An improved electro-mechanical detection system for sensing the time of capture for noctuid male moths in Texas pheromone cone traps is described. Operational factors which were found to adversely affect detection accuracy and require design modifications during three seasons of field testing with cotton bollworm, Heliothis zea Boddie, and tobacco budworm, Heliothis virescens (F), moths are discussed. The improved system incorporates the use of a two-stage infrared light beam detection technique, an air pressure capture assist device and a one-way capture passage to enhance detection accuracy. Time of capture data indicated that cotton bollworm nightly capture profiles tended to be bimodal with early peaks occurring before midnight and late peaks occurring one to two hours before dawn throughout the season. Tobacco budworm profiles tended to be unimodal with peak capture occurring two to four hours after midnight. Other facets of cotton bollworm and tobacco budworm capture timing will be discussed.