Of these remaining MSS. one is directly copied from F, and another seems to be certainly derived from the same source, though perhaps not immediately. In the case of H₃ (MS. Harl. 7184) the question of origin is not quite so simple. Its text generally seems to suggest ultimate dependence on F, but it is very unequal as regards accuracy, and in one part it regularly follows the early first recension readings and seems to belong for the time to the ERCLB₂ group. In addition to this it has a Latin marginal note at the beginning of the Prologue, which is wanting in F. The problem is perhaps to be solved by means of the Keswick MS. This is written in several hands, varying greatly in accuracy, and exactly in that place where H₃ seems to follow a first recension copy the Keswick MS. is defective, having lost several leaves. It also contains the marginal note referred to above, and on examination we find that a whole series of corruptions are common to the two MSS. There seems to be very little doubt that K is the source of H₃, the inequality of the latter MS. being to a great extent in accordance with the change of hands in K, and the variation of H₃ in a portion of the third book to a different type of text being exactly coincident with the gap left in K by loss of leaves, a loss which must apparently have taken place in the first forty or fifty years of its existence[AJ]. As to the text of K itself, in the parts which are most carefully written it reproduces that of F with scrupulous exactness, giving every detail of orthography and punctuation, and for the most part following it in such small errors as it has. It is impossible for one who places these MSS. side by side, as I have been able to do, to avoid the conviction that in some parts at least the exemplar for K was the Fairfax MS. itself. On the other hand, the Latin marginal note at the beginning was derived from some other copy, and setting aside the many mistakes, which possibly are due to mere carelessness on the part of some of the scribes, the Keswick MS. does undoubtedly contain some readings which seem to be derived from a different source. In form of text generally it corresponds exactly with F, reproducing all the additions and corrections made by erasure or otherwise, and containing the same Latin and French pieces in the same order at the end, so far at least as it is perfect. The Magdalen College MS. must be derived ultimately from the same source as H₃, and it has the same lapse from the third recension to the first, coinciding with the gap in the Keswick book. On the other hand W, though in form of text it corresponds with these and with F, is quite independent of the group above mentioned, and probably also of the Fairfax MS. It is late and full of corruptions, but in several instances it assists in the correction of errors which appear in F, and it is apparently based on a copy which retained some of the variants of the earlier text still uncorrected.