Determining the relative importance of dilution and detection for zebra foraging in mixed-species herds

http://dx.doi.org/10.1016/j.anbehav.2014.08.012 0003-3472/© 2014 The Association for the Study of A For many animals, a key benefit of group living is lowered predation risk. With increasing group size, individuals commonly reduce vigilance. This group size effect can arise from both dilution of risk and increased collective detection. To determine which was more important, we compared vigilance levels of plains zebra, Equus quagga, in areas inhabited by their main predator, lion, Panthera leo, as a function of herd size and composition (zebra-only versus mixed-species herds). For zebra-only herds, vigilance declined with increasing zebra numbers. In contrast, in mixed-species herds, zebra vigilance levels were significantly lower and did not vary with the total number of ungulates and/or zebra in the herd. This lower vigilance can be explained by detection benefits, since we controlled for the influence of herd size (i.e. dilution) statistically. Furthermore, we found that zebra in smaller herds were half as vigilant when they co-occurred with another preferred prey of lion, blue wildebeest, Connochaetes taurinus, compared with zebra-only herds. However, when zebra foraged with impala, Aepyceros melampus, a nonpreferred prey, there was only a slight reduction in vigilance compared with similar-sized zebra-only herds. Interestingly, the decrease in zebra vigilance when herding with wildebeest was influenced by the presence/absence, but not number, of wildebeest. In contrast, impala only reduced zebra vigilance when they comprised about 75% of the herd. Ultimately, our results indicate that for zebra in mixed-species herds, detection benefits are relatively more important than dilution over a larger range of herd sizes. However, detection only reduces vigilance when zebras herd with species that share a common predator (i.e. diluting partner). Potentially, detection by species that do not share a common predator (i.e. lowdiluting partner) is not as reliable as detection by diluting species. © 2014 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.