In 1935, a group of mathematical statisticians in the United States announced the founding of a new professional organization, the Institute of Mathematical Statistics. In the coming years, this institution became a significant source of support for the growing community of mathematical statisticians. Thus it would seem that the formation of the Institute marked the beginning of the process of the creation of that community. This paper, however, argues that those who initially participated in the Institute constituted a small, loosely (...) organized, but wellconnected group within the broader scientific world even before they joined together formally to promote their discipline. We will examine some aspects of their education, employment, and interests in order to determine how American mathematical statisticians set themselves apart from other scientists and helped to define their discipline before the community acquired its official status. (shrink)