When the tumour is circumscribed, whether meningeal, cortical, or subcortical, it may be shelled out of its bed with greater or lesser ease according to its nature and position. This shelling out process is carried out with an ordinary tea-spoon or scoop. Hæmorrhage may be severe though generally readily controlled by lightly packing with dry gauze. More rarely one or more of the superficial vessels will require to be underrun with a small needle threaded with the finest catgut. Muscle grafts (see p. 18) may be of considerable assistance. If a cyst be found it may be possible to shell it out entire, failing which the parietal wall is freely dissected away, and the cavity drained for two or three days. [Illustration: FIG. 72. COMBINED FLAP FORMATION AND DECOMPRESSION. After osteoplastic resection, the tumour has been found irremovable. The dura mater is therefore sewn back in position, after which a portion of the bone is nibbled away from the bone-flap--as depicted in the illustration--and the underlying dura mater freely incised.] If the tumour be extensive and ill-defined in margin, no attempt should be made at removal, the surgeon remaining content with the second desideratum of brain tumours in general--the production of a general decrease of intracranial pressure. This might be readily effected by leaving the dura open and by removing at the same time the osseous portion of the osteoplastic flap. The bone is readily dissected away and free decompression would be permitted. In such cases, however, the hernial protrusion is usually excessive, and insomuch as an osteoplastic flap is more often than not framed over the Rolandic region, the protrusion would include the motor area with disastrous results on the contra-lateral extremities. This course, therefore, should never be adopted. In such cases it is infinitely preferable to follow Cushing’s method of combined exploration and decompression. This is done as follows: ‘From under the portion of temporal muscle which has been turned back with the flap, a roughly semicircular area of bone is cut away with heavy rongeurs, which remove bone without jar, and so without risk of stripping the remainder of the resected bone from the soft parts. This accomplished, a similar area is rongeured away from the side of the skull well down the temporal fossa under the tourniquet, the temporal muscle being held away by a retractor. If the base of the bone has been made sufficiently broad, a margin possibly a centimetre in width can be left on each side as a support for the flap after its replacement. A subtemporal bone defect is thus secured with even less difficulty than is experienced in making the usual subtemporal opening from without through a split muscle incision. The dura is then carefully opened and incised in a stellate fashion to the margin of what promises to be a sufficient circle of denuded cortex for a generous decompression.’