Factors affecting the precision of age determination of sperm whales (Physeter macrocephalus)

Abstract

Teeth from 92 sperm whales were prepared by etching for age determination. The total number of growth layer groups (GLGs) in the dentine of each tooth was determined from three to five reading sessions by a single reader. Four other readers, as part of a cross-reading experiment, read a subset of these teeth (n = 5). This study investigated: (1) intra- and (2) inter-reader precision in GLG counts; (3) possible variation in growth structure deposition between different teeth within the same individual; (4) the use of photographs to identify and count GLGs and the effect of this technique on the precision of counts; and (5) mineralisation anomalies in tooth sections and the possible effects these may have on GLG count precision. Intra- and inter-reader precision was determined using coefficients of variation (CV) and indices of precision (D). Total numbers of GLGs estimated from individual teeth ranged from 0.75-64 (¯x = 32.8, n = 92). Intra-reader mean CV was 10.6 and mean D was 4.8. Inter-reader mean CV ranged from 4.8-12.3 and mean D ranged from 2.8-7.1. Differences in final counts between readers appeared to be the result of differing interpretation of GLGs and this was the largest factor affecting the precision of GLG counts. While GLG counts between teeth in the same individual varied, it is possible that this variation was due to within reader variation rather than variation in the development of growth structures, but establishment of this cause is confounded by differential tooth wear. Use of photographs increased the definition of growth structures, decreasing the variation between GLG counts within reading sessions. The incidence of mineralisation anomalies and the closure of the pulp cavity increased with increasing GLG counts in individuals, but were not consistent between teeth from the same individual. These factors, while potentially affecting the accuracy of GLG counts in relation to age estimates, had little effect on the precision of GLG counts. The lack of an ability to validate age estimates in this species and the large inter-reader variation seen in this study suggests that age estimates based on GLG counts in this species are subjective and can only be regarded as relative. High-quality photographs of tooth sections should be used to verify GLG counts with other readers, resulting in ‘consensus counts’ generated by a number of readers, ensuring interpretation of the same structures and confidence in comparing GLG counts produced in different studies.