¹ Published in three successive editions, each more complete than the previous: (i) Kenyon, F. G., Chester Beatty Biblical Papyri, Introduction (London: Emery Walker, 1933)Google Scholar and Kenyon, F. G., Chester Beatty Biblical Papyri Fasciculus III (Text), Pauline Epistles and Revelation (London: Emery Walker, 1934)Google Scholar; (ii) Sanders, H. A., A Third-Century Papyrus Codex of the Epistles of Paul (Ann Arbor: University of Michigan, 1935)Google Scholar; (iii) Kenyon, F. G., Chester Beatty Biblical Papyri, Fasciculus III Supplement, Pauline Epistles (London: Emery Walker, Text 1936; Plates 1937).Google Scholar

² Aland, B. and Aland, K., Novum Testamentum Graece (27th ed.; Stuttgart: Deutsche Bibel-gesellschaft, 1993) 684–9.Google Scholar The attempt of Kim to date 46 to the late first century (Kim, K. Y., ‘Palaeographical Dating of P46 to the Later First Century’, Biblica 69 [1988] 248–57Google Scholar) has met with little response among other palæographers, and so I will continue to assume a date of around 200 AD. I will refer to Hebrews as a letter of Paul since it appears to have been considered as such by the scribe of 46.

³ E.g., Easton, B. S., The Pastoral Epistles (London: SCM, 1948) 31Google Scholar; Houlden, J. L., The Pastoral Epistles (London: Penguin Books, 1976) 40Google Scholar; MacDonald, D. R., The Legend and the Apostle (Philadelphia: Westminster, 1983) 85Google Scholar; Young, F., The Theology of the Pastoral Letters (Cambridge: CUP, 1994) 142CrossRefGoogle Scholar; McDonald, L. M., The Formation of the Christian Biblical Canon (Peabody: Hendrickson, 1995) 196Google Scholar; Gamble, H. Y., Books and Readers in the Early Church (New Haven: Yale University, 1995) 59 n. 74Google Scholar; Davies, M., The Pastoral Epistles (Sheffield: Sheffield Academic, 1996) 106.Google Scholar

⁴ Dibelius, M. and Conzelmann, H., Die Pastoralbriefe (4th ed.; Tübingen: J. C. B. Mohr, 1966) 2Google Scholar; Köster, H., Einführung in das Neue Testament (Berlin: Walter de Gruyter, 1980) 736Google Scholar. There is room for confusion here: the extant 46 certainly does not contain the Pastorals (nor 2 Thessalonians and Philemon). However the use to which these authors put ‘ 46’ makes clear that they are using this term to refer to the original papyrus, not just the extant parts of it.

⁵ Objections to this assessment were expressed in a brief paragraph in Trobisch, D., Die Ent– stehung der Paulusbriefsammlung (Göttingen: Vandenhoeck &Ruprecht, 1989) 27–8CrossRefGoogle Scholar. This paper identifies many of the same objections as Trobisch did, though marshals them into a more detailed argument.

⁶ As becomes important later (below, p. 588) this only shows that the extant pages come from a single quire: it is pure conjecture to claim that the whole codex, including the pages we no longer have, comprised a single quire (as do Metzger, B. M., The Text of the New Testament [3rd ed.; Oxford: OUP, 1992] 37Google Scholar; Epp, E. J., ‘The Papyrus Manuscripts of the New Testament’, The Text of the New Testament in ontemporary Research [ed. Ehrman, B. D. and Holmes, M. W.; Grand Rapids: Eerdmans, 1995] 5Google Scholar; Aland, K., Repertorium der griechischen christlichen Papyri I: Biblische Papyri [Patristische Texte und Studien 18; Berlin: Walter de Gruyter, 1976] 275Google Scholar, and almost all other commentators). A clear description of the construction of a codex is given in Robinson, J. M., ‘On the Codicology of the Nag Hammadi Codices’, Les Textes de Nag Hammadi (ed. Ménard, J.; Leiden: Brill, 1975) 15–131Google Scholar and Turner, E. G., The Typology of the Early Codex (Philadelphia: University of Pennsylvania, 1977) 43–88.CrossRefGoogle Scholar

⁷ This figure and the following calculation is only approximate for three reasons, (i) Gaps occur in the manuscript around the titles of epistles, thus the inclusion of a text in the manuscript would require slightly more space than a letter-count implies, (ii) The number of letters in the epistles which are not present in the manuscript has been calculated from a modern critical text, adjusted to reflect the spelling used in 46. (iii) Although it is written in a basically square hand, certain letters do take up more space than others. The errors introduced by these three factors are extremely unlikely to be significant in the calculations.

⁸ Metzger, Text, 37–8; Epp, ‘Papyrus Manuscripts’, 5.

⁹ Tertullian, Adversus Marcionem, 5.21.

¹⁰ This date has recently been strongly challenged by Hahneman, G. M., The Muratorian Fragment and the Development of the Canon (Oxford: Clarendon, 1992)CrossRefGoogle Scholar, but still has many defenders, e.g. E. Ferguson′s review of Hahneman's book in JTS 44 (1993) 691–7.Google Scholar

¹¹ Mentioned as a possibility in Jeremias, J., Die Briefe an Timotheus und Titus (Güttingen: Vandenhoeck & Ruprecht, 1963) 4Google Scholar; argued for by Quinn, J. D., ‘P46 – The Pauline Canon’, CBQ 36 (1974) 379–85Google Scholar and favoured in Hanson, A. T., The Pastoral Epistles (London: Marshall, Morgan and Scott, 1982) 12.Google Scholar

¹² The report that Tatian living just before the composition of 46 accepted Titus but rejected 1 Timothy and 2 Timothy (Jerome, Commentary on Titus, preface) should caution us against imputing to those responsible for producing 46 the modern assumption that 1 Timothy, 2 Timothy and Titus should be seen as a group.

¹³ E.g. Kenyon, Chester Beatty III Supplement (Text), ix, and Aland, Repertorium, 274.

¹⁴ Two factors in the way I calculated these numbers need to be noted, though neither is likely to have produced significant inaccuracies, (i) Where a title occurs on a page, spaces are left. In these cases the mean of the previous and the following page were taken to represent how many letters there would have been on that page, had there not been a title, (ii) Counting the number of letters on every page of 46 was impossible since the top lines of many pages are damaged. Hence the first word on the fifth line (which is always extant) on each page was noted. The number of letters between these words was then counted in a computerised version of Nestle-Aland's 27th edition of the NT, adjusted to reflect the spelling used in 46 and the major textual variation at the end of Romans). Hence when I refer to the number of letters on say, page 40 of 46, I am actually referring to the number of letters in an adjusted modern text between the first letter of the first word of the fifth line of page 40 of 46 and the last letter of the last word of the fourth line of page 41 of 46, inclusive. The fragmentary nature of the codex prevents the number of letters per page being calculated for any page after page 184 except for 192.

¹⁵ This line depicts the general trend once the variations between individual pages are smoothed out. It is a sixth-order polynomial approximation: its r-squared value (0.91) indicates that is a good fit.

¹⁶ Turner, Typology, 58.

¹⁷ At this point Quinn 's thesis (that Philemon was also missing: see p. 582 above) is even less satisfactory since it implies an even greater number of blank pages.

¹⁸ Turner, , Typology, 73–4.Google Scholar

¹⁹ Kenyon, Chester Beatty HI Supplement (Text), xi.

²⁰ Kenyon, F. G., Chester Beatty Biblical Papyri, Fasciculus VI (Text), Isaiah, Jeremiah, Ecclesiasticus (London: Emery Walker, 1937) v–vi.Google Scholar

²¹ Ziegler, J., Septuaginta: Isaias (Vol 14) (Güttingen: Vandenhoeck & Ruprecht, 1939).Google Scholar

²² Martin, V. and Kasser, R., Papyrus Bodmer XIV-XV: Evangiles de Luc et Jean (Cologny-Genève: Bibliothèque Bodmer, 1961) 9–13.Google Scholar

²³ Such a false deduction is often implied, and is even stated clearly in Köster, , Einfiihrung, 451Google Scholar, ‘die Pastoralbriefe fehlen völlig [from 46] … waren also hier noch nicht zum Bestandteil des Corpus Paulinus geworden’.

²⁴ Furthermore, the corrections and stichometry found in the codex seem to point to some use being made of the codex. Unfortunately nobody has, as of yet, been able to identify any correlation between the stichometry and the number of lines used in the papyrus for each epistle which might help determine the codex's contents. Trobisch, Entstehung, 27 n. 65, suggests that the gross inaccuracies of the stichometry might help explain how the scribe miscalculated the space required in the codex.

²⁵ The possibility that more pages were added was mentioned in Sanders, Codex, 10 and reiterated in Guthrie, D., The Pastoral Epistles (London: Tyndale, 1957) 13–14Google Scholar; Jeremias, Briefe, 4; Kelly, J. N. D., A Commentary on the Pastoral Epistles (London: A. & C. Black, 1963) 4Google Scholar and Giversen, S., ‘The Pauline Epistles on Papyrus’, Die paulinische Literatur und Theologie (ed. Pedersen, S.; Göttingen: Vandenhoeck & Ruprecht, 1980) 211Google Scholar. However, generally this possibility has been ignored.

²⁶ Codices normally consist of sheets – that is pieces of papyri folded in two whose middle is fixed in the centre of the codex to give two leaves; one leaf in the first half of the codex, the other in the second. A single leaf is a piece of papyrus half the size of a sheet, one edge of which is fixed in the centre of the codex giving a leaf in just one half of the codex.

²⁷ Details in Robinson, J. M., The Facsimile Edition of the Nag Hammadi Codices, Introduction (Leiden: Brill, 1984) 41–1, 52–3.Google Scholar

²⁸ Robinson, , Introduction, 33.Google Scholar

²⁹ Turner, , Typology, 64.Google Scholar

³⁰ See note 6 above.

³¹ Unfortunately the bindings of the codex, which might have revealed how many quireswere in the codex, were never recovered.

³² Guéraud, O. and Nautin, P., Origène Sur la Pâque (Paris: Beauchesne, 1979) 23.Google Scholar

³³ Robinson, , Introduction, 40.Google Scholar

³⁴ As Grant, R. M., A Historical Introduction to the New Testament (London: Collins, 1963) 209–11Google Scholar does after a careful analysis.

³⁵ A selection are given in note 3 above.

³⁶ Goodspeed, E. J., Christianity Goes to Press (New York: Macmillan, 1940) 86Google Scholar. Goodspeed evidently ignored the use of the Pastorals by Clement of Alexandria.

³⁷ Gamble, , Books, 59 n. 74.Google Scholar