Comparing longitudinal and cross-sectional sampling methods on growth variables and age estimation: lessons from postnatal growth of the Geoffroy's bat, Myotis emarginatus.

Studies on the postnatal growth of free-ranging bats can be based on multiple recaptures of marked individuals of known age (longitudinal sampling) or derived from measurements of captured pups with unknown birth dates (cross-sectional sampling). We therefore used individuals of Geoffroy's bat Myotis emarginatus in a case study to test the hypothesis that these two methods yield similar data on growth rates and growth derivatives. We followed postnatal growth in body mass, forearm length, and total epiphyseal gap in 24 tagged neonates sampled 10 times (via 143 cases of longitudinal sampling) and compared finding with those deriving from cross-sectional sampling carried out on the same occasions among 138 non-tagged neonates of the same colony. A generalized estimating equation relating to the first 3 weeks of postnatal growth indicated that initial values (y-intercepts) for forearm length and body mass did not differ significantly between groups sampled in the two ways (P > 0.05), while rates of growth characterized by these parameters were found to differ significantly (P < 0.05). In relation to the length of total epiphyseal gap obtained for days 12–40, tests for equal intercepts and parallelism confirmed significant differences between the two sampling methods at P < 0.05. Specifically, cross-sectional sampling of the three growth parameters resulted in significant overestimation of ages among the studied bats. However, given the greater logistical challenges with longitudinal sampling, our demonstration of the relative value of the two sampling methods, and of their dependence for accuracy on the amount of variation found in growth variables, allows us to persist in advocating cross-sectional sampling in postnatal studies on bats. [ABSTRACT FROM AUTHOR]