||Despite being among the most intensively studied and managed species of wildlife in the world, northern bobwhite (Colinus virginianus) populations have declined over most of their range. Although macrohabitat features associated with high bobwhite densities have been described, few data exist linking habitat attributes with individual survival or productivity. We compared survival, clutch success, and home range size of radiotagged bobwhite during summer between 2 areas in Kansas with different land uses (cropland dominated [CSA] vs. rangeland dominated [RSA]), and compared macrohabitat surrounding surviving versus depredated adults and clutches within areas at 5 spatial scales. Female survival (CSA = 0.26 [SE = 0.01], n =46; RSA = 0.36 [SE = 0.02], n = 32), home range size (CSA = 75 +- 15 ha; mean +-SE; RSA = 54 +- 16 ha), and clutch success (CSA = 0.58; RSA = 0.67) were similar (P > 0.05) between areas, whereas male survival was higher (P = 0.021) on the RSA (0.51 [SE = 0.01], n = 70) than on the CSA (0.26 [SE = 0.01], n = 61). Within the RSA, surviving adults had more cropland and greater mean distances to grassland in areas surrounding them at small to intermediate scales than those that were depredated (P<=0.1). Habitat surrounding surviving versus depredated adults was similar (P > 0.1) on the CSA. Successful clutches on the CSA were surrounded by less rangeland and more hayland at small to intermediate scales than those that were depredated (P<=0.1). Habitat surrounding successful versus depredated clutches was similar (P > 0.1) on the RSA. Our results suggest the amount of breeding (i.e., grassland) habitat as well as its composition and distribution influenced some aspects of bobwhite survival and productivity, but habitat selection could not be used to reliably identify habitat attributes associated with higher probabilities of adult survival or clutch success. Reversing bobwhite population declines at large scales depends on identifying macrohabitat attributes that specifically influence individual fitness rather than population distributions.