The mantis prey recognition algorithm: Species variations on a theme

The praying mantises are a diverse group of charismatic, predatory insects. Their allure is evidenced in a rich body of folklore and natural history literature spanning four millennia but their scientific appeal, though strong, has been kept in check by a dearth of information regarding all but a few species. However, recent discoveries have provided insights into the remarkable heuristic potential of these insects. In the wild, mantises are opportunistic predators that capture a wide variety of prey ranging from small arthropods to same-sized conspecifics and, occasionally, small vertebrates. The mix of prey eaten is limited only by the rate at which it is encountered and can be successfully captured. Given the variety of retinal images created by such diverse prey, it is clear that mantises do not identify prey by a matching-to-template strategy. Instead, they identify an object as prey if it falls within a broad but well circumscribed perceptual envelope. If threshold levels of a sufficient number of key stimulus parameters are met, and motivational state is appropriate, the mantis will attempt a capture. Extensive behavioral and psychophysical studies in our lab have produced a coherent body of data demonstrating that the perceptual envelope (or “schema,”) that defines prey for mantises is based an assessment of ten stimulus parameters: (1) overall size (if the stimulus is compact); (2) leading edge length (if the stimulus is elongated); (3) contrast to the background; (4) location in the visual field; (5) speed; (6) direction; (7) geometry; (8) retinal distance traversed; and, (9,10) the degree to which sub-threshold stimulus elements are summed over space and time. Although this algorithm is similar across species, there are interesting differences which seem not to be related to overall mantis size. Rather, they appear to be related to species differences in compound eye morphology.