In this paper, I use a unique proprietary dataset from the foreign exchange market to examine the existing hypotheses on price clustering. I find that market uncertainty plays an important role in price clustering. Moreover, since trading behavior changes under different market conditions, market timing also affects the likelihood of price clustering. The results support both the price resolution hypothesis (Ball et al. J Futures Mark 5:29-43, 1985) and the negotiation hypothesis (Harris Rev Financ Stud 4:389-415, 1991). Since the data covers the interbank foreign exchange market, which is the market for the professional bank dealers, the attraction hypothesis is less likely to be a plausible explanation for price clustering in the foreign exchange market. © 2009 Springer Science+Business Media, LLC.