Testing the Precision and Accuracy of Age Estimation Using Lines in Scutes of Chelydra serpentina and Chrysemys picta

Abstract

Counts of lines on carapacial and plastral scutes are widely used to age turtles because the technique is convenient and non-destructive. However, the procedure has rarely been tested for accuracy or precision against known-aged turtles because verification of age is difficult in such long-lived animals. We tested the accuracy and precision of counts of scute lines using juvenile painted turtles (Chrysemys picta) of known age (1-7 yr), and juvenile and adult snapping turtles (Chelydra serpentina) whose ages were not known. Lines on plastral scutes of C. picta were counted from a marked population in the field, and un- derestimated known age as early as age 4 yr. The accuracy of these counts decreased with known age and by age 6 and 7 yr, only 11% of turtles were aged correctly. In C. serpentina, we used plaster casts and black and white photographs to make permanent records of the fourth vertebral scutes of a set of marked indi- viduals. Counts were made by both naive and experienced observers from casts made between 1-10 yr apart and were compared with one another and with counts made from enlarged black and white photo- graphs of the same individual. We tested whether counts of scute lines were affected by the age of the turtle, the experience of the observer, observer expectancy bias, and the method of recording data. We concluded that all these factors influenced both the accuracy and precision of counts, and that estimates became increasingly unreliable with age. We suggest that counts of scute lines must be tested carefully against known-aged turtles and conducted so as to eliminate expectancy bias. Even then, variability among and within observers and among methodologies may produce age estimates that are at best approximations and at worst misleading, except in the youngest specimens.