Adult cats develop spontaneous airstepping (walking motions without ground contact) 4-6 weeks after spinal transection (Giuliani and Smith, 1985). This unique preparation provides an in vivo model for studying the effects of hindlimb deafferentation on stepping behaviors without rostral input to lumbosacral segments. The primary purpose of this study was to characterize airstepping and, to a lesser extent, bipedal treadmill locomotion in chronic spinal cats after deafferentation. Five cats were spinalized at T12, and EMG electrodes were implanted in selected knee and ankle muscles. After spontaneous airstepping developed, one hindlimb was deafferented extensively. Movements of both hindlimbs were depressed following unilateral deafferentation. By the third week, spontaneous airstepping recovered in the normal hindlimb (N-Limb), but in the deafferented limb (D-Limb) airstepping was not spontaneous during the 3-4 months of testing. By the seventh week, bilateral airstepping was facilitated by tonic exteroceptive simulation (tail-pinch). During bouts of bilateral airstepping, assessed during the third month of deafferentation, D-Limb airstep cycles were characterized by erratic rhythm. Further, cycle periods and extensor burst durations were prolonged, but flexor burst durations were unmodified. In contrast, N-Limb cycles were rhythmic and of normal duration, but extensor burst durations were short and unrelated to cycle periods. Interlimb coordination was markedly unstable, showing only brief periods of alternating steps. When erratic bilateral airstepping was accompanied by micturition, rhythmic and alternate stepping emerged, with normalized intralimb synergies. During hindlimb treadmill locomotion, tested in 2 cats, the N-Limb took weight-bearing steps and followed changes in the belt speed. The D-Limb, however, stepped erratically, making contact on the dorsum of the paw; consistent bouts by bilateral stepping were not obtained. Lumbosacral afference of some type, either from the hindlimb or from regions remote from the limb, such as the bladder, appears to be essential for stabilizing the coupling between hindlimb generators for stable interlimb coordination. Future modeling of interlimb coordination should consider the role of afference.