Field capability of dogs to locate individual human teeth

Cablk, M.E. and J.S. Sagebiel. 2011. Field capability of dogs to locate human teeth. Journal of Forensic Sciences. 56(4):1018-1024.

This work was funded by Desert Research Institute’s competitive Institutional Project Assignment fund.
For a copy of the published paper please email Dr. Cablk: This email address is being protected from spambots. You need JavaScript enabled to view it.

cablk dog locates teeth1Grid lines established for tooth placement. Grid lines were removed post-placement of teeth and no false alerts were reported on these factors.
Avulsed teeth can be difficult if not impossible to recover in the outdoor environment, yet are important for victim identification. Dogs have an advantage as a tool to locate teeth in that they rely primarily on olfactory rather than visual cues and their olfactory sense exceeds man-made equipment. Each of three dog teams searched two separate (10m)2 plots containing ten teeth each. Results of showed that dog team capability to recover individual teeth was as high as 79 percent and as low as 25 percent. Dog teams worked between 27 and 50 minutes in each plot, with longer time spent working yielding higher recovery rates. We also analyzed training data over the seventy-eight days immediately preceding the trials. Dog team capability in the field trials correlated with capability in training. Results show that dog teams are capable of recovering individual human teeth in the field setting with high precision and accuracy and that training records demonstrate a team’s expected field performance.

Study Discussion and Conclusions:
cablk dog locates teeth2Numbered flags were placed by handlers and pink flags are recoverable teeth. Distance between flags was measured. Given that many of the components in teeth are the same tissues found elsewhere in the body, it is reasonable to expect that dogs trained in human remains detection would find teeth. Furthermore, extracted human teeth main contain blood or closely-associated tissue that may also contribute to the scent picture. However, many of these teeth were quite clean and on visual inspection contained no other tissues. The teeth had been extracted, most displayed gross disease states, and many had dental work; teeth with extensive dental work where the majority of the tooth appeared to be artificial were excluded. Even though these tissues are similar and HRD trained dogs should be able to detect them, the training process was specifically designed to expose the dogs to the same type of sources they would be exposed to in the field trials. It is common in HRD training to expose the dogs to as many different sources as reasonable to help expand their scent picture as to what constitutes human remains. Nevertheless, the challenge of finding such small and limited scent sources as single human teeth should not be underestimated.

cablk dog locates teeth3A tooth both marked by handler as found by dog and identified as recovered.

Using dog teams certified to an accepted, recognized standard was important from a research perspective in that the dog teams were all working from least at the same minimum standard of capability; however our results showed that this criteria yielded teams that had variable recovery rates in the field (25-79% success). Training records for the three teams were consistent, but their field performances varied tremendously. Therefore the criteria for selecting teams to find teeth in the field may merit additional requirements. Because the three teams’ training records and cumulative recovery curves were similar it could be that the lowest performing team lacked enough blind problems to be consistent under the additional pressure of participating in a research context (or a real search scenario with law enforcement and detectives present). However the similarity of cumulative recovery curves coupled with the difference in this team’s search time for plots and the fact that this same team was not skilled at in-field calibration suggests that it is the handler that affects the team’s performance; particularly given that the teams each demonstrated a minimum level of proficiency at human remains detection through testing to a recognized standard and through training records.

The findings of the relationship between the cumulative recovery rate in blind problems and the performance in the field trials strongly suggests that training with blind problems is an essential component of the training regimen for any detector dog. The further finding that it was not the total number of targets encountered, but the total number of days over which training sessions were held highlights the need for training accountability to be incorporated in to detector dog training guidelines, particularly for such forensically important components as teeth. Additionally, the importance of blind problem training and blind problem testing is clearly shown in these results. Training and testing programs for detector dogs should rely heavily on both. Maintaining accurate training records such as those kept in this study would also allow teams to have an on-going assessment program and an ability to more accurately present their probability of detection to the search managers, detectives, and investigators.

While this study clearly demonstrated that dogs are capable of finding single human teeth in a field setting, it also showed the variability in capability in dog/handler teams. Much of this could be connected with both the type and amount of training the teams did prior to the field trials. Blind problems, where both the number and location of sources was unknown to the handler, are the best at developing the skills of the team to find sources in a simulated search environment.