Assessment of faecal output in geese.
The faecal output of geese (Anseriformes) is often measured in the course of bioenergetic or flock management studies. The first objective of this investigation was to examine the influence of method on the field assessment of defaecation rate in greater snow geese (Chen caerulescens atlantica). The 'hourly block method' was the only one which produced unbiased estimates of this parameter: 'direct interval estimate' and 'short-cut' methods (in particular Owen's half-time interval) were found to produce biased and/or high variance estimates and were deemed unreliable in most circumstances. Computer modelling confirmed these conclusions. The second objective of the investigation was to identify the sources of bias and variance affecting dropping-count estimates obtained in sample plots. The observation method chosen (observer standing or crouched, observer single or paired) had a great effect on the size of the bias encountered. Variance due to observer, plot and habitat was also assessed. Suggestions to stabilize and/or remove causes of bias and to improve the precision of dropping-count estimates are given.