• feeding behaviour;
  • herbivory;
  • kangaroo;
  • plant-animal interaction;
  • seedling selection


Investigations into the mechanisms underpinning plant selection by herbivores have often yielded conflicting results. Such inconsistency might stem from whether selection experiments are conducted with captive or wild populations, and upon the different measures of plant selection used to determine herbivore preference. Here we compared the feeding preferences of captive and wild kangaroos (Macropus fuliginosus) using a standard set of plant species (14 Hakea spp., Proteaceae) and several measures of herbivore selection to examine how environment influenced relative consumption. Three indices of herbivore consumption were measured: number of plants (NP), total shoot volume per plant (TV) and percentage available shoot volume (PV) consumed. NP and TV were closely correlated in the wild and captive populations and consistently correlated with six morphological and chemical plant attributes examined, the most notable being a strong negative correlation with shoot phenolic content. This uniformity suggests that plant selection by captive kangaroos is broadly consistent with that observed in field trials, and consequently that for macropods, at least, captive trials offer a valid way to determine the relative acceptability of different plant species. However, the fact that our third measure of herbivore selection PV was weakly correlated in captive and wild populations and showed no relationship with shoot phenolic content highlights the importance of which measure of plant selection is applied. We suggest that, while NP and TV are potentially confounded by plant size and availability, they offer the clearest insight into plant selection from the point of view of the herbivore, while PV is more suitable for plant-centred studies.