Carbon isotope variability in the bone collagen of red kangaroos (Macropusrufus) is age dependent: Implications for palaeodietary studies

Isotopic assessment of bone collagen is often used as an environmental trac er in both contemporary and palaeoenvironmental studies. However, variabili ty in the isotopic composition of this tissue remains poorly understood for naturally occurring and wild populations of animals. In this study the sta ble carbon isotope composition of both diet and bone collagen was assessed for a population of red kangaroos (Macropus rufus). Animals sampled ranged in age from approximately 10 months to 15 years. The diet of this populatio n, estimated from faeces collected in the field, varied from predominantly C-4 grasses in late summer ((sic)C-13-16,5%) to mostly C-3 herbage in late winter ((sic)C-13-22,5%), with a long-term average (sic)C-13 of between -19 and -20%. Bone collagen was enriched in(13)C by 3 to 4% in older animals r elative to pouch young. Isotopic analysis of hair, used to assess more rece nt diet in individuals, indicated that diet selection was similar in all an imals that had been weaned. We suggest that the most likely explanation for the age-dependent relationship in the (sic)C-13 of bone collagen occurs be cause milk (the only source of nutrition in suckling kangaroos) is not frac tionated in the same manner as plant-derived carbon during its assimilation into skeletal tissue. If this is the case, then such a relationship should be most predominant in mammals that have low birth weights (relative to th e adult mother) and gain significant weight from milk. Whatever the precise mechanism(s) for the observed fractionation, bone collagen of kangaroos se ems to retain an isotopic memory of the carbon laid down prior to weaning, which takes several years to be diluted and replaced with carbon derived fr om an "adult" herbaceous diet. These results have implications for palaeoec ological research where fossil skeletal tissue is used as dietary or enviro nmental tracers particularly if the relative age of the animals sampled is unknown. Copyright 2001 Academic Press