A critical component of efficient foraging behavior is leaving a feedi
ng site when prey density in that site declines. If ''real time'' hypo
theses are to explain patch-leaving mechanisms, we will need to examin
e the influence that moment-to-moment experience within a current prey
patch has on persistence in that patch. We used linear regression to
investigate how current experience of patch quality influenced homing
pigeons' patch persistence. We did this by examining the amount of var
iance in persistence that was accounted for by different measures of w
ithin-session patch quality. The importance of several measures of the
precise sequence of events in individual sessions were assessed with
selected averaging algorithms. Mean inter-capture interval and mean nu
mber of inter-capture pecks accounted for a significant amount of the
variance in giving-up time in three of four, and four of four birds, r
espectively. Conversely, only one rate measure in one bird showed an i
nfluence on patch persistence. In three birds, recent information had
a strong influence on giving-up time. Current-visit experience did inf
luence the patch persistence of the pigeons, but the birds' behavior i
ndicated that different individuals used different measures of patch q
uality, different mechanisms of assessing those measures, and differen
t patch-exit mechanisms. (C) 1998 Elsevier Science B.V. All rights res
erved.