We compared estimates of occupancy of northern bobwhite (Colinus virginianus) between areas with relatively low and high abundance using single-survey and multiple-survey protocols, with and without accounting for detection probability, and investigated how time during the breeding season affected detection probability in Oklahoma, USA, in 2009-2011. Estimates of occupancy and detection probability increased as the number of survey occasions increased. Detection probability was significantly higher in the area of high abundance (P £ 0.001), and increased as the breeding season advanced from mid-May to late July. Accounting for detection probability increased occupancy estimates by 31% in the low-abundance area but only 1.9% in the high abundance area when using 3 survey occasions per year. Managers using occupancy to detect changes in bobwhite populations should use ³ 4 survey occasions per year to ensure accurate estimates of both occupancy and detection probability. © 2012, The University of Tennessee, National Bobwhite Conservation Initiative.