Military Working Dogs (MWDs) are the most effective and versatile means of identifying explosives in combat and low-intensity conflicts. These dogs are exceptionally sensitive sensors for all known threat substances and existing dogs can be trained to meet new threats by detecting new substances in very rapid order. Temperamental factors (such as emotional stability and boldness) rather physical abilities (such as olfaction and endurance) are the primary determinant of MWD effectiveness in the field. However, several elements of the current temperament assessment program are suboptimal from the standpoint of scientific assessment: Little attention is paid to psychometric criteria, measurement methods are unrefined, little is known about tasks be predicted, and little is known about the performance of individual MWDs in the field. This work will develop and psychometrically validate a set of temperament measures designed to improve the effectiveness with which MWDs are selected, trained, and utilized in combat situations. Specifically, the work will implement more sensitive measurement methods, will undertake a task analysis of MWD behavior in the field, and will create a system to track how well individual dogs perform. The work will build on earlier research, which has shown that temperament can be assessed in dogs and other species with high degrees of reliability and validity. The work will follow three cohorts of dogs through selection, training, and utilization to evaluate (1) the psychometrics of the temperament tests, (2) the basic structure underlying differences in temperament, (3) the relative merit of different assessment methods, and, most crucially, (4) the degree to which performance in the field can be predicted from tests conducted when the MWDs are acquired and before and after training. The findings will be used to devise a new selection, training, and evaluation strategy to improve the Military Working Dog program.