An experimental approach is used to compare bidding behavior and auction performance in uniform-price and discriminatory auctions when there is incomplete information concerning the common value of the auctioned good. In a symmetric information environment, the different auction formats provide the same average revenue. However, when information is asymmetric the discriminatory auction results in higher average revenue than the uniform-price auction. The volatility of revenue is higher in the uniform-price auctions in all treatments. The results, therefore, provide support for the use of the discriminatory format. Subject characteristics and measures of experience in recent auctions are found to be useful in explaining bidding behavior.