THE familiar absorption of water from the air by cotton has been studied in our laboratories for some years from various aspects. The phenomena are valuable as clues for interpreting the physico-chemical structure of cotton cellulose; and in this connexion we have naturally examined the capillary hypothesis, using Anderson's formula for calculation of capillary radii. We recently initiated a simple series of observations which indicate that this hypothesis is inadequate, by showing that the weight of vapour absorbed is dependent on the chemical nature of the vapour, with a suggestion that there is a constant “low level” group, including definite types of liquids, with which the values are due to simpler causes than with water or acetic acid, the values for the two latter being far higher. As the differences for various liquids are very marked and do not seem to have been previously recorded, the following particulars may be of interest, in spite of the primitive methods used in what was merely a preliminary examination.