Milk supply elasticities were estimated for 70 farms who participated in the New York Dairy Farm Business Summary Program from 1985 through 1993. Technology was modeled as a single output, single composite input Cobb-Douglas function. The resultant supply function is the natural log of milk quantity as a function of the log of milk price to prices paid for all inputs, with a time trend added. Since random output shocks to each farm may have occurred in any of the nine data years, a dummy year variable was modeled sequentially for each year for each farm. Ignoring 12 negative estimates, elasticities averaged .65. Pooling the data with a fixed effects model produced an elasticity estimate of .47. A geometric lag pooled model produced a short-run elasticity of .2 and a long-run supply elasticity of 1.00 using instrumental variables, but an OLS short-run elasticity of .25 and long-run elasticity of .64.