Before leaving the consideration of stem structures there is another feature to which attention must be drawn. On the casts shown in fig. 82 there is a circle of small oval scars situated just below the nodes, these are clearly shown at _c_, _c_, _c_. Each of the scars is in reality a slight projection from the upper end of an internodal ridge. As the ridges correspond to the broad inner faces of medullary rays, the small projection at the upper end of each ridge is a cast of a depression or canal which existed in the medullary tissue of the living plant. There have been various suggestions as to the meaning of these oval projections; several writers have referred to them as the points of attachments of roots or other appendages, but Williamson proved them to be the casts of canal-like gaps which traversed the upper ends of principal medullary rays in a horizontal direction. In a tangential section of a Calamite stem the summit of each primary medullary ray often contains a group of smaller elements which are in process of disorganisation, and in some cases these cells give place to an oval and somewhat irregular canal. In the diagrammatic tangential section represented in fig. 83, _A_ the upper end of each ray is perforated by a large oval space, which has been formed as the result of the breaking down of a horizontal band of cells. Williamson designated these spaces _infranodal canals_. While proving that they had nothing to do with the attachment of lateral members, he suggested that they might be concerned with secretion; but their physiological significance is still a matter of speculation. The casts of infranodal canals are especially large and conspicuous in the subgenus _Arthrodendron_, a form of Calamite characterised by certain histological features to be referred to later. Williamson[625] originally regarded the presence of infranodal canals as one of the distinguishing features of _Arthrodendron_, but they occur also in the casts of the commoner type _Arthropitys_. As a rule we have only the cast of the inner ends of the infranodal canals preserved as slight projections like those in fig. 83, _A_; but in one exceptionally interesting pith-cast described by Williamson, these casts of the infranodal canals have been preserved as slender spoke-like columns radiating from the upper ends of the ridges of the infranodal region of a pith-cast. [Illustration: FIG. 82. Branched rhizome of _Calamites_. ½ nat. size. _C_, _C_, nodes showing casts of infranodal canals. From a specimen in the Manchester Museum, Owens College.]