There is much interest in explaining the persistent ethnic gaps in education among Israeli Jews; specifically, the much lower attainments of those from Asian and African countries compared to the rest - Mizrahim vs. Ashkenazim, respectively. Some explanations (especially early ones) have stressed premigration immigrant characteristics, particularly the relatively lower level of educational attainment among Mizrahim. More recent interpretations have tended to focus on discrimination of various sorts that took place after the immigrants arrived in Israel. Crucial evidence for the discriminatory effect was introduced by Yaakov Nahon (1987), who demonstrated a shift toward a Mizrahi-Ashkenazi dichotomy in educational attainment between birth cohorts of adult immigrants and birth cohorts of adults born in Israel. From this evidence, a wide range of scholars concluded that the premigration educational characteristics of immigrants could not explain Israeli educational patterns, and that, consequently, the explanation based on discrimination was thereby greatly strengthened. In this paper, we use the 1961 Israel census public-use dataset to refine Nahon's analysis. Instead of using age cohorts as proxies for fathers and children, we focus on actual fathers and their children. Our results vary substantially from Nahon's. In fact, we find that the educational attainment of immigrant fathers clusters quite closely around the Ashkenazi-Mizrahi dichotomy, and conclude that it is no longer reasonable to rule out the premigration hypothesis. This outcome leaves researchers with a more challenging explanatory task than before, because they are now faced with the notoriously difficult situation of having to determine the relative influence of premigration characteristics, on the one hand, and of discriminatory processes, on the other.