Accuracy and Precision of Estimating Age of Gray Wolves by Tooth Wear

 USGS Northern Prairie Research Center




We used skulls and teeth of 27 wolves from Alaska of known-minimum age, and for which estimates of actual age were available (Ballard et al. 1987, 1995), to develop criteria for assigning wolves to yearly age classes based on tooth wear. The known-minimum age of Alaska wolves differed by 0-7 years from estimated age. We evaluated the accuracy and precision of tooth wear by testing the ages of 20 wolves of known-age from Minnesota and Ontario. Ages were also estimated by Matson's Laboratory (Milltown, Montana, USA) by sectioning canine and-or premolar teeth and counting cementum annuli.

The Alaska skulls were arranged in 1-year age classes from youngest to oldest, <1-13 years old (Table 1). Progressive wear on canines, incisors, and carnassials was described for wolves of each age class. Three of the authors (Ballard, Gipson, and Nowak) then independently estimated the ages of 20 known-age wolves from Minnesota and Ontario by comparing wear on their incisors, canines, and carnassials to the Alaska collection. Next, we collectively compared tooth wear on the 20 wolf skulls and arrived at a consensus age estimate of each wolf. We were not aware of the actual ages or origins of the wolves which included 12 wild wolves and 8 captives from Minnesota and Ontario.

Table 1.  Skulls of wolves from Alaska used to define year age classes.
Assigned age class Skull number Basis for assigned age Known-minimum age in yr
 <1 122027 Killed as pup <1
  122044 Killed as pup <1
  122065 Killed as pup, cementum annuli <1
  122127 Killed as pup, cementum annuli <1
  122151 Killed as pup, cementum annuli <1
  122421 Killed as pup <1
  122456 Killed as pup <1
  122559 Killed as pup <1
1-2 122073 Cementum annuli n.a.a
  122152 Tagged as pup, cementum annuli 1.8
  122252 Cementum annuli n.a.
  122440 Tooth wear n.a.
2-3 122148 Cementum annuli, tooth wear n.a.
  122170 Cementum annuli, tooth wear n.a.
3-4 122135 Tagged as adult, cementum annuli 1.8
  122143 Cementum annuli, tooth wear n.a.
  122368 Tagged as adult, tooth wear 3.0
4-5 61977 Tooth wear


5-6 122009 Tagged as adult, cementum annuli 3.8
  122018 Tooth wear n.a.
  122038 Tagged as adult, cementum annuli 2.5
  122081 Cementum annuli n.a.
  122251 Tagged as adult, cementum annuli 1.8
6-7 122255 Tooth wear n.a.
7-8 122136 Tagged as adult, cementum annuli 3.5
9-8 122174 Cementum annuli, tooth wear n.a.
13-14 122094 Tagged as adult, tooth wear 8.0
a n.a. = Not available.

Both accuracy (i.e., the proximity of the age estimates to the true ages) and precision (i.e., the repeatability of age estimates) of our age estimates and those of the commercial aging service were evaluated. We followed the recommendations of Campana et al. (1995), who examined the value of statistical and graphical methods for determining the consistency of fish age estimates. First, age-bias graphs were developed for diagnosing systematic differences between age determinations based on tooth wear and by the commercial aging service that counted cementum annuli. The age-bias graphs show known ages of the wolves along the horizontal axis and estimates of age by each reader on the vertical axis. Age estimates that are parallel but separate from the 1:1 equivalent line, or that diverge as the lower or upper age limit is approached, indicate systematic bias.

Next, to compare the precision of the 4 biologists, we estimated the coefficient of variation for their estimates relative to the known ages. We calculated individual coefficients of variation of the age estimate for each wolf, then averaged across wolves to produce a mean coefficient of variation for the reader (Campana et al. 1995). An estimate of the coefficient of variation was expressed as the ratio of the standard deviation to the mean (Campana et al. 1995). Higher coefficient of variation values indicate lower precision.

We used linear regression of known ages and estimated ages to test for significant differences from a slope of 1 and an intercept of zero. A slope other than 1 would reflect inconsistency in the age estimate compared to known age. An intercept other than zero would indicate a systematic bias between the estimate of age by a reader and known age. We then used the 20 wolves of known age to describe wear on incisors, canines, and carnassials characteristic of each yearly age class from 1-14 years of age. We also developed charts illustrating typical tooth wear in 2-year increments on incisors, canines, and carnassials that can be compared to the teeth of live wolves or museum specimens to estimate their age (Figs. 1, 2).

Fig. 1.
Progressive wear on wolf incisors and canines in 2-year increments from 1 to 12 years of age. Wear on incisors typically progresses beyond the lobes on the first 2 upper and lower incisors at 8 years of age, leaving approximately 5 mm of enamel. At 10 years of age, 2-4 mm of enamel remain on the first and second incisors. Length of canines is reduced 30-50% with 10-16 mm of enamel remaining. Beyond 12 years of age, incisors may be worn to the roots, with a few peg-like stumps projecting above the gum line, or the gums may cover the roots. Length of canines is reduced 50% with 10 mm of enamel remaining.
Fig. 2.
Progressive wear on wolf carnassials (upper premolar 4 and lower molar 1) in 2-year increments from 6 to 12 years of age. Wear is visible on tips of major prominences at 5 years of age and profiles flatten slightly by 6 years. Deep wear on the posterior cusp of the lower carnassial after 10 years of age results from occlusion with the first upper molar, not the upper carnassial.