ABSTRACT We start out from Harriet Zuckerman’s study of the US scientific ultra-elite of Nobel laureates, in which Robert Merton's idea of ‘Matthew effects’ as a key mechanism in the creation of social inequalities was first introduced. We then consider two issues arising from critical commentary on this study by Elisabeth Crawford, a historian of science. First, how far can a scientific ultra-elite be shown to exist as a collectivity that is socially distinctive? Second, how far is Zuckerman’s account of the formation of the US ultra-elite trough ‘bilateral associative selection’ between scientific masters and their would-be apprentices historically specific to the US? In the UK case, we compare the social origins and educational careers of members of two possible scientific ultra-elites, defined by differing degrees of stringency, with those of other elite scientists. We find that as one moves from the elite to the less stringently defined ultra-elite, there is little evidence of increasing social stratification but that such evidence does emerge in moving to the more stringently defined ultra-elite. We also show through two contrasting Cambridge case studies, that the underlying social processes that Zuckerman identifies in ultra-elite formation in the US are also present in these UK contexts.