discovered by Mrs Gibson after her catalogue was published. It is a palimpsest, the upper script of which is Arabic of about the end of the ninth century; the under text is that the Peshitto Syriac; and Mrs Gibson, after examining it, believes the writing to be not later than the sixth century. The other specimens are all from dated MSS. They extend from the eighth to the eighteenth century A.D., one specimen at least being given from each century.

In addition to the interest with which those whose attention has been called to the fund of Christian literature in the Sinai convent will view these reproductions from the MSS., such a series of dated specimens of writing is of great value from a palæographical point of view. It is, however, very difficult so to trace development in Arabic writing as to be able to fix by this criterion the period of undated MSS. Specimens ii. and iii., from MSS. in the British Museum, belonging, one to the end of the ninth and the other to the beginning of the tenth century, do show a style of writing very different from that of any of the others. Some of the letters are distinctly Cufic in shape, and it is tempting to suppose that we have here examples of the transition from the ancient Cufic style of writing to the modern style; but more probably, as Professor Margoliouth hints, it is the writing of Syriac which has influenced the writing of these MSS.

These specimens will also be very useful as an introduction to the reading of Arabic MSS. Facsimiles of Arabic writing hitherto published are not very accessible. Here the student will find in convenient form a sufficient number of specimens, each furnished with an accurate transliteration and a careful translation.

The selected pages must have been photographed with great care. The reproductions are almost as legible as the originals can be. Altogether the work is well and carefully done. The only criticism one feels inclined to make is that the volume is worthy of a better binding than the usual bluish paper cover of the Studia Sinaitica.

Wamphray.

RICHARD BELI.

SPRÜCHE UND REDEN JESU, von Adolf Harnack. (Beiträge zur Einleitung in d. Neue Testament, II.) Leipzig: Hinrichs. Pp. iv., 219. M. 4.20; geb. 5 M.

IN the second number of his Beiträge zur Einleitung in das neue Testament Harnack attacks the problem of the source of the passages common to Matthew and Luke. This source was formerly (in England, at least) generally called "the Logia," which implied a rather rash, and probably wrong, interpretation, of the famous passages from Papias about the origin of the Gospels quoted by Eusebius. The Dean of Westminster uttered a protest against this custom in his little book on the study of the Gospels, and there are signs that the leading English scholars are changing their practice and conforming to that of the Germans, who have generally adopted the custom of calling this source Q (Quelle). This is so simple and colourless a title that it is desirable that it should become universal.

Q has not up to the present time been studied so fully as Mark: this is natural, for the material is less promising, and and one of the most valuable data must necessarily be the information supplied by the previous study of Mark as to the methods of Matthew and Luke in dealing with their sources. But in the last few years there has been a tendency to turn to Q, and first Wellhausen's Einleitung and now Harnack's study are valuable contributions to our knowledge.

Harnack divides his work into two main chapters: (1) the identification of passages taken from Q, and the establishment of the most probable text. (2) A literary and historical discussion of Q. The first part is again divided into two sub-divisions-passages in which the agreement between Matthew and Luke is very close, and those in which it is less marked, with an appendix dealing with a few passages in which there is a great difference, though there is some reason for thinking that they may belong to Q.

It would obviously be impossible to discuss the details of

this part; Harnack's conclusion is that Luke made many more alterations of a stylistic nature than Matthew (eight to ten times as many), but alterations of a dogmatic character are rarer in Luke than in Matthew. By removing these alterations the original text of Q can be partially restored, and it is pointed out that the result is a document which is almost entirely free from the linguistic characteristics of Matthew or Luke. To facilitate study, Harnack gives the text of this reconstruction in numbered paragraphs, together with a German translation, the necessity for which may be doubted.

It is this document which forms the basis of the second part of Harnack's book. The result of his investigation into the style of Q seems to show that it has definitely marked characteristics, though not to quite the same extent as Mark. This is not, I think, strange. The special characteristics of Mark are just the points which have disappeared in Matthew and Luke, which sometimes removed them in the same and sometimes in divergent manners. If Mark did not exist in a separate form we should not know its most typical characteristics. It is fair to argue that it was probably the same with Q, and that its most marked characteristics have been disguised by the common dislikes of the redactors of Matthew and Luke. If this be so, it is important to notice the points which remain to bear witness to an originally characteristic style. Harnack has collected on Pp. 102-15 the available evidence: taken in the mass it is interesting, and if not absolutely convincing, strongly corroborates his view. The points which are perhaps most striking are that ù and Tapà are not found at all, although each of them is found six times in Mark and often in Matthew and Luke, and that Q has a preferencc for xai over d, which is also characteristic of Mark, but not of Matthew or Luke.

According to Harnack's reconstruction, Q began with an account of John's preaching, and the temptation of Jesus, and then went on to the kernel of the Sermon of the Mount, and the story of the Centurion of Capernaum, followed by a series of parables and sayings, and ending with instructions

regarding the Parousia. There are also seven incidents which are at least partly of the nature of narrative, but Harnack insists, apparently rightly, that in each narrative the whole emphasis is on the teaching and not on the incidents. This is the case even with the Centurion at Capernaum, where the important point is not healing but faith. Thus Q was neither a complete Gospel on the one hand, nor a formless collection of sayings on the other, but a more or less ordered body of teaching, beginning with the preaching of John and ending with the prophecies of the Parousia. As it stands, it omits all mention both of the Baptism and of the Passion and Resurrection. Harnack thinks that the account of the Temptation implies that there was also an account of the Baptism, which Luke used, in order to modify the Marcan account, but Matthew ignored; but he agrees that it omitted the Passion and Resurrection. It was a translation from Aramaic, but both Matthew and Luke knew only the Greek, and the passages which have been thought to imply the different translations are capable of other explanations. This point deserves emphasis, for the matter is really much simpler than it is sometimes made. Either Q was written in Aramaic and translated separately by Matthew and Luke, or Matthew and Luke both used Greek translations. If the former be true, we should expect general agreement in meaning, and comparatively little verbal agreement. If the latter be true, we should expect considerable verbal agreement. No one doubts which we have the verbal agreement is far too close to be explained except by a Greek original. Therefore the theory of direct translation by Matthew and Luke from a common Aramaic source is excluded so far as Q is concerned. The possibilities which remain are that the variations are textual in Qor due to the redactorial work of Matthew and Luke. The former means that there were already differences of readings in the Greek MSS. of Q used by the Evangelists. Such differences might be accidental, or might be due to the influence of a knowledge of the original Aramaic on various Greek MSS. This is by no means an impossible suggession, but it probably ought to be used with caution, for

so large a percentage of the changes from Q are clearly redactorial alterations by Matthew or Luke, that it is probable that the same explanation ought to be applied to many of those ehanges which nevertheless do not admit of its adoption with certainty. In any case the number of those variants which really strongly suggest the secondary influence of an Aramaic original is so small that one may well ask if they are not accidental.

In the next section Harnack compares Q with Mark, and differs very definitely from Wellhausen. It will be remembered that Wellhausen thinks that Q was inferior to and probably dependent on Mark. Harnack thinks that Q is certainly not dependent on Mark, that it largely agrees in the picture which it gives of our Lord, and that it is probably older. The agreement between the two is due chiefly to the historical faithfulness of both to the original preaching of the Apostles, but Harnack thinks that it is probable that Mark implies acquaintance with much of the contents of Q. As to the connection of Q with the Logia referred to by Papias, Harnack adopts the view that it is probable that Papias himself meant our Matthew, but that it is possible that his informant was really referring to Q, which has therefore some claim to be connected with Matthew. Personally, I must admit that his interpretation of the meaning of the informant of Papias seems less probable than the view, first suggested by Prof. Burkitt, that the original Logia were a collection of Messianic oracles. The evidence that "Logia" was ever used by the earliest Christians of a collection of sayings is as weak as the testimony that it was used of the Old Testament, regarded as Oracles, is strong. When Christians of the earliest period wished to speak of "sayings," they said Logoi, though no doubt by the time of Papias the words of the Lord had been raised to the same authoritative level as the Old Testament. This conscious co-ordination with the Old Testament seems to be implied by the word Logia. Is it probable that Q would have been lost as a separate book if it had been so co-ordinated? This objection does not apply to a collection of Old Testament sayings, which would