21In water, transparency seems an ideal concealment strategy, as testified by the variety of transparent 22 aquatic organisms. By contrast, transparency is nearly absent on land, with the exception of insect 23 wings, and knowledge is scarce about its functions and evolution, with fragmentary studies and no 24 comparative perspective. Lepidoptera (butterflies and moths) represent an outstanding group to 25 investigate transparency on land, as species typically harbour opaque wings covered with coloured 26 2 scales, a key multifunctional innovation. Yet, many Lepidoptera species have evolved partially or 27 fully transparent wings. At the interface between physics and biology, the present study investigates 28 transparency in 123 Lepidopteran species (from 31 families) for its structural basis, optical 29 properties and biological relevance in relation to thermoregulation and vision. Our results establish 30 that transparency has likely evolved multiple times independently. Efficiency at transmitting light 31 is largely determined by clearwing microstructure (scale shape, insertion, colouration, dimensions 32 and density) and macrostructure (clearwing area, species size or wing area). Microstructural traits -33 density, dimensions -are tightly linked in their evolution, with different constraints according to 34 scale shape, insertion, and colouration. Transparency appears highly relevant for vision, especially 35 for camouflage, with size-dependent and activity-rhythm dependent variations. Links between 36 transparency and latitude are consistent with an ecological relevance of transparency in 37 thermoregulation, and not so for protection against UV radiation. Altogether, our results shed new 38 light on the physical and ecological processes driving the evolution of transparency on land and 39 underline that transparency is a more complex than previously thought colouration strategy. 40 41

In water, transparency seems an ideal concealment strategy, as testified by the variety of transparent aquatic organisms. By contrast, transparency is nearly absent on land, with the exception of insect wings, and knowledge is scarce about its functions and evolution, with fragmentary studies and no comparative perspective. Lepidoptera (butterflies and moths) represent an outstanding group to investigate transparency on land, as species typically harbor opaque wings covered with colored scales, a key multifunctional innovation. Yet, many Lepidoptera species have evolved partially or fully transparent wings. At the interface between physics and biology, the present study investigates wing transparency in 123 Lepidoptera species (from 31 families) for its structural basis, optical properties, and biological relevance in relation to visual detection (concealment), thermoregulation, and protection against UV. Our results suggest that transparency has likely evolved multiple times independently. Efficiency at transmitting light is largely determined by clearwing microstructure (scale shape, insertion, coloration, dimensions, and density) and macrostructure (clearwing area, species size, or wing area). Microstructural traits, scale density and dimensions, are tightly linked in their evolution, with different constraints according to scale shape, insertion, and coloration. Transparency appears highly relevant for concealment, with size-dependent variations. Links between transparency and latitude are consistent with an ecological relevance of transparency in thermoregulation, and not so for protection against UV radiation. Altogether, our results shed new light on the physical and ecological processes driving the evolution of transparency on land and underline that transparency is a more complex coloration strategy than previously thought.

In opaque butterflies and moths, scales ensure vital functions like camouflage, thermoregulation, and hydrophobicity. Wing transparency in some species - achieved via modified or absent scales - raises the question of whether hydrophobicity can be maintained and of it dependence on scale microstructural (scale presence, morphology, insertion angle, and coloration) and nanostructural (ridge spacing and width) features. To address these questions, we assessed hydrophobicity in 23 clearwing species differing in scale micro and nanofeatures by measuring static contact angle (CA) of water droplets in the opaque and transparent patches of the same individuals at different stages of evaporation. We related these measures to wing structures (macro, micro, and nano) and compared them to predictions from Cassie-Baxter and Wenzel models. We found that overall, transparency is costly for hydrophobicity and this cost depends on scale microstructural features: transparent patches are less hydrophobic and lose more hydrophobicity with water evaporation than opaque patches. This loss is attenuated for higher scale densities, coloured scales (for erect scales), and when combining two types of scales (piliform and lamellar). Nude membranes show lowest hydrophobicity. Best models are Cassie-Baxter models that include scale microstructures for erect scales, and scale micro and nanostructures for flat scales. All findings are consistent with the physics of hydrophobicity, especially on multiscale roughness. Finally, wing hydrophobicity negatively relates to optical transparency. Moreover, tropical species have more hydrophobic transparent patches but similarly hydrophobic opaque patches compared to temperate species. Overall, diverse microstructures are likely functional compromises between multiple requirements.