1501 1737 Therefore B is the finer. Purdy in 1905 (Trans. Amer. Ceram. Soc., 7, 441) pro- posed a modification of Jackson's factor, and the Jackson- Purdy surface factor is the one used in the States. But it makes precisely the same assumptions as its predecessor, and differs chiefly in not inverting the diameters. If Jackson's factor were to trace a rectangular hyperbola, so also would Purdy's. Cushman (Journ. Amer. Chem. Soc., 29, 589) shows the inaccuracy of assuming that the mean diameter of a group of particles can be obtained by averaging the extreme diameters of that group. Mellor (Trans. Eng. Ceram. Soc., 1909, 9, 94) discusses the Jackson -and Purdy factors, and suggests a new one : CERAMIC CHEMISTRY. 29 Where s is the mean specific gravity of the powder, W x . . * the respective weights of the fractions per unit weight of powder, and dj . . . the corresponding average diameters, He calculates the average diameter from the mean dia- meters dj and d 2 as l x + d,) (d^ + d,*) It will be seen, then, that no one factor has been generally accepted. 30 CERAMIC CHEMISTRY. CHAPTER IV.