The space

same line with the old, is 15 centim. in n, but 12 to 1 in a2. between the columns seems about the same in each MS., viz. about 2.5 centim.; the columns in n, however, measure about 8.3 centim. in length, against the 7 to 7.6 of a2.

A simple way of proving that the two MSS. cannot have been originally part of the same is to apply the left-hand top corner of the first column in the photograph of ʼn to the same corner of the first column in the photograph of a: it will then be found that the second columns do not at all correspond in position, but that the second column in n projects considerably beyond that in a.

It will be seen from this that n is throughout on a larger scale than a,: still, though the two MSS. cannot be identical, they are in all probability of the same age, and closely related to each other. It is quite possible that both may be products of the same scriptorium.

§ 8. FRAGMENTA AMBROSIANA (s).

These fragments were first published by Dr. Ceriani in his Monumenta Sacra et Profana, Tom. i, Fasc. 1 (Milan, 1861). The following account is taken mainly from his preface, supplemented from Peyron, Knoell, and the description in the Palaeographical Society's Series mentioned below.

The fragments belonged originally, like the MS. which stands first in this volume, to the monastery of St. Columban at Bobbio, and passed from it into the Ambrosian Library at Milan. There are indications that, after being bound up with another codex, they had been transferred more recently to the end of the volume numbered C. 73. Inf., of which an account was given by Peyron in his preface to Ciceronis Orationum pro Scauro, &c. (Stuttgart and Tübingen, 1824), p. 131 sqq. The greater part of the volume is occupied by a palimpsest containing in the upper writing, which is of the seventh century, 'Eugyppii Thesaurus,' a collection of extracts from St. Augustine (this is the MS. designated A in the recent Vienna edition, Eugippii Opera, ed. Knoell, 1885, pp. x-xiii), and in the lower writing portions of a commentary on St. Luke, edited by Mai in Script. Vet. Nov. Coll. Tom. iii, Part ii, p. 186 sqq., and fragments of the apocryphal work Parva Genesis, published by Ceriani in the same Fasciculus with the fragments of St. Luke.

The fragments themselves are not palimpsest. They occupy folios 73-76 of the MS. in its present condition. They consist of four leaves, two only of which are consecutive: after the first and after the third two leaves are missing. There are two columns to a page, and 26 lines to a column; the pages being 23 centim. (9 in.) high by 20 centim. (71⁄2 in.) broad: the text is 18 centim. (7 in.) high, and averages about 7 centim. (2 in.) broad. The parchment is thin, and was originally very white,

but is now much stained. The third folio has been torn in the middle, giving rise to several lacunae. Besides this, on the reverse side the letters of the first column are extremely faint and difficult to read, while the second column has been further discoloured by the use of chemicals. The fourth folio has had six lines cut off from the bottom.

The heading 'secundum lucanum' was written in characters similar to those of the text, but somewhat smaller.

The character is an extremely regular uncial, sometimes slightly reduced and contracted at the end of a line, where combinations are found of the letters an, nt, ul, um, unc, unt, ur, us. The abbreviations are those ordinarily used, ihs, ihu, ihm, ds, di, do, dm, dms, dmi, dmo, dom, dme: Christus is written xps: m or n at the end of a word is represented by a horizontal stroke turned, or not, at the ends, and with, or without, a dot under it.

The text is divided into paragraphs of some length. These begin with a letter rather larger than the rest, but of the same shape: a larger letter also sometimes heads the column, not projecting beyond the line. The words are, for the most part, not divided. A slight space is sometimes left to indicate a pause. Numbers are placed between middle points, which also follow the proper names Noe (in xix. 27, not apparently in v. 26) and Lot.

Dr. Ceriani is of opinion that the MS. is not later than the sixth century, and perhaps earlier. A good facsimile and description (also in part by Dr. Ceriani) are given in the Palaeographical Society's Series, pl. 54.

§ 9. FRAGMENTA BERNENSIA (†).

The first to discover and edit these fragments was Dr. Hermann Hagen, Professor of Classical Philology and Director of the Philological Seminary at Berne. They are obtained from a palimpsest forming Nos. 26 and 27 in a miscellaneous collection of 28 treatises, put together in the eighth or ninth centuries, and now numbered 611. These two treatises (described by Dr. Hagen as Physiologus sive de Natura Animalium, and Excerpta Canonica) are written in a Lombard hand over the remains of two older MSS. The larger of these, which is assigned to the fifth or sixth century, Dr. Hagen proposes to describe later; the smaller he has ascertained to belong to the Old Latin Version of St. Mark, and has edited from it in Hilgenfeld's Zeitschrift für Wissenschaftliche Theologie, tom. xxvii, p. 470 sqq., the text which is now reprinted.

It was evidently a matter of much difficulty to decipher the text, which was only brought out by the use of ammonium hydro-sulphuratum. The difficulty was increased by the fact that it was written not like the other portions of the palimpsest at right angles to the upper writing, but parallel to it. The portion preserved makes up folios 143, 144 of the existing MS. Originally it formed the second double leaf in a quaternion, but it has lost out of the four columns which made up each of its two double pages. The text as it now appears is in inverted order, so that fols. 144 b col. 2, 143, and 144 a col. contain St. Mark i. 2 mid. -23 mid., while fol. 144 a col. 2 contains St. Mark ii. 22 mid.—28 mid., and fol. 144 b col. 1 St. Mark iii. 11 mid.—18 mid.

Dr. Hagen does not give any further description of the MS. As, however, the full length of all the columns but the first, and the full breadth of four out of the six has been preserved, we may infer that it was written in 23 lines to a column, and 14-22 letters to a line.

A few words should be said about the reconstruction of the missing half columns in fol. 2. Here we have not been able to follow Dr. Hagen entirely. The question turns upon the criticism of the text, and the critical materials which Dr. Hagen had at his command were imperfect.

He made use only of Sabatier, whose texts were for the most part either late copies of the Old Latin or Vulgate. But the section that will follow on the text of the Berne fragments will show that its affinities are rather with the early MSS. a b d f. These, then, are the real criteria that we should use, and the results that they suggest are different. It may be well to note here the principal points in which Hagen's text appears to be wrong or doubtful.

P. 91, 1. 1 (Mark ii. 22). Hagen reads alioquin, for which we have ventured, though with some hesitation, to substitute nouellum. It is true that alioquin occurs in many other texts, but always before, not after disrumpet (or its equivalent), and it seems out of place in any other position. On the other hand if nouum were supplied from f, with which has some affinity, it would give a line of only 13 letters, which is one letter shorter than the shortest (not being an end-line) of those that are complete. Now d, though reading uinum only without any addition here, has nouellum two lines above at the beginning of the verse. If we might supply this we should get a perfectly satisfactory line, which rests, however, it must be admitted, on rather precarious conjecture. We can hardly be wrong in 'supplying et uinum et utres in the line that follows.

P. 91, ll. 4, 5. eum... ambulare. Hagen reads cum ... ambularet; but et in the next line is much more favourable to eum ... ambulare, which is found in a d. There remains, however, the possibility that had ut ... ambularet with f.

P. 91, l. 13. cum. Hagen reads quando, which is found in f only, whereas cum is the reading of abdei. As keeps closely to the main body, especially when it includes d, cum is more probable, though quando is also possible.

P. 91, l. 15. eo. Hagen reads illo. The balance of evidence (illo fi: eo ab d

e) and of probability is much the same as in the last instance.

P. 92, l. 1. exclamabant: so Hagen, with dei; there is, however, equal authority for et clamabant which is found in a bƒ.

P. 92, 1. 12. dedit illis. The balance of evidence (b dƒi) is in favour of this, against haberent of a e.

P. 92, l. 15. petrus: so Hagen with b dfi, and this is confirmed by Iacobus following, which must have been written under the influence of the preceding word.

There are other expressions besides those above noted which are not absolutely certain, but as there is always a distinct preponderance of evidence it has seemed best to keep (?) for really doubtful cases.

f