In this paper, we examine changes in the behavior of ex-dividend stock prices when the exchanges changed from pricing stocks in discrete intervals to decimal pricing. Based on prior models of ex-dividend behavior and price discreteness of Dubofsky and of Bali and Hite, we anticipate that the move to trading in decimals would decrease the variance of returns on all exchanges and increase the level of ex-dividend-day returns on the NYSE while reducing them on the Amex and Nasdaq. Our sample of ex-dividend-day returns covers periods slightly longer than one year before and after decimalization. For the overall sample and for each of the individual exchanges (Amex, Nasdaq and NYSE), the variances of ex-dividend returns experience a significant decrease after decimalization while the mean returns increase by a positive and significant amount. To account for the increase in ex-day returns on the Amex and Nasdaq, we develop an alternative model to explain the effect of discreteness on ex-day returns. Tests of the three models (Dubofsky's, Bali and Hite's, and ours) indicate that prior to decimalization, as expected, Dubofsky's model is better for explaining NYSE ex-day returns and ours fits the Nasdaq better. Bali and Hite's model, however, is unable to explain any of the pre-decimalization ex-day returns, including those of the Nasdaq where the Bali-Hite model might provide a reasonable description of ex-day market behavior. After decimalization, ex-dividend-day returns do not appear to follow either the scenario described by Dubofsky or by us. The most likely cause of this is that traders in the market are placing ex-dividend-day orders with limits somewhere between prices indicated by Dubofsky and by us. We also provide evidence that ex-dividend returns attributable to factors other than discreteness and the dividend yield actually declined following decimalization. Since the most obvious factor is transactions costs, we interpret this to be evidence of a reduction in ex-day returns caused by a reduction in transactions costs. We also find that the dividend yield is a significant influence on ex-dividend-day returns.