Synopsis We raise and explore possible answers to three questions about the evolution and ecology of silent flight of owls: (1) do owls fly silently for stealth, or is it to reduce self-masking? Current evidence slightly favors the self-masking hypothesis, but this question remains unsettled. (2) Two of the derived wing features that apparently evolved to suppress flight sound are the vane fringes and dorsal velvet of owl wing feathers. Do these two features suppress aerodynamic noise (sounds generated by airflow), or do they instead reduce structural noise, such as frictional sounds of feathers rubbing during flight? The aerodynamic noise hypothesis lacks empirical support. Several lines of evidence instead support the hypothesis that the velvet and fringe reduce frictional sound, including: the anatomical location of the fringe and velvet, which is best developed in wing and tail regions prone to rubbing, rather than in areas exposed to airflow; the acoustic signature of rubbing, which is broadband and includes ultrasound, is present in the flight of other birds but not owls; and the apparent relationship between the velvet and friction barbules found on the remiges of other birds. (3) Have other animals also evolved silent flight? Wing features in nightbirds (nocturnal members of Caprimulgiformes) suggest that they may have independently evolved to fly in relative silence, as have more than one diurnal hawk (Accipitriformes). We hypothesize that bird flight is noisy because wing feathers are intrinsically predisposed to rub and make frictional noise. This hypothesis suggests a new perspective: rather than regarding owls as silent, perhaps it is bird flight that is loud. This implies that bats may be an overlooked model for silent flight. Owl flight may not be the best (and certainly, not the only) model for “bio-inspiration” of silent flight.

Two hypotheses have been proposed for the evolution of structures that reduce flight sounds in birds. According to the stealth hypothesis, flying quietly reduces the ability of other animals (e.g. prey) to detect the animal’s approach from its flight sounds. This hypothesis predicts that animals hunting prey with acute hearing evolve silencing features. The self-masking hypothesis posits that reduced flight sounds permit the animal itself to hear better (such as the sounds of its prey, or its own echolocation calls) during flight. This hypothesis predicts that quieting features evolve in predators that hunt by ear, or in species that echolocate. Owls, certain hawks and nightbirds (nocturnal Caprimulgiformes) have convergently evolved a sound-reducing feature: a velvety coating on the dorsal surface of wing and tail feathers. Here we document a fourth independent origin of the velvet, in the American Kestrel (Falco sparverius). Among these four clades (hawks, falcons, nightbirds, owls), the velvet is longer and better developed in wing and tail regions prone to rubbing with neighboring feathers, apparently to reduce broadband frictional noise produced by rubbing of adjacent feathers. We tested whether stealth or self-masking better predicted which species evolved the velvet. There was no support of echolocation as a driver of the velvet: Oilbird (Steatornis caripensis) and Glossy Swiftlet (Collocalia esculenta) each evolved echolocation but neither had any velvet. A phylogenetic least squares fit of stealth and self-masking (to better hear prey sounds) provided support for both hypotheses. Some nightbirds (nightjars, potoos and owlet-nightjars) eat flying insects that do not make much sound, implying the velvet permits stealthy approach of flying insects. One nightbird clade, frogmouths (Podargus) have more extensive velvet than other nightbirds and may hunt terrestrial prey by ear, in support of self-masking. In hawks, the velvet is also best developed in species known or suspected to hunt by ear (harriers and kites), supporting the self-masking hypothesis, but velvet is also present in reduced form in hawk species not known to hunt by ear, in support of the stealth hypothesis. American Kestrel is not known to hunt by ear, and unlike the other falcons sampled, flies slowly (kite-like) when hunting. Thus the presence of velvet in it supports the stealth hypothesis. All owls sampled (n = 13 species) had extensive velvet, including the Buffy Fish-owl (Ketupa ketupu), contrary to literature claims that fish-owls had lost the velvet. Collectively there is support for both the self-masking and stealth hypotheses for the evolution of dorsal velvet in birds.