干旱区生态环境知识资源中心

We analysed the effects of the design of feeding-preference experiments in a study of seed selection by Rufous-collared Sparrows Zonotrichia capensis. We presented eight seed options in multiple- and simple-offer experiments. Statistical results indicated that in multiple-offer tests there were three preferred, and two avoided seed species. The other three species could not be assigned to any of those groups. In contrast, results of simple-offer tests showed that all seed species were preferred, except the two avoided ones. Using a graphical combination of the results of both approaches, we obtained an easily interpretable assessment of the seed preferences of Rufous-collared Sparrows. In the graph space, seed species make up three groups: preferred, avoided, and less-preferred (i.e., heavily consumed in simple-offer trials, but avoided or consumed to a lesser extent in multiple-offer trials). With the combination of both protocols we avoided the biases caused by using one experimental design alone and could detect secondary-preferred food items. This seems crucial when the bird under study is plastic in its feeding preferences.