As a result of the examination of fractured tracheids of Pinus insignis it is shown that an outer lamella (presumed to be the primary layer) of the tracheid wall is composed of cellulose chains which lie at an angle of some 11 degrees to the transverse plane. This orientation is quite distinct from that in the secondary layer, where the cellulose chains lie in steep spirals and in which, as shown in the present paper by an optical method, the chain direction is more or less uniform throughout the whole thickness of the layer. There is no significant correlation between these two directions of orientation. This pseudo-transverse orientation in the primary wall necessitates re-examination of the connexion between secondary wall structure and tracheid length, established by the writer in 1934 and at that time ascribed to changes in primary wall structure which the present paper shows cannot occur. The opportunity is taken of examining some of the current ideas concerning the mechanism of orientation in primary walls, and it is pointed out that the mechanism may involve the cytoplasmic surface. In the cambial initial, as in growing cells investigated by other workers, elongation does not apparently involve any change in the orientation of the cellulose chains. The implications of this are briefly discussed, and it is shown that much needs to be known about the detailed organization of these walls before this maintenance of pseudo-transverse orientation can be completely understood.