Weather and land use are important factors influencing the population dynamics of northern bobwhites (Colinus virginianus) in Texas and elsewhere. Using an artificial neural network, we studied the effects of these factors on an index of bobwhite abundance (hereafter, index) in 6 ecoregions in Texas. We used roadside-count data collected by the Texas Parks and Wildlife Department (TPWD) during 1978-1997. Weather variables were June, July, and August mean maximum temperatures, and winter (Dec-Feb), spring (Mar-May), summer (Jun-Aug), and fall (Sep-Nov) rainfall. We also included the proportion of county area in cultivation, the number of livestock per hectare of noncultivated land, and the previous year's bobwhite count in the analyses. The data were partitioned into training and validation data sets prior to analyses. The neural model explained 65% of the variation in the training data (n = 72) and 61% of the variation in the validation data (n = 17). The most important variables contributing to network predictions were July temperature, fall rainfall, cattle density, and the previous year's bob-white count. State-level simulation results indicated that the bobwhite index decreased with increasing June temperature and livestock density. The bobwhite index increased with July and August temperature, fall rainfall, and the previous year's bobwhite count. Bobwhite abundance increased with the proportion of county area in cultivation up to approximately 20% cultivation and then declined. Winter, spring, and summer rainfall had little effect on the bobwhite index. Although many relationships appeared approximately linear or were decelerating, proportion of county area in cultivation and livestock density on noncultivated land showed strongly curvilinear responses. Therefore, cultivation up to approximately 20% of county area was beneficial, but the benefits disappeared as cultivation increased beyond this level. Further, at low livestock densities, between 0.15 and 0.40 head/ha, small increases in head/ha resulted in a decrease in the bobwhite index of 156.4%/head/ha. The results also indicated that a potential bias might exist in the survey protocol resulting in artificially inflated counts under some weather conditions.