^{page 453 note 2} These theories are henceforth referred to collectively as ‘the two-document hypothesis’.

^{page 453 note 3} Streeter, B. H., The Four Gospels (revised edition 1930, reprint 1961), pp. 160 f.Google Scholar; Burkitt, F. C., The Gospel History and its Transmission (1911), pp. 40 ff.Google Scholar

^{page 453 note 4} One can best observe the extent of these agreements, both positive agreements and agreements in omission, by trying to find a triple-tradition pericope which does not have any agreements between Matthew and Luke against Mark; there are only one or two. The agreements, especially the positive ones, are discussed in the literature referred to in notes 1 and 3 above. See further Argyle, A. W., ‘Agreements Between Matthew and Luke’, Exp. T. LXXIII (1961/1962), 19–22Google Scholar and Turner, N., ‘The Minor Verbal Agreements of Mt. and Lk. against Mk.’, Studia Evangelica (TU, LXXIII, ed. K. Aland, and others, 1959), 223–34.Google Scholar Turner's article is especially persuasive on the point that Matthew and Luke agree against Mark in the triple tradition in ways which cannot be explained by their independently altering the text of Mark. A list of agreements in omission is given by Hawkins, , Horae Synopticae ² (1909), pp. 125 ff.Google Scholar, and they may be seen very readily in the first column in Abbott and Rushbrooke's, The Common Tradition of the Synoptic Gospels (1884)Google Scholar, Which is headed ‘St. Mark. [Portions not found in Matthew or Luke.].

^{page 454 note 1} See, e.g., Streeter, , ‘St. Mark's Knowledge and Use of Q’, Oxford Studies in the Synoptic Problem (1911), pp. 166 f.Google Scholar; The Four Gospels, p. 191.

^{page 454 note 2} For criticisms of Streeter's method, see the works by Butler and Farmer cited above.

^{page 454 note 3} In addition to ‘St. Mark's Knowledge and Use of Q’, See The Four Gospels, pp. 186–91; 305 f.Google Scholar

^{page 454 note 4} ‘St. Mark's Knowledge and Use of Q’, pp. 171, 173; The Four Gospels, p. 187.

^{page 454 note 5} In the essay of 1911.

^{page 454 note 6} The Four Gospels, p. 306.

^{page 455 note 1} Ibid. p. 243.

^{page 455 note 2} Ibid. p. 259.

^{page 455 note 3} The Marcan overlaps with L, according to Streeter, are these: the rejection at Nazareth, the call of Peter, the Anointing, and the entirely of the Passion narrative. Mark and M overlap threee passages: on divorce, the man with the withered hand, and the Syrophoenician woman. The overlaps of Mark and Q are given below.

^{page 456 note 1} In the former passage, in fact, E. Schweizer argued that Luke had a special source. See ‘Eine hebraisierende Sonderquelle des Lukas?’, Th.Z. VI (1950), 161–85.Google ScholarJürgen, Roloff, Das Kerygma und der irdische Jesus (1970), p. 147,Google Scholar nothing that Matthew and Luke's agreements against Mark in the story of the epileptic child cannot be explained by the theory that they independently changed Mark, suggests that both Matthew and Luke used a special source.

^{page 456 note 2} ‘St. Mark's Knowledge and Use of Q’, p. 173.

^{page 456 note 3} Streeter, , The Four Gospels, p. 183;Google ScholarWoods, F. H., ‘The Origin and Mutual Relations of the Synopitic Gospels’ (1886), Studia Biblica et Ecclesiatica II, 61–2.Google Scholar

^{page 457 note 1} See below.

^{page 457 note 2} On agreements in order between Matthew and Luke againse Mark, See ‘The Argument from Order and the Relationship between Matthew and Luke’, N.T.S. xv (1969), 249–61.Google Scholar

^{page 457 note 3} In the early essay, Streeter argued that in certain passages Matthew and Luke maintained primitive tradition from Q while Mark abbreviated it. Thus he stressed the similarities among all three and had no occasion to speak of ‘obviously different versions’. When he later argued that Mark, while overlapping with Q, was independent of it, he stressed the dissimilarities bewtween Mark on the one hand and Matthew and Luke on the other and he made the point about different versions.But in both instances he remained concerned to explain the agreements of Matthew and Luke against Mark, and in both instances he dealt with the same passages.

^{page 457 note 4} See The Tendencies of the Synoptic Tradition (1969), p. 69 and p. 462 n. 3 below.

^{page 457 note 5} These statements would also apply to Matthew xxi. 10–17 and parallels.

^{page 457 note 6} The figures are by my own count. They obviously depend on what text is used as well as on what are counted as agreements. I have used the text printed in the ninth edition of Huck's Synopsis. I have counted as agreements instances in which the same root appears in two or more Gospels, even if in another form. The articles is counted whrn it precedes a word which is itself counted.

^{page 458 note 1} This is the list given in The Four Gospels, pp. 305 f. In ‘St. Mark's Knowledge and Use of Q’, Streeter also listed Mark ix. 42–50; Mark xii. 38–40; and all of Mark iv. 21–5. Having established the case in eight instances, he proceeded to suggest another eight possibilities (pp. 176 f.). In The Four Gospels, p. 191, he said that only portions of Mark iv. 21–5 and ix. 42–50 are paralleled in Q, but he added two examples of small overlaps in Mark viii. 34 and viii. 38. J. P. Brown, op. cit., listed several other passages as overlaps of Mark and Q. This indicates the tendency to expand Q to explain agreements between Matthew and Luke against Mark.

^{page 459 note 1} ‘St. Mark's Knowledge and Use of Q’, p. 168: ‘Matthew iii. 12 = Luke iii. 17 has no meaning appart from the preceding verse, which therefore must have stood in Q and not have been derived by editors of Matthew and Luke from Mark’.

^{page 459 note 2} Ibid.

^{page 459 note 3} The Four Gospels, pp. 188 f.

^{page 459 note 4} Streeter, , The Four Gospels, p. 162.Google Scholar

^{page 459 note 5} The Synoptic Problem, pp. 211–15.

^{page 459 note 6} See David, Dungan, ‘Mark – the Abridgement of Matthew and Luke’, Jesus and Man's Hope I (Pittsburgh, 1970), 60–5.Google Scholar

^{page 459 note 7} In one accepts the Western text for Luke iii. 22, as Streeter did. See The Four Gospels, p. 188.

^{page 459 note 8} Streeter, B. H., ‘The Original Extent of Q’, Oxford Studies in the Synoptic Problem, p. 187.Google Scholar

^{page 459 note 9} The absence from Mark of the ‘body’ of the passage is somewhat reminiscent of Mark xi. 11. It is either a different tradition, or Mark's from has been so severely abbreviated as to alter the nature of the story.

^{page 460 note 1} Streeter, ‘St. Marks's Knowledge and Use of Q’, pp. 169–71; The Four Gospels, pp. 189, 211.

^{page 460 note 2} B. C. Butler, op. cit. pp. 8 ff. Butler was generally sceptical about all theories which require conflation; see p. 5. Lamar, Cope, Matthew: A Scribe Trained for the Kingstom of Heaven (unpublushed dissertation, Union Theological Seminary, N.Y. 1971), pp. 64–70,Google Scholar agrees with Bulter that the passage in Matthew is too coherent structurally to be the result of conflation, but his description of the structural unity differs from Butler's. Brown, op. cit. pp. 35 f., also thinks that Matthew's form is coherent and original: ‘Mark thus presupposes … precisely the arrangement we find in Matthew’. His general view is that, in the overlaps, ‘Mark gives a mostly inferior version of Matthew's Source for sayings’, although Matthew depends directly on Mark for narrative material (pp. 43 f.).

^{page 460 note 3} ‘The Argument from Order and the Relationship between Matthew and Luke’, 260.

^{page 461 note 1} If the Baptism is not included in Q, then Matthew and Luke would have agreed in placing Marcan material after the Q passage Matthew iii. 12 = Luke iii. 17. Luke, however, has an intervening pericope.

^{page 461 note 2} Brown's arguments (op. cit) to the same effect are also frequently telling.

^{page 461 note 3} It is noteworthy that Kümmel, , Introduction to the New Testament (E., T. by Mattill, A. J., 1966), p. 55,Google Scholar emphasizes the ‘strong linguistic difference between Mark and Q’ in the overlaps, without mentioning the strong verbal agreements in some of the passages.

^{page 461 note 4} On Matthew as the Middle term, see also Butler, op. cit. ch. I, especially pp. 3, 5, and 10.

^{page 462 note 1} F. H. Woods, op. cit. pp. 61 f.

^{page 462 note 2} Streeter, , The Four Gospels, p. 162:Google Scholar Matthew and Luke ‘practically never agree together against Mark’.

^{page 462 note 3} Streeter, Ibid. p. 163: ‘the tendency to abbreviate is especially noticeable in Matthew’. See also Hawkins, , Horae Synopticae ², pp. 158 ff.Google Scholar For a discussion of Mathew as an abbreviator, see The Tendencies of the Synoptic Tradition, pp. 84 ff. Sparks, H. F. D., J.T.S. n.s. XXI (1970), 471–3,Google Scholar has correctly objected to my use of word totals in arguing the Matthew was not an abbreviator. The general points made against that view on p. 85 still seem to me persuasive, however.

^{page 463 note 1} One of the principal objections to the theory that Mark used Matthew is that he would not have left out so much excellent material. See Streeter, , The Four Gospels, p. 158;Google Scholar Kümmel, op. cit. p. 45. But if one supposes that Mark knew Q, one must admit that Mark omitted large bodies of excellent material. For a list of those who have argued for a literary relationship between Mark and Q, see Kümmel, p. 55.

^{page 463 note 2} ‘St. Marks's Knowledge and Use of Q’, p. 176.

^{page 464 note 1} See ‘The Argument from Order and the Relationship between Matthew and Luke’, 252 f.

^{page 465 note 1} It does not necessarily follow that Matthew composed his version by conflating Mark and Q lsquo;represented by Luke’. The history of the pericope could be much more complex than either of the two simple solutions: Matthew conflated Mark and Q; Mark and Luke independently copied alternate verses from Matthew. In this case, as in others, neither alternative seems to account adequately for the agreements and variations among the Gospels.