A trade-off between hydrodynamic performance and morphological bias limits the evolution of symmetric lattice animal wings

Abstract

Bristled and membranous insect wings have co-evolved despite apparently serving the same functionality. We emulate flight physics using an automated free-fall experiment to better understand how and why several distinct wing forms may have developed. Biomimetic two-dimensional lattice animals were laser cut from a continuous sheet of paper, and their descent in a settling tank was tracked using a camera. Data from 31 generic symmetric polyominos (1,692 experiments) reveal that morphology impacts the drag coefficient $C_D$ and hence flight efficiency. Some polyominos rapidly sediment while others remain suspended for longer. Positioning the search for an optimal shape within an evolutionary context, we relinquished control of the automated setup to a genetic algorithm. Hereditary information passes between generations in proportion to fitness and is augmented by rare mutations. High-performing morphologies are observed, but {experimental repeats do} not re-evolve the same shapes. Adaptation rates also differ if the selective pressure is reversed from suspension to sedimentation. Our data (50,086 experiments) reveal several physical sources of indeterminism in the simulated natural selection process, including fluid flow variability and morphological entropy. This provides experimental support for the idea that the lack of convergence in insect wing shape may partly reflect the competition between fitness and entropy, making it nearly impossible to achieve unique optimal forms.