Abstract

Turtles are long lived and demographic models requiring estimates of age, growth, fecundity and survival are central for management. Most studies that estimate age and growth of freshwater turtles use annuli as an index of age without estimating its error and very few studies that use growth models include many juveniles, where growth is often large and variable. In this paper, I compare the reliability of growth annuli and common models in determining age and growth of two widely distributed turtles in Australia. Most turtles are carnivorous during the juvenile stage but many species shift to a lower-quality omnivorous diet prior to maturing. Patterns of growth are often characterised by this dietary shift and I compared the growth of a common omnivorous turtle (Emydura macquarii) and a vulnerable sympatric species that is an obligate carnivore (Chelodina expansa). Mark–recapture programs were established in three lagoons on the Murray River. In total, 1218 hatchling E. macquarii were released into two of the lagoons and growth annuli were found to be unreliable in estimating their age by Year 2. The von Bertalanffy and logistic growth models can reliably estimate age of both male and female E. macquarii and C. expansa respectively. Growth is extremely rapid during the juvenile stage of E. macquarii, but is highly variable in C. expansa, with rapid growth occurring only beyond three years of age. Hence growth models fail to predict age when juveniles are excluded from the analyses. Female E. macquarii delay maturity until 9–12 years of age because clutch size is positively related to body size and they can produce only one large clutch per year. Female C. expansa mature later (at ~14 years) than female E. macquarii and both species are sexually dimorphic, as males mature earlier at smaller sizes than females. Common growth models describe the growth of two widely distributed freshwater turtles, but different patterns of growth and age at maturity relate to quality of diet and reproduction.