The surgeon cuts (Plate II. figs. 3, 3) in a semilunar direction (with the convexity downwards) from one side of the deltoid to the other, viz., from the root of the acromion to near the coracoid process; he then raises the large flap upwards and throws it back, opens the joint, disarticulates, passes the knife behind the head of the bone, and cuts out without attempting to save any flaps below, in a transverse direction. By this means the artery is still almost the last structure to be divided, and can be secured by a ready assistant. In cases where much injury has been done to the floor of the axilla and wall of chest, the deltoid flap must be made large in proportion, and triangular rather than semilunar in shape. _N.B._--The statistics of amputation at the shoulder-joint bring out some interesting facts: 1. That the primary amputations here are far more successful than secondary ones. Guthrie records nineteen cases of the former out of which only one died, while out of a similar number in which the amputation was secondary, fifteen died. In the Crimea, British surgeons had thirty-nine cases, with thirteen deaths; of thirty-three primary, nine died; and of six secondary, four were fatal. S.W. Gross's[34] statistics confirm this: of one hundred and seventy-eight primary, forty-six died--25.8 per cent.; ninety-five secondary, sixty-one died--64.2 per cent. AMPUTATIONS ABOVE THE SHOULDER-JOINT.--Under this head we may group the comparatively rare cases in which, from accident or disease, the removal of portions of the scapula and clavicle, or even the entire bones, is rendered necessary. That it is quite possible to survive such injuries has been frequently shown in cases of accident when the scapula along with the arm has been torn off, and yet the patient recovered. Encouraged by such cases, Gaetani Bey of Cairo removed the whole of scapula and part of the clavicle in a case where he had amputated at the shoulder for smash. The patient recovered. Heron Watson has had a similar case. Dr. George M'Lellan amputated arm and scapula in a youth of seventeen for an enormous encephaloid tumour. Fifty-one such cases are now on record. Syme amputated with success the arm along with the scapula and outer half of clavicle, in a case in which he had previously excised the head of the humerus for a tumour.[35] Gilbert, Mussey, Rigaud, Fergusson, and others have performed similar operations, secondary to amputation at the shoulder-joint, for cases of caries and malignant tumour. It is impossible to give any exact directions for the incisions which must be planned for individual cases, with two chief aims, to avoid hæmorrhage as far as possible, and to leave abundance of skin. In operations on the scapula, it should be freely exposed by large enough incisions. (See _Excisions_.)