A broken cross-recessed die was examined. Examination of the unetched, polished section for impurities revealed several coarse streaks of slag. The purity did not therefore correspond to the requirements set for a high speed tool steel of the given theoretical quality DMo 5. After etching with 5% nital the polished surface exhibited a pronounced, easily-visible, fibrous structure. Microscopic examination revealed that this etch pattern was produced by marked segregation bands. The very unfavorable structure for a high speed steel tool of these dimensions and subject to such stresses together with the low purity favored the fracture of the tool.