The third edition of Stephanus' Greek New Testament (ΤΗC ΚΑΙΝΗC ΔΙΑΘΗΚΗC ΑΠΑΝΤΑ: Paris, 1550), known as the editio regia, is held in high regard in English Protestantism. It was this text which underlay the English translation (by W. Whittingham and others) published in Geneva in 1557 that greatly influenced the Geneva Bible published three years later. In effect, Stephanus' edition was the Textus Receptus of the Greek New Testament for over three hundred years.

Stephanus' 1550 New Testament was a handsome folio edition (8⅝ × 13 inches) and was the first to print variant readings consistently. (Later editions of Erasmus' text had recorded a few variants in the margin.) Stephanus cites readings from fifteen manuscripts.Footnote ¹ They are identified in the inner margins with a Greek numeral (from β´ to ις´).Footnote ²

Our normal port of call when needing to translate earlier systems of denoting manuscripts (by numbers, letters or other sigla) into the currently agreed classification known as the Gregory–Aland system is the appendix ‘Frühere Zeichnen und Nummern’ (in C. R. Gregory, Die griechischen Handschriften des Neuen Testaments [Leipzig: Hinrichs, 1908] 172–259; hereafter Gregory 1908), where his newly devised registration was first published. Unfortunately, there are errors in his translation of Stephanus' sigla.Footnote ³ (The Sigelkonkordanzen found in Kurt Aland, Kurzgefaßte Liste der griechischen Handschriften des Neuen Testaments [ANTF 1; Berlin and New York: de Gruyter, 2nd ed. 1994], do not refer to or convert the distinctive numbering systems found earlier than Tischendorf.)

The following is a table giving the cross-references from Stephanus' number to the equivalent Gregory-Aland number.Footnote ⁴ Column 1 is Stephanus' Greek numeral, column 2 the contents (using the conventions e = gospels, a = Acts and/or the Catholic Epistles, p = Pauline corpus, including Hebrews, r = Revelation), column 3 contains the generally agreed identifications with the normal number in use today (and, where relevant, the former, Tischendorf, numbers in brackets), the final column specifies the library and its call-number reference.

Table 1.

It will be seen that all but two of the identifiable manuscripts are in Paris. Minuscule manuscript 398, now in Cambridge, seems to have once belonged to Stephanus' friend Vatablus in Paris, so an earlier link for this manuscript with Paris is explicable. In Stephanus' day most of them had been in the Royal Library there; some were in private hands, although they too were deposited later in the National Library. It was natural for Stephanus, as the king's printer, to celebrate Garamond's large Greek-type with an edition of the most important Greek text and to use for it manuscripts predominantly from the Royal collection. Two manuscripts are unidentified.

Stephanus claims in his prefatory Epistle to Readers that he had had access to the readings of D 05 through collations made by ‘friends in Italy’. Our knowledge of the whereabouts of D 05 prior to its having being deposited by Beza in Cambridge University Library is usually based on Beza's statement in his letter of December 6th, 1581 to the Senate of Cambridge University accompanying the gift of the manuscript to the University, namely, that he acquired it from the Monastery of S. Irenaeus in Lyons; his supplementary note prefixed to the manuscript itself implies it had come as booty during the sack of Lyons in 1562. Scrivener in the introduction to his edition of the manuscriptFootnote ⁷ (p. viii) indicates that the manuscript may well have been in Trent in 1546 and was used to demonstrate Greek support for the Vulgate reading sic eum volo at John 21.22: D alone among Greek witnesses adds ουτως before εως. Perhaps Stephanus did indeed receive collations of D 05 made by friends in Italy who had worked on it there on his behalf.

The errors in Gregory 1908 include the following:

1. (a) ι´ is said on page 174 to be 567ap (= Paris, Bibl. Nat., Gk 103A).Footnote ⁸ Scrivener makes the same identification.Footnote ⁹ But see Gregory 1908 Nachtrag p. 364 where ι´ = 2298 (= Paris, Bibl. natl. de France gr 102);

2. (b) ιε´ (a manuscript containing a p r) is said on page 175 to be ‘10’ (i.e. 10e). 10e = Paris, Bibl. natl. de France gr 91. (What of course was intended was the old 10a [plus 12p and 2r] that became 82 in Gregory 1908.)

3. (c) There is a problem with the manuscript numbered ζ´. As 8e (originally Paris Reg. 49) there is no problem in the Gospels. According to Scrivener, Vansittart's investigations had established that identification.Footnote ¹⁰ Gregory 1894 page 623 says that 50a–8p, a manuscript now lost, was cited as ζ´ by Stephanus in Acts 17.5, seven times in the Catholic Epistles and 27 times in Paul, an identification confirmed by Mill.Footnote ¹¹ Gregory 1894 therefore assumes that ζ´ was used to refer to two different manuscripts: 8 for e and the now lost 50a–8p for a and p. (Gregory 1908 refers to ζ´ only as having formerly been 50a and he ignores 8p on pages 174, 186.) But could Stephanus have used Paris Bibl. nat. de France gr 47 (formerly Reg. 47 and now Gregory–Aland 18 eapr) for sections a and p? (Scrivener says such an identification had been made, without saying by whom.Footnote ¹²) More work on Stephanus' use of ζ´ seems therefore to be warranted.

Stephanus himself as a pioneer is, unsurprisingly perhaps, not always very accurate. ε´ is cited at Rev 19.14; η´ at Acts 24.7 (bis); 25.14; 27.1; 28.11; ι´ at Luke 5.19; John 2.17; ια´ at Matt 10.8, 10; 12.32; John 2.17; Rev 13.4; ιγ´ at Matt 26.64; John 2:17 (yet again!); ις´ is cited at Luke 22.30, 67; 2 Cor 12.11; 1 Tim 3.3. Scrivener, A Plain Introduction, 1.190 is particularly scathing about Stephanus' misquoting of the Polyglot (cf. his general assessment in A Plain Introduction. 2.197 that Stephanus was negligent and capricious). However, Scrivener in his introduction to the edition of Bezae p. ix assesses that Stephanus was more accurate in his citing of Bezae (the readings from which were second-hand from his friends) than his references to the Complutensian Polyglot which he had actually consulted. Scrivener claims that 309 out of 389 references to D are correctly reported (including the extensive readings at, e.g., John 6.56; Acts 5.15; 6.10; 16.35, 38, 39, as well as the famous apocryphal agraphon about the man working on the Sabbath signalled after Luke 6.4) and in addition forty other places where D is merely ‘loosely’ reported. Wettstein had already analysed many of the wrongly cited references in 1730.Footnote ¹³

Nonetheless, modern readers may find it useful to examine (with care) Stephanus' apparatus for access to and information on the readings of the now lost manuscripts cited (8a–10p, 3r, and, perhaps 50a–8p). In particular ις´ is worth investigating: it is cited some 77 times mainly up to Rev 17.8. There are comparatively fewer manuscripts of Revelation than other parts of the New Testament, so this is worth adding to the evidence. Also interesting to explore is ια´, found some 400 times with, perhaps significantly, 276 of them in support of Latin witnesses.

In addition, not all extant manuscripts used by Stephanus are to be found regularly cited in the apparatus to modern editions (e.g. 8, 9, 120), so access to some of their readings, not readily or regularly available elsewhere, may be found in an edition printed four hundred and fifty years ago. One is ζ´ in the Gospels (minuscule 8). As far as I can see, no modern edition cites 8 as a separate witness. Stephanus gives it on its own at Mark 6.20; 14.15; Luke 1.37 (and elsewhere alongside other manuscripts). Another rarity is minuscule 9, seen in Vogels' edition but not in Nestle–Aland as a separate witness; so to obtain a flavour of its distinctive readings Stephanus' apparatus may be worth combing. θ' is not used in Nestle although it is found in the sparse apparatus in the UBS editions and in Merk and Souter. Yet it is cited some 38 times in Stephanus (e.g. Matt 26.45; Luke 8.18; 19.26; and note its support for αγγɛλια at 1 John 1.5).

Access to Stephanus' variants may be had from Beza. Beza's 1565 New Testament incorporated the Greek text of Stephanus' fourth edition of 1551 albeit with changes; he did not repeat Stephanus' apparatus. Beza's annotations do, however, refer to variants and to manuscripts, identified by numbers corresponding to Stephanus' Greek sigla. Almost all of the references are also to be found in the notes to his earlier, Latin, edition of 1556. It is interesting to see that Beza had access not only to Stephanus' printed text but to the collations themselves and he drew on some ignored by Stephanus himself.Footnote ¹⁴ This source is referred to in his Preface of 1556. Those collations (excluding Codex Bezae) had been assembled by Robert Stephanus' son, Henri, by 1546. Although his second edition (1565) may give the impression that he had actually seen the manuscripts themselves,Footnote ¹⁵ he had not in fact consulted the manuscripts. It is also interesting to observe that Beza had access to readings in D 05 through Stephanus prior to his acquisition of the manuscript itself.

Beza did not always correctly repeat Stephanus' variants; sometimes discrepancies between Stephanus and Beza may be explained as Beza's corrections of Stephanus, or are due to his making his own errors or to his following Stephanus' collations rather than the different reading found in Stephanus' printed edition. As those collations are no longer extant we cannot verify such explanations for the differences. Beza is particularly inconsistent in citing D! Sometimes he reports the reading as Stephanus' β´, sometimes from D 05 itself (which he had to hand, alongside D 06), sometimes from both without always apparently recognising that these were one and the same witness.

Despite Beza's claim that he had access to up to 25 manuscripts it is improbable that he knew more than the ones cited by Stephanus (apart from the addition of D 06)—the higher figures in his prefaces (25 in the first, second and fourth editions; 17 in the third; 19 in the fifth!) may either be mere exaggeration or typographical slips.Footnote ¹⁶