but also in arrangement and in contents. Its division into three books, and some other indications, seemed to show that it was really a specimen of that older Roman Sacramentary which writers of the ninth century alleged to have been recast by S. Gregory, and which they connected with the name of S. Gelasius. But there were no other 'Gelasian' Sacramentaries known to its editors: and although here and there the Gregorian books which they were able to employ might throw light on the text of the Vatican manuscript, where this was faulty or obscure, they were of less value for such a purpose than books more nearly approaching to the date of the Vatican manuscript, and more nearly allied to it in structure.

Before Muratori issued his reprint, or Vezzosi issued his new edition, Pierre Le Brun, the learned French Oratorian, had published, under the title of Explication de la Messe1, his valuable dissertations on the Liturgy. He also recognizes, in the manuscript published by Tommasi, a representative of the Sacramentary of S. Gelasius; but he regards it as only one specimen of a class of manuscripts of which he knew other examples. Some of these, he tells us, he proposed to publish in his projected Bibliotheca Liturgica; and while he specifies several points of distinction between the 'Gelasian' and the Gregorian' Sacramentaries, his language suggests that his statements as to the former class were based upon a considerable number of instances 2. Unhappily the Bibliotheca Liturgica never appeared, and Le Brun's mention of 'Gelasian' books had not the effect of producing an enquiry as to their place or their contents.

In 1777 Dom Martin Gerbert, Abbot of S. Blaise in the Black Forest, published, in the first volume of his Monumenta Veteris Liturgiae Alemannicae3, a text which was in part professedly 'Gelasian,' and concerning which there could be little, if any, doubt that it was in part actually drawn from a Sacramentary of the class which was in the ninth century known as 'Gelasian.' At first sight, therefore, it would seem that Gerbert's text must have a high value for the purpose of comparison with the text of the Vatican manuscript. His work has certainly

been of much service in the preparation of the present edition, and it seems almost ungrateful to criticise the method and execution of a scholar of the last century, to whom students of Liturgy owe so much as we owe to Gerbert. But it must be said that his text is one which requires to be used with caution, that the plan which he followed in this part of his work was chosen with an unfortunate want of judgement, and that his mode of handling his materials, and of explaining what those materials were, is at times exceedingly confused and misleading.

Of the three principal manuscripts used by Gerbert in editing, or compiling, the text of his Sacramentary, two, the Codex Rhenaugiensis and Codex Sangallensis antiquior, belonged to the class of books known as 'Gelasian.' The third, a manuscript of a peculiar character, of which he sometimes speaks by the title Sangallensis recentior, sometimes simply by the title Sangallensis, was, when he used it, not at S. Gallen, but at Zürich. He calls it in one passage 'Sangallensis olim nunc Turicensis, ex triplici ritu Gelasiano, Gregoriano, et Ambrosiano compositus. To this third manuscript, of much later date

1 The two 'Gelasian' manuscripts, the Rheinau and the S. Gallen, have been collated for this edition, and are frequently referred to by the symbols R. (for the Codex Rhenaugiensis) and S. (for the Codex Sangallensis). The third of Gerbert's principal MSS. is sometimes indicated in the following pages by the symbol T. (for Turicensis): but, for reasons which will be apparent, it is not often mentioned. Even if it had been of more value for the purposes of this edition than there is any reason to suppose, it cannot now be traced. Its presence at Zürich is most probably to be explained by supposing that it was one of the manuscripts which fell to the share of the town of Zürich, and were removed to its Town Library, when the Library of S. Gallen was plundered in the religious war of 1712 by the forces of Zürich and Bern, and that it was not among those which were returned to S. Gallen a few years later. Both in a Zürich list of manuscripts brought from S. Gallen, drawn up in 1713, and in a S. Gallen

list of losses suffered by the Library of the monastery in 1712, there appears a volume described as Collectae Missales: the S. Gallen list adds the further description 'seu Missae Gregorianae et Ambrosianae,' a phrase which seems to point to a compound Sacramentary, and may not impossibly be a description of the Sacramentary which Gerbert, later in the eighteenth century, found at Zürich. (Weidmann, Gesch. der Bibliothek von S. Gallen, pp. 435, 440.) In the Catalogue of the S. Gallen MSS. edited by G. Scherrer (Halle, 1875), in the notice of the S. Gallen MS. 348, reference is made to Gerbert's work, and the triple Sacramentary used by him is described as 'jetzt Zürcher codex C. 389.' It does not appear whence Scherrer derived this information. But the manuscript now bearing the mark C. 389 in the Town Library of Zürich is not a Sacramentary, but a collection of miscellaneous frag

than either of the others, Gerbert gave the first place in his estimation; and it was this manuscript which he chose, rather than either of the other two, as the basis of the greater part of his text. The reason for this preference, and some of its consequences, must be very briefly stated.

Gerbert considered the Codex Sangallensis (S.) to be a manuscript of the latter part of the eighth century ('vix assurgit ad mille annos'); the Codex Rhenaugiensis (R.) he supposed to be earlier than S. ' aliquot annorum decadibus.' The third manuscript was of the tenth century, and he believed it to have been written at S. Gallen. It was, as has been said, a Sacramentary of an unusual type. The compiler had had for his object the combination in a single volume of the services for the various days of the Church year according to the rites which were known to him as 'Gelasian,' 'Gregorian,' and 'Ambrosian.' In carrying out this task he appears to have used the Codex Sangallensis (S.) as one of his authorities, transcribing its contents, or great part of its contents, and distinguishing as 'Gelasian' the portions of his triple text which were derived from this source1.

The special interest attaching to the combination of rites presented by T. gave to that manuscript, no doubt, part of its special importance in Gerbert's eyes: a Sacramentary which included not only Gelasian forms but also Gregorian and Ambrosian missae for the same days appeared to him to possess a character of completeness which the older manuscripts could not claim but he was influenced also by other reasons. He conceived that the resemblance between R. and S. was so close

Librarian of Zürich :-' Remitto codicem Turicensem incomparabilem, cui forte parem in re liturgica non fert orbis litterarius. Si otium fuerit, edam cum singulari commentario.'

1 This makes it clear that S. (and therefore also R., which is certainly a book of the same type as S.) would in the tenth century have been described as 'Gelasian.' It was, presumably, on the connexion between S. and T. that Gerbert, in part at least, relied in supposing that T. was written at S. Gallen. He seems to have had no doubt as to the relation of the two manuscripts; indeed the comparison of S. with his printed text leaves very little

room for hesitation in accepting his view. The 'Gelasian' portions of his text are in very close agreement with S., as the latter now stands; and the case is still further strengthened by the fact that the compiler of T. seems, in almost all cases, to have given effect to certain marginal notes which appear in S., intended for the guidance of a copyist, as to the order of missae and prayers. In one case a marginal gloss in S., explanatory of a word in one of the prayers, has been incorporated with the text of the prayer as it appears in Gerbert's Sacramentary (see note 18 on p. 12 of this volume).

as to leave no doubt that both had been copied from the same source, though S. contained some missae for festivals and some other matter which had been added in the interval between the dates at which the two had been written. It was true, he thought, that the latter part of R. contained a good deal which was not to be found in S.: but so far as the first part, containing the services for the yearly round of Sundays and festivals, was concerned, S. seemed to him to contain practically everything that was to be found in R., with valuable additions. As between these two, therefore, he would have given the preference to S. as an authority, even if it had not possessed what he regarded as an advantage of special importance, Originally, he says, R. and S. had agreed 'ad apicem usque . . in corrupto etiam dicendi scribendique genere, erroribus grammaticalibus ac sphalmatis aevo Merovingico propriis 1.' But the original text of S. had been revised and corrected by a later hand, which had 'restored the true sense' by emending the errors of the original scribe: and these improvements in the text were to be found also in T., since the compiler of this manuscript had followed the corrected text of S. Thus, in Gerbert's view, while S. contained, in its round of services, all, or nearly all, that was to be found in the parallel portion of R., T. also contained the whole, or nearly the whole, of what was to be found in S., in a corrected and improved form, and with further additions.

He accordingly determined to reproduce the text of T., distinguishing by special type those parts of the 'triple text' which were to be found in S., and marking by brackets those portions which, though contained in S., were not to be found. in R. Where T. was defective, he used one or other of the two older manuscripts as the basis of his text, and he also reproduced from R. a good deal of matter not contained in S. or T., while he occasionally added, in his footnotes to the text, some particulars as to the readings or arrangement of the two older manuscripts. Unfortunately, in carrying out his plan, he

1 An examination of the two MSS. does not bear out Gerbert's statement on this point. The text of the two agrees, in the common matter, very closely, but they do not by any means exactly accord with one another in the matters of grammar and spelling. The grammar of R. is more

faulty than that of S., and its spelling much more variable and incorrect. As to the nature and extent of the agreement which Gerbert alleges to exist between them in respect of the matter contained in the two Sacramentaries, more will be said at a later stage.

was not always careful in the employment of his marks of differentiation. Some entire missae and many portions of missae which are contained in S. are not distinguished by the type which ought to mark the matter found in the older books, while a large number of missae which are absent from R. appear in his text without the brackets which ought to have indicated their absence from that manuscript 1.

But besides this defect of accuracy in following out the plan he had laid down for himself (a defect which seriously diminishes the value of the information conveyed by his method) Gerbert's way of regarding his materials had another unfortunate result. As he was inclined to minimise the differences of text between R. and S., he was also inclined to underrate, or to misunderstand, the importance of the changes introduced into the latter manuscript by the hand of the corrector. These changes were, in many cases, much more than mere corrections of mistakes in grammar or variations in spelling. The effect, if not the purpose, of the corrector's work has generally been the alteration of the original text into closer agreement with the text of the later Sacramentaries of the type known as 'Gregorian.' Hence it comes to pass that the Gelasian' portions of the 'triple text' of T. (and therefore of Gerbert's printed text) do not represent the original text of the 'Gelasian' books, but a revision of that text, apparently based upon the text of the Sacramentaries of the 'Gregorian' type 2 The evidence for this statement will be found in the notes of the present volume, where it will again and again appear that the text of R. and the original text of S. are in agreement with the readings of the Vatican manuscript, while the readings of the later hand of S., followed by Gerbert's printed text, are in agreement with the

1 Gerbert himself apologizes, in his preface, for the omission, in two specified cases, of these distinguishing brackets: but the instances of such omission are far more numerous. As a matter of fact he has failed to mark in this way more than sixty entire missae, and a large number of Collects and Prefaces. It is quite impossible to gather from his text the real state of the case as to the amount of matter common to R. and S.

2 This does not of course apply to the portions of Gerbert's text where R. has

been employed as his authority. In these portions he reproduces the readings of the original text, not literally, but with substantial accuracy for the most part. He has, however, occasionally failed to read. the manuscript correctly, or to notice the existence of a gap occasioned by the loss of a leaf, and has thus perpetrated one or two amazing blunders. In anything which he extracts directly from S., he appears as a rule to follow the readings of the second hand.