Abstract
Age estimates for turtles are usually made without measuring the precision of the estimates and without calibration from known-age individuals. To estimate ages of adult female snapping turtles (Chelydra serpentina) from Algonquin Park, Ontario, Canada, we counted growth lines on the fourth vertebral scute of the carapace. We quantified error arising from variability within and among observers by recording several counts by each of a series of observers. Measurements of the distances between the focus and each growth line on turtles possessing all lines were used to estimate numbers of lines missing through wear on other turtles. Ninety-five percent confidence limits on the estimated ages averaged ±17% of the age. A long-term mark-recapture study indicated that counting growth lines underestimated age, but insufficient numbers of known-age adults prevented calibration for this bias. Comparisons of aging techniques on the same individuals indicated a significant underestimate of age if the number of growth lines lost due to wear was not estimated.

This publication has 3 references indexed in Scilit: