Aging Coyotes Using Dental Characteristics

The accepted methods of age determination in the coyote (Canis latrans) are either highly subjective and unquantifiable or expensive and require the extraction of the canine tooth. Since neither of these methods are ideal, their limitations have impeded research on this species. Therefore, it was my objective to (1) develop and test the accuracy and precision of a descriptive key based on tooth wear patterns on the lower canine tooth, (2) develop and test the reliability of multiple regression models for aging coyotes using measurements from extracted teeth, and (3) suggest criteria for improving the consistency of results using these techniques. From a sample of 996 teeth collected from coyotes that had been previously aged by counting cementum annuli, a subsample of 303 teeth were carefully examined for characteristic tooth wear patterns. These characteristics were used to develop an illustrated tooth wear key that could be used to assign coyotes to 1 of 7 age classes: 0.5, 1.5, 2.5, 3.5, 4.5, 5.5, and ≥ 6.5 years. Using the illustrated key, I estimated the age of a subset of 203 of these teeth. I correctly aged 138 of the 203 (68%) teeth and of the remaining 65 teeth 58 (89%) were aged within one year. My estimated ages were highly correlated to the assigned ages (r = 0.882). Four other readers using the key and composite estimated the age of 20 teeth. The four readers had a mean coefficient of variation (CV) of 27.9, ranging from 10.8-35.6. The most accurate reader aged 16 of 20 (80%) teeth correctly and the least accurate 10 of 20 (50%). Older individuals tended to be underaged. The second age determination technique came from multiple regression models. Multiple regression models were developed based on a series of 12 measurements taken from the 303 teeth. The measurements taken were: total tooth length, minimum root length, maximum root length, maximum root width, maximum root thickness and crown width, crown thickness, maximum crown width, maximum crown thickness, anterior crown length, posterior crown length and pulp width. No single measurement could be used to determine age or sex because of overlap in the ranges of measurements. However, there were significant differences (p ≤ 0.001) between measurements for the sexes, except for pulp width (p = 0.689). Therefore, 3 multiple regression models were developed: one for males, one for females and one for both sexes combined. All were significant predictors of age (p ≤ 0.001). The male model was the least accurate and the least precise. These models cannot be used on living coyotes without extracting the canine because each model required at least 2 root measurements. To investigate the precision of measuring teeth, 5 people measured 11 of the 12 measurements (excluding pulp width) on 20 teeth. The measurements of posterior crown length and maximum crown width were the least precise with CVs of 8.6 and 8.8. As expected, total tooth length was the easiest to measure with a CV of 1.5. Because there is some overlap in age-classes using either of these techniques (tooth wear or multiple regression), they probably are not as accurate as ages determined from cementum annuli aging which is usually cited as the most accurate method. But, they can be an alternative when either the tooth cannot be extracted from living coyotes or when time and expense are concerns