In this paper it is emphasized that density is a typically macroscopic concept which cannot be transferred to small particles or bodies with small pores. This fact has not been duly accounted for in earlier work on the density of cellulose fibers. The usual interpretation of density determinations in various media is criticized and the point of view that cellulose fibers should be considered as porous bodies is dropped. The pore system occurring in native fibers is coarse enough not to interfere with density determinations in indifferent organic liquids.
It is shown experimentally that even substances like water and glycerol penetrate very slowly into dry cellulose and that indifferent organic liquids do not penetrate at all. In the experiments with water vapor a new phenomenon, the heterogeneous course of absorption, is observed, giving rise to sharp microscopically visible boundary lines between the wet and the dry portion of the fiber.