2 Yet one motive in publishing this paper is to call the attention especially of younger scholars to the extreme caution which is necessary in trying to draw deductions from text-critical statistics.
3 Wordsworth and White's text, 1905, has been used for collation.
4 For the different conclusion of Wordsworth and White, see below, p. 79. The point ought to be tested by someone operating with a larger body of material than the two chapters used in the present article.
5 See H. von Soden, Die Schriften des Neuen Testaments, vol. 1, pp. 877, 1662–1664, 1928; J. H. Ropes, The Text of Acts, pp. liii f., cclxviii.
6 The cases in this series of variations where the five Old Uncial witnesses present three different readings are so few as to be negligible, and they have in fact been neglected in summarizing the conclusions.
7 The Text of Acts, pp. cclxi f.; cf. pp. cclxxii-cclxxiv.
8 This is one of the many examples of studies that must be made before it is attempted to construct a fresh critical Greek text of the New Testament. That task involves more kinds of preliminary and incidental research than is sometimes imagined.
9 The Text of Acts, pp. cclvii f., cclxvi, cclxix.
10 That this fact was not brought out by the examination of ‘singular’ readings in “The Text of Acts” probably indicates that in small variants (including the use of the article), where the Latin makes no distinction, A has a larger proportion of aberrant readings than either B or 81. These small variants were included in the figures used in “The Text of Acts,” but not in the tabulations of the present article. If we consider all kinds of ‘singular’ readings in those portions of Acts where all five witnesses are extant, the figures are: B, 96; Codex 81, 101; A, 120; N, 158; C, 186.
11 Of the fifteen instances of agreement with H L P adduced by Wordsworth and White (‘Praefatio in Actus,’ p. xii, note 2, concluding paragraph) one (Acts 26, 28 fieri) is a case where the ‘Western’ reading is not known, while of the others all but four (16, 6 transeuntes; 19, 37 vestram; 20, 25 dei; 24, 21 in me) are readily explained as ‘Western’ survivals in the Vulgate and the Antiochian alike but independently; and as to the four exceptions just named (no one of which can be called an important or striking reading) there is in every instance something to be said against the necessity of assuming direct Antiochian influence. One of the most noteworthy of the fifteen (15, 18) is, in the light of all the facts, very distinctly a case where the reading is not due to an Antiochian Ms. As has been remarked above, a perfectly satisfactory conclusion would require an examination of the whole body of material.
12 The text used for collating was the edition of the British and Foreign Bible Society, 1920.
13 For a comparison of similar percentages for all four versions, including the Sahidic and Bohairie, see Table F below, p. 93.
14 Homer's text, Oxford, 1922, has been used for collation.
15 Homer's text, Oxford, 1905, has been used for the collation.