Now, comparing these versions together, the right reading becomes perfectly clear. The scabbard is red, for it was made (at least one half of it was) of the wood of the Tree of Life, which, as we are distinctly told, turned red at the death of Abel; and the inscription 'memoire de sang' was intended to keep this event in mind. The confusion, in the case of Q. and D. L., clearly arose from the MS. at the root of the first having had the reading sans for sanc or sang (a reading often met with); a careless copyist, heedless of the sense of his transcription, wrote sens and this was correctly translated by the compiler of D. L. as sinne; a reading which, however unintelligible in itself, would probably not strike the compiler (who was certainly an intelligent writer with a very good knowledge of French) as absolute nonsense, inasmuch as it was connected with the 'calling to mind' of the death of Abel. Q., who omits this qualifying passage, does make nonsense of it. In M.'s case the mistake was in the first word, and probably arose from a confusion between m and uv, which may very well be due to Caxton; otherwise M. appears to have had the same version as 1533, which, alone, has preserved it free from error. W. omits the inscription altogether.
The 'erle Hernox' in M., Ernous in Q., is in D. L. and 1533 Arnout and Arnoul. Ernoulf in W.
Both D. L. and 1533 state that the maiden who shall cure the lady by her blood must be not only a virgin and a king's daughter but Perceval's sister. This is neither in Q. nor in M., and may perhaps indicate that, as I have suggested, these two versions belonged to an original Perceval-Lancelot redaction, from which they introduced occasional additions to Perceval's share of the Queste, as in the previous allusion to his having recovered his kingdom in D. L.[184]
In the account of Lancelot's visit to Corbenic, after being struck down at the sight of the Grail, Q. says he is discovered seated (seant) before the door, while the other three all represent him as lying (lyinge—licgende—gisant), which is certainly more in harmony with the general situation. D. L. says that when Lancelot recovers and knows he has lain unconscious fourteen days he bethinks him:
'Hoe hi hadde gedient den viant
.xiiij. jaer, ende pensede te hant,
Dat hem onse here daerti dede
Die macht verlisen in sine lede
.xiiij. dage.'—ll. 9919-23.
Whereas Q. only says 'qu'il avoit servi l'anemi.' A meaningless phrase, as it stands. M. agrees with D. L. with the exception that he says twenty-four instead of fourteen, in which he is certainly correct, as Lancelot's liaison with Guinevere had begun long before the birth of Galahad. The number may have been altered by the compiler of D. L. for the exigencies of the rhyme, which would not admit the original form. 1533 omits the passage altogether, condensing considerably at this point.[185] W. does not specify whether he were lying or seated, but agrees with D. L. in giving fourteen years, which rather looks as if that number may have been in the source of this latter.
In the account of the questers at Castle Corbenic D. L. and 1533 alike clear up a passage which, as it stands, is obscure in M. and utterly unintelligible in Q. Nine stranger knights arrive at the castle,[186] three being of Gaul, three of Ireland, and three of Denmark. When they separate the next day, Q. has this unintelligible passage—Galahad has asked the strangers' names—'et tant qu'il trouverent estrois de gariles, que claudins li fieus le roi claudas, en ert li uns et li autre de quel terre qu'il fuissent, erent asses gentile homme et de haut lignage.'[187] M. renders, without any mention of names being asked, 'But the thre knyghtes of Gaule, one of them hyghte Claudyne, kynge Claudas sone / and the other two were grete gentylmen' (which should surely have given Dr. Sommer a clue to the right rendering of the passage).
D. L. runs thus:
'Ende alsi buten den castele quamen
Vragede elc oms sanders namen,
Soedat si worden geware das,
Dat van den drien van Gaule was
Claudijn Claudas sone die een,
Ende si vonden van den anderen tween
Dat si waren van groter machte
Ridders, ende van groten geslachte'—ll. 10601-8.
1533 has, 'Si trouverent que des trois de Gaulle Claudius le filz au roy Claudas en estoit ung et les autres estoiet assez vaillans.' It seems clear that M.'s text is that of D. L. carelessly abridged.[188]