Evaluation of techniques for indirect measurement of body composition in a free-ranging large herbivore, the southern hairy-nosed wombat

Several indirect methods for measuring body composition in a large herbivore, the southern hairy-nosed wombat (Lasiothinus latifrons), were evaluated. Body composition was determined by whole-body chemical analysis of 15 wild-caught wombats, and compared with several indices of body fat: total body water measured by isotope dilution, bioelectrical impedance analysis (BIA), body-mass index, and a body condition score. Total body water and total body fat (by soxhlet analysis) were highly correlated (4² = 0.97, intercept s.e. = 1.00). Total body water measured by desiccation was highly correlated with isotope dilution space (r² = 0.97, intercept s.e. = 0.43 for deuterium; r² = 0.95, intercept s.e. = 0.44 for H₂¹⁸O). Percentage body fat by soxhlet analysis was highly correlated with total body water measured as deuterium dilution space (r² = 0.83, intercept s.e. = 2.46). Multiple linear regression models using BIA plethysmograph measurements (resistance and impedance) and total body mass, were successful in predicting body fat (r² = 0.90, s.e. = 1.99) and total body water (r² = 0.90, s.e. = 1.64). Isotope-dilution techniques are the most accurate means of indirectly measuring total body water and total body fat, but at considerable expense of time and money. BIA offers reduced accuracy but at less cost and may be useful for measuring changes in body composition in populations of herbivores. Body-condition indices and scores correlate poorly with body fat, suggesting that their application as a means to predict body fat is limited.