Not all fields, nor even portions of fields, have the same economically optimal corn plant density. However, until the recent introduction of precision farming, producers could not benefit from these accepted intrafield differences. This field study was conducted on 170 cooperating farmer fields throughout the Midwestern U.S. Corn Belt between 1987 and 1996 and consisted of over 42 000 individual experimental units. At each location, corn (Zea mays L.) was overplanted and thinned to 44 000 to 104 000 plants ha -1 . The objective of our field research was to estimate the economic value, to the farmer, of variable rate seeding (VRS) as compared with uniform rate seeding (URS). We first estimated the correlation between field quality and economically optimal plant density. The economically optimal uniform plant density for the Midwest Corn Belt was 67 900 plants ha -1 . For every tonne per hectare increase in site quality, as measured by yield potential, the predicted value of the site-specific economically optimal plant density increased by approximately 1200 plants ha -1 . We compared differences in revenues minus seed costs on four simulated fields. The value of VRS, ignoring the costs of VRS equipment and services, ranged from $12.83 ha -1 for farmers with VRS technology and full information to $0.15 ha -1 for farmers with VRS technology but only partial information. Profitable implementation of VRS will require detailed and expensive information regarding site characteristics, production inputs, and stochastic factors. Therefore, VRS will remain economically infeasible for most commercial corn growers until the cost of obtaining such information decreases considerably.