This study relies on the Ricardian method to estimate the damages of climate change to US agriculture. The study uses repeated cross sectional analyses of US Census data collected at the county level from 1978-2002. Regressions of farmland value on climate and other control variables reveal that climate consistently affects farmland values across the US. The 1978 and 1982 data imply that warming is beneficial to farming whereas the data from more recent years implies warming is harmful. Linear, enhanced linear and log linear regressions all confirm these results. With the linear and enhanced linear models, projections of future impacts from climate models lead to similar findings. With the Hadley and uniform climate scenarios, coefficients from the 1978 and 1982 data generally imply benefits whereas the coefficients from the later data imply damages. The loglinear model leads to different results. The loglinear model implies that these climate scenarios will generally be beneficial, although the benefits do vary across years. All the regression models suggest the mild PCM scenarios will generally be beneficial. The climate results are mixed. They generally imply that mild warming is beneficial but more extensive warming is harmful to the US. However, the results are not stable across time, suggesting the repeated cross sections are still not capturing all the important factors that change annually in the farmland market.