1. The section of certain vegetable cell structures takes normally the form of a net of hexagons. When a departure from this form occurs it is observed to be such that three edges still meet at each vertex of the net and the average number of sides per polygonal mesh is still six. After noting this fact, Graustein shows in a recent paper that under these conditions the average number of sides is necessarily precisely six. He then generalises the geometrical result. The present note derives the same results by an entirely different process. The arguments are of a physical character and do not pretend to be mathematically rigorous. But it seems worth while to put them forward as indicating just why the effect should occur in Nature.