A potential multivariate index of condition for small mammals

Abstract

Body condition in small mammals is often investigated by examining the residuals of a regression of mass on a linear measure of body size. Such indices include a number of assumptions, which, in most cases, are not tested. This study presents a residual condition index, that uses the regression of body mass on the Axis 1 factor scores from a Principal Components Analysis, to estimate body condition in free‐living ship rats, Rattus rattus Linneaus 1758, caught in mixed forest in New Zealand. The results from this index and a simpler ratio index were compared to determine how well they met the assumptions of the technique, and how well they identified biological differences in the population. Of the indices tested, the residual index based on an Ordinary Least Squared regression proved to be the most statistically robust method. The condition indices calculated were probably correlated with significant differences between individuals, but were not compared with true condition as estimated by fat reserves. Further applications and the need for the validation of the residual index are discussed.

Keywords:

• body condition
• residual index
• linear regression
• Principal Components Analysis
• ship rat
• Rattus rattus