APPENDIX A
CALCULATION OF MANUSCRIPT PERCENTAGES
The calculation of manuscript percentages is not as straightforward as one might assume. There are four primary factors that must be taken into consideration: (1) the length of the variant unit, (2) whether corrections and additions by later scribes are counted, (3) how misspellings and alternate spellings are handled, and (4) how long omissions are handled. Because the manuscript percentages in this volume are calculated from the collations presented in the Text und Textwert volumes, those volumes guide how these four factors are handled in the present volume. The length of the variant unit is simply the length as presented in Text und Textwert.*It should be noted that the length of any given variant unit in Text und Textwert may sometimes be longer than what is cited in the footnotes of The Text-Critical Greek New Testament. For example, the first variant in Mark 2:16 reads as follows:

και οι φαρισαιοι 96.6% ¦ των φαρισαιων CT 0.2%

However, the length of the variant unit in Text und Textwert is actually as follows:

αυτω και οι γραμματεις και οι φαρισαιοι ¦ αυτω και οι γραμματεις των φαρισαιων

This difference is due to the fact that some manuscripts have further variations that are not included in the main text of any of the editions of the Greek New Testament compared in the footnotes presented in this volume. Nevertheless, the percentages listed in the footnotes correspond to the variant units as they are presented in Text und Textwert because it provides a truer picture of the manuscript evidence as a whole.

The same sort of rule applies for Wasserman's collation of Jude, Solomon's collation of Philemon, and Morrill's collation of John 18. For John 7:53–8:11 and the Revelation variant units calculated from the Editio Critica Maior and Hoskier, the length of the variant unit generally matches the length of that particular variant unit as listed in the footnotes of this edition.
Corrections and additions by later scribes are counted in addition to the original reading of a manuscript. Thus, a single manuscript can be counted more than once.Percentages calculated from Tommy Wasserman's collation of Jude are based on Joey McCollum's tabulations, which take into account only the original text for each manuscript. Percentages calculated from Maurice Robinson's collation of John 7:53–8:11 take into account only the original text for each manuscript. In the same way, percentages calculated from the combined collations of the Editio Critica Maior and Hoskier take into account only the original text for each continuous text manuscript (ignoring lectionary manuscripts). The effect of this different calculation method on the overall percentages is extremely minimal. Apart from this difference, the percentages are calculated in much the same way as the percentages based on the Text und Textwert volumes. Consequently, it is more accurate to speak of the percentage of manuscript readings than the percentage of manuscripts. When Text und Textwert groups alternate spellings under one variant, they are counted as one variant in the calculation of manuscript percentages in this volume. Similarly, when Text und Textwert groups alternate spellings under separate variants, they are counted as separate variants in the calculation of manuscript percentages in this volume.In Revelation minor spelling differences are usually not counted as separate variants. This is in accordance with the presentation of the data in the Revelation volume of the Editio Critica Maior. In Revelation, the difference between numbers represented by words (for example, τεταρτου) and numbers represented by letters (for example, Δ) is usually ignored unless it is the primary difference under consideration for the variant unit. Finally, manuscripts that have long omissions due to factors such as homoioteleuton are included in the total number when calculating percentages.§The one exception is John 7:53–8:11. For this passage only the manuscripts that contain the passage are considered in calculating percentages.
The process of calculation is best illustrated by example. Below is a summary of the Text und Textwert collation for 2 John 9. This variant unit occurs after the words ο μενων εν τη διδαχη (‘whoever abides in the teaching’). The readings are as follows:
ReadingTextSubtotal
1του χριστου (‘of Christ’)458
223
3του θεου (‘of God’)1
4του κυριου (‘of the Lord’)1
U1long omission (homoioteleuton)4
U2long omission (homoioteleuton)25
Vlong omission (other)1
Xillegible3
Yfilm error1
Zlacuna52
Although the grand total is 569, the manuscripts labeled as X, Y, and Z are not included in the calculations. This reduces the total to 513. Thus the percentages are as follows:
Reading Percentage
189.3% (458/513)
24.5% (23/513)
30.2% (1/513)
40.2% (1/513)
U10.8% (4/513)
U24.9% (25/513)
V0.2% (1/513)
In summary, 89.3% of the manuscript readings support the Byzantine text (Reading 1), while 4.5% of the manuscript readings support the critical text (Reading 2).
1 Corinthians 2:1 provides another helpful example. This variant unit occurs after the words καταγγελλων υμιν (‘proclaiming to you’).
ReadingTextSubtotal
1το μαρτυριον του θεου (‘the testimony of God’)556
1Bτο μαρτυριον θεου (‘the testimony of God’)2
2το μυστηριον του θεου (‘the mystery of God’) 26
3το μαρτυριον του θεου {υμιν} (‘the testimony of God’ but placed before the words ‘to you’)1
4το μαρτυριον του ιησου (‘the testimony of Jesus’)1
5το μαρτυριον του χριστου (‘the testimony of Christ’)3
6το σωτηριον του θεου (‘the salvation of God’)1
7το ευαγγελιον του θεου (‘the good news of God’)6
Ulong omission (homoioteleuton)1
Wuncertain as to whether it is Reading 1, Reading 2, or Reading 61
X illegible 3
Zlacuna145
Although the grand total is 746, the manuscripts labeled as W, X, and Z are not included in the calculations. This reduces the total to 597. Thus the percentages are as follows:
Reading Percentage
193.1% (556/597)
1B0.3% (2/597)
24.4% (26/597)
30.2% (1/597)
40.2% (1/597)
50.5% (3/597)
60.2% (1/597)
71% (6/597)
U0.2% (1/597)
In summary, 93.1% of the manuscript readings support the Byzantine text (Reading 1), while 4.4% of the manuscript readings support Nestle-Aland and Westcott-Hort (Reading 2). (In this case SBL and Tyndale House agree with the Byzantine text.) It should be noted that although Reading 1B and Reading 3 have the same English translation as Reading 1, it is the percentage for Reading 1 that is given in the footnote at 1 Corinthians 2:1. This is because the Greek text of Robinson and Pierpont matches Reading 1 and not Reading 1B or Reading 3.
As a rule, manuscripts labeled as U or V are included in the calculations, while manuscripts labeled as W, X, Y, or Z are not. Readings labeled as 1B, 1C, 1D, 2B, 2C, 2D, etc., are always treated separately.*Two exceptions are 1 Cor. 14:24 and 2 Cor. 11:3 in which the percentage for Reading 1B is combined with the percentage for Reading 1 since the only difference is the inclusion or exclusion of sigma for the word ουτω(ς). Other exceptions are Mark 16:9–20, John 5:4, John 7:53–8:11, Romans 14:24–26, and Romans 16:25–27 in which the totals from different readings are combined because the differences in the readings are based upon the presence or absence of asterisks or obeli in the margin and other similar notations or distinctions.
The Text und Textwert collations for the Gospel of John present manuscript totals in a peculiar way that requires additional explanation. While some readings are labeled 1B, 1C, 1D, 2B, 2C, 2D, etc., other readings are labeled 1-f or 1-f1, 1-f2, 1-f3, etc. Readings with the -f label are not presented with a subtotal. Usually (but not always) they are included in the prior subtotal. For example, in John 4:37 the first three readings are listed as follows:
ReadingTextSubtotal
1ο αληθινος1,333
1-fο αληθοςno subtotal
1Cο αληθης1
In this case, the actual subtotals are as follows:
ReadingTextSubtotal
1ο αληθινος1,332
1-fο αληθος1
1Cο αληθης1
In order to be consistent with the calculation methods used with the other Text und Textwert volumes, the subtotal of 1,332 is what would be used to calculate the percentage of manuscripts for Reading 1.
Nevertheless, there are times when the -f label is not included in the prior subtotal. Reading 13 and Reading 14-f in John 7:40(1) provide a good example. Below is an exact representation of the Text und Textwert data.
13 OM. των λογων τουτων
030* 106 164 494 516* 1349* 1356 1474 2649*
ANZAHL DER ZEUGEN:That is, “NUMBER OF WITNESSES.” 9
14-f των οχλων λογων
335
Reading 13 is found in nine manuscripts, which are each listed. The subtotal is 9 and cannot therefore include Reading 14-f. Reading 14-f, which is supported only by manuscript 335, must be given a separate subtotal of 1 even though no subtotal is listed. The rule of thumb is that if the label preceding -f is the same as the prior label, it is included in the prior subtotal, but if the label preceding -f is not the same as the prior label, it is not included in the prior subtotal. Thus in John 4:37 Reading 1-f is included in the subtotal for Reading 1 because the label preceding -f is the same as the prior label. However, in John 7:40(1) Reading 14-f is not included in the subtotal for Reading 13 because the label preceding -f is not the same as the prior label. While -f is the most common label that is not presented with a subtotal, the same rule applies for any readings that have a hyphen followed by a lowercase letter or lowercase letters (such as -o and -of).
By following the model for calculation described above, anyone with access to the Text und Textwert volumes should be able to replicate the manuscript percentages found in the footnotes of this volume.It should be noted, however, that the collations for five variant sets (Mark 2:14; 12:25; Acts 2:7; 18:21; 27:5) are unreliable, with many manuscripts classified incorrectly.
Additional Percentages
The Text und Textwert volumes present collations for a total of 1,043 variant units. Of those 1,043 variant units, 872 are listed in the footnotes of this volume and 5 are ignored because the collation data is incorrect. For the remaining 166 variant units, all the editions of the Greek New Testament compared in this volume are in agreement, and so it is not feasible to present them in the footnotes. Instead, the percentage for each variant unit is presented below. The mean of these percentages is 94%, and the median is 96.4%.
Matt. 1:1186.9%
Matt. 1:1697.5%
Matt. 1:25(1) 99.5%
Matt. 3:1275.6%
Matt. 5:4–598.3%
Matt. 7:2199.1%
Matt. 12:297.0%
Matt. 14:398.8%
Mark 1:699.6%
Mark 1:797.4%
Mark 1:1198.2%
Mark 1:1579.3%
Mark 1:25(1) 99.5%
Mark 1:25(2) 88.5%
Mark 1:3570.3%
Mark 1:4599.2%
Mark 2:26(1) 99.8%
Mark 3:599.1%
Mark 3:1899.2%
Mark 3:21(1) 96.3%
Mark 3:21(2) 94.7%
Mark 4:698.4%
Mark 5:21(1) 95.3%
Mark 5:21(2) 96.6%
Mark 5:2298.1%
Mark 5:3398.3%
Mark 6:2(1) 98.8%
Mark 6:14(2) 91.9%
Mark 6:4497.6%
Mark 6:4798.7%
Mark 6:5094.2%
Mark 7:1399.0%
Mark 7:3391.5%
Mark 7:3599.0%
Mark 8:10(1) 97.9%
Mark 8:10(2) 96.0%
Mark 8:15(1) 94.5%
Mark 8:15(2) 96.7%
Mark 8:35(2) 98.8%
Mark 9:2598.9%
Mark 9:3492.1%
Mark 9:3576.1%
Mark 10:19(2) 84.1%
Mark 10:19(3) 87.7%
Mark 10:43(2) 93.8%
Mark 12:14(2) 94.5%
Mark 12:4188.5%
Mark 13:2(2) 97.9%
Mark 14:2596.3%
Mark 15:1296.6%
Mark 16:1–2099.8%
Mark 16:14(2) 99.8%
Luke 19:2593.8%
Luke 22:43–4498.7%
Luke 23:3489.3%
John 1:499.7%
John 1:27(3) 90.5%
John 1:3697.7%
John 2:15(1) 82.3%
John 3:5(2) 98.9%
John 3:2082.1%
John 3:27(2) 97.7%
John 3:3196.8%
John 3:3296.8%
John 4:1(2) 94.9%
John 4:986.6%
John 4:11(2) 97.2%
John 4:2391.2%
John 4:2494.2%
John 4:5390.2%
John 5:798.8%
John 5:2593.1%
John 5:3992.2%
John 5:44(2) 97.7%
John 6:190.4%
John 6:499.8%
John 6:1083.3%
John 6:1183.4%
John 6:14(2) 98.1%
John 6:2193.4%
John 6:23(2) 91.5%
John 6:42(1) 99.5%
John 6:4696.3%
John 6:5284.2%
John 6:5699.8%
John 6:6492.0%
John 7:497.0%
John 7:47(1) 94.8%
John 7:47(2) 97.4%
John 8:1497.1%
John 8:16(2) 99.2%
John 8:25(1) 97.3%
John 8:25(2) 98.1%
John 8:3499.0%
John 8:38(3) 97.4%
John 8:38(6) 98.8%
John 8:4697.2%
John 8:5798.5%
John 9:4(2) 99.3%
John 9:1896.4%
John 9:3395.4%
John 9:38–39a99.8%
John 10:1599.1%
John 10:16(1) 61.0%
John 10:34(1) 99.0%
John 10:34(2) 98.7%
John 10:3693.6%
John 10:40(1)89.4%
John 10:40(2)93.2%
John 10:4187.5%
Acts 2:43–4487.3%
Acts 2:4678.0%
Acts 5:2196.3%
Acts 8:3992.6%
Acts 12:398.5%
Acts 13:33(2) 99.8%
Acts 15:294.8%
Acts 16:3595.4%
Acts 18:2776.1%
Acts 20:3298.5%
Acts 23:25(1) 98.6%
Acts 23:25(2) 97.9%
Rom. 3:1299.0%
Rom. 5:295.1%
Rom. 5:890.7%
Rom. 8:23(2) 97.7%
Rom. 13:191.5%
Rom. 14:1997.2%
Rom. 15:1390.8%
Rom. 16:399.2%
Rom. 16:4–594.9%
Rom. 16:798.3%
Rom. 16:2199.5%
Rom. 16:2795.3%
1 Cor. 1:8(1) 95.2%
1 Cor. 1:8(2) 95.7%
1 Cor. 5:697.5%
1 Cor. 9:1698.8%
1 Cor. 10:1796.4%
1 Cor. 14:34–3595.0%
1 Cor. 15:596.3%
1 Cor. 15:1094.6%
1 Cor. 15:1593.8%
1 Cor. 16:1999.2%
2 Cor. 2:389.4%
2 Cor. 4:1796.9%
2 Cor. 12:792.1%
Gal. 1:1589.7%
Gal. 3:1497.5%
Gal. 4:1785.0%
Eph. 1:196.0%
Eph. 6:1999.3%
Phil. 1:397.2%
Phil. 1:11(2) 96.7%
Phil. 3:1298.8%
Col. 3:2178.5%
2 Thess. 3:479.5%
Heb. 7:196.7%
Heb. 13:21(2) 96.3%
James 1:1794.5%
James 4:991.9%
1 John 2:1093.7%
1 John 5:675.4%
Jude 1:193.3%
Rev. 13:271.7%
Rev. 18:2284.1%

*^ It should be noted that the length of any given variant unit in Text und Textwert may sometimes be longer than what is cited in the footnotes of The Text-Critical Greek New Testament. For example, the first variant in Mark 2:16 reads as follows:και οι φαρισαιοι 96.6% ¦ των φαρισαιων CT 0.2%However, the length of the variant unit in Text und Textwert is actually as follows: αυτω και οι γραμματεις και οι φαρισαιοι ¦ αυτω και οι γραμματεις των φαρισαιωνThis difference is due to the fact that some manuscripts have further variations that are not included in the main text of any of the editions of the Greek New Testament compared in the footnotes presented in this volume. Nevertheless, the percentages listed in the footnotes correspond to the variant units as they are presented in Text und Textwert because it provides a truer picture of the manuscript evidence as a whole. The same sort of rule applies for Wasserman's collation of Jude, Solomon's collation of Philemon, and Morrill's collation of John 18. For John 7:53–8:11 and the Revelation variant units calculated from the Editio Critica Maior and Hoskier, the length of the variant unit generally matches the length of that particular variant unit as listed in the footnotes of this edition.

^ Percentages calculated from Tommy Wasserman's collation of Jude are based on Joey McCollum's tabulations, which take into account only the original text for each manuscript. Percentages calculated from Maurice Robinson's collation of John 7:53–8:11 take into account only the original text for each manuscript. In the same way, percentages calculated from the combined collations of the Editio Critica Maior and Hoskier take into account only the original text for each continuous text manuscript (ignoring lectionary manuscripts). The effect of this different calculation method on the overall percentages is extremely minimal. Apart from this difference, the percentages are calculated in much the same way as the percentages based on the Text und Textwert volumes.

^ In Revelation minor spelling differences are usually not counted as separate variants. This is in accordance with the presentation of the data in the Revelation volume of the Editio Critica Maior. In Revelation, the difference between numbers represented by words (for example, τεταρτου) and numbers represented by letters (for example, Δ) is usually ignored unless it is the primary difference under consideration for the variant unit.

§^ The one exception is John 7:53–8:11. For this passage only the manuscripts that contain the passage are considered in calculating percentages.

*^ Two exceptions are 1 Cor. 14:24 and 2 Cor. 11:3 in which the percentage for Reading 1B is combined with the percentage for Reading 1 since the only difference is the inclusion or exclusion of sigma for the word ουτω(ς). Other exceptions are Mark 16:9–20, John 5:4, John 7:53–8:11, Romans 14:24–26, and Romans 16:25–27 in which the totals from different readings are combined because the differences in the readings are based upon the presence or absence of asterisks or obeli in the margin and other similar notations or distinctions.

^ That is, “NUMBER OF WITNESSES.”

^ It should be noted, however, that the collations for five variant sets (Mark 2:14; 12:25; Acts 2:7; 18:21; 27:5) are unreliable, with many manuscripts classified incorrectly.