Discussion
The reflectance of European butterflies followed the ecogeographical patterns predicted by Bogert’s rule: butterfly species in colder regions showed lower reflectance than species in warmer regions in both VIS and NIR wavebands. This pattern was consistent for dorsal and ventral reflectance of all body regions. The consistent pattern for both VIS and NIR wavebands is not surprising because reflectance in these two parts of the spectrum was highly correlated. However, even after removing the effect of this correlation, residual NIR reflectance of the ventral thorax-abdomen and basal wings still showed thermally adaptive patterns. Thus, our results clearly demonstrate that thermal benefits drive ecogeographical patterns of reflectance in European butterflies.
We also found evidence for Gloger’s rule in European butterflies. Delhey proposed two different definitions of Gloger’s rule: a simple version states that animals are darker in a more humid environments while a more complex version includes differential effects of humidity and temperature on different types of melanin pigments (Delhey 2019). Our results generally follow the patterns predicted by the simple version of Gloger’s rule: after accounting for the effect of temperature-related variables (PC1), reflectance of most body regions was lower in species found in a more humid area (i.e. with higher mean precipitation). The trends were consistent for all body regions (coefficients of PC2 in Tables 1 and S5-6 are all negative; although this relationship was not statistically significant for the ventral thorax). This suggests that not only thermal environments but the degree of humidity also affects the ecogeographical patterns of butterfly reflectance. Camouflage in warm and humid environments could drive this relationship because such environments are usually covered by darker vegetations and under low light conditions, favouring the occurrence of darker species (Xinget al. 2016; Cheng et al. 2018).
In accordance with a previous study on Australian butterflies (Munroet al. 2019), we found similar high correlations between VIS and NIR reflectance. This is not surprising because reflectance varies continuously and often gradually across the spectrum and the degree of VIS reflectance generated by pigments, such as melanins, often correlate with their NIR reflectance (Alla et al. 2009). However, structural colour, which is common in butterflies, can produce a wide diversity of spectral shapes with multiple peaks in different parts of the spectrum, potentially enabling VIS and NIR properties to respond differently to selection. Our results suggest that selection for thermal benefits has shaped both VIS and NIR reflectance in European butterflies because both showed patterns consistent with Gloger’s rule. However, the ventral basal wing and thorax regions also showed thermally-adaptive variation independent of their VIS reflectance. This implies that butterfly reflectance might be tuned to modulate signalling or camouflage needs in VIS reflectance and thermoregulatory needs in NIR reflectance despite the constraints imposed by the correlations between them (Munro et al. 2019).
Though entire wing reflectance also showed thermally adaptive ecogeographical patterns, the strength of this relationship was weaker than for the thorax and basal wing regions. The evolution of butterfly reflectance is likely to be affected by multiple competing functions, including camouflage and signalling (Silberglied 1984; Kapan 2001; Chenget al. 2018; van der Bijl et al. 2020). The stronger climate-reflectance relationships for the thorax and basal wing area suggest that the relative importance of thermoregulation is greater for these body regions, consistent with their more critical role in thermoregulation (Wasserthal 1983). Basal wing and thorax regions comprise a smaller area than the entire wing and are pivotal for thermoregulation due to haemolymph circulation and proximity to flight muscles (Arnold 1964), thus they may be less affected by competing selective pressures other than selection for thermal benefits.
Our results show that larger species have lower entire wing reflectance than smaller species in the NIR but not VIS wavebands. In other words, size correlates with NIR reflectance, but not colour. Why have larger butterflies evolved lower NIR reflectance of the wings independent of both climate and colour? Cryptic NIR adaptations of wings could contribute to thermoregulation. Although heat transfer from the wings to the thorax through conduction may be limited, heat transfer may be greater for larger than smaller wings. Larger butterflies have been shown to perform better at controlling temperature and affording elevated body temperature (Gilchrist 1990; Bladon et al. 2020). Alternatively, the wings may function as solar concentrators and reflect the solar energy radiated across the entire wings towards the thorax (Shanks et al. 2015). This effect should be greater for larger species because larger wings can reflect more solar radiation towards the thorax than smaller ones. Thus, larger species may have evolved lower NIR reflectance than smaller species to modulate the amount of reflective energy from the wings to the thorax. A third possibility is that larger butterflies may prefer to be active in the shade and crepuscular hours which could also drive the evolution of lower NIR reflectance (Xing et al. 2016). The underlying reason for the observed size-reflectance relationship remains to be tested.
Butterflies use both dorsal and ventral basking, and both dorsal and ventral reflectance can contribute to the process of heat transfer, depending on basking behaviour (Clench 1966; Kingsolver 1985). However, ventral regions are additionally exposed during cooling down when butterflies close their wings tightly to minimise the absorption of solar radiation (Clench 1966). To avoid the absorption of unnecessary heat during cooling down, it may be equally important to have high ventral reflectance, especially for species in warmer climates. Thus, the reflectance of ventral regions in butterflies may be a result of evolutionary modulation between two conflicting selective pressures: absorbing light energy when heating up and reflecting it when cooling down. In cold climates, selection for low reflectance to enable rapid warming may prevail, while in hot climates, there may be stronger selection for high ventral reflectance to facilitate cooling. Indeed, our results demonstrate that ventral surfaces had higher reflectance than dorsal surfaces in most species and the difference was larger in warmer climates. Notably this relationship was only present for the thorax and basal wing regions that are crucial for thermoregulation. This suggests that the evolution of the ventral surfaces of butterflies is affected by thermoregulatory pressures related to both heating and cooling.
Thermal benefits have been considered as one of the major selective agents that operate on butterfly reflectance (Kingsolver 1988; Hegnaet al. 2013). Our findings provide the most comprehensive evidence to date that climatic gradients have shaped both visible and near-infrared reflectance of butterflies consistent with both Gloger’s and Bogert’s rules. We also show that not all body regions were equally affected, but the observed climate-reflectance relationship was stronger for body regions that play a greater role in thermoregulation. This highlights that the relative strength of competing selective pressures (e.g. signalling, camouflage, heating up, or cooling down) may vary between different body parts and these collectively have affected the evolution of the reflectance properties of butterflies.