Wet markets have been implicated in multiple zoonotic outbreaks, including COVID-19. They are also a conduit for legal and illegal trade in wildlife, which threatens thousands of species. Yet wet markets supply food to millions of people around the world, and differ drastically in their physical composition, the goods they sell, and the subsequent risks they pose. As such, policy makers need to know how to target their actions to efficiently safeguard human health and biodiversity without depriving people of ready access to food. Here, we propose a taxonomy of wet markets, oriented around the presence of live or dead animals, and whether those animals are domesticated or wild (either captive-reared or wild-caught). We assess the dimensions and levels of risk that different types of wet markets pose to people and to biodiversity. We identify six key risk factors of wet markets that can affect human health: (1) presence of high disease-risk animal taxa, (2) presence of live animals, (3) hygiene conditions, (4) market size, (5) animal density and interspecies mixing, and (6) the length and breadth of animal supply chains. We also identify key factors informing risk to biodiversity. Finally, we recommend targeted, risk-adjusted policies to more efficiently and humanely address the dangers posed by wet markets. Key messagesMany wet markets do not sell live or wild animals Wet markets are often incorrectly conflated with live-animal or wildlife markets. Wet markets sell consumption-oriented, perishable goods in a non-supermarket setting. By contrast, wildlife markets sell non-domesticated wild animals, either captive-bred or wild-caught, dead or alive. Live-animal markets sell live animals. The Huanan Seafood Wholesale Market in Wuhan, China, a possible source of the COVID-19 pandemic, was a wet market, live-animal market, and wildlife market.

Predation is widely believed to exert strong selective pressure on primate behavior and ecology but is difficult to study and rarely observed. In this study, we describe seven encounters between lone wild leopards (Panthera pardus) and herds of geladas (Theropithecus gelada) over a 6‐year period in an intact Afroalpine grassland ecosystem at the Guassa Community Conservation Area, Ethiopia. Three encounters consisted of attempted predation on geladas by leopards, one of which was successful. All three attacks occurred in low‐visibility microhabitats (dominated by tussock graminoids, mima mounds, or tall shrubs) that provided leopards with hidden viewsheds from which to ambush geladas. An additional four encounters did not result in an attempted attack but still document the consistently fearful responses of geladas to leopards. In encounters with leopards, geladas typically gave alarm calls (n = 7 of 7 encounters), reduced interindividual distances (n = 5), and collectively fled towards or remained at their sleeping cliffs (n = 7), the only significant refugia in the open‐country habitat at Guassa. Geladas did not engage in mobbing behavior towards leopards. Encounters with leopards tended to occur on days when gelada herd sizes were small, raising the possibility that leopards, as ambush hunters, might stalk geladas on days when fewer eyes and ears make them less likely to be detected. We compare the behavioral responses of geladas to leopards at Guassa with those previously reported at Arsi and the Simien Mountains and discuss how gelada vulnerability and responses to leopards compare with those of other primate species living in habitats containing more refugia. Lastly, we briefly consider how living in multilevel societies may represent an adaptive response by geladas and other open‐country primates to predation pressure from leopards and other large carnivores.

No abstract