Letters from W.S. Gosset to R.A. Fisher are used to describe Gosset's first contacts and growing friendship with Fisher, their collaboration over the tabulation of Student'st, and the wider applications of Gosset's work that were made possible by Fisher's extension and generalization of the distribution theory in the early 1920's. These letters reveal Gosset's practical approach to the statistical problems of his brewery work that “naturally required” the solutions for which the name of Student is famous; and they exhibit his appealing character, modest yet self-reliant, in relationship not only to Karl Pearson and R.A. Fisher but to statistical assistants still in training.