WITH reference to the particularly interesting article by H. C. H. C. in NATURE of November 17, p. 728, might I mention the following amongst many other examples which have come under my notice illustrating the tendency of cementite to form cell walls or a network under conditions where the occurrence of pearlite is more commonly anticipated? In dead mild steels the occurrence of cementite in either network or comparatively massive formation has been recognised by a number of investigators. Fig. 1 illustrates an exceptional case in which isolated cell walls were found only near the edge of a dead mild-steel plate, in a region otherwise microscopically carbonless. The only apparent explanation of this occurrence was that the plate must have become carburised locally during the processes of manufacture.