date for ecological causation of sexual dimorphism, but also raise the question of whether this evidence is simply an isolated case resulting from ecological release on islands or alternatively is more widespread than previously imagined. The purple-throated carib is absent from the islands of Grenada, Tobago, and Trinidad, even though one of the heliconias it visits, H. bihai, occurs there. Another aspect of our work is to determine what hummingbird(s) replace purple-throats geographically as pollinators of H. bihai (and H. caribaea). Heliconia bihai, on these three islands is a generalist with extremely short flowers pollinated by 3 to 6 hummingbird species per island. We discovered that two of these hummingbirds, the rufous-breasted hermit, Glaucis hirsutus, and the green hermit, Phaethornis guy, have sexual dimorphism in bill curvature similar to that of the purple-throated carib. This discovery prompted a phylogenetic analysis of sexual dimorphism within the hermit hummingbirds (Phaethornithinae) in which we determined that sexual dimorphism in bill curvature is widespread within this subfamily (Fig. 7).
Fig. 7. Sexual dimorphism in bill curvature in Phaethornithinae. Values in boxes are bill curvature dimorphisms calculated for extant taxa or reconstructed for ancestral species. Black shading indicates bill curvature dimorphism greater than 30%, whereas white boxes represent moderate (approx. 25%) or no bill curvature dimorphism. Note that the non-hermit Eriocnemis luciani is the only taxon with male-biased bill curvature dimorphism (grey shading). Arrows represent transitions (greater than 5%) in bill curvature dimorphism for hermit hummingbirds, which are either towards increasing female bias (arrows up) or decreasing female bias (arrows down). Photographs of species depict female above and male below. (Phylogeny adapted from McGuire et al. (2007, 2008), figure from Temeles et al. 2010 Phil Trans Roy Soc Lond B).
Moreover, our own field observations together with those of Taylor and White (2007 Ornithol Neotrop.) indicate that the sort of sexual partitioning of heliconias we have documented for purple-throats also occurs in many species of hermit hummingbirds, and that ecological causation of sexual dimorphism may be more common than has been previously thought. (Fig. 8). This work was published as Temeles et al. 2010 Phil. Trans. Roy. Soc. B with honors student Joanna L. Rifkin and molecular phylogenetics wizard Jill S. Miller.
