Genuine explosive materials are traditionally employed in the training and testing of explosive-detecting canines so that they will respond reliably to these substances. However, challenges arising from the acquisition, storage, handling, and transportation of explosives have given rise to the development of "pseudo-explosive" training aids. These products attempt to emulate the odor of real explosives while remaining inert. Therefore, a canine trained on a pseudo-explosive should respond to its real-life analog. Similarly, a canine trained on an actual explosive should respond to the pseudo-explosive as if it was real. This research tested those assumptions with a focus on three explosives: single-base smokeless powder, 2,4,6-trinitrotoluene (TNT), and a RDX-based plastic explosive (Composition C-4). Using gas chromatography-mass spectrometry with solid phase microextraction as a pre-concentration technique, we determined that the volatile compounds given off by pseudo-explosive products consisted of various solvents, known additives from explosive formulations, and common impurities present in authentic explosives. For example, simulated smokeless powders emitted terpenes, 2,4-dinitrotoluene, diphenylamine, and ethyl centralite. Simulated TNT products emitted 2,4- and 2,6-dinitrotoluene. Simulated C-4 products emitted cyclohexanone, 2-ethyl-1-hexanol, and dimethyldinitrobutane. We also conducted tests to determine whether canines trained on pseudo-explosives are capable of alerting to genuine explosives and vice versa. The results show that canines trained on pseudo-explosives performed poorly at detecting all but the pseudo-explosives they are trained on. Similarly, canines trained on actual explosives performed poorly at detecting all but the actual explosives on which they were trained.