Experimental designs developed to address mixtures are suited ideally to many areas of experimental biology, including pheromone blend studies, because such designs address the confounding of proportionality and concentration intrinsic to factorial and one-factor-at-a-time designs. Geometric multivariate designs coupled with response surface modeling allowed us to identify optimal blends of a two-component pheromone for attraction and trap disruption of the leafminer moth, Phyllocnistis citrella, a major pest in citrus growing areas around the world. Field trials confirmed that the natural 3:1 blend of (Z,Z,E)-7,11,13-hexadecatrienal:(Z,Z)-7,11-hexadecadienal was most effective as an attractant for male moths. However, the response surface generated in mating orientation trials revealed that the triene component alone was more effective than the natural blend in disrupting trap catch. Each individual component was effective at disrupting orientation in field trials, but (Z,Z,E)-7,11,13-hexadecatrienal was approximately 13 times more effective, at the same concentration, compared with (Z,Z)-7,11-hexadecadienal alone. In addition, the application of geometric design and response surface modeling to field studies provided insight into a possible mechanism of mating disruption and supported sensory imbalance as the operating mechanism for this species.