Fifty bridge players and 50 nonplayers, between the ages of 55 and 91, were given tests of working memory, reasoning, reaction time, and vocabulary. Data were analyzed using multivariate and univariate analyses of variance with age as a covariate. Results indicated that the players outperformed nonplayers in measures of working memory and reasoning, but not vocabulary and reaction time. Results were consistent with the hypothesis that bridge, which provides specific experience in working memory and reasoning, should enhance performance in tasks tapping these abilities and not enhance performance in unrelated abilities. Because the data were correlational, the rival hypothesis that bridge playing selects for individuals who perform better at working memory and reasoning tasks could not be rejected.