xix. 36, and because there is nothing like it in the narratives of the

three earlier Evangelists. If we decide to retain κλώμενον, it must be in opposition to the four chief manuscripts אABC, though אC insert it by the third hand of each. Cod. D, like its namesake of the Gospels and Acts, is somewhat inclined to paraphrases, and has θρυπτόμενον(420) by the first hand, κλώμενον by the second. Only two cursives here side with the great uncials (17, and the valuable second hand of 67), as do Zohrab’s Armenian, Cyril of Alexandria and Fulgentius in the fifth century, and Theodoret’s report of Athanasius. The word κλώμενον is found in EFGKLP, all other cursives, the Latin versions of DE (_quod frangitur_), with Ambrosiaster: G and the interlinear Latin of F, which, as has been already shown under that MS., is taken from G, prefer _quod frangetur_, with both Syriac, the Gothic, and the Armenian of Uscan. The Latin Vulgate has _tradetur_ (but _traditur_ in _harl._2, even in the parallel column of F and against its Greek, and so Cyprian); the Bohairic renders _traditur_; but the Sahidic and Ethiopic _datur_, after the διδόμενον of Zacagni’s Euthalius, derived from Luke xxii. 19. Theodoret himself knew of both forms. The main strength of κλώμενον rests on Patristic evidence. Mr. Forster has added to our previous store the “conclusive testimony” of Basil (Forster, p. xxvi) and of Athanasius himself (_ibid._ p. xvii), which is better than Theodoret’s report at second hand; and thus too Chrysostom in three places, one manuscript of Euthalius, John Damascene, the Patriarch Germanus (A.D. 715, _ibid._ p. xix), Œcumenius and Theophylact. Mr. Forster is perfectly justified also in pressing the evidence of the Primitive Liturgies, in all of which κλώμενον occurs in the most sacred words of Institution (_ibid._ pp. xx, xxi). Whatsoever change these services have received in the course of ages, they have probably been little altered since the fourth century, and very well established must the word have then been to have found a place in them all. On the whole, therefore, we submit this important text as a proof that the united readings of אABC are sometimes at variance, not only with the more modern codices united, but with the text of the oldest versions and most illustrious Fathers. We confess, however, that in ver. 29 ἀναξίως (compare ver. 27) and τοῦ _κύ_ look too much like glosses to be maintained confidently against the evidence of א*ABC*, 17, (67**) and some manuscripts of the Ethiopic.

35. 1 COR. xiii. 3. ἐὰν παραδῶ τὸ σῶμά μου ἵνα καυθήσωμαι, “though I give my body to be burned.” Here we find the undoubtedly false reading καυχήσωμαι in the three chief codices אAB and in 17, adopted by Drs. Westcott and Hort(421), and it is said to have been favoured by Lachmann in 1831, by Tregelles in 1873 (A. W. Tyler, Bibl. Sacra, 1873, p. 502). Jerome testifies that in his time “apud Graecos ipsos ipsa exemplaria esse diversa,” and preferred καυχήσωμαι (though all copies of the Latin have _ut ardeam_ or _ut ardeat_), which is said to be countenanced by the Roman Ethiopic: the case of the Bohairic is stated by Bp. Lightfoot (Chap. IV)(422). Tischendorf cites Ephraem (ii. 112) for καυχήσομαι. This variation, which involves the change of but one letter, is worth notice, as showing that the best uncial MSS. are not always to be depended upon, and sometimes are “blemished with errors” (Wordsworth, N. T., _ad loc._). As a parallel use, Theodotion’s version of Dan. iii. 8 (παρέδωκαν τὰ σώματα αὐτῶν εἰς πῦρ) is very pertinent: and for the punishment of burning alive, as practised in those times, consult (if it be thought needful) Joseph., Antiq. xvii. 6, 4 (Hort). Καυχήσωμαι may have obtained the more credit, inasmuch as each of the other principal readings, namely Tischendorf’s καυθήσομαι (DEFGL, 44, 47, 71, 80, 104, 113**, 253**, 254, 255, 257, 260, 265, with nine of Matthaei’s, and some others: καθήσομαι 244) and καυθήσωμαι (CK, 29, 37, and many others, Chrysostom, Theodoret, &c.) of Lachmann and Tregelles, are anomalous, the former in respect to mood, the latter to tense. The important cursive 73 has καυθήσεται with some Latin copies: Codd. 1, 108*, Basil (perhaps Cyprian) adopt καυθῇ: the Syriac (ܕܢܐܘܕ or ܕܘܐܢܕ), and I suppose the Arabic, will suit either of these last. Evidence seems to preponderate on the side of καυθήσομαι, but in the case of these itacisms manuscripts are very fallacious we know. Such a subjunctive future as καυθήσωμαι, however, I should have been disposed to question, had it not passed muster with much better scholars than I am: but to illustrate it, as Tregelles does (An Account of the Printed Text, p. 117, note), from ἵνα δώσῃ Apoc. viii. 3, is to accomplish little, since δώσηι is the reading of אAC, 1 (although Erasmus has δώσῃ with BP, 6, 7, 91, 98, and the Complutensian), 13, 28, 29, 30, 37, 40, 48, 68, 87, 94, 95, 96 (δωσι 8, 26, 27: δω 14), together with the best copies of Andreas, and is justly approved by Lachmann and Tischendorf, nay even by Tregelles himself in his second revision (1872). It seems most likely that in both places ἵνα, the particle of design, is followed by the _indicative_ future, as (with Meyer and Bp. Ellicott) I think to be clearly the case in Eph. vi. 3. In John xvii. 3 even Tregelles adopts ἵνα γινώσκουσιν(423).

36. 1 COR. xv. 51. We have now come to a passage which has perplexed Biblical students from St. Jerome’s time, and has exercised the keen judgement of Bp. Pearson in his Exposition of the seventh article of the Apostles’ Creed. There is but little doubt that the Received text, as rendered in our English versions, is the true reading: (a) Πάντες μὲν οὐ κοιμηθησόμεθα, πάντες δὲ ἀλλαγησόμεθα. Some of the leading authorities omit μέν, a few put δέ or γάρ in its place, but, with this trifling exception, the clause stands thus in B, the third hand of D, and consequently in EKLP, 37, 47, 265, and indeed nearly all the cursives, as in some manuscripts known to Jerome, and has the support of Theodore of Heraclea and Apollinarius: and so the two Syriac, the Bohairic (the Sahidic not being extant), the Gothic, and one edition of the Ethiopic version. For the same form may be cited Ephraem the Syrian, Caesarius, Gregory of Nyssa, and Chrysostom (often) in the fourth century; Theodoret and Euthalius in the fifth century; Andreas of Caesarea in the sixth; John Damascene in the eighth. A modification of this main and true reading (b) Οὐ πάντες κοιμησόμεθα, πάντες δὲ ἀλλαγησόμεθα is supported only by Origen and some copies known to Jerome: it is only a clearer way of bringing out the foregoing sense. The next form also hardly enters into competition, (c) Πάντες [μὲν] ἀναστησόμεθα, οὐ πάντες δὲ ἀλλαγήσομεθα: it is supported by the first hand of D, by the Vulgate (whose manuscripts vary between _resurgimus_ and _resurgemus_, while _m_ omits the negative), by Tertullian and Hilary. Even the Latin versions of EF maintain it against their own Greek, while Jerome and Augustine note it as a point wherein the Latin copies diverge from the Greek. A fourth variation is due to Cod. A alone, (d) Οἱ πάντες μὲν κοιμησόμεθα, οἱ πάντες δὲ ἀλλαγησόμεθα, the second οι being altered by the first hand, and ου by the same or a very early hand super-added after οἱ πάντες δέ: but this is only a correction of transcriptional error. The real variation consists in the transfer of the negative from the first clause to the second, (e) Πάντες [μὲν] κοιμηθησόμεθα, οὐ πάντες δὲ ἀλλαγησόμεθα of אC(F)G, 17, and apparently of A also by intention. This last is discussed by Jerome, who alleges in its favour Didymus and Acacius of Caesarea; it appears also in Origen, Cyril of Alexandria, and in copies known to Pelagius and Maximus, but their testimony fluctuates. In its favour are quoted the Armenian and one form of the Ethiopic, but all the Latin prefer (c) except the interlinear version of G, and the rendering set above the Vulgate text of F, which is assimilated to the latter. The Complutensian margin in a special note chronicles one other change, Πάντες μὲν οὖν κοιμηθησόμεθα, ἀλλ᾽ οὐ πάντες ἀλλαγησόμεθα, but this is bye-work. “The objection made in ancient times to the Received reading was, that the _wicked_ would not be changed, namely, glorified; but St. Paul is here speaking only of the resurrection of the Just” (Bp. Chr. Wordsworth): compare 1 Thess. iv. 14-17. Thus Cod. B and the cursives for once unite to convict of falsehood a change which men were pleased to devise in order to evade a difficulty of their own making.

37. EPH. v. 14. It is instructive to observe how a reading, pretty widely diffused in the fourth century, though not obtaining much acceptance even at that period, has almost entirely disappeared from extant codices. In the place of ἐπιφαύσει σοι ὁ χριστός the first hand of D, followed of course by E (Sangermanensis) and the Latin versions of both, exhibits an interesting variant ἐπιψαύσεις τοῦ χριστοῦ, _continges Christum_. Jerome had heard of it in the form ἐπιψαύσει, id est _continget te Christus_, but refused to vouch for it, as do Chrysostom and Theodoret, though they treat it with somewhat more consideration. The Latin interpreter of Origen (against his own Greek twice, and the Latin once), with Victorinus and the writer cited as Ambrosiaster, adopt it as genuine. Augustine (on Psalm iii) has _et continget te_ once, but once elsewhere the common reading. Theodore of Mopsuestia, in the Latin version of his Commentary on St. Paul’s Epistles, recently edited by Dr. Swete from two manuscripts, one at Amiens (Cod. 68) brought from Corbey [x], a second from Cuza, now Harleian. 3063 [ix], after translating _inluminabit tibi Christus_, goes on to say “alii _continget te Christus_ legerunt; habet autem nullam sequentiam” (Swete, vol. i. p. 180). The variation of D* is surely too curious to be lost sight of altogether. “The two imperatives [ἔγειρε and ἀνάστα] doubtless suggested that the following future would be in the second person, the required σ stood next after ἐπιφαύσει, easily read as ἐπιψαύσει, and then the rest would follow accordingly.” Hort, Notes, p. 125. Such are the harmless recreations of a critical genius.

38. PHIL. ii. 1. εἴ τις κοινωνία πνεύματος, εἴ τινα σπλάγχνα. For τινα, to the critic’s great perplexity, τις is found in אABCD EFGKLP, that is, in _all_ the uncials extant at this place. As regards the cursives nearly the same must be said. Of the seventeen collated by Scrivener, eleven read τις (29, 30, 252, 254, 255, 257, 258, 260, 265, 266, 277), and six τι (31, 104, 221, 244, 253, 256). Mill enumerates sixteen others that give τις, one (40) that has τι: Griesbach reckons forty-five in favour of τις, eight (including Cod. 4) for τι, to which Scholz adds a few more (18, 46, 72, 74). Thus _am. fuld. tol._ of the Vulgate render _si quid viscera_, for the more usual _si qua viscera_. One cursive (109) and a manuscript of Theodoret have τε. Basil, Chrysostom (in manuscript) and others read τις, as do the Complutensian, the Aldine (1518), Erasmus’ first four, and R. Stephen’s first two editions. In fact it may be stated that no manuscript whatever has been cited for τινα, which is not therefore likely to be found in many. Theodore of Mopsuestia alone, in his Latin version published by Dr. Swete (vol. i. p. 214), has _si qua et viscera_ against the Vulgate. In spite of what was said above with regard to far weaker cases, it is impossible to blame editors for putting τις into the text here before σπλάνχνα: to have acted otherwise (as Tischendorf fairly observes) would have been “_grammatici quam editoris partes agere_.” Yet we may believe the reading to be as false as it is intolerable, and to afford us another proof of the early and (as the cursives show) the well-nigh universal corruption of our copies in some minute particulars. Of course Clement and later Fathers give τινα, indeed it is surprising that any cite otherwise; but, _in the absence of definite documentary proof_, this can hardly be regarded as genuine. Probably St. Paul wrote τι (the reading of about nineteen cursives), which would readily be corrupted into τις, by reason of the σ following (ΤΙΣΠΛΑΓΧΝΑ), and the τις which had just preceded. See also Moulton’s “Winer,” p. 661, and note 3.

39. COL. ii. 2. τοῦ μυστηρίου τοῦ Θεοῦ καὶ πατρὸς καὶ τοῦ χριστοῦ, “of the mystery of God the Father, and of Christ.” The reading of B (approved by Lachmann, by Tischendorf in his eighth edition, by Tregelles, Westcott and Hort, Bp. Chr. Wordsworth, and Bp. Ellicott), τοῦ μυστηρίου τοῦ θεοῦ χριστοῦ (“ita cod. nihil interponens inter θεοῦ et χριστοῦ,” Mai, 2nd ed.(424)), has “every appearance of being the original reading, and that from which the many perplexing variations have arisen” (Canon II). At present it stands in great need of confirmation, since Hilary (de Trin. ix) alone supports it (but καὶ χριστοῦ Cyril), though the Scriptural character of the expression is upheld by the language of ch. i. 27 just preceding, and by the Received text in 1 Tim. iii. 16. Some, who feel a difficulty in understanding how χριστοῦ was removed from the text, if it ever had a place there, conceive that the verse should end with θεοῦ, all additions, including χριστοῦ the simplest, being _accretions_ to the genuine passage. These alleged accretions are τοῦ θεοῦ ὅ ἐστι χριστός, manifestly an expansion of χριστοῦ and derived from ch. i. 27; τοῦ θεοῦ πατρὸς τοῦ χριστοῦ: τοῦ θεοῦ καὶ πατρὸς καὶ τοῦ χριστοῦ, the final form of the Received text. Now, of these four readings, τοῦ θεοῦ the shortest, and, according to Griesbach, Scholz, Tischendorf in his seventh edition, Alford, and Dr. Green, the true one, is found only in the late uncial P, and in a few, though confessedly good, cursives: 37, 71, 80*, 116 (καὶ θεοῦ 23), and the important second hand of 67; witnesses too few and feeble, unless we consent to put our third Canon of internal evidence to a rather violent use. Of the longer readings, ὅ ἐστιν χριστός is favoured by D (though obelized by the second hand, which thus would read only τοῦ θεοῦ), _d_ _e_ (whose parallel Greek speaks differently), by Augustine and Vigilius of Thapsus, but apparently by no cursives. The form best vouched for appears to be that of א*AC, 4, of the Sahidic according to one of the readings of Griesbach, and of an Arabic codex of Tischendorf, τοῦ θεοῦ πατρὸς τοῦ (א* omits τοῦ) χριστοῦ. To these words “_ihu_” is simply added by _f_ (FG, _g_ are unfortunately lost here) and by other manuscripts of the Vulgate (_am. fuld._, &c.), though the Clementine edition has “Dei patris et Christi Jesu,” the Complutensian in the Latin “dei et patris et C.J.” With the Clementine Vulgate agree the Bohairic, and (omitting ἰησοῦ) the Peshitto Syriac, Arabic, 47, 73, Chrysostom; while 41, 115, 213, 221, 253* (τοῦ θ. καὶ π. τοῦ χ.), so far strengthen the case of אAC. The Received text is found in (apparently) the great mass of cursives, in D (_tertiâ manu_), EKL, the Harkleian Syriac (but the καί after πατρός marked with one of Harkel’s asterisks), Theodoret, John Damascene and others. The minor variations, τοῦ θεοῦ ἐν χριστῷ of Clement and Ambrosiaster, τοῦ θεοῦ τοῦ ἐν χριστῷ of 17, uphold D*, as may the Ethiopic (“domini quod de Christo”): to the reading of Cod. 17 Zohrab’s or the Venice Armenian (A.D. 1789) simply adds “Jesu.” We also find “dei Christi Jesu patris et domini” in _tol._, “dei patris et domini nostri Christi” in _demid._, “dei patris in Christo Jesu” in Uscan’s Armenian; but these deserve not attention. Theodore of Mopsuestia (Swete, vol. i. p. 283), has _mysterii Dei Patris et Christi_, which need not imply the omission of καί before πατρός.

On reviewing the whole mass of conflicting evidence, we may unhesitatingly reject the shortest form τοῦ θεοῦ, some of whose maintainers do not usually found their text on cursive manuscripts almost exclusively. We would gladly adopt τοῦ θεοῦ χριστοῦ, so powerfully do internal considerations plead in its favour, were it but a little better supported: the important doctrine which it declares, Scriptural and Catholic as that is, will naturally make us only the more cautious in receiving it unreservedly. Yet the more we think over this reading, the more it grows upon us, as the source from which all the rest are derived. At present, perhaps, τοῦ θεοῦ πατρὸς τοῦ χριστοῦ may be looked upon as the most strongly attested, but in the presence of so many opposing probabilities, a very small weight might suffice to turn the critical scale.

40. 1 THESS. ii. 7. We have here a various reading, consisting of the prefix of a single letter, which seems to introduce into a simple verse what is little short of an absurdity. Instead of ἤπιοι of the Received text, of Tischendorf and Tregelles, we find νήπιοι adopted by Lachmann as a consequence of his own stringent rules, and by Westcott and Hort of their own free will, unless indeed it be said that they also are working in chains of their own forging. How St. Paul can compare himself to a babe in one clause of the verse and to its nurse in the other would be quite unintelligible if Origen, who read νήπιοι, had not instructed us that the nurse is playing at baby for the babe’s amusement (ἐγένετο νήπιος καὶ παραπλήσιος τροφῷ θαλπούσῃ τὸ ἑαυτῆς παιδίον καὶ λαλούσῃ λόγους ὡς παιδίον διὰ τὸ παιδίον, iii. 662). It needs but the exercise of common sense to brush away such a fancy as this, and the state of the evidence will show us how the best authorities are sometimes hopelessly in the wrong; for νήπιοι is the form favoured by א*BC*D*FG, 5, 23, 26, 31* 37, 39**, 74, 87, 109**, 114, 115, 137, 219*, 252, and is easily accounted for by the accidental reduplication of the letter after Ν in ΗΜΕΝΗΠΙΟΙ (_see_ p. 10). The Vulgate and the Latin versions accompanying DEFG (_e_ testifying against its own Greek) have _parvuli_, and so the Bohairic, Ethiopic, Clement of Alexandria (ἤπιος οὖν ὁ νήπιος), Ambrosiaster, Jerome, and Augustine very expressly. On the other hand ἤπιος is vouched for by א**AC**D**EKLP, 17, 47, 61, 260, and by all cursives not named above, by both Syriac versions, by the Sahidic and by its follower the Bashmuric, by the Armenian, by Clement and Origen elsewhere (but their inconsistency means nothing but carelessness), Basil, Chrysostom, Theodore of Mopsuestia(425), Theodoret, Euthalius, Œcumenius, John Damascene and the catenae. Theophylact knew of and expounds both readings. It is almost pathetic to mark Dr. Hort’s brave struggle to maintain a cause which in this instance is simply hopeless. “The second ν might be inserted or omitted with equal facility; but the change from the bold image to the tame and facile adjective is characteristic of the difference between St. Paul and the Syrian revisers (cf. 1 Cor. iii. 1, 2; ix. 20, &c.). It is not of harshness that St. Paul here declares himself innocent, but of flattery and the rhetorical arts by which gain or repute is procured, his adversaries having doubtless put this malicious interpretation upon his language among the Thessalonians” (Notes, p. 128). For his alleged Syrian revision, _see_ above, p. 287.

41. 1 TIM. iii. 16. Θεὸς ἐφανερώθη ἐν σαρκί. This text has proved the _crux criticorum_. The Vatican has now failed us, but all manuscripts (D _tertiâ manu_, KLP, 300 cursives) read Θεός with the common text, except א*A*? C*? FG, 17, 73, which have ὅς, D* which (after the Latin versions) has ὅ: the Leicester codex, 37, gives ὁ _θς_ (_see_ facsimile No. 40, l. 1), as if to combine two of the variations(426). In the abridged form of writing usual in all manuscripts, even the oldest, the difference between ΟΣ and _ΘΣ_ consists only in the presence or absence of two horizontal strokes; hence it is rather to be regretted than wondered at that the true reading of each of the uncial authorities for the former is more or less open to question. Respecting Cod. א we have the statement of Tischendorf, a most consummate judge in such matters: “_corrector aliquis, qui omnium ultimus textum attigit, saeculi ferè duodecimi_, [_pro_ ος _primae manûs_] reposuit θεος, _sed hoc tam cautè ut antiquissimam scripturam intactam relinqueret_” (Notitia Cod. Sinait. p. 20), which is unequivocal enough: _see_ facsimile No. 13 in Scrivener’s “Collation of Cod. Sin.,” and Introd., p. xxv: also Plate iv, facsimile No. 11 c of this volume, wherein the twelfth century θε above the line, the new accent over ΟΣ, and the triple points to denote insertion, are very conspicuous. Nor is there any real doubt respecting the kindred codices FG. From the photographed title-page of the published “Cod. Augiensis” (F) l. 9, and Matthaei’s facsimile of G (N. T., vol. i. p. 4)(427), it will be seen that while there is not the least trace of the horizontal line within the circle of omicron, the line above the circle in _both_ (_ΟΣ_) is not horizontal, but rises a little towards the right: such a line not unfrequently in F, oftener in G, is used (as here) to indicate the rough breathing: it sometimes stands even for the _lenis_ (e.g. ἱδιον 1 Cor. vi. 18; vii. 4; 37; ἱσσα Phil. ii. 6). Those who never saw Cod. C must depend on Tischendorf’s Excursus (Cod. Ephraemi, pp. 39-42) and his facsimile, imitated in our Plate x. No. 24. His decision is that the primitive reading was ΟΣ, but he was _the first to discern a cross line within_ Ο (facsimile, l. 3, eighth letter); which, however, from the colour (“_subnigra_”) he judges to belong to the second or third hand, rising upwards (a tendency rather exaggerated than otherwise in our Plate); while the coarse line above, and the musical notes (denoting a word of two syllables) below, are plainly of the third hand. This verdict, especially delivered by such a man, we know not how to gainsay, and merely point to the fact that the cross line in Θ, the ninth letter further on, which is certainly _primâ manu_, also ascends towards the right. Cod. A, however, I have examined at least twenty times within as many years, and yet am not quite able to assent to the conclusion of Mr. Cowper when he says “we hope that no one will think it possible, either with or without a lens, to ascertain the truth of the matter by any inspection of the Codex” (Cod. Alex., Introd. p. xviii). On the contrary, seeing (as every one must see for himself) with my own eyes, I have always felt convinced with Berriman and the earlier collators that Cod. A read _ΘΣ_, and, so far as I am shaken in my conviction at all, it is less by the adverse opinion even of Bp. Ellicott(428), than by the more recently discovered fact that ΟΣ (which is adopted by Griesbach, Lachmann, Tischendorf, Davidson, Tregelles, Alford, Ellicott, Wordsworth, Hort and Westcott), was read in א as early as the fourth century.

The secondary witnesses, versions, and certain of the Fathers, also powerfully incline this way, and they deserve peculiar attention in a case like the present. The Peshitto (ܕ) and Harkleian (text and ܗܘ in margin) Syriac have a relative (whether ὅς or ὅ); so have the Armenian, the Roman Ethiopic, and Erpenius’ Arabic. The Gothic supports ὅς; the Sahidic, Bohairic, and Platt’s Ethiopic favour ὅς or ὅ: all Latin versions (even _f_ _g_ whose Greek is _ΟΣ_) read “quod,” while θεός appears only in the Slavonic (which usually resembles KL and the later copies) and the Polyglott Arabic. Of ecclesiastical writers the best witness for the Received text is Ignatius, Θεοῦ ἀνθρωπίνως φανερουμένου (“Ephes.” 19), both in the Greek and Old Latin, although the Syriac abbreviator seems to have τοῦ υἱοῦ: the later interpolator expanded the clause thus: θεοῦ ὡς ἀνθρώπου φαινομένου, καὶ ἀνθρώπου ὡς θεοῦ ἐνεργοῦντος. Hippolytus (Adv. Not. 17: fl. 220) makes a “free reference” to it in the words Οὗτος προελθὼν εἰς κόσμον, θεὸς ἐν σώματι ἐφανερώθη, and elsewhere with ὁ before προελθών. The testimony of Dionysius of Alexandria (265) can no longer be upheld (Tregelles, Horne, iv. p. 339), that of Chrysostom to the same effect is by some deemed precarious, since his manuscripts fluctuate, and Cramer’s catena on 1 Tim. p. 31 is adverse(429). The evidence borne for θεός by Didymus (de Trin.) and Gregory Nyssen(430) is beyond all doubt; that of later writers, Theodoret, John Damascene, Theophylact, Œcumenius (as might be looked for) is clear and express. The chief Latins, Hilary, Jerome, Augustine, &c., exhibit either _qui_ or _quod_: Cyril of Alexandria (for so we must conclude both from manuscripts and his context)(431), Epiphanius (_twice_), Theodore of Mopsuestia (in Latin)(432), and others of less weight, or whose language is less direct, are cited in critical editions of the N. T. in support of a relative; add to which that θεός is not quoted by Fathers (e.g. Cyprian, p. 35; Bentleii Critica Sacra, p. 67) in many places where it might fairly be looked for; though this argument must not be pushed too far. The idle tale, propagated by Liberatus the Deacon of Carthage, and from him repeated by Hincmar and Victor, that Macedonius Patriarch of Constantinople (A.D. 506) was expelled by the Emperor Anastasius for corrupting Ο or ΟΣ into ΘΣ, although lightly credited by Dr. Tregelles (An Account of the Printed Text, p. 229) and even by Dr. Hort (Notes, p. 133), is sufficiently refuted by Bp. Pearson (On the Creed, Art. ii. p. 128, 3rd edition).

On a review of the whole mass of external proof, bearing in mind too that ΟΣ (from which ὅ of D* is an evident corruption) is grammatically much the _harder_ reading after μυστήριον (Canon I), and that it might easily pass into ΘΣ, we must consider it probable (indeed, if we were sure of the testimony of the first-rate uncials, we might regard it as certain) that the second of our rules of Comparative Criticism must here be applied, and θεός of the more recent many yield place to ὅς of the ancient few(433). Yet even then the force of the Patristic testimony remains untouched. Were we to concede to Dr. Hort’s unproved hypothesis that Didymus, de Trinitate, abounds in what he calls Syrian readings, and that they are not rare with Gregory Nyssen (Notes, p. 133), the clear references of Ignatius and Hippolytus are not thus to be disposed of. I dare not pronounce θεός a corruption.

This decision of Dr. Scrivener would probably have been considerably strengthened in favour of θεός, if the above passage had been written after, instead of before, the composition and appearance of Dean Burgon’s elaborate and patient examination of all the evidence, which occupies seventy-seven pages in his “Revision Revised” (pp. 424-501). Dean Burgon shows at length that after about 1770 the passage in A became so worn that it has been since that time increasingly difficult to see it; he casts much doubt upon the witness of C for ὅς, which Mr. Hoskier (Cod. 604, Appendix J), after a long examination of the MS., not only confirms, but actually removes in the opposite direction by claiming C as a witness for θεός; he maintains with reason that the transverse line in F and G is the sign of contraction; he exhibits the consentient testimony of the cursives; he claims upon the testimony of the scholar who was editing the Harkleian that version, as also the Georgian and Slavonic; and he adds to the Fathers enumerated above, besides doubtful testimonies, Gregory of Nazianzus, Cyril of Alexandria, Severus of Antioch, Diodorus of Tarsus, Euthalius, Macedonius, Epiphanius of Catana, Theodorus Studita, Euthymius, some scholia, the author of Περὶ θείας σαρκώσεως, and an anonymous author,—making some fifty testimonies in all.

42. 1 TIM. vi. 7. By omitting δῆλον of the Received text, Lachmann, Tischendorf, Tregelles, Westcott and Hort, produce a Greek sentence as inconsequential as the most thorough votaries of the “harder reading” can wish for: “For we brought nothing into the world, because neither can we carry anything out.” Dr. Hort sees, of course, that St. Paul could not reason in this fashion, and says that “The text [i.e. _his_ text, without δῆλον] is manifestly the parent of all the other readings, which are futile attempts to smooth away its difficulty. A primitive corruption must lurk somewhere,”—and then ventures on the awkward suggestion that ΟΤΙ arose from the transcriptional repetition of the last syllable of κοσμον (ΟΝ being read as ΟΤΙ), a guess which we observe that Dr. Westcott does not care to vouch for (Notes, p. 134). But why create a difficulty at all? Cod. B, which ends in Heb. ix. 14, is now lost to us, and of the rest δῆλον is omitted in א*AFG and its Latin version _g_ with copies of the Vulgate referred to by Lachmann, the Bohairic (καί for ὅτι), Sahidic; the Armenian and both Ethiopic varying with the Bohairic. Instead of δῆλον D*, _m_, _fuld._, Cyprian and the Gothic have ἀληθές, and the printed Vulgate with its codices (even _f_) and Ambrosiaster _haud dubium_, which will suit δῆλον well enough, as will ܘܕܝܥܐ (or ܐܥܝܕܘ) (_et notum est_) of the Syriac versions. For δῆλον itself stand א**D** (_hiat_ E) KLP, all the cursives save one, and of the Fathers Basil, Macarius, Chrysostom, Euthalius, Theodoret, and John Damascene, evidence which we should have liked to see a little stronger.

43. PHILEM. 12. For ὃν ἀνέπεμψα; σὺ δὲ αὐτόν, τουτέστι τὰ ἐμὰ σπλάγχνα, προσλαβοῦ of the Received text, the critics, Lachmann, Tischendorf, Tregelles (but not his margin), Bp. Lightfoot, Westcott and Hort read ὃν ἀνέπεμψά σοι, αὐτόν, τουτέστι τὰ ἐμὰ σπλάγχνα, omitting προσλαβοῦ, which they judge to have been interpolated from ver. 17. Tregelles and Bp. Lightfoot, moreover, put a full stop after σοι, so that αὐτόν is regarded as an “accusative suspended; the sentence changes its form and loses itself in a number of dependent clauses; and the main point is not resumed till ver. 17 προσλαβοῦ αὐτὸν ὡς ἐμέ, the grammar having been meanwhile dislocated.” So Lightfoot, who vindicates the emphatic place he has assigned to αὐτόν by the not very close parallels John ix. 21, 23; Eph. i. 22. Manuscripts, of course, will not help us much in punctuation, but Codd. א*A, 17 are very good witnesses for σοι in the room of σὺ δέ and for the omission of προσλαβοῦ, a simple, although somewhat rude, construction well worthy of attention. For σοι, with or without σὺ δέ following, we have the additional support of C*DE, _d_ _e_ and _g_ against its own Greek, the Clementine Vulgate and such Vulgate codices as _demid. harl._2**, the Peshitto Syriac, Bohairic, Armenian, Ethiopic, &c. For the omission of προσλαβοῦ, which is of course the chief variation, besides א*A, 17 are cited F and G in the Greek but not in their Latin versions, 37 and others setting it before αὐτόν. It is found in all the rest, D**E**KLP, all other cursives, and (as might have been anticipated) the versions, as well Latin as Syriac, Bohairic (which reads as Cod. 37), Gothic, and Ethiopic: _g_, the Armenian and Theodoret put it after αὐτόν.

Fourth Series. Catholic Epistles.

44. JAMES iv. 4. Μοιχοὶ καί should be omitted before μοιχαλίδες on the testimony of א*AB, 13. The Peshitto, Bohairic, Latin, Armenian, and both Ethiopic versions have “adulterers” (_fornicatores ff_) only, but since no Greek copy thus reads, we must suppose that their translators were startled by the bold imagery so familiar to the Hebrew prophets (Isa. liv. 5; Jer. ii. 2; Ezek. xvi. 32 are cited from a host of similar passages by Wordsworth) and endeavoured to dilute it in this way. Tischendorf would join μοιχαλίδες with δαπανήσητε ver. 3, alleging the point or stop placed after it in Cod. B: but this point is not found in Vercellone’s edition, although he leaves a small space before οὐκ. The full form Μοιχοὶ καὶ μοιχαλίδες of אcKLP, the later Syriac, and all other known copies, is evidently a correction of early scribes.

45. JAMES iv. 5. The variation between κατῴκισεν and κατῴκησεν is plainly to be attributed to a mere itacism, whichsoever is to be regarded as the true form. We find ι in אAB, 101, 104 only, nor is it quite accurate to say with Tischendorf that collators are apt to overlook such points. In KLP, and apparently in all other manuscripts of every class, η is read, and so the catenas, with Theophylact and Œcumenius, understand this difficult passage. That all the versions (Latin, Syriac, Egyptian, &c.) thus render seems decisive in favour of η. The combination of אAB, however strong, has repeatedly been seen not to be irresistible; and while it must be confessed that in our existing Greek copies the interchange of ι and η (though found in Cod. A) is not an itacism of the very oldest type (p. 10), yet here the testimony of the versions refers it back to the second century. Lachmann, Tischendorf, Tregelles, Westcott and Hort, combine in reading κατῴκισεν.

46. 1 PET. i. 23. Here we have a remarkable example to illustrate what we saw in the cases of Rom. viii. 20; 2 Cor. iii. 3, Phil. ii. 1, that the chief uncials sometimes conspire in readings which are unquestionably false, and can hardly have arisen independently of each other. For σπορᾶς φθαρτῆς Codd. אAC have φθορᾶσ φθαρτῆς, the scribe’s eye wandering in writing σπορᾶς to the beginning of the next word: Cod. B is free from this vile corruption. When Mill records the variation for Cod. A, he adds (as well he might), “dormitante scribâ:” but that the same gross error should be found in three out of the four oldest codices, _and in no other_, is very suggestive, and not a little perplexing to false theorists.

47. 1 PET. iii. 15. Κύριον δὲ τὸν θεὸν ἁγιάσατε ἐν ταῖς καρδίαις ὑμῶν. For θεόν we find χριστόν (a change of considerable doctrinal importance)(434) in אABC, 7, 8 (Stephen’s ια´), 13, 33 (_margin_), 69, 137, 182, 184 (but not 221: _see_ p. 310, note 2), Apost. 1 (_ιν_ _χν_ ἡμῶν) with its Arabic translation. Thus too read both Syriac versions, the Sahidic, Bohairic, Armenian (τὸν αὐτὸν καὶ χριστόν), Erpenius’ Arabic, the Vulgate, Clement of Alexandria, Fulgentius, and Bede. Jerome has “Jesum Christum:” the Ethiopic and one other (Auctor de promiss., fourth century) omit both words. Against this very strong case we can set up for the common text only the more recent uncials KLP (not more than seven uncials contain this Epistle), the mass of later cursives (ten out of Scrivener’s twelve, also Wake 12, or Cod. 193), the Polyglott Arabic, Slavonic, Theophylact, and Œcumenius, authorities of the ninth century and downwards. It is a real pleasure to me in this instance to express my cordial agreement with Tregelles (and so read Lachmann, Tischendorf, Westcott and Hort), when he says, “Thus the reading χριστόν may be relied on _confidently_” (An Account of the Printed Text, p. 285). I would further allege this text as one out of many proofs that the great uncials seldom or never conspire in exhibiting a really valuable departure from the later codices, unless supported by some of the best of the cursives themselves. See, however, Acts xiii. 32.

48. 2 PET. ii. 13. The resemblance between the second epistle of St. Peter and that of St. Jude is too close to be unobserved by the most careless reader, and the supposition that the elder Apostle’s letter was in Jude’s hands when he wrote his own is that which best meets the circumstances of the case. The σπῖλοι of the present verse, for example, looks like the origin of σπιλάδες in Jude 12, where the latter word is employed in a signification almost unprecedented in classical Greek, though the Orphic poems have been cited for its bearing the sense of “spots,” which all the ancient versions rightly agree with our Authorized Bible in attributing to it. Bearing in mind the same verse of St. Jude, it seems plain that ἀπάταις of the Received text cannot be accepted as true, as well because it affords so poor a meaning in connexion with ἐντρυφῶντες and συνευωχούμενοι, as because the later writer must have seen ἀγάπαις in his model, when he paraphrased it by οἱ ἐν ταῖς ἀγάπαις ὑμῶν σπιλάδες συνευωχούμενοι. For this change of two letters we have the support of Cod. A (as corrected by the first hand) and B alone of the manuscripts, but of the versions, the Latin Speculum _m_ which in these later epistles is strangely loose, yet cannot be misunderstood in the present place, the Vulgate, the Sahidic version, the Ethiopic, the Syriac printed with the Peshitto(435), and the margin of the Harkleian version. Add to these Ephraem and the Latin author of the tract “de singularitate clericorum,” both of the fourth century. The little group of cursives 27, 29, and the second hand of 66 read ἀγνοίαις; but ἀπάταις, _nescio quo sensu_(436), still cleaves to the text of Tischendorf and of Westcott and Hort, and to the margin of Tregelles, who in the text prefers ἀγάπαις with Lachmann and Westcott and Hort’s margin. Codd. אA (in its original form) CKLP, all other cursives, the catenas (Cod. 36, &c.), the Bohairic, Armenian, and Harkleian versions also have ἀπάταις, and so Theophylact and Œcumenius, but hardly Jerome as cited by Tischendorf.

49. 1 JOHN ii. 23. The English reader will have observed that the latter clause of this verse, “_but he that acknowledgeth the Son hath the Father also_,” is printed in italics in our Authorized version, this being the only instance in the New Testament wherein variety of reading is thus denoted by the translators, who derived both the words and this method of indicating their doubtful authenticity from the “Great Bible” of 1539(437). The corresponding Greek ὁ ὁμολογῶν τὸν υἱὸν καὶ τὸν πατέρα ἔχει (which appears to have been lost out of some copies by Homoeoteleuton), was first inserted in Beza’s Greek Testament in 1582(438), it is approved by all modern editors (Griesbach, Scholz, Lachmann, Tischendorf, Tregelles, Westcott and Hort), and, though still absent from the _textus receptus_, is unquestionably genuine. This is just such a point as versions are best capable of attesting. The “Great Bible” had no doubt taken the clause from the Latin Vulgate, in whose printed editions and chief manuscripts it is found (e.g. in _am. fuld. demid. tol. harl._), as also in both Syriac, both Egyptian (the Sahidic not for certain), the Armenian, Ethiopic, and Erpenius’ (not the Polyglott) Arabic version. Of manuscripts the great uncials אABC (with P) contain the clause, the later KL omit it. Of the cursives only two of Scrivener’s (182, 225) have it, and another (183) _secundâ manu_: from twelve or more of them it is absent, as also from seven of Matthaei’s: but of the other cursives it is present in at least thirty, whereof 3, 5, 13, 66** (_marg._), 68, 69, 98 are valuable. It is also acknowledged by Clement, Origen (_thrice_), Eusebius, both Cyrils, Theophylact, and the Western Fathers. The younger Cyril, possibly Euthalius, and one or two others have ὁμολογεῖ for the final ἔχει: the Old Latin _m_, Cyprian, and Hilary repeat τὸν υἱὸν καί before τὸν πατέρα ἔχει. The critical skill of Beza must not be estimated very highly, yet in this instance he might well have been imitated by the Elzevir editors.

50. 1 JOHN v. 7, 8. Ὅτι τρεῖς εἰσιν οἱ μαρτυροῦντες [ἐν τῷ οὐρανῷ, ὁ Πατήρ, ὁ Λόγος, καὶ τὸ Ἅγιον Πνεῦμα; καὶ οὗτοι οἱ τρεῖς ἕν εἰσι. καὶ τρεῖς εἰσιν οἱ μαρτυροῦντες ἐν τῇ γῇ], τὸ πνεῦμα, καὶ τὸ ὕδωρ, καὶ τὸ αἷμα; καὶ οἱ τρεῖς εἰς τὸ ἕν εἰσιν.

The authenticity of the words within brackets will, perhaps, no longer be maintained by any one whose judgement ought to have weight; but this result has been arrived at after a long and memorable controversy, which helped to keep alive, especially in England, some interest in Biblical studies, and led to investigations into collateral points of the highest importance, such as the sources of the Received text, the manuscripts employed by R. Stephen, the origin and value of the Velesian readings, and other points. A critical _résumé_ of the whole discussion might be profitably undertaken by some competent scholar; we can at present touch only upon the chief heads of this great debate(439).

The two verses appear in the early editions, with the following notable variations from the common text, C standing for the Complutensian, Er. for one or more of Erasmus’ five editions. Ver. 7.—ἐν τῷ οὐρανῷ _usque ad_ τῇ γῇ ver. 8, Er. 1, 2.—ὁ _prim. et __ secund_. Er. 3. [_non_ C. Er. 4, 5]. + και (_post_ πατήρ) C.—τό Er. 3. πνεῦμα ἅγιον Er. 3, 4, 5.—οὗτοι C. + εισ το (_ante_ εν) C. Ver. 8, επί της γης C.—τὸ _ter_ Er. 3, 4, 5 [_habent_ C. Er. 1, 2].—καὶ οἱ τρεῖς _ad_ _fin. vers._ C. They are found, including the clause from ἐν τῷ οὐρανῷ to ἐν τῇ γῇ in no more than three Greek manuscripts, and those of very late date, one of them (Cod. Ravianus, Evan. 110) being a mere worthless copy from printed books; and in the margin of a fourth, in a hand as late as the sixteenth century. The real witnesses are the Codex Montfortianus, Evan. 61, Act. 34 (whose history was described above, p. 187(440)); Cod. Vat.-Ottob. 298 (Act. 162), and, for the margin, a Naples manuscript (Act. 83 or 173, q. v.). On comparing these slight and scanty authorities with the Received text we find that they present the following variations: ver. 7. ἀπὸ τοῦ οὐρανοῦ (_pro_ ἐν τῷ οὐρανῷ) 162.—ὁ _prim. et secund._ 34, 162.—τό 34, 162. _πνα_ ἅγιον 34, 162.—οὗτοι 162. + εἰς τό (_ante_ ἕν) 162. Ver. 8. εἰσί 73 _marg._ ἐπὶ τῆς γῆς 162.—τό _ter_ 34.—καί (_post_ _πνα_) 34, 162.—καὶ οἱ τρεῖς _ad fin._ _vers_. 34, 162, _fin_. εἰσι 173. No printed edition, therefore, is found to agree with either 34 or 162 (173, whose margin is so very recent, only differs from the common text by dropping ν ἐφελκυστικόν), though on the whole 162 best suits the Complutensian: but the omission of the article in ver. 7, while it stands in ver. 8 in 162, proves that the disputed clause was interpolated (probably from its parallel Latin) by one who was very ill acquainted with Greek.

The controverted words are not met with in any of the extant uncials (אABKLP) or in any cursives besides those named above(441): the cursives that omit them were found by the careful calculation of the Rev. A. W. Grafton, Dean Alford’s secretary (N. T. _ad. loc._), to amount to 188 in all (to which we may now add Codd. 190, 193, 219-221), besides some sixty Lectionaries. The aspect of things is not materially altered when we consult the versions. The disputed clause is not in any manuscript of the Peshitto, nor in the best editions (e.g. Lee’s): the Harkleian, Sahidic, Bohairic, Ethiopic, Arabic do not contain it in any shape: scarcely any Armenian codex exhibits it, and only a few recent Slavonic copies, the margin of a Moscow edition of 1663 being the first to represent it. The Latin versions, therefore, alone lend it any support, and even these are much divided. The chief and oldest authority in its favour is Wiseman’s Speculum _m_ and _r_ of the earlier translation; it is found in the printed Latin Vulgate, and in perhaps forty-nine out of every fifty of its manuscripts, but not in the best, such as _am. fuld. harl._3; nor in Alcuin’s reputed copies at Rome (_primâ manu_) and London (Brit. Mus. Add. 10,546), nor in the book of Armagh and full fifty others. In one of the most ancient which contain it, _cav._, ver. 8 precedes ver. 7 (as appears also in _m. tol. demid._ and a codex at Wolfenbüttel, _Wizanburg._ 99 [viii] cited by Lachmann), while in the margin is written “_audiat hoc Arius et ceteri_,” as if its authenticity was unquestioned(442). In general there is very considerable variety of reading (always a suspicious circumstance, as has been already explained), and often the doubtful words stand only in the margin: the last clause of ver. 8 (_et hi tres unum sunt_), especially, is frequently left out when the “Heavenly Witnesses” are retained. It is to defend _this_ omission by the opinion of Thomas Aquinas, not to account for the reception of the doubtful words, that the Complutensian editors wrote a note, the longest and indeed almost the only one in their New Testament. We conclude, therefore, that the passage from ἐν τῷ οὐρανῷ to ἐν τῇ γῇ had no place in ancient Greek manuscripts, but came into some of the Latin at least as early as the sixth century.

The Patristic testimony in its favour, though quite insufficient to establish the genuineness of the clause, is entitled to more consideration. Of the Greek Fathers it has been said that no one has cited it, even when it might be supposed to be most required by his argument, or though he quotes consecutively the verses going immediately before and after it(443): [but a passage occurs in the Greek Synopsis of Holy Scripture of uncertain date (fourth or fifth century), which appears to refer to it, and another from the Disputation with Arius (Ps.-Athanasius)]. The same must be said of the great Latins, Hilary, Lucifer, Ambrose, Jerome(444), and Augustine, with others of less note. On the other hand the _African_ writers, Vigilius of Thapsus, at the end of the fifth century, and Fulgentius of Ruspe (fl. 508) in two places, expressly appeal to the “three Heavenly Witnesses” as a genuine portion of St. John’s Epistle; nor is there much reason to doubt the testimony of Victor Vitensis, who records that the passage was insisted on in a confession of faith drawn up by Eugenius Bishop of Carthage and 460 bishops in 484, and presented to the Arian Hunneric, king of the Vandals [or of Cassiodorus, an Italian, in the sixth century]. From that period the clause became well known in other regions of the West, and was in time generally accepted throughout the Latin Church.

But a stand has been made by the maintainers of this passage on the evidence of two African Fathers of a very different stamp from those hitherto named, Tertullian and Cyprian. If it could be proved that these writers cited or alluded to the passage, it would result—_not by any means that it is authentic_—but that like Acts viii. 37 and a few other like interpolations, it was known and received in some places, as early as the second or third century. Now as regards the language of Tertullian (which will be found in Tischendorf’s and the other critical editions of the N. T.; advers. Prax. 25; de Pudic. 21), it must be admitted that Bp. Kaye’s view is the most reasonable, that “far from containing an allusion to 1 John v. 7, it furnishes most decisive proof that he knew nothing of the verse” (Writings of Tertullian, p. 550, second edition); but I cannot thus dispose of his junior Cyprian (d. 258). One must say with Tischendorf (who, however, manages to explain away his testimony) “_gravissimus est_ Cyprianus _de eccles. unitate_ 5.” His words run, “Dicit dominus, _Ego et pater unum sumus_ (John x. 30), et iterum de Patre, et Filio, et Spiritu Sancto scriptum est, _Et tres unum sunt_.” And yet further, in his Epistle to Jubaianus (73) on heretical baptism: “Si baptizari quis apud haereticos potuit, utique et remissam peccatorum consequi potuit,—si peccatorum remissam consecutus est, et sanctificatus est, et templum Dei factus est, quaero cujus Dei? Si Creatoris, non potuit, qui in eum non credidit; si Christi, nec hujus fieri potuit templum, qui negat Deum Christum; si Spiritus Sancti, cum tres unum sunt, quomodo Spiritus Sanctus placatus esse ei potest, qui aut Patris aut Filii inimicus est?” If these two passages be taken together (the first is manifestly much the stronger(445)), it is surely safer and more candid to admit that Cyprian read ver. 7 in his copies, than to resort to the explanation of Facundus [vi], that the holy Bishop was merely putting on ver. 8 a spiritual meaning; although we must acknowledge that it was in this way ver. 7 obtained a place, first in the margin, then in the text of the Latin copies, and though we have clear examples of the like mystical interpretation in Eucherius (fl. 440) and Augustine (contra Maximin. 22), who only knew of ver. 8.

Stunica, the chief Complutensian editor, by declaring, in controversy with Erasmus, with reference to this very passage, “Sciendum est, Graecorum codices esse corruptos, nostros [i.e. Latinos] verò ipsam veritatem continere,” virtually admits that ver. 7 was translated in that edition from the Latin, not derived from Greek sources. The versions (for such we must call them) in Codd. 34, 162 had no doubt the same origin, but were somewhat worse rendered: the margin of 173 seems to be taken from a printed book. Erasmus, after excluding the passage from his first two editions, inserted it in his third under circumstances we have before mentioned; and notwithstanding the discrepancy of reading in ver. 8, there can be little or no doubt of the identity of his “Codex Britannicus” with Montfort’s(446). We have detailed the steps by which the text was brought into its present shape, wherein it long remained, unchallenged by all save a few such bold spirits as Bentley, defended even by Mill, implicitly trusted in by those who had no knowledge of Biblical criticism. It was questioned in fair argument by Wetstein, assailed by Gibbon in 1781 with his usual weapons, sarcasm and insinuation (Decline and Fall, chap. xxxvii). Archdeacon Travis, who came to the rescue, a person “of some talent and attainments” (Crito Cantab., p. 335, note), burdened as he was with a weak cause and undue confidence in its goodness, would have been at any rate—_impar congressus Achilli_—no match at all for the exact learning, the acumen, the wit, the overbearing scorn of Porson(447). The “Letters” of that prince of scholars, and the contemporaneous researches of Herbert Marsh, have completely decided the contest. Bp. Burgess alone, while yet among us [d. 1837], and after him Mr. Charles Forster [d. 1871], clung obstinately to a few scattered outposts after the main field of battle had been lost beyond recovery(448).

On the whole, therefore, we need not hesitate to declare our conviction that the disputed words were not written by St. John: that they were originally brought into Latin copies in Africa from the margin, where they had been placed as a pious and orthodox gloss on ver. 8: that from the Latin they crept into two or three late Greek codices, and thence into the printed Greek text, a place to which they had no rightful claim. We will close this slight review with the terse and measured judgement of Griesbach on the subject: “Si tam pauci, dubii, suspecti, recentes testes, et argumenta tam levia, sufficerent ad demonstrandam lectionis cujusdam γνησιότητα, licet obstent tam multa tamque gravia, et testimonia et argumenta: nullum prorsus superesset in re criticâ veri falsique criterium, et _textus Novi Testamenti universus planè incertus esset atque dubius_” (N. T., _ad locum_, vol. ii. p. 709).

51. 1 JOHN v. 18. In this verse, according to the Received text, we have the perfect γεγεννημένος of continued effects and the aorist γεννηθείς of completed action used for the same person, although elsewhere in the same Epistle the man begotten of God is invariably γεγεννημένος (ch. ii. 29; iii. 9 _bis_; iv. 7; v. 1, 4). Hence the special importance of the various reading αὐτόν for ἑαυτόν after τηρεῖ, since, if this were to be accepted, ὁ γεννηθείς could be none other than the Only-begotten Son who keepeth the human sons of God, agreeably to His own declaration in John xvii. 12(449). In behalf of αὐτόν we can allege only AB, 105 (a cursive collated by Matthaei), and the Vulgate (_conservat eum_), the testimony of A, always so powerful when sanctioned by B, being nothing weakened by the fact that it is corrected into ἑαυτόν by the original [?] scribe(450), who in copying had faithfully followed his _exemplar_, and on second thoughts supposed he had gone wrong. _All_ other authorities, including copies, versions, and Fathers, א and the rest (C being lost here), have ἑαυτόν, the Peshitto very expressly [and Origen thrice, Didymus four times, Ephraem Syrus and Severus twice each, besides Theophylact and Œcumenius(451)]. We venture to commend this variation as one of a class Dean Vaughan speaks of, which, seeming violently improbable at first sight, grows upon the student as he becomes familiar with it. It must be confessed, however, that St. Paul makes but slight distinction between the two tenses in Gal. iv. 23, 29, and that we have no other example in Scripture or ecclesiastical writers of ὁ γεννηθείς being used absolutely for the Divine Son, though the contrast here suggested is somewhat countenanced by that between ὁ ἁγιάζων and οἱ ἁγιαζόμενοι in Heb. ii. 11. [So that Dr. Scrivener’s view demands considerable sacrifice for its acceptance.]

52. JUDE 5. Here we have a variation, vouched for by AB united, which it is hard to think true, however interesting the doctrinal inference would be. Instead of ὁ κύριος λαὸν ἐκ γῆς Αἰγύπτου σώσας, the article is omitted by אAB, and perhaps by C*, so that it must at any rate resign its place; while for _ΚΣ_ of א (apparently of C*) and the mass of copies, with the Harkleian, we find _ΙΣ_ in AB, 6, 7, 13, 29, 66 (_secundâ manu_), the Vulgate, Sahidic, Bohairic, and both Ethiopic versions. The Bodleian Syriac has yet another variation, ὁ Θεός, in support of which we have the important second hand of C, 5, 8, 68, tol. of the Vulgate, the Armenian (with _ισ_ in the margin), the Arabic of Erpenius, Clement of Alexandria, and Lucifer. The Greek of Didymus has _κσ_ _ισ_, but his Latin translation _ισ_, which Jerome also recognized, although he wrongly supposed that Joshua was meant. While we acknowledge that the Person who saved Israel out of Egypt was indeed the Saviour of the world, we should rather expect that He would be called the Christ (1 Cor. x. 4) than Jesus. There is a similar variation between _χν_, _κν_, and _θν_ in the parallel passage 1 Cor. x. 9.

Lachmann alone reads Ἰησοῦς here, though Tregelles gives it a place in his margin. Westcott and Hort would be acting on their general principle if they received it, but, while setting Κύριος in the text and Ἰησοῦς in the margin, they brand the passage as corrupt, and would be inclined to believe that the original words were ὁ ... σώσας, without either of the nouns. Dr. Hort (Notes, p. 106) points out how slight the change would be from ΟΤΙΟ to ΟΤΙΣ (one Ι being dropped) in the simple uncials of early times.

Fifth Series. Apocalypse.

53. APOC. xiii. 10. Εἴ τις αἰχμαλωσίαν συνάγει, εἰς αἰχμαλωσίαν ὑπάγει. This reading of the Received text is perfectly clear; indeed, when compared with what is found in the best manuscripts, it is too simple to be true (Canon I, Chap. VIII). We read in Codd. אBC: ει (C) τις εις αιχμαλωσιαν ὑπαγει (ὑπάγῃ B), the reading also of those excellent cursives 28, 38, 79, 95, and of a manuscript of Andreas: εἰς is further omitted in 14 (_sic_), and in 92 its echo, in 32, 47, the Bohairic (?), Arabic (Polyglott), and a Slavonic manuscript: and so Tregelles in 1872. The sense of this reading, if admissible at all, is very harsh and elliptical; that of the only remaining uncial A, though apparently unsupported except by a Slavonic manuscript and the best copies of the Vulgate (_am. fuld._ and another known to Lachmann), looks more probable: εἴ τις εἰς αἰχμαλωσίαν, εἰς αἰχμαλωσίαν ὑπάγει: “if any one _is_ for captivity, into captivity he goeth” (Tregelles, Kelly: the latter compares Jer. xv. 2, LXX): the second εἰς αἰχμαλωσίαν being omitted by Homoeoteleuton in the above-mentioned codices. Tregelles (in 1844), Lachmann, Tischendorf, Kelly, Westcott and Hort follow Cod. A, and it would seem rightly.

All other variations were devised for the purpose of supplying the ellipsis left in the uncials. For συνάγει of the common text (now that it is known not to be found in C) no Greek authority is expressly cited except Reuchlin’s Cod. 1, after Andreas (whence it came into the text of Erasmus) and the _recent_ margin of 94. The favourite form of the cursives is that printed in the Complutensian Polyglott: εἴ τις ἔχει αἰχμαλωσίαν, ὑπάγει, after P, 2, 6, 8, 13, 26, 27, 29, 30, 31, 37, 40, 41, 42, 48, 49, 50, 89, 90, 91, 93, 94*, 96, 97, 98, perhaps some six others, a Slavonic manuscript, Andreas in the edition of 1596. The Vulgate, the Latin version printed with the Peshitto Syriac, and Primasius in substance, read “Qui in captivitatem duxerit, in captivitatem vadet,” but (as we stated above) _am. fuld._ (not _demid._) and the best codices omit “duxerit” and have “vadit” (Syr. ܡܘܒܥ ... ܐܙܥ or ܥܙܐ ... ܥܒܘܡ), _which brings the clause into accordance with Cod. A._ The Greek corresponding with the _printed_ Vulgate is εἴ τις εἰς (33 omits εἰς) αἰχμαλωσίαν (ὑπάγει 87), εἰς (ἐς 87) αἰχμαλωσίαν ὑπάγει, 33, 35, 87. Other modes of expression (e.g. εἴ τις αἰχμαλωτίζει εἰς αἰχμαλωσίαν ὑπάγει, 7; εἴ τις αἰχμαλωτιεῖ, αἰχμαλωτισθήσεται, 18; εἴ τις αἰχμαλωτησεῖ, εἰς αἰχ. ὑπ. 36, &c.) resemble those already given, in their attempt to enlarge and soften what was originally abrupt and perhaps obscure.

We submit the two following as a pair of readings which, originating in the pure error of transcribers, have been adopted by eminent critics in their unreasonable and almost unreasoning admiration for Bengel’s canon, “Proclivi orationi praestat ardua.”

54. APOC. xv. 6. In the transparently clear clause ἐνδεδυμένοι λίνον καθαρόν Lachmann, Tregelles in his text, Westcott and Hort, present the variation λίθον for λίνον “arrayed with stone,” i.e. precious stone, for which καθαρόν “clean” would be no appropriate epithet. Dr. Hort (Notes, p. 139) justifies what he rightly calls “the bold image expressed by this well-attested reading” by Ezek. xxviii. 13 πάντα λίθον χρηστὸν ἐνδέδεσαι (or ἐνδέδυσαι), σάρδιον καὶ τοπάζιον κ.τ.λ., but that was said of a king of Tyre, not of the angelic host. The manifestly false λίθον is only too “well-attested” for the reputation of its advocate, AC, 38 in the margin, 48, 90, the best manuscripts of the Vulgate (_am. fuld. demid. tol. lips._4.5.6, &c.), though not the printed editions. Andreas knew of the variation without adopting it: Haymo and Bede also mention both readings. Cod. א reads καθαροὺς λίνους with the Bohairic, and so helped to keep Tischendorf right: Tregelles sets this form in his margin. For λίνον or λινοῦν or λην- we have all the other manuscripts and other authorities, including BP, that excellent cursive Cod. 95, Primasius. Between the two forms with ν we should probably choose λινοῦν of B, [7], 14, 18, 92, 97, as λίνον seems to belong to the raw material in a rough state. The later Syriac has ܒܬܢܐ (or ܐܢܬܒ) (χιτῶνα), which admits of no ambiguity.

55. APOC. xviii. 3. For πέπωκε of the Received text, or πέπωκαν of Lachmann and Tischendorf, Tregelles (whose margin has πεπτώκασιν), Westcott and Hort in their text (not margin) have πέπτωκαν. Dr. Hort has no note on this place, but treats it in his index of “Quotations from the Old Testament” as a reference to Isa. li. 17, 22 (ἡ πιοῦσα τὸ ποτήριον τῆς πτώσεως) and to Jer. xxv. 27 (πίετε καὶ μεθύσθητε ... καὶ πεσεῖσθε), with the notion of stumbling through drink. What is required to complete the parallel is some passage in the Septuagint wherein πέπτωκαν stands alone, whether τοῦ οἴνου be in the text or not, and, in the absence of such parallel, πέπτωκαν must be regarded as incredible on any evidence. Yet πέπτωκαν or the virtually identical πεπτώκασιν is found in אAC, in B, 7, 8, 14, 25, 27, 29, 91, 92, 94, 95 (πέπτωσι _primâ manu_), the Bohairic and Ethiopic. The alternative reading πέπωκαν or πεπώκασιν (πέπωκε 96) occurs in P, 1, 18, 31, 32, 36, 37, 38, 39, 47, 48, 49, 50, 79, 87, 90, 93, 97, 98, the Latin and later Syriac. Thus the very versions are divided in a case where the omission of a single letter produces so great a change in the sense.

56. APOC. xxi. 6. Καὶ εἶπε μοι, Γέγονε. ἐγώ εἰμι τὸ Α καὶ τὸ Ω. Here the true reading Γέγοναν “They are done” (adopted, with or without εἰμι after ἐγώ, by Lachmann, Tischendorf, Tregelles, Kelly, Archdeacon Lee in the “Speaker’s Commentary,” Westcott and Hort) is preserved by Cod. A, whose excellency is very conspicuous in the Apocalypse: its compeer C is defective here. The very valuable Apoc. 38 confirms it (γεγόνασιν), as did אc, but the whole word was afterwards erased: the interpreter of Irenaeus renders _facta sunt_, and this is all the support A has. The first hand of א with BP, 1, 7, 8, 11, 12, 13, 26, 27, 31, 32, 33, 35, 47, 48, 79, 87, 89, 91, 92 (_hiat_ 14), 93, 96, 97, 98, the Armenian, Origen (_quod mireris_), Andreas, Arethas, with the Complutensian, read γέγονα, most of them omitting either the ἐγώ or the ἐγώ εἰμι which follows. Erasmus was too good a scholar to adopt from Apoc. 1 a meaning for γίγνομαι which it cannot possibly bear, and seems to have got his own reading Γέγονε (though he recognizes that of Apoc. 1 in his Annotations) from the Vulgate _factum est_, which is confirmed by Primasius: it probably has no Greek authority whatsoever. The Syriac printed with the Peshitto (commonly assigned to the sixth century) appears, like the hand which followed אc, to omit γέγονα, as do the Bohairic and Ethiopic versions, with _lux._ of the Vulgate. Those which read γέγονα yet retain the following ἐγώ (אBP, 7 and some others) obviously differ from the true reading γέγοναν by the single stroke which in uncial manuscripts was set over a letter to represent _nu_, especially at the end of a line, and so avoid the monstrous rendering necessarily implied in 1, 8, 93, 96, 97, 98, “I have _become_ alpha and omega, the first and the last.” P accordingly puts the proper stop after γέγονα.

‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐‐

God grant that if these studies shall have made any of us better instructed in the letter of His Holy Word, we may find grace to grow, in like measure, in that knowledge which tendeth to salvation, through faith in His mercy by Christ Jesus.

APPENDIX A. ON SYRIAC LECTIONARIES.

A very interesting group of Syriac manuscripts is found in the collections of Syriac MS. Lectionaries which have descended to us. That the number of them is large may be inferred from the fact that thirty-five may be found in the British Museum alone (Catalogue, i. pp. 146-203).

Syriac Lectionaries are of two classes, (i) those according to the Greek Use, and (ii) those according to the native Syriac Use. The former, or _Malkite_ Lectionaries, may be dismissed from the present enquiry. They are only Greek works in a Syriac dress, and their value is historical rather than critical(452).

The true Syriac Lectionaries, whether Jacobite or Nestorian, follow as to their main features the Greek Lectionaries which have been described in our first volume, coming under two main classes, Evangelistaries and Apostolos(453). But they present one important contrast. In both families of Syriac descent, the Ecclesiastical year begins with Advent, and not, as in Greek Lectionaries, with Easter; and in general the arrangement is similar in both, so that the system must at least be of considerably greater antiquity than the days of the schism. In some of the Jacobite copies the text of the Harkleian revision has been substituted for the ancient Peshitto. Some include Lessons from the Old Testament. Some contain a Menology. In a few instances the Lessons for special festivals form a separate volume.

The majority of the Syriac MS. Lectionaries are comparatively late, but others possess an antiquity which, in the case of some MSS., would be considered remarkable. The British Museum copies, Add. 14,485 and 14,486, are each dated A. GR. 1135 = A.D. 824. Others must be referred to the same century. Add. 14,528, foll. 152-228 (an Index), and the leaf in Add. 17,217, appear to be three centuries older. Another sixth century MS., Add. 14,455 (the Four Gospels), contains many Rubrics, a pr. m. in the text, besides those in the margins by later hands, such as occur in MSS. of all ages. When to these facts we add the consideration already mentioned, that the same system was in use in both branches of the Syrian Church, we see the importance of the testimony of works of this class. They are very ancient ecclesiastical records from the unchangeable East. Like Greek Lectionaries, they are difficult to use, because of their arrangement of Lessons in the succession ordered by the calendar: they are of course public documents, and in consequence possess an importance above that of copies which were in many cases the property of private persons, and may have been carelessly and cheaply prepared. Yet it would not be right to claim for copies of a version a position quite as important as that held by the Greek service-books, since the evidence of versions, as well as of quotations in ancient writers, is only subsidiary. Nevertheless, in the fact that the number of ancient Greek copies of the New Testament is relatively small as compared with the early copies of the Peshitto version, we are warned not to underrate Syriac Lectionaries, though they are of less value for the Syriac, on account of the large number of very ancient and well-written copies which have come down to us, such as those which have been enumerated in our account of the materials for ascertaining the text of the Peshitto.

APPENDIX B. ADDITIONAL BOHAIRIC MANUSCRIPTS IN EGYPT (1893).

Cairo 1 [1184] attributed and possible date, fol., _chart._, ff. 290, 27 × 18·6 (23), κεφ., Copt. Gr., _Am._, _Eus._, _pict._ Evann., Copt., restored under patronage of Athanasius, Bp. of Abutij, 1794, whose statement gives date 900 of the martyrs. Dedication to monastery of St. Antony in the eastern desert; now in the library of the Patriarch in Cairo, numbered 12 and 14.

Ancient writing begins St. Matt. v. 25, continues to St. Luke x. 2. begins St. Luke x. 27, continues to St. Luke xxii. 52. begins St. Luke xxii. 66, continues to St. Luke xxiv. 53. begins St. John i. 31, continues to St. John xix. 24.

Cairo 2 [1291], fol., _chart._, ff. 409, 26·9 × 18 (24, 25), κεφ., Copt. Gr., _Am._, _Eus._, _pict._ (pictures of SS. Mark, Luke, and John). Evann. Copt. Arab. Written by Deacon Barsuma, mended by Michael of Akhmîm, monk of monastery of Siryani (Nitrian), under patronage of Cyril, 112th Patriarch, 1878. Dedication to monastery of St. Barsuma, called Al Shahrân, 1329; now in the library of the Patriarch in Cairo, numbered 12 and 14. Quires numbered in Syriac. Same text as Paris 15.

Cairo 3 [xviii], fol., _chart._, ff. 342, 22·8 × 13 (29), _Carp._ and _Eus. t._ at end of St. Mark, _proll._, κεφ. _t._, κεφ., Copt. Gr., _Am._, _Eus._, _pict._ Evann. Copt. Arab. Written by Michael Pilatos, who gives his name in the duplicate book at Alexandria, and who wrote the Epistles and Acts below in 1714. In the library of the Patriarch in Cairo. Text same as Curzon 126.

Cairo 4 [1327], fol., _chart._, ff. 395, 27·5 × 17·8 (27), κεφ., Copt., _Am._, _Eus._, _pict._ Evann. Copt. Written by Thomas. Dedication to the Church of St. Mercurius in old Cairo, where it now rests. Text of St. Matt. is same as Brit. Mus. 3381.

Cairo 5 [1257], fol., _chart._, ff. 382, 26·4 × 19 (25), _prol._ St. Luke, Capp. Copt. _Am._, _Eus._, _pict._, _mut._ Evann. Copt. Arab. _Mut._ St. Matt. i-iv. 5, St. Mark i. 1-7, St. John i. 1-21; a few leaves restored. Written by monk and priest Gabriel, who wrote in the house of Ibn ´Assâl; now in the Church of Al Moallaqah in old Cairo. Text similar to manuscript of Göttingen.

Cairo 6 [1272], fol., _chart._, ff. 328, 24·9 × 17 and 25·7 × 18. Epilogue to St. Matt. Κεφ., Copt., _Am._, _Eus._, _pict._, _mut._ Evann. Copt. St. Matt. by more recent writer. SS. Mark, Luke, and John written by original scribe, Simon Ibn Abu Nasr. Text of St. Matt. similar to Bodl. vii. In the Patriarchal Library in Cairo.

Cairo 7 [xiv], 4to, St. Luke, restored under Bp. Athanasius of Abutij. Text unimportant.

Besides several which are too late to have any critical importance.

APOCALYPSE.

1. [xix], folio.

ALEXANDRIA 1 [xviii], fol., paper, duplicate of Cairo 3, by same writer. Evann.

2. [xix], SS. Matt. and Mark.

3. [1861], St. John, Copt.

DAYR AL MOHARRAQ, nr. Manfalût on the Nile (station and telegraph Nasâli Gânûb).

1. [1345], fol., _chart._, 22·5 × 14·2 (27), _Carp._ at end. _Mut._, but fairly perfect, _pict._, and richly glossed. Text unimportant. Evann. Copt. Arab.

ST. PAUL, CATH., ACTS.

1. [xii?], probably of same date as Evann., Cairo 1, fol., _chart._, ff. 432, 25·6 × 18·2 (24), κεφ., Copt. Gr. Thess., Heb., Tim., _pict._, Copt.: restored Rom. and 1 Cor. i-xvi. 12, copious glosses in Arabic.

2. [xiv], fol., _chart._, 26 × 18·5 (25), κεφ., Copt. Gr., _pict._ Philemon, Hebr., Copt.

INDEX I. TEXTS OF THE NEW TESTAMENT ILLUSTRATED IN THIS TREATISE.

(Where the page is given alone, the reference is to the first volume. _n_ indicates _note_.)

ST. MATTHEW. i. 18 II. 321-3