The earliest in date, and certainly one of the best from an art point of view, is a sketch in profile done by Haydon preparatory to introducing Keats's head into the picture of Christ's Entry into Jerusalem. The sketch dates in November 1816, just after Keats had come of age. The picture is in Philadelphia, and I cannot speak of the head as it appears there. In the sketch we see abundant wavy hair; a forehead and nose sloping forward to the nasal tip in a nearly uniform curve; a dark, set, speaking eye; a mouth tolerably well moulded, the upper lip being fully long enough, and noticeably overhanging the lower lip, upon which the chin--large, full, and rounded--closely impinges. The whole face partakes of the Raphaelesque cast of physiognomy. At some time, which may have been the autumn of 1817, some one, most probably Haydon, took a mask of the face of Keats. In respect of actual form, this is necessarily the final test of what the poet was like--but masks are often only partially true to the _impression_ of a face. This mask confirms Haydon's sketch markedly; allowing only for the points that Haydon has rather emphasized the length of the nose, and attenuated (so far as one can judge from a profile) its thickness, and has given very much more of the overhanging of the upper lip--but this last would, by the very conditions of mask-taking, be there reduced to a minimum. On the whole we may say that, after considering reciprocally Haydon's sketch and the mask, we know very adequately what Keats's face was--he had ample reason for acquitting himself of being "a fright." We come still closer to a firm conclusion upon taking into account, along with these two records, two of the portraits left to us by Severn. One is a miniature, which was exhibited at the Royal Academy in 1819, and which we may surmise to have been painted in that year, or late in 1818: the well-known likeness which represents Keats in three-quarters face, looking earnestly forwards, and resting his chin upon his left hand. Here the eyes are larger than in Haydon's sketch, and the upper lip shorter, while the forehead seems straighter; but, as to those matters of lip and forehead, a profile tells the plainer tale. The whole aspect of the face is not greatly unlike Byron's. There is also the earlier charcoal drawing by Severn, the best of all for enabling us to judge of the beautiful rippling long hair; it is a profile, and extremely like Haydon's profile, except for the greater straightness of the forehead, and the decided smallness of the chin, points on which the mask shows conclusively that Haydon was in the right. Most touching of all as a reminiscence is the Indian-ink drawing which Severn made of his dying friend on "28 Jan^y. 1821, 3 o'clock morn^g.," as he lay asleep, with the death-damp on his dark hair. It exhibits the attenuation of disease, but without absolute painfulness, and produces, fully as much as any of