The predatory behavior of free ranging praying mantises, Sphodromantis lineola (Burmeister), in response to prey at various positions in the midsagittal plane, was examined using high speed (200 frames per second) videography. Predatory strikes fell neatly into two categories based on the elevation of the prey from the surface on which the mantises stood: high strikes and low strikes. When the prey was 35° or more above the surface (measured from the mesothoracic tarsus), mantises assumed a posture that elevated and pointed the body upwards (high strikes). When prey was near or below the surface on which the mantises stood, they assumed a posture that lowered the body and shifted its center of gravity forward (low strikes). Each of these two initial postures was followed by distinctly different constellations of movements, which included a rapid grasping movement of the raptorial forelegs and, if the prey was sufficiently distant, a displacement of the body upwards (high strikes) or forwards (low strikes). Our analyses suggest that prothoracic angle and, to a lesser degree, head angle and the degree to which the mesothoracic legs are extended provide the critical proprioceptive cues used in programming the appropriate attack sequence. Based on our results, we hypothesize that mantises process visual and proprioceptive information indicating prey location in ''pterothorax-centered space''.

The predatory behavior of free ranging praying mantises, Sphodromantis lineola (Burmeister), in response to prey at various positions in the horizontal plane was examined using high speed (200 frames per second) videography: We found that the movements of the meso- and metathoracic legs over the course of the strike were analogous in many respects to those made by the cockroach Periplaneta americana during escape turns. When mantises struck at prey directly ahead of them, they were propelled forward by extensions of the metathoracic femur-tibia, and the meso- and metathoracic coxa-femur joints (changes in the latter were determined indirectly via changes in the femur-pterothorax angles). This pattern of movements is similar to that of cockroach Type 1 turns. However, when prey lay to either side of the pterothorax-abdomen axis, mantises turned toward the prey as they stuck. These turning movements were the result, primarily, of changes in the femur-thorax angles. Specifically, as the mantises turned toward the prey, contralateral mesothoracic femora and metathoracic tibiae and femora extended, and the corresponding ipsilateral joints extended to a lesser degree or flexed. This pattern of movements is similar to that of cockroach Type 2 turns. In addition, these leg movements were accompanied by flexion of the prothorax-abdomen angle which turned the prothorax even further in the direction of the prey. We found a stronger relationship between mantis leg movements and the position of the prey in relationship to the pterothorax than between the leg movements and the position of the prey in the visual field. Our data suggest that the praying mantis'' central nervous system integrates proprioceptive and visual information in order to determine the location of prey in ''pterothorax-centered’ rather than ''head-centered'' space.