We have a general knowledge of the external appearance of _Sphenophyllum_ from the numerous impressions of leaves attached to twigs which are found in the rocks of the Carboniferous period. These impressions present a good deal of variety, but all have rather delicate stems with whorls of leaves attached at regular intervals. The specimens are generally easy to recognize from the shape of the leaves, which are like broad wedges attached at the point (see fig. 112). In some cases the leaves are more finely divided and less fanlike, and it may even happen that on the same branch some may be wedge-shaped like those in fig. 112, and others almost hairlike. This naturally suggests comparison with water plants, which have finely divided submerged leaves and expanded aerial ones. In the case of _Sphenophyllum_, however, the divided leaves sometimes come at the upper ends of the stems, quite near the cones, and so can hardly have been those of a submerged part. The very delicate stems and some points in their internal anatomy suggest that the plant was a trailing creeper which supported itself on the stouter stems of other plants. [Illustration: Fig. 112.—Impression of _Sphenophyllum_ Leaves attached to the Stem, showing the wedge-shaped leaflets arranged in whorls] The stems were ribbed, but unlike those of the Calamites the ribs ran straight down the stem through the nodes, and did not alternate there, so that the bundles at the node did not branch and fuse as they did in _Calamites_. The external appearance of the long slender cones was not unlike that of the Calamite cones, though their internal details showed important distinctions. In one noticeable external feature the plants differed from those of the last two groups considered, and that was in their _size_. Palæozoic Lycopods and Equisetaceæ reached the dimensions of great trees, but hitherto no treelike form of _Sphenophyllum_ has been discovered, and in the structure-petrifactions the largest stems we know were less than an inch in diameter. In the internal anatomy of these stems lies one of the chief interests and peculiarities of the plants. In the very young stage there was a sharply pointed solid triangle of wood in the centre (fig. 113), at each of the corners of which was a group of small cells, the protoxylems. The structure of such a stem is like that of a root, in which the primary wood all grows inwards from the protoxylems towards the centre, and had we had nothing but these isolated young stems it would have been impossible to recognize their true nature. [Illustration: Fig. 113.—_Sphenophyllum_, Transverse Section of Young Stem _c_, Cortex, the soft tissue within which has decayed and left a space, in which lies the solid triangle of wood, with the small protoxylem groups _px_ at each corner. (Microphoto.)]