THE so-called ‘crossed fibrillar’ structure observed in the walls of some plant cells1,2,3 and inferred in a variety of others4 still awaits explanation, and although the evidence for its presence in the latter cases can scarcely be considered satisfactory5, this type of structure is sufficiently widespread to present a problem of the first magnitude in cell-wall physics. Several attempts have been made in these columns and elsewhere to define the underlying mechanism, and it is the purpose of this note to point out a new possibility which arises from the recent work of Astbury and Bell6. The importance of the organization of the protoplasmic surface in considerations of wall structure is now slowly being realized, and recent evidence1 suggests that it is in this surface that we are to seek the mechanism involved in the regular alternation of chain orientation implied by the crossed fibrillar structure. Now in the recent emendation put forward by Astbury and Bell of what may be called the ‘classical’ model of protein structure, the fibrous proteins keratin and myosin are considered always to be effectively in the β-configuration whatever the state of folding [see Figs. 1, 2 and 3, p. 698 of this issue for diagrams of the α- and β-configurations; the supercontracted form would resemble the diagram given by Astbury and Bell for the globular proteins]. The contraction of the α- or of the β- form to the supercontracted condition now corresponds in this model to an effective rotation of the constituent protein chains to the transverse direction.