58. The intention of this additional piece of apparatus, was to draw down the superior fragment, to prevent the circular casts of the roller from separating by their relaxation, and, by that means, to retain the fragments in apposition. But, on the one hand, may not the superior fragment be drawn by the hand, as well as by a roller employed for the purpose? And, on the other, if the circular casts of the roller be liable to become relaxed, why not the strip of linen also? These considerations induced Desault to lay it aside, and use the bandage in the form just described. 59. The advantages it offers are far from being equivocal. 1st, The limb is kept in a state of invariable extension by the anterior splint, and, on this account, there can be no displacement on the part of the inferior fragment. 2dly, The bandage, which accurately envelopes the whole limb, restrains the action of the muscles by compressing them, and prevents in part the contractions of the triceps; while the casts in the form of the figure of 8, applied with skill and precision, hold down the superior fragment, and render it difficult for it to be displaced. 3dly, Without the application of a roller over the whole limb, a swelling, more or less considerable, would probably be the effect of the constriction at the elbow, which must necessarily be somewhat tight, because, as the turns of the roller, in form of the figure of 8, act on the olecranon obliquely, if they be too loose, they will slip and not perform the office of retention. 60. Like all kinds of apparatus composed of rollers, this ought to be frequently examined, lest, by becoming relaxed, it should not make sufficient resistance to the triceps, which is always disposed to draw itself upwards. There can be no period fixed on for the reapplication of the apparatus; the moment it begins to become slack, it ought to be renewed: three or four times during the course of the treatment are generally sufficient. Should a considerable swelling give reason to suspect that the constriction is too great, it will be necessary to remove the bandage in order to apply it anew. 61. The period necessary for the reunion of fractures of the olecranon varies, according as the bandage is more or less exactly kept in its place. Among ten cases of this kind, collected in the Hotel-Dieu, four united in twenty-four days, three in twenty-eight, and three in thirty-two. Hence, taking the mean term, all other circumstances being alike, the process of cure requires about twenty-six days. 62. When this is completed, it is necessary to impress on the limb motions of flexion and extension, gradually increased every day. This is, as David properly observes, the most certain method of avoiding a stiffness, and even an anchylosis, too often the consequence of this fracture.