An investigation of bone histology as a potential age indicator in Roe Deer
The histological structure of 72 modem roe deer mandibles was examined to see if those features that showed age-related development patterns in humans (Kerley, 1965), and laboratory rats (Singh & Gunberg, 1971), could be quantified and used to predict age in a species which poses particular ageing problems to zooarchaeologists. The roe deer sample consisted of 30 bucks and 42 does ranging in age from kids through to 14 years. Bone pieces, 5mm thick, were removed from the base of each right mandibular ramus and prepared for analysis in a scanning electron microscope operated in back-scattered electron detector mode. Four fields of view per section were photographed at a magnification of x44, and the numbers of secondaiy osteons and non-Haversian canals, together with the area of periosteal circumferential lamellae in each was recorded. Histological feature counts were regressed on age to produce equations which could be reversed and used to predict new ages for each specimen. The average accuracy and bias of the new ages were calculated. Statistical analysis revealed that circumferential lamellae area did not show any correlation with age in roe deer, but doe total and average non-Haversian canal counts showed highly significant correlations with age (r = -0.75, P = 0.00) and doe average non-Haversian canal counts provided the most accurate age estimates (average inaccuracy = 2.50 years, average bias = -0.04 years). However, tooth wear scores were found to have a stronger correlation with age, for bucks and does together, (r = 0.88, P = 0.00) and a greater level of accuracy of age prediction (average inaccuracy = 1.31 years, average bias = 0.16 years). Still, in the absence of the dentition, histology-based ageing remains the only technique that can be applied to roe deer over their entire age range.