asked why they are not regarded as having been originally one. Here we must turn to the subject of doublets for a reply. Scattered throughout all the gospels are occasional repetitions of a saying twice in one gospel; these are naturally termed doublets. A list of some of those which are most obvious is given in Table VII, with each side of the references tabulated in order. To see the utility of these, turn to the longest document, the Nucleus. In that there are most doublets in Luke, and we can see their meaning best there. Of the 11 doublets in all Luke, 7 are in the Nucleus as regards one side only; but there is not a single instance of both sides of a doublet occurring within the Nucleus. Their absence shows that when a passage was once used in a document, it was not repeated in the same document. Such a conclusion might be expected a priori, but we are here able to prove it. Thus we see that a doublet serves to shew that its two contexts are parts of different documents.

Now, in the sermon A, and in the document C, there is a doublet, vii 17 and 20 (see the parallel in sequence, Luke vi 43) similar to

xii 33 and 37. This indicates that the documents were independent.

Similarly, in the sermon A and in E there is a doublet, Matt. v 34 and xxiii 22, which shews these were separate documents.

12. The Disintegrated Documents (IV). The fourth class is that of documents which have been in all cases disintegrated and used for extracts. At first sight it seems farfetched to suppose that scattered verses in one gospel had any connection with scattered parallels in another. But we have learned much from the treatment of the Sermon on the Mount by Luke, who has a scattered chain of fragments shewing where it has been used. If one writer could thus break up and utilise a document, two or three writers may have done so; and thus there may only be left a chain of verses in the same order in each gospel, but widely scattered in quite different contexts. We have therefore to look for the occurrence of several episodes or sayings repeated in different gospels, in the same sequence though differing in context. Such a chain we can identify as portions of a

document which has been differently treated by each writer.

We can therefore trace thus two scattered documents, the parallel quotations of which are given in the Appendix, Table VI, lettered B and D.

The block Matt. ix 32 to xi 27, from which seven sayings are parallel in Mark and Luke, might be thought to be due to a misplaced part of the Nucleus. But there is a doublet with the Nucleus in x 38-9 and xvi 24-5, which shews that this is an independent document, D. It seems to be an early form of a commission to the apostles.

The other document, B, is very scattered, from Matt. viii 23 to xxiv 41, which comprise fifteen passages parallel to Luke in the same order throughout, except a transposition of two in Luke xvii 25, 33. Where the Marcan parallel is already reckoned as parallel to either of the others in the Nucleus or twogospel sequences, it is marked =. These passages in the two gospels indicate that B had been incorporated by both Matthew and Luke before Mark drew from either of them.

To thus find fifteen scattered passages

remaining yet unchanged in their consecutive order in different gospels, is a very astonishing result, and it throws a flood of light on the working system of accretion in the gospels.

13. The Isolated Episodes (V).

The fifth class is that of isolated episodes or sayings, which are not in the same order or connection in different gospels. Of these there are about thirty in Matthew, mostly of one or two verses in length. The whole amount is not more than about seventy-five verses, or onefourteenth of the gospel. This shews that not much use was made of short sayings or detached logia, from oral tradition which had no context.

14. The Unrepeated Episodes (VI).

The sixth class is of episodes which are only preserved in a single gospel, the material which is peculiar to each evangelist. In Matthew this is the largest class, two-sevenths of the whole; and it is much the same in Luke. But in Mark there are only two short passages, nine verses in all, which are peculiar to that gospel. This indicates that the second

gospel was soon isolated from the fountainhead of material in Judaea, while Matthew seems to have long continued to add to his gospel, and Luke collected a large body of documents which he carried away out of reach of Mark to build up his extensive new chapters, quite independently of the gospel of Matthew, which remained in Jerusalem.

We have now seen how far purely structural criticism will lead us, without looking at the subjects of the documents or the connections of ideas involved in the arrangement.

Everything that has been brought to light is entirely clear of personal opinions or suppositions. It is difficult to see how any different conclusions could be reached from the mechanical facts of the sequences and of the distribution of episodes which lie before us.