A critical aspect of canine explosive detection involves the animal's ability respond to novel, untrained odors based on prior experience with training odors. In the current study, adult Labrador retrievers (N = 15) were initially trained to discriminate between a rewarded odor (vanillin) and an unrewarded odor (ethanol) by manipulating scented objects with their nose in order to receive a food reward using a canine-adapted discrimination training apparatus. All dogs successfully learned this olfactory discrimination task (≥80 % correct in a mean of 296 trials). Next, dogs were trained on an ammonium nitrate (AN, NH4NO3) olfactory discrimination task [acquired in 60-240 trials, with a mean (±SEM) number of trials to criterion of 120.0 ± 15.6] and then tested for their ability to respond to untrained ammonium- and/or nitrate-containing chemicals as well as variants of AN compounds. Dogs did not respond to sodium nitrate or ammonium sulfate compounds at rates significantly higher than chance (58.8 ± 4.5 and 57.7 ± 3.3 % correct, respectively). Transfer performance to fertilizer-grade AN, AN mixed in Iraqi soil, and AN and flaked aluminum was significantly higher than chance (66.7 ± 3.2, 73.3 ± 4.0, 68.9 ± 4.0 % correct, respectively); however, substantial individual differences were observed. Only 53, 60, and 64 % of dogs had a correct response rate with fertilizer-grade AN, AN and Iraqi soil, and AN and flaked aluminum, respectively, that were greater than chance. Our results suggest that dogs do not readily generalize from AN to similar AN-based odorants at reliable levels desired for explosive detection dogs and that performance varies significantly within Labrador retrievers selected for an explosive detection program.
Keywords: Canine; Cognition; Dog; Olfactory discrimination.