MIRACLE CYCLES. 187 five. A further inference is possible. If the diver- gence of the textual tradition in the cyclic manu- scripts goes back to the fifteenth century, the tradition of P, or rather of its cyclic original, which contained twenty-one plays only, must be considerably older. In that case P itself, which belongs to the very end of the fifteenth century, cannot have been transcribed from a contemporary 'original,' but is probably a copy of an older prompt- book belonging to the guild. This would lead to the conclusion that the common ancestor of P and the cyclic manuscripts may well be as early as 1400. In no other cycle will the textual tradition take us back anything like as far as this. 1 There are two other instances of a divergence in the tradition which might throw light on the general history of the transmission of the text. It must be borne in mind that the group B D W K, and the sub-group W K, are well established by general textual considerations. Now the two earliest manuscripts, D and W, give a text of the Resurrection play extending to 432 lines, and finish it off in their usual manner. Clearly their proto- type ended at that point ; but the play is incomplete, and in the interval between 1592 and 1600 George Bellin, the scribe of W and K, discovered that there was another tradition, for his later manuscript 1 I ought possibly to state that owing to the very faded condition of the original writing in MS. P, it is impossible to be absolutely certain that the number in the heading is * xx ma ' and not ' xx iiia ,' though the latter would be an exceedingly unusual form. If P's original contained twenty-four pageants, it brings it very close indeed to H ; on the other hand, the further inference as to the antiquity of the whole tradition would collapse.
