Hummingbirds represent a special case of Rensch's rule, as they display mixed allometries and a predominance of male-biased body mass, which is the general trend for most bird groups (Székely et al. 2007). Using a similar analysis for 154 hummingbird species, Colwell (2000) reports mixed allometry and a slope value (0.84) nearly identical to the one found here (0.8630). Abouheif and Fairbairn (1997) propose that if 80% or more of the species in a group have males larger than females, the taxon could be characterized as male biased, which applies in this study since 80% of the 45 hummingbird species examined here showed male-biased SSD in body mass.
Does SSD vary with body size?
In contrast to our hypothesis, our analysis did not follow a trend of increasing variation in SSD from small to large hummingbirds. We expected that greater SSD would be more common in large species. The smaller and more energetically limited species would show a lower capacity for high variation in SSD. Implicit in our hypothesis was the expectation that high variation in SSD could foster a higher differentiation in intersexual niche divergence. In contrast, we observed a relatively modest variation in SSD (Székely et al. 2007). Although we might expect differences in SSD to translate into changes in functional performance, such as competition for resources (Maglianesi et al. 2022) and in the intensity of intersexual and intrasexual selection (Payne 1984), hummingbirds can show highly plastic behavior buffering the effects of morphological differences. Intra and interspecific differences in size, while important in determining competitive hierarchies and resource access, can be buffered through behavioral changes reflected in opportunistic foraging as well as nectar robbing (Ornelas 1994; Boehm 2018). Behavioral plasticity makes it possible to circumvent morphological limitations in accessing resources and this takes place across the whole range of body mass.
However, in the small dataset, the magnitude of the SD in body mass, wing chord, and wing area increased with trait magnitude and hummingbird size only in males. This suggests that selective pressures act differentially on the sexes and potentially influence sexual niche segregation. The primary causes of this variation need to be further explored through field studies that quantify intraspecific and intersexual morphological variation. SSD has been analyzed using the ratio of average values of morphological characters for males and females, usually obtained from museum specimens and very small samples, which ignores the intraspecific and intersexual variation of the actual character. This flaw has been pointed out before (Smith 1999), and as found here, the variation associated to one trait and sex is lost when the analysis is done using the ratio of average values, which does not incorporate the variation around the mean.
Causes of variation in SSD
Although hummingbird species fit Rensch's rule by showing mixed allometry, 48% of the species displayed a modest variation in SSD for body mass (less than 10%). Only 5 species, including the small bees and coquettes, and the Stripe-throated Hermit, Phaethornis striigularis (mean body mass 2.6 g), showed female-biased SSD, or reverse sexual size dimorphism (RSD), ranging 10 to 17 %. The largest species, C. hemileucurus, exhibited a large variation in SSD biased towards males. In this species, the male is, on average, 36% heavier than the female. Body size in this case provides a competitive advantage to access food resources. We have observed that male C. hemileucurus dominate competitive interactions around artificial feeders (displacing C. hemileucurus females, as well as males and females from other 6 species).
However, the average variation in the absolute value of SSD for body mass across species was 12%, which we considered moderate (Fairbairn 2007). Instead of scaling with body size, the hypothesis of physiological limits to SSD could apply to hummingbirds as a group, regardless of body size. This is because the high metabolic rates of hummingbirds in general, their high dependence on access to high-energy food (which must be supplied almost constantly), and the high energy costs imposed by a small size and hovering flight, may limit the range of variation in body size altogether, as well as between sexes, despite of different selective pressures leading to SSD in males and females.
Mating competition and Aerial agility hypotheses
Eighty percent of the hummingbird species examined here showed male-biased SSD. Large males are usually better competitor for food resources and are better in combat for access to females. However, for the rest of the species with RSD, aerodynamics, and maneuverability (aerial agility) could affect SSD and favor intersexual niche divergence since these characters are subject to different selective pressures between males and females. Males of smaller species do complex, acrobatic displays for females, and thus, morphological characters related to aerodynamics, such as wing area and wing loading, could determine their reproductive sucess in small species such as S. flammula and S. scintilla (Clark et al. 2011) and could affect SSD in these species. In small species, the cost of reproduction (e.g., egg production) would be higher for females, favoring larger size in this sex relative to males (Wheeler and Greenwood 1983).
As in hummingbirds, bustards (Otididae) display mixed allometries, with species having species in which males can be three times heavier than females, species that are monomorphic (isometry), and species showing RSD (i.e., males < females). In this group, smaller males are common in species showing more agile sexual displays, supporting the aerial agility hypothesis (Raihani et al. 2006). This is also the case of Falco tinnunculus where males perform complex acrobatics to attract females and are 10% smaller than females (Hakkarainen et al.1996). In shorebirds and gulls where males do more acrobatic displays, males are smaller than females (Székely et al. 2000).
Other alternative explanations, such as the intersexual niche divergence hypothesis, where sexes have evolved different sizes to lessen intersexual competition for food, and the small-male hypothesis, where small males are more efficient foragers (Krüger 2005), could combine to explain the differences in SSD observed here for large and small hummingbird species. There is a dearth of field experiments necessary to test these hypotheses, but the results found here could orient future research to examine the causes of SSD in species of hummingbirds showing large and small variation in SSD. For instance, in small species such as S. flammula (SSD = 9.6%), we have observed habitat segregation. Males are more common in the Páramo whereas females prefer forest edges and the interior of oak forests. Intersexual habitat segregation has been reported in S. sasin and S. rufus in California, United States, where males and females forage in different habitats (Howell and Gardali 2003).
Other morphological traits influencing SSD
Body mass is not the only morphological trait with a significant effect on competitive performance, foraging behavior, or resource access despite of being the most analyzed morphological character used to summarize differences in SSD (i.e., Fairbairn 1997; Colwell 2000). Body mass is regularly recorded in field studies as well as in museum specimens. The influence of other traits on SSD should be further explored. For instance, the size of the hallux allows mountain gem species with relatively large legs, such as Panterpe insignis, to perch while foraging, which saves energy by avoiding hovering flight (R. Colwell, pers. com.) In addition, long-billed hummingbirds could access a wider variety of corolla lengths than short-billed ones, including legitimate and non-legitimate visits. Opportunistic and non-specialized foraging could be common in many pollination networks (Simmons et al. 2019). The length and morphology of the bill can be of great importance affecting foraging for nectar and insects (Feinsinger and Colwell 1978; Rico-Guevara et al. 2019), but so is the presence of bill serrations that favor nectar robbing by piercing corollas, as well as insect foraging (Ornelas 1994). These bill serrations are secondary sexual characters influencing competitive interactions between males (Rico-Guevara and Araya-Salas 2015).
We found significant heterogeneity in the SSD of bill length that was not associated to body mass, and that did not conform to Rensch´s rule, despite that many species were clearly dimorphic in bill length. In addition to bill length, sexual dimorphism in bill curvature was inversely related to the overlap in the use of flower resources in 31 species of hummingbirds (Maglianessi et al. 2022), although the pattern was highly variable and affected by territorial behavior.