The embryonic development of the foliar organ among the Compositor is in general too much abbreviated to give much evi dence in the consideration of the present question, and it should be so expected from the position which the family holds at the head of the vegetable kingdom.
Petioles of the kind seen in this type of leaf-development are very often short and usually more or less margined or winged by the contracted basal parts of the lateral portions of the primitive leaf. They are evidently genetically different from the petioles of stipulate leaves which are developed by the elongation of the axial portion alone. Sessile leaves also are of this type, hence the ab sence of stipules, the stipular tissue being incorporated into the basal part of the blade. But even where stipules are present, the lateral basal portions of the leaves are often in the closest anatomical relation with the stipules. This may be seen in Viola obliqua Hill (fig. 18) in which, near the bundle which passes into the stipule, a similar one arises, takes its course up the petiole supporting its narrow wing and is distributed to a small part of the basal portion of the lamina. We shall find several cases similar to this when we come to the consideration of the Rosacee. There is in this a suggestion of the occasional separation of only a part of the lateral portions to form the stipules and the incor poration of the remainder into the petiole and blade.
The second case is that of the sheathing petiole as it occurs in the Graminoe, A racefe and Umbelliferie. In this case the central basal portion of the primitive leaf is very largely developed and with it the lateral portions which form the margins of the sheath ing petiole. The lamina and true petiole are later developments of the apical and axial tissues. We are strongly supported in this view by the fact that the sheathing petiole is interchangeable with petioles of the ordinary type accompanied by stipules. This occurs in the ilmbelliferw. In Hydrocotyle and a few other genera the sheathing petiole is wanting and stipules are present. The closely related Aralia racemosa L. also has stipules. Still more striking is the case of Comarum palustre L. in which the basal leaves have the sheathing petiole remarkably developed with no indication of stipules (fig. 19), while the upper leaves possess
well developed stipules adnate for not more than half their length (fig. 20).
But the identity of the marginal tissue of sheathing petioles is perhaps best shown in the Ranunculaceie. In the upper basal leaves of Ranunculus bulbosus L., the separation of the lateral por tions is seen actually to have begun, presenting exactly the ap pearance of adnate stipules. The development can be clearly traced from below upward. The first leaf has a short sheathing petiole of the ordinary type (fig. 21). This is slowly modified till in the fourteenth leaf (fig. 22) the vascular bundles have drawn closer together, the sheath has grown shorter and the broad lateral portions, hyaline in texture and requiring no special support other than that of the surrounding leaves, are rounded oil distinctly at the top at the point of beginning of the true petiole. In the fifteenth leaf (fig. 23) there is a further reduction in size and the tips of the lateral portions are free. Another case among the Ranunculacere is that of Thaliclrum poly g a ;num in which the sheathing petiole is of a very generalized type 24). The lateral portions are chiefly hyaline, though sometimes faintly netted-veined and their margins turn in at the apex and meet in the central dorsal channel of the petiole at its base, form ing a ridge between the sheathing and true petioles. This ridge supports a very narrow hyaline membrane which appears to me as the rudiment of a ligule. It would become typical by a little further development of marginal tissue. I believe this to be the origin of the ligule wherever it occurs, though it does not appear so clearly evident in highly specialized groups, nor should we expect such to be the ease. There is also present at the first and second forkings of the petiole a transverse hyaline scale very much like a ligule.
It is noteworthy that the lignle always occurs in connection with the sheathing petiole, as in the erramine:-e and Cyperacea•, or where there is evidence that there has been a sheathing petiole which has disappeared by degeneration, leaving the lignle axillary as in some of the Naiadacete which we shall presently consider.