The reliability of flaked tool function studies in New Zealand archaeology

We report the results of a blind test of flaked tool function. Thirty freshly knapped obsidian flakes were used in a range of tasks. Tool function was identified employing a low power (< 40x) microscope and commonly used edge variables. We show that flakes used on pliable materials such as flax could not be identified as tools, whereas ca. 50% of those used in woodworking and all those used in boneworking were so recognised. However, most boneworking tools were wrongly assigned to woodworking and half of the woodworking implements were assigned a wrong function. Our results suggest an unacceptably high level of error in analyses of New Zealand flaked tool function. Studies to determine the functions of flaked tools used in pre-European Maori society have nearly all involved a 'lowpower' approach employing microscopic techniques at less than lOOx magnification (e.g. Bellwood 1969, Fredericksen