A theoretical analysis has been made of the longitudinal behavior of an automatically controlled supersonic interceptor during the attack phase against a nonmaneuvering target, and for a target which had a constant acceleration normal to its flight path. First-order lead collision navigation was assumed in the investigation, and characteristics of this navigation when used against a maneuvering target are discussed. The flight path of the interceptor was controlled by commanding either a pitching velocity or normal acceleration proportional to the vertical steering error. Computed attack runs are presented which demonstrate some of the advantages and disadvantages of using high gain or integration in the tracking system to minimize or eliminate bias errors in the system which result from target acceleration or interceptor trim changes.