Nam si plaga in molli parto est, sui debet, maximeque si discissa auris ima est, vel imus nasus, vel frons, vel bucca, vel palpebra, vel labrum, vel circa guttur cutis, vel venter. Si vero in carne vulnus est hiatque, neque in unum orae facile attrahuntur, sutura quidem aliena est; imponendae vero fibulae sunt; ἀγκτῆρας Graeci nominant; quae oras paulum tamen contrahant, quo minus lata postea cicatrix sit. 'Suture is indicated if the lesion is in a soft part, especially in the lobule of the ear, or the ala nasi, or the forehead, or cheek, the edge of the eyelid, or the skin over the throat, or the abdominal wall. But if the wound is in a muscular part and gape, and the edges cannot easily be opposed, suture is contraindicated, and fibulae (Graece ἀγκτῆρας) are to be used in order that the cicatrix afterwards may not be wide.' We have here contrasted two methods of closing a wound, and the conclusion is readily arrived at that sutures in the first case and some metal contrivance in the second are intended. Celsus goes on to say, however: Utraque optima est ex acia molli, non nimis torta, quo mitius corpori insidat. Utraque neque nimis rara, neque nimis crebra iniicienda. 'Both are best made of soft thread, not too hard twisted that it may sit easier on the tissues, nor are too few nor too many of either of them to be put in.' A consideration of various passages in which the Greek authors use the term leaves a distinct impression on one's mind that a metal clasp is intended. Thus Paul (VI. cvii), in treating of compound fractures, says that if a large portion of the bone is laid bare we use fibulae and sutures (ἀγκτῆρσι καὶ ῥαφαῖς). It must be confessed, however, that the words of Celsus render it difficult for us to assert with certainty that fibulae were metal clasps, and we find ancient commentators in equal difficulty. Fallopius and Fabricius d' Aquapendente think fibulae mean interrupted sutures. Guido de Cauliac thinks they mean metal clasps. There is just the possibility that a contrivance like our harelip pin with a figure of eight thread may be indicated. This would satisfy both sides of the question. If fibulae were metal clasps, however, we have several varieties of ancient fibulae that might have been used for closing wounds. That most suited for the purpose in hand seems to me to be one consisting of a small bar terminating in two hooks. Several of these from Roman London are in the Guildhall Museum (Pl. LII, figs. 5, 6, 7). They represent a useful form of 'clip' still in use by cyclists, and they could be applied to wounds to act on the principle of Malgaigne's hooks for the patella. A modicum of support for this view may be derived from the fact that whereas Galen, from whom the above passage on compound fractures is quoted by Paul, uses the word ἀγκτῆρσι, the codices of Paul almost unanimously have ἀγκίστροις. Fourteen out of fifteen give the latter