This illuminating paper by Stone is the latest in a series of papers drawing attention to possible disadvantages of using nonintegrable prior densities. In these comments, I shall attempt to clarify the role of nonintegrable prior densities in Bayesian inference, thereby providing also a delayed answer to James and Stein [5], Dempster [1], Stone [9], Klotz,et al.[6], and Portnoy [7]. More extensive discussion of the ideas introduced will be given in a subsequent paper.