The determinants and labor market consequences for immigrants of proficiency in speaking the dominant language (English in the United States and English of French in Canada) are explored, with focus on adult men using self-reported data from the 1980 and 1981 censuses of the United States and Canada, respectively. It is shown that the determinants of earnings among immigrants are remarkably similar in the two countries. Fluency in the dominant language has a large positive effect on earnings, independent of other personal characteristics and country of origin. The study shows the importance of explicitly incorporating dominant language fluency, and the determinants of dominant language fluency, in the criteria for allocating immigrant visas, if immigrant economic success is a policy objective. It also indicates that, because of the questions asked and the coding procedures, the data related to language in the U.S. census are superior to the data from the Canadian census for both statistical and public policy analysis. (Author/LB)