This paper experimentally investigates the bidding behavior in two-stage sequential ascending clock auction (i.e., English auction). The equilibrium given by Milgrom and Weber (2000) predicts that the prices of the two goods sold are exactly the same regardless of the value realization. We, in addition, show that this equilibrium can be obtained in our setting by two rounds of iterated deletion of weakly dominated strategies. Despite such a strong prediction, we find little support for equilibrium-like behavior. Instead, we find on average the same prices in three-bidder auctions and slight on average price declines in five-bidder auctions. While these results do not sound like serious violations of what the theory predicts, we find that the main equalizer of the prices is not the information revealed by the dropouts as the theory predicts, instead it is mainly driven by the interaction of different types of off-equilibrium behaviors. We discuss the implication of the results to the declining price anomaly discovered by Ashenfelter (1989).