COOK, A.P.; TRIBLEHORN, J.D.; LORSONG, A.; LENNON, M.; YAGER, D.D.*; Univ. of Maryland, College Park; Georgia State Univ., Atlanta; Univ. of Maryland, College Park; Univ. of Maryland, College Park; Univ. of Maryland, College Park: Praying Mantis Evasive Response: Descending Control and Context Gating

Most praying mantids have a single midline ear in the ventral metathorax. Flying mantids that hear ultrasound perform a bat evasion behavior that comprises a head roll, complete prothoracic leg extension, wing beat changes, and abdomen dorsiflexion. The first movements begin 50-70 ms after the stimulus and are completed by 250-350 ms. Control of the behavior could be local (�reflex�) or require brain processing. Using normal and decapitated mantids in tethered flight, high-speed video showed that the head is required, thus sacrificing speed for greater control. Relative latencies (neural + biomechanical) measured from high-speed videos show that component activation is not sequential. The arms most often move first followed by the head or the wings. This result suggests multiple descending commands, which is further supported by neural recordings of ultrasound-triggered descending activity that show complex, multi-unit responses. The ultrasound-triggered evasive behavior is strictly context-dependent. It only occurs if the flight system is activated and never in standing or non-flying tethered animals. Mantids perform a defensive display when standing that contains all the components of the in-flight evasive behavior and is probably its evolutionary precursor. However, the defensive display is triggered by vision or touch, never by sound. Headless mantids perform the defensive display normally when prodded. Thus, context determines not only the effective stimuli, but also the mode of neural control (descending vs. local).