The increasing scale and intensity of modern blast-furnace operation, combined with a simultaneous reduction in coke rate, have led to more stringent coke quality requirements with a greater emphasis on the coke's (1 2) role as a burden spacer ' Accordingly, coke quality data limited to chemical composition, initial size grading and ambient size stability are insufficient to predict the coke's behaviour in use. A realistic assess ment of the likely performance of coke should also take into consideration its resistance to degradation in the chemical and thermal environment of the blast furnace. A test is, therefore, needed which simulates as far as possible the mechanical, chemical and thermal conditions of the smelting process and measures the resultant changes in size distribution of the coke, and hence in its bulk permeability to gas flow, using representative samples of the feed coke as charged to the furnace. Moreover, new developments in coke production make it necessary to reassess coke testing so as to guide the coke producer in making the best use of available coals. The mode of mechanical coke breakage at ambient temperature has been extensively characterised and related to changes in the coke's bulk permeability(3). In contrast, the causes of deterioration in the hot strength and size of bulk coke arising from its partial gasification with carbon dioxide and reheating to high temperature are relatively less explored. The present paper describes the v.