In the statements of evidence I have included the percentage of manuscript attestation for each variant within either ( ) or [ ]. I have used ( ) for the evidence taken from TuT, which I take to be reasonably precise. For the variant sets that are not covered there, I mainly depended on Legg (Novum Testamentum Graece Secundum Westcotto-Hortianum. Evangelium Secundum Marcum [Ed. S.C.E. Legg, Oxonii: E. Typographeo Clarendoniano, 1935]). I occasionally supplemented from Swanson, Scrivener and von Soden—the percentages offered, I have used [ ] for these, are extrapolations based on a comparison of these sources.
I venture to predict, if complete collations ever become available, that for any non-Byzantine variants listed with 5 to 1% support (in my apparatus) the margin of error should not exceed ±2%; for non-Byzantine variants listed with 10 to 6% support the margin of error should hardly exceed ±4%; where there is some division among the Byzantine witnesses the margin of error should rarely exceed ±15%—since my sources had collated a lower percentage of the extant MSS than ECM for the General Epistles, for example, my guesses as to percentages are more tentative than they were there, except that I guarantee the witness of f35. Please see the last footnote for Matthew for further information.
*16:1 μαρια f35 [88%] HF,RP,CP ¦ 1 η του A,B [2%] TR,OC[NU] ¦ 1 η ℵC,W [10%] ¦ D has a long omission
†16:1 ηγορασαν rell ¦ ηγωρασαν CP
‡16:1 τον ιησουν f35 [30%] CP ¦ αυτον ℵA,B,C(D)W [70%] HF,RP,TR,OC,NU
§16:2 της μιας f35 A,C [97%] HF,RP,CP,TR,OC ¦ 2 D ¦ 12 των [1%] ¦ τη μια των ℵ [1.5%] NU ¦ μια των B,W ¦ μια [0.5%]
*16:7 αλλ rell ¦ αλλα ℵA,B,C,D,W [2%] NU
†16:8 εξελθουσαι rell ¦ 1 ταχυ [5%] TR
‡16:8 δε f35 A,C [99.5%] HF,RP,CP,TR,OC ¦ γαρ ℵBD(W) [0.5%] NU
§16:8 have verses 9-20 rell [[NU]] ¦ — ℵB (For a full discussion of this variant, please see Appendix E in my, The Identity of the New Testament Text IV.)
*16:9 ο ιησους f35 [30%] CP ¦ — A,C,D,W [70%] HF,RP,TR,OC,NU (ℵB omit verses 9-20 and NU puts them in double brackets.)
†16:9 αφ f35 A [98%] HF,RP,CP,TR,OC ¦ παρ C,D,W [2%] NU
‡16:9 ης rell ¦ 1 και CP
§16:11 [This was either the third or the fourth appearance; the 2nd was to the women; Jesus also appeared to Peter, but the Text does not permit us to know for sure if it was before or after Emmaus.]
*16:14 υστερον f35 C,W [98%] HF,RP,CP,TR,OC ¦ 1 δε A,D [2%] [NU]
†16:14 εγηγερμενον rell ¦ 1 εκ νεκρων A,C (10.7%)
‡16:17 καιναις f35 A,W (97.8%) HF,RP,CP,TR,OC ¦ και εν ταις χερσιν 1 C (2.2%) [NU]
§16:18 οφεις αρουσιν f35 A,W (97.4%) HF,RP,CP,TR,OC ¦ και εν ταις χερσιν 12 C (2.2%) [NU] ¦ — (0.2%) ¦ omit verses 9-20 ℵcB
*16:18 βλαψη f35 A,C,W [95%] HF,RP,CP,NU ¦ βλαψει [5%] TR,OC
†16:19 κυριος f35 A (94.5%) HF,RP,CP,TR,OC ¦ 1 ιησους C(W) (5.2%) NU ¦ ιησους (0.2%) ¦ one other variant
‡16:20 αμην f35 C,W (97.9%) HF,RP,CP,TR,OC ¦ — A (2.1%) NU
§16:20 have verses 9-20 rell [[NU]] ¦ — ℵB (For a full discussion of this variant, please see Appendix E in my, The Identity of the New Testament Text IV.)
*16:20 In the colophons, f35 [50%] have, “published ten years after the ascension of Christ”. For 50% of the MSS to have this information probably means that the tradition is ancient; and of course, I have demonstrated, to my own satisfaction at least, that f35 goes back at least to the 3rd century. If this information is correct, then Mark was ‘published’ in 40/41 AD. The same sources have Matthew published two years earlier (38/39) and Luke five years later (45/46), while John was ‘published’ thirty-two years after the ascension, or 61/62 AD. Not only were the authors eyewitnesses of the events, but many others were still alive when the Gospels appeared. They could attest to the veracity of the accounts, but could also be the source of textual variants, adding tidbits here and there, or ‘correcting’ something that they remembered differently.
†16:20 The citation of f35 is based on the following fifty-three MSS—18, 35, 128, 141, 204, 510, 547, 553, 586, 645, 689, 789, 824, 928, 1023, 1040, 1072, 1075, 1111, 1117, 1133, 1145, 1147, 1199, 1251, 1339, 1384, 1435, 1461, 1496, 1503, 1572, 1628, 1637, 1652, 1667, 1705, 1713, 2122, 2221, 2253, 2261, 2323, 2352, 2382, 2466, 2503, 2554, 2765, 2875, 2876, Iviron 2110 and Leukosia 65 [the last two do not yet have a GA number, so far as I know]—all of which I collated myself. None of them is a ‘perfect’ representative of f35 in Mark, as it stands [an unreasonable expectation, presumably, for a book this size, besides being a Gospel]. But 586 is only off by one letter, and its exemplar, and that of 35 and 2382, probably were perfect! And several other exemplars come close—that of 1628 was off by one variant, those of 510 and 2253 were off by two variants, those of 824, 1435, 1503 and 1637 were off by three, several by four, and so on. [This refers to the MSS I have collated—there may be even better ones out there! In fact, since I have collated perhaps 20% of the family representatives for this book, there probably are better ones out there.] The uniformity is impressive. Since these MSS come from all over the Mediterranean world (Sinai, Jerusalem, Cyprus, Patmos, Constantinople, Bucharest, Aegean, Tirana, Mt. Athos [nine different monasteries], Corinth?, Athens, Grottaferrata, Vatican, Leukosia, Kalavryta, Meteora, Lesbos, Sparta, etc.) they are certainly representative of the family, giving us the precise family profile—it is reflected in the Text without exception.In the statements of evidence I have included the percentage of manuscript attestation for each variant within either ( ) or [ ]. I have used ( ) for the evidence taken from TuT, which I take to be reasonably precise. For the variant sets that are not covered there, I mainly depended on Legg (Novum Testamentum Graece Secundum Westcotto-Hortianum. Evangelium Secundum Marcum [Ed. S.C.E. Legg, Oxonii: E. Typographeo Clarendoniano, 1935]). I occasionally supplemented from Swanson, Scrivener and von Soden—the percentages offered, I have used [ ] for these, are extrapolations based on a comparison of these sources.I venture to predict, if complete collations ever become available, that for any non-Byzantine variants listed with 5 to 1% support (in my apparatus) the margin of error should not exceed ±2%; for non-Byzantine variants listed with 10 to 6% support the margin of error should hardly exceed ±4%; where there is some division among the Byzantine witnesses the margin of error should rarely exceed ±15%—since my sources had collated a lower percentage of the extant MSS than ECM for the General Epistles, for example, my guesses as to percentages are more tentative than they were there, except that I guarantee the witness of f35. Please see the last footnote for Matthew for further information.