Estimations of sizes of fish from subfossil bones with a logarithmic regression model

Abstract

Measurements of bones of ancient fish from archaeological excavations allow estimates of the sizes of the original fish, which may add important aspects to the archaeological story. Currently, these estimates are calculated using regression models derived from data collected from samples of recent fish of different sizes. The linear and power regression models which are in common use have some drawbacks. In this preliminary report, we propose four criteria for selection of the most appropriate available regression analyses, and recommend the use of a logarithmic regression model.