Correlated evolution between colour conspicuousness and drum speed in woodpeckers

Abstract

Sexual selection drives the evolution of many spectacular animal displays that we see in nature. Yet, how selection combines and elaborates different signal traits remains unclear. Here, we investigate this issue by testing for correlated evolution between head plumage colour and drumming behaviour in woodpeckers. These signals function in the context of mate choice and male-male competition, and they may appear to a receiver as a single multimodal display. We test for such correlations in males of 132 species using phylogenetic linear models, while considering the effect of habitat. We find that the plumage chromatic contrast is positively correlated with the speed of the drum, supporting the idea that species evolving more conspicuous plumage on their head also evolve faster drum displays. By contrast, we do not find evidence of correlated evolution between drum speed and head colour diversity, size of the head's red patch, or extent of the plumage achromatic contrast. Drum length was not correlated with any of the plumage coloration metrics. Lastly, we find no evidence that habitat acts as a strong selective force driving the evolution of head coloration or drumming elaboration. Coevolution between different signal modalities is therefore complex, and probably depends on the display components in question.

Keywords: correlated evolution; drumming; plumage coloration; woodpeckers.

Figure 1.

(a) Neighbour-joining tree representing the variation of the head plumage coloration of male woodpeckers obtained by calculating colour distances between each pair of images. Note that this tree does not depict the phylogenetic relationship among species, but rather shows how species clump together as a function of colour quantification through our analyses (see electronic supplementary material, methods). Units are Euclidean distance in CIE Lab colour space (a.k.a. units of ‘ΔE') with a range between 0 and 256, where 256 reflects the most different. (b) Colour segmentation performed using the ‘recolorize' package. A PNG with all non-plumage components masked using transparency (original) is loaded into the package and run through the ‘recolorize2' function, which bins all pixels into a fixed set of colour bins and then combines any colour clusters which are less than 55 ΔE apart in CIE Lab colour space, resulting in a discrete set of distinct colour patches (recolorized), along with a colour palette (palette). Use of modified illustrations was granted with permission by ©Lynx Edicions.

Figure 2.

PGLS regressions between drum speed and plumage head coloration in woodpeckers. Dashed regression line represents the model for which significance was found. (a–d) represent the models between colour diversity, chromatic contrast, achromatic contrast and red area with residuals (res) of drum speed (after controlling for body mass). Data for all variables was log₁₀ transformed. Use of modified illustrations was granted with permission by ©Lynx Edicions.

Figure 3.

PGLS regressions between drum length and plumage head coloration in woodpeckers. (a–d) represent the models between colour diversity, chromatic contrast, achromatic contrast and red area with drum length. Data for all variables were log₁₀ transformed. Use of modified illustrations was granted with permission by ©Lynx Edicions.