It is, nevertheless, a source of serious danger both in marine and land warfare. Defective ventilation in the boiler rooms of ships and fires below decks, both in and out of action, are especially dangerous because of the carbon monoxide which is produced. In one of the naval engagements between the Germans and the English, defective high explosive shell, after penetrating into enclosed portions of the ship, evolved large quantities of carbon monoxide and thus killed some hundreds of men. On shore, machine gun fire in enclosed spaces, such as pill boxes, and in tanks, liberates relatively large quantities of carbon monoxide. Similarly, in mining and sapping work, the carbon monoxide liberated by the detonation of high explosives constitutes one of the most serious of the difficulties connected with this work and necessitated elaborate equipment and extensive military training in mine rescue work. The removal of carbon monoxide from the air is difficult because of its physical and chemical properties. Its low boiling point and critical temperature makes adequate adsorption at ordinary temperatures by the use of an active absorbent out of the question. Its known insolubility in all solvents similarly precludes its removal by physical absorption. After extensive investigation two absorbents have been found.[24] The first of these consists in a mixture of iodine pentoxide and fuming sulfuric acid, with pumice stone as a carrier. Using a layer 10 cm. deep and passing a 1 per cent carbon monoxide air mixture at the rate of 500 cc. per minute per sq. cm. cross section, a 100%-90% removal of the gas could be secured for two hours at room temperature and almost as long at 0° C. The reaction is not instantaneous, and a brief induction period always occurs. This may be reduced to a minimum by the addition of a little iodine to the original mixture. [Footnote 24: Complete details of this work may be found in _J. Ind. Eng. Chem._, =12=, 213 (1920).] The sulfur trioxide given off is very irritating to the lungs, but by the use of a layer of active charcoal beyond the carbon monoxide absorbent, this disadvantage was almost completely eliminated. However, sulfur dioxide is slowly formed as a result of this adsorption and after prolonged standing or long-continued use of the canister at a high rate of gas flow gives serious trouble. Considerable heat is given off in the reaction and a cooling attachment was required. The most satisfactory device was a metal box filled with fused sodium thiosulfate pentahydrate, which absorbed a very considerable amount of the heat. Still a further disadvantage was the fact that the adsorbents became spent by use, even in the absence of carbon monoxide, since it absorbed enough moisture from the air of average humidity in several hours, to destroy its activity. The difficulties mentioned were so troublesome that this absorbent was finally supplanted by the more satisfactory oxide absorbent described below.