Based on previous studies of odor-modulated flight where track parameter data was lumped and averaged, the speed and orientation of the moths' movement along their flight tracks have been said to be controlled to maintain certain "preferred" values. The results from our fine-scaled analysis of this behavior show that none of the track parameters typically measured are held constant. The moths' speed along the flight track is modulated substantially and predictably: fastest along the straight legs and slowest around the turns. In addition, about half of the individuals studied progressively reduced the peak speed along the straight legs as they approached the pheromone source. While most of the track legs between the turns were directed upwind, their orientations were widely distributed, indicating no preferred direction. Small fluctuations of orientation along some straight legs suggest corrective maneuvers to stabilize flight direction about an internal set point. The visual inputs hypothesized to control steering and speed, transverse and longitudinal image flow, changed continuously during upwind flight in pheromone, but no regular relationship between them was observed. We found that the orientation of the longitudinal body axis and the direction of thrust (course angle) were only rarely coincident during upwind flight to the odor source, suggesting that moths receive sensory input which differs quantitatively from that calculated by conventional methods. Our results strongly suggest that the long-accepted hypothetical mechanisms of control for this behavior do not operate in the manner in which they have been proposed.