beginning of the thirteenth century, the present writer in his edition of the Mort Artu, pp. XXXIIf. used this supposed fact as affording an approximate terminus a quo for the romance. He has since discovered, however, mention of elaborate horse-armour being worn as early as 1187 by followers of Count Baldwin of Flanders. Cp. A. Cartellieri, Philipp II. August, König von Frankreich, I, 255f. (3 vols., Leipzig and Paris, 1899-1910).

6

One might, of course, query whether these allusions to the Mort Artu in the Estoire may not be late insertions especially, since there are other grounds for suspecting that the original text of the Estoire underwent subsequent changes and additions. Cp. p. 371 note 1 above and RR, IX, 380. As long, however, as no cogent reasons are offered for regarding the allusions in question as late interpolations, we may accept them as belonging to the original Estoire.

The Mort Artu plainly presupposes the Queste, but is, itself, referred to in the Estoire (if the above-mentioned allusions are not late additions to the Estoire). This then would confirm the conclusion which we reached pp. 428ff., above, that the last-named romance was composed after the Queste.

Chapter VI.

Development of the Vulgate Cycle.

It will be observed from the above discussions that according to general agreement,1 the Lancelot (in a shorter and earlier form) was the oldest romance in the Vulgate cycle and that the other branches postulate its existence. The evidence which we have presented as to the priority of the Queste over the remaining branches of the cycle prove that it was composed next to the Lancelot. Then followed the Mort Artu, which is dependent on the Queste, but is referred to in the Estoire next expansions and extensions of the original Lancelot that connect it with the Queste and Mort Artu, and after that, the Estoire. That the Merlin was the last of the branches to be composed has never been disputed.

2

If the scheme of relative dates which is set forth in this summary is correct, it is manifest that the cyclic character of the five great romances with which we are dealing does not spring from the fact that they were composed in execution of a preconceived design, whether on the part of a single author or of more than one author, working in collaboration, but rather from

1

Lot constitutes a partial exception to this statement, inasmuch as he supposes (pp. 122f.) that, after the author of the Lancelot had written a large part of that romance, he turned aside, in order to compose the Estoire, and only resumed his work on the Lancelot after he had completed the Estoire. Cp. p. 378, note 4, above.

2

The passages here referred to occur especially in the later portions of the Lancelot, beginning with Gawain's visit to the Grail castle. They are particularly numerous in the so-called Agravain division of the Lancelot. On these subjects cp. pp. 402, note 71, above. Strictly speaking, the Vulgate Merlin-continuation, which its author combined with the old prose-rendering of Robert's Merlin to make up the Vulgate Merlin. On its relative date, cp. p. 395, above. These are the theories of Lot and Brugger, respectively, and will be discussed below. Brugger propounded his first.

4

the romancers' practise of attaching their inventions concerning any particular phase of Arthurian tradition to those of their predecessors concerning some other phase of that tradition. In a period, then, when the sense of literary property was non-existent and the sense of personal authorship not very strong, such a practise was entirely natural, and the result was the creation of the vast composite cycle before us. We face, then, virtually, the same conditions that have called into existence the other great cyclic works of the past for example, the Greek cyclic poems concerning the siege of Troy or the cycle of Guillaume d'Orange among the Old French chansons de geste. An incitement to the composition of a connected series of romances on the Arthurian theme was supplied by Robert de Boron's Joseph-Merlin and its proposed sequels, but how little Robert's series really influenced the Vulgate cycle as a model is patent from the fact that the author of the Queste of this cycle threw overboard that poet's Grail hero and attached his work to a romance (the Lancelot) to which there is nothing even remotely corresponding in Robert's series. By this act an element that was totally alien to the Joseph

The cyclic connection is, largely, of a general kind, and, in the case of the Lancelot, especially, the connecting episodes are most often cyclic in effect, not in aim. Cp. Bruce, RR, X, 121.

6

Cp. RR, IV, 4631. The interval that separated the composition of the various members of the Vulgate cycle was not so long as in the case of the other cyclic works mentioned above. The romances that make up the Guillaume d'Orange cycle are also combined in some cyclic MSS. e. g. MS. 24369 (fonds français, Bibl. Nat.), which contains 17 romances of the total of 24 extant and the Boulogne MS. which contains 11. The latter MS. ends with the words, Explicit li Roumans de Guillaume d'Orange, on which Bédier, Legendes Épiques, I, 8, remarks: "cette rubrique nous est temoin que les hommes du XIII et du XIV siecle, lisant nos chansons, croyaient lire un seul roman." The romances in such a collection, however, were, of course, by different authors. Gröber, Grundriss, II, 9971. tried to trace the history of the development of the Arthurian Vulgate cycle by an examination of the grouping of the romances in the extant MSS, but these MSS. are too late to throw any light on this subject and Gröber's effort is a failure.

е

Merlin, etc., in every way, became incorporated into the history of the Grail. Moreover, if Robert's cycle had, in any strict sense, furnished the model of a pre-conceived plan for the Vulgate cycle, the execution of this plan would, of course, have begun with the composition of the Estoire (which corresponds to Robert's Joseph), but, to say nothing of the priority of the Lancelot, the whole weight of evidence goes to show that the Estoire is later than the Queste, and even than the Mort Artu.

The cyclic character of the romances, it should be said in conclusion, was finally still further strengthened by the occasional cross-references which were inserted in the different branches by individual scribes or by assembleurs. It was the latter themselves, doubtless, scribes who started the manuscript tradition of our extant MSS. by bringing copies of these romances together and editing them in a very rudimentary fashion."

7

Brief anticipatory cross-references in any one branch to later members of the cycle must have got into the manuscript tradition in this manner occasionally, also, somewhat longer interpolations of considerable importance e. g. the interpolations in the Lancelot which are based on the Vulgate and MS. 337 Merlin-continuations and Perlesvaus, respectively. Cp. RR, X, 393f. The passages, remarked on above, pp. 402f., which serve to connect the Lancelot_with the Estoire probably belong to a somewhat earlier stage in the development of the Lancelot. How artless an interpolator could be is shown by the instance in the Lancelot, III, 117, where King Pelles is spoken of as dead although his whole share in the story falls later in the romance. Of course, this interpolator was writing with the complete romance before him. Another speaks unequivocally, III, 429, of the Lancelot being "adjusted" to the Grail story. It seems likely that the insertions in the Lancelot and Queste of passages taken virtually verbatim from the Estoire (cp. p. 376, note 3, and p. 402, note 72, above), was the work of assembleurs. So, too, with the recapitulation, here and there in the Lancelot, of preceding adventures. On the general conditions under which our extant manuscript tradition was established, cp. Bruce, MPh., XVI, 113 ff.

Chapter VII.

The Pseudo-Robert de Boron cycle of the Prose Romances.

The second of the prose cycles, as far as it has been preserved, is ascribed in the MSS. to Robert de Boron. This ascription, however, is unquestionably false, and we, therefore, retain the manuscript designation only in the modified form which will be noted above and which has latterly become general.

The feature of this second cycle which distinguishes it essentially from the Vulgate (Walter Map) cycle is the fact that in the former the characters of the Tristan story play a part, which is not the case with the latter. These characters, Tristan, Marc, Iseult, etc., have evidently been introduced into the cycle from the prose Tristan.3

Of the Pseudo-Robert cycle mere fragments have survived. Curiously it seems to have been more popular abroad than in the country of its origin (France), so that for important portions of the cycle we have only versions in foreign languages especially Portuguese and Spanish -the French originals having been lost. In fact, the very existence of this cycle was first pointed out by G. Paris in 1886. From indications in the Huth-Merlin - i. e.

1

For the passages concerned, cp. G. Paris, Huth-Merlin, pp. XXVIII.

2

8

Everything shows this its dependence on the Galahad Queste, the difference of style, the character of the incidents secular and often fantastic, etc. Cp., especially, A. Pauphilet, Romania, XXXVI, 591 ff. The prose Tristan, in its earlier forms, is undoubtedly anterior to the Pseudo- Robert cycle. We treat the latter first, however, immediately after the discussion of the Vulgate because of its close connection with the Vulgate.

--

4 Cp. his Introduction to the Huth-Merlin particularly, pp. 1ff. This Introduction, although it needs correction in some important