Absorbents must be _mechanically strong_ in order to retain their structure and porosity under conditions of transport and field use. Further, they must not be subject to abrasion for the production of a relatively small amount of fines would tend to plug the canister or to cause channels through which the gas would pass without being absorbed. Since the canister is filled several months before it is first used in the trenches, and since the canister may be used over a period of months before it is discarded, it is obviously the ultimate activity and capacity (not the initial efficiency) which determines the value of an absorbent. It must therefore have a very considerable degree of _chemical stability_. By this is meant that the absorbent itself is not subject to chemical deterioration, that it does not react with carbon dioxide, that it does not disintegrate or become deliquescent even after being used and that it has no corrosive action on the metal container. In a good general absorbent there must be a proper balance between its various essential qualities, and hence the most suitable mixture will probably always be a compromise. CHARCOAL The fact that charcoal would condense in its pores or adsorb certain gases, holding them firmly, had been known for a long time.[28] In general, it was known that so-called animal charcoal was the best for decolorizing sugar solutions, that wood charcoal was the best for adsorbing gases and that coke had very little adsorbing or decolorizing power. No one knew the reason for these facts and no one could write a specification for charcoal. The ordinary charcoal used in the scientific laboratory was cocoanut charcoal, since Hunter had discovered more than fifty years ago that this was the best charcoal for adsorbing gases. RAW MATERIALS[29]