We tested the Glock-Stark model of the influence of particular Christian beliefs in producing anti-Semitism, using a survey of Protestants in the Atlanta SMSA and Philadelphia SMSA. Our items and indices were almost identical to those of Glock and Stark in their 1963 California study. In table analysis we tested for spuriousness of the relationship between Christian beliefs and anti-Semitism, and we found evidence of it. A path model built from the Glock-Stark variables turned out similar to that constructed by Middleton from nationwide data. A more elaborate model including other religious, social-psychological, and background variables accounted for 28 percent of the variance in anti-Semitism; the most important factors were social-psychological factors—anomie, dogmatism, and status concern. The religious factors alone accounted for about 5 percent of the variance. We agree with Middleton that Glock and Stark overestimated the role of Christian beliefs in producing anti-Semitism and that they overlooked important social-psychological factors.