Arial (1990) showed that the average returns of U.S. market indices on trading days prior to holidays are 9 to 14 times higher, a phenomenon that was independent of other calendar effects and the small-firm effect. We first confirm his results. Extending the sample to 1983 to 2019, we find that the pre-holiday effect now exists only among small firms. For large firms, the differences in returns between pre-holidays and non-pre-holidays have become insignificant, and especially after 1990.