New Zealand Journal of Ecology (1998) 22(2): 181- 188

Comparison of two techniques for assessing possum (Trichosurus vulpecula) diet from stomach contents

Research Article
P. J. Sweetapple  
G. Nugent  
  1. Manaaki Whenua - Landcare Research, P.O. Box 69, Lincoln, New Zealand

Two techniques for assessing possum (Trichosurus vulpecula Kerr) diet from stomach contents ("point-sampling" and "layer- separation") are described and compared. Point-sampling involves sieving stomach contents, systematically selecting fragments from the retained material then, identifying and weighing these. Layer-separation involves separation, identification, and weighing of the discrete layers apparent in most possum stomach contents. In 41 of 43 stomachs examined, we were able to separate discrete layers that nearly always comprised a single food item. To compare the two techniques both were applied to these 41 stomachs, with the point-sampling technique applied as two separate treatments using 1.4-mm and 2.0-mm sieves. There were major differences in diet composition estimates between layer-separation and point-sampling but with few differences between the two point-sampling treatments. Relative to layer-separation, point-sampling underestimated the proportions of food groups with small average fragment size and overestimated those with large fragment size. However, both techniques gave similar frequencies of occurrence for 8 of 10 food groups tested, although the apparent importance of foods based on ranking by frequency of occurrence did not accurately match the ranking based on percent composition data. Identification of material was usually easier and more complete with layer-separation than with point-sampling (i.e., there were virtually no unidentifiable stems and fibre after layer- separation). Layer-separation therefore appears likely to provide a simple technique for diet assessment in possums. Although the technique requires formal validation the existence of layers shows that there can have been little mixing (or digestion) of stomach contents, and therefore, that the layer- separation estimates cannot differ greatly from what was eaten. Techniques that involve sieving possum stomach contents appear to have serious limitations, but may be useful as a last resort when layers contain a mixture of foods, or for stomachs in which the layers are not distinguishable.