Two discrete choice models relate the probability of choosing two water-saving irrigation technologies - sprinkler and tailwater recovery pits- to the underlying physical and economic attributes of the farming using a national cross section of farm level data. The results show that small farm size, high water or labor costs, and soils with low water-holding capacity increase the likelihood of adopting sprinkler irrigation. For gravity irrigators, large farms, high water costs, and solid with high water-holding capacity increase the probability of recirculating field runoff. In both models soil characteristics and, to a lesser extent, climate dominate the selection probabilities.