LARGE defects in the roof of an orbit may allow herniation of the intracranial contents with displacement of the globe. Such a defect may occur as a congenital deformity, and a method of repair was advocated more than 20 yr. ago by Dandy.1 The same author also popularized the intracranial approach for the removal of orbital neoplasms. When this route is utilized, the thin part of the roof of the orbit is removed, but usually the defect is not large enough to allow the prolapse of the frontal lobe, and the healed dura, together with the orbital contents, has sufficient strength to prevent an encephalocele. However, if the orbital contents should require removal at a later time, this may not be the case, and protrusion of the brain through the defect may occur. Such a situation occurred in Case 1, reported here. Had sufficient foresight been available, a prolonged