Loss of limb, in man, is often followed by an illusion\p=m-\the conviction that the limb, or parts of it, continues to be present. Wier Mitchell30 coined the term "phantom limb" to designate this phenomenon, but descriptions can be found for centuries before him.* The more recent literature has assumed forbidding proportions,\s=d\yet the phenomenon itself remains puzzling. Perhaps all attempts at explanation are premature until we know more about the phenomenon itself, for the great mass of studies tells us little about such simple questions as the incidence of phantom limbs after amputation, the relative frequency of different phantom sensations, or the changes in the illusion with time.To answer these questions, one needs systematic surveys of groups of cases; most of the available accounts are based, instead, on particularly striking aspects found in selected instances. Thus, for obvious reasons, painful phantoms (first described by Par\l=e'\ in 1552) have been discussed at the expense of "natural" painless phantoms, which are actually much more frequent.:}:A good deal of the literature on phantom limbs is concerned with theories about their origin. These theories are based on conjec¬ tures but are nevertheless useful in guiding one's observations. There are essentially three views on etiology: a peripheral view, a central view, and a third approach, com¬ bining the other two. theories Peripheral Theory.-This theory ascribes phantom limbs to sensations arising from nerve endings in the stump. § Impulses are said to be triggered off by irritants, such as neuromata, mechanical defects, scar tissue, or anoxia. Sensations due to the propagated impulses are assigned to parts originally in¬ nervated. Thus the phantom is "projected" to those areas which had been supplied by the nerve trunk, even though the distal fibers were lost with the amputation.Attempts at direct confirmation of the peripheral theory have been made by claim¬ ing that presence of neuromata or excessive scarring of the stump are associated with especially persistent and painful phantoms; by claiming beneficial effects of peripheral surgery, and by claiming that phantom sensa¬ tions are readily influenced by stimulation of the stump or its nerves.Evidence on all these points is far from convincing. The impression of association between neuroma and persistent phantom is based on unsystematic observations and is rejected by many neurosurgeons.|| Similarly, there are few that accept without question the claims for beneficial effects on phantoms of corrective surgery on the stump, removal of neuroma, proximal cutting of sensory nerves, dorsal root sections, or blocking or \ s=s\ References 10 11, 23, and 34. \ m= p a r \ References 9, 23, 27, 28, and 43.