The use and misuse of public information by foraging red crossbills

Group foragers may assess patch quality more efficiently bypayingattention to the sampling behavior of group members foragingin the same patch(i.e., using "public information"). To determinewhether redcrossbills (Loxia curvirostra) use public informationto aid theirpatch departure decisions, we conducted experimentsthat compared the samplingbehavior of crossbills foraging ona two-patch system (one patch was alwaysempty, one patch containingseeds) when alone, in pairs, and in flocks ofthree. When foragingalone, crossbills departed from empty patches in a waythatwas qualitatively consistent with energy maximization. We foundevidencefor the use of public information when crossbills werepaired with two flockmates, but not when paired with one flockmate. When foraging with two flockmates, crossbills sampledapproximately half the number of cones on the emptypatch beforedeparting as compared to when solitary. Furthermore, as expectedifpublic information is used, the variance in both the numberof cones andtime spent on the empty patch decreased when crossbillsforaged with two flockmates as compared to when alone. Althoughhigh frequencies of scroungingreduce the availability of publicinformation, scrounging is usually uncommonin crossbills, apparentlybecause they exploit divisible patches.Consequently, publicinformation is likely to be important to crossbills inthe wild.We also show that feeding performance is greatly diminishedwhen thefeeding performances of flock mates differ. This providesa mechanism thatwill favor assortative grouping by phenotypewhen phenotypes affect feedingperformance, which may in turnpromote speciation in some groups ofanimals.