inflexion and syntax, which occur in this version: not a few will also be met with, though much more rarely, in the other Old Latin Versions, including the Vulgate.

(1) Nouns of the fourth declension, though often used with their proper terminations, are frequently turned into the second: actus Acts xix. 18; adcubitus Luke xi. 43; xiv. 7; xx. 46 (addub.); concursus Acts xix. 40; conspectus Luke iv. 7; xiii. 26; xv. 19; 21; xix. 27; xxi. 36; Acts x. 30 (not vv. 31, 33); cornum Luke i. 69; cubitus Mark xii. 39; fructus Matth. xii. 33; xxi. 34 (both); 43; Luke vi. 43 bis; 44; xii. 17; 18; xx. 10; gradus Acts xii. 10 (graduus acc. pl. ibid. xxi. 35); habitus Acts xii. 21; intellectus Luke ii. 47; intercessus ibid. xxii. 59; magistratus Acts xvi. 19; 22; 35; porticus John v. 2; spiritus Matth. i. 20; Luke i. 67; ii. 27 (spo, but v. 26 spu); iii. 16; Mark i. 25; v. 8; ix. 25 (both); xii. 36; Acts i. 5; iv. 8; 31; vi. 10; vii. 51; x. 38; xi. 16; 24; xiii. 4; 9; xv. 7; 29; 32; xvi. 6; xix. 21; xx. 22; tonitrum John xii. 29; vultus Luke xxiv. 5. We find the genitive of such nouns in -ui Luke ix. 55; Mark iii. 17; xi. 21; Acts xx. 19, and in -u Acts vi. 5; in -um in the plural Mark xi. 13. Notice also the ablatives caelu, Luke xvii. 24; domu Acts xvi. 34; tyru Luke x. 13; and the genitive dolus Acts xiii. 10; somnus John xi. 13.

(2) Neuter nouns of the second declension are sometimes made masculine: as donum Acts xi. 17; foros Matth. xxiii. 7; sabbatum once (Acts xiii. 27); signum Luke ii. 34; templum once (Mark xiv. 58); verbus John xxi. 23; Luke iv. 32; Acts xix. 38; domus is masc. Luke ix. 4; manus masc. ibid. xxii. 21; apex fem. xvi. 17; porticus masc. Acts iii. 11 (with am.); valetudo masc. Acts v. 15: on the other hand cibus is neuter Mark vii. 4, and kumerus Matth. xxiii. 4. We find for κωμη castellus Luke ix. 56; xvii. 12; xix. 30; Mark xi. 2 masc.; but castellum neut. certainly seven times and probably eight more. Illum appears to be neut. Mark x. 15; Acts xii. 1; also eum Matth. xxiii. 18; 20: we read quendam civitatem Acts xxi. 16; marem Mark iii. 7; salem ibid. ix. 50. For rete we find retia and of the first declension in all places, viz. Matth. iv. 18; 20; 21; John xxi. 6; 8; 11 bis; Luke v. 2; 4; 6; Mark i. 16; 19: sidona Matth. xv. 21, only: tenebra John vi. 17 only: baptismus is preferred to baptisma Luke xii. 50; xx. 4; Mark i. 4; vii. 4; x. 38 (not v. 39); Acts x. 37, but not in the ten other places. Lystra is neut. pl. Acts xiv. 6 even against the Greek, but fem. sing. in its four other places (xiv. 20; 21; xvi. 1; 2), once (xvi. 2) against the Greek: we have socra Matth. x. 35 only: columbus Luke ii. 24 only. In the accusative of the third declension we have securem Luke xiii. 7, but testim Acts iv. 33: in the ablative rude Luke v. 36 bis; nave Mark v. 2; mare Luke xxi. 25; Mark v. 13 semel; vi. 47: even ae for e in regae Acts vii. 10; but i for e in sidoni Luke x. 13; peccatori ibid. xv. 10; veteri Mark ii. 21; corpori ibid. v. 29; morti vii. 10; sermoni Acts xv. 32: in the plural nominative -ae is put for -es in lampadae Matth. xxv. 8: civitatium is gen. pl. in Luke v. 12; vi. 17. Moysi is the genitive form in Matth. xxiii. 2; John ix. 28; Luke ii. 22; xxiv. 44 (-e); Mark xii. 26; Acts xiii. 39; xv. 1; 5; iohanni Luke vii. 24: alio is dative Luke xiv. 31; xvi. 7: the unusual quemquem Mark xii. 14; Acts ii. 3: ipsud Acts xvi. 35: a few adjectives in us and -er sometimes are formed in is, or vice versa, as austeris Luke xix. 21 (not v. 22); infirmis Matth. xxv. 43 (not v. 44); xxvi. 41; John v. 13 only; paupera Luke xxi. 2; 3; pleres Acts xvii. 12; subdoles Acts xvii. 5; uberam Luke viii. 8, but uberes ibid. xii. 16; unanimes Acts i. 14. Merely barbarous are fratrorum Matth. xxv. 40; salutarem Luke iii. 6; hominorum ibid. xiv. 24; stadios xxiv. 13 (not John vi. 19), so Cod. a; interfecti (dat.) Acts viii. 1; progeniebus ibid. xv. 21; novius xvii. 21: perhaps altari (gen.) Luke i. 11.

(3) In verbs the chief anomalies occur in the compounds of eo, which make -iebam &c. in the imperfect, iam -ies &c. in the future: such are exiebat Luke vi. 19; exiebant John viii. 9;

f

Luke iv. 22; 41; Acts viii. 7; exies Luke xii. 59; exiet Matth. ii. 6; John x. 9; Mark x. 12; periet John x. 28; xi. 50; Luke xxi. 18 (but peribunt Luke v. 37); transiebat ibid. v. 15; transiebant Mark ix. 30; pertransiebat Luke xix. 1; pertransiebant Acts xvi. 6; transiet Matth. v. 18; Mark xiii. 30 (not v. 31) and in six other places; pertransiet Luke ii. 35; rediebant (but also ibant) Mark vi. 31; abientes Luke viii. 14; xxii. 13; exientes v. 2; Mark vi. 34; veniunt Luke xii. 6; venitum est John xii. 5; veniri Mark xiv. 5. Possum, fero and odi are also conjugated incorrectly: potebat Luke xix. 3; poterint ibid. xxi. 15; differitis xii. 7; 24; conferitis xxiv. 17; adfers (imperat.) John xx. 27 bis; Luke xiii. 7; offers (imperat.) Mark i. 44; adferi (infin. pass.) ibid. vi. 27 (adferent for -unt vii. 32, as often elsewhere, may be a clerical error); odies Matth. v. 43; odiet Luke xvi. 13; odient Matth. xxiv. 10; both odit and odivit John xv. 18; odierunt ibid. xv. 24; 25; odierint Luke vi. 22; odientibus Matth. v. 44; Luke vi. 27; even fieretur John xiii. 2. The simple pareo for appareo occurs Matth. vi. 5; ix. 34; xiii. 26; xvii. 3 (paretur depon. xxiii. 27, but parent Vulg.); xxiii. 28 (so Vulg.); xxiv. 30 (so Vulg.); xxvii. 53; Luke xxiv. 11; (but conparuit v. 31). Other anomalies in conjugation are lugunt Matth. v. 4; fodiit Matth. xxi. 33; nubor pass. Matth. xxii. 30; Luke xvii. 27; xx. 34 (joined with pariuntur); 35; Mark xii. 25; loquor pass. Matth. xxvi. 13; linuit John ix. 6; 11; pariret Luke i. 57; ii. 6; habibat (from habeo) ibid. vi. 8; custodiabatur viii. 29; stupuebant ix. 43; egeri xv. 14; consolatur pass. xvi. 25; paeniteor xvi. 30; xvii. 4; certabatur depon. xxii. 59; coxerunt (from cogo) xxiv. 29; proficebat Mark v. 26; obstipuerunt ibid. v. 42; secuntur vii. 5 (so Cod. e sometimes); petieremus x. 35; respondite xi. 29 (not v. 30); taediari xiv. 33; respondis xv. 4; possidit and crepavit Acts i. 18; locuntur ibid. ii. 7; serpiat iv. 17; obstupiscebat viii. 13; quaesire xiii. 7; resistabat xiii. 8; decedisset xiii. 13; silerunt xv. 13; perconfirmor depon. xv. 32; extorsuit xvi. 15; vetatus xvii. 15; vellit xvii. 18; 20; conventi xxi. 18; adsistans xxii. 20. In Luke xv. 6 ovvxapnтe is barbarously rendered cum gaudete (but not in v. 9), so Acts xvi. 13; compare Luke vi. 4.

(4) The most remarkable peculiarity of the syntax in this version is the frequent habit of omitting the antecedent to a relative: such cases are seen in Matth. xxiii. 31; John iv. 34; v. 10; vi. 39 (not v. 38); Luke i. 45; iii. 7; vi. 4; xiv. 10; 15; xix. 24; xxi. 26; xxii. 21; xxiii. 25; 43; xxiv. 33; Mark iii. 34; v. 40; x. 23; Acts iv. 21; vi. 1; xvi. 10; 13; xvii. 11; xxii. 11: the relative is omitted in Mark xi. 21; Acts xiii. 2. In expressing prohibitions non is more frequent than ne, e. g. John xiv. 1; 27: vuwv is vestris John xiii. 21, nostrorum renders nuwv Acts xvii. 27. Otherwise, though there is a rudeness in the whole style approaching to barbarism (e. g. Luke xxiv. 31 fin.; Acts xiii. 1; 10; 29), yet there is seldom found any notable violation of the rules of Latin grammar, except to accommodate it to the parallel Greek, on which point we have already spoken at large (see p. xxxii). We find however suaserunt turbas Mark xv. 11 against the Greek: so vetare eis Luke xviii. 16. Sometimes the Latin softens down a loose construction of the original (e. g. Acts iii. 13; iv. 3), while in a few instances it approaches nearer to classical propriety than does the Vulgate (e. g. Mark xiii. 20 ulla caro, yet it is omnis in Matth. xxiv. 22). It is even elegant at times, e. g. opai facile puto John xxi. 25. In the Acts it nicely discriminates throughout the Jewish from the Christian Tрeσẞureрot, where the Vulgate completely fails.

(5) As in other specimens of provincial Latin, we find h very often omitted, and as often inserted, improperly. Instances of the former are eroden Matth. ii. 12; umerus, ibid. xxiii. 4; Luke xv. 5; ypocrytae Matth. xxiii. 29; Luke xi. 39; ebraice John v. 2 (-aeice); xx. 16; ora ibid. v. 35; Mark vi. 35 bis; xiv. 35; 41; xv. 33 semel; umorem Luke viii. 6; ac (i. e. hac) ibid. xii. 20; abeo xii. 50, and both abet and habet xix. 26; umido xxiii. 31; abetis Mark iv. 40; aesitaverit ibid. xi. 23; exortor Acts ii. 40; xx. 2; aebraeos ibid. vi. 1; ospitor x. 6; 18; xxi. 16,

peribent x. 43; ymnum xvi. 25; exibere xvii. 31; esitassent xvii. 34; ellada xx. 2: of the latter exhortus Matth. xiii. 5; Mark iv. 5; 6; harunt Matth. xiii. 6; haruit ibid. xxi. 19; 20 (not John xv. 6); hostendite xxii. 19; haridam xxiii. 15; habundabit xxv. 29; habe (i. e. ave) xxvi. 49; xxvii. 29; xxviii. 9; Luke i. 28; xxiii. 37; Mark xv. 18; harundinem Matth. xxvii. 29; 30; 48 (not xi. 7); Luke vii. 24; heliam Matth. xxvii. 47; 49, and in all the 24 other extant places; hosteis John xx. 19; 26; hieris Luke ix. 57; holus xi. 42; honeratis and honus xi. 46; Acts xv. 28; hiericho Mark x. 46; hebrii Acts ii. 15; habire ibid. iv. 15; hopus v. 38; hemulati vii. 9; harena vii. 24; horabit x. 9; hiconio xiii. 51 (not xiv. 1); hemulatores xxi. 20.

(6) Other peculiarities of spelling, which prevail indeed through every page of this version, are the interchange of b and v (more rarely b is turned into p, e. g. Matth. ix. 32; xiv. 35; xxii. 44; Acts xvi. 20, where b is washed out under p, and both are by the first hand) which the corrector G has emended in two large portions of the work (see p. xxvi), and the placing of ƒ for ph and d for t in such words as at (ad ille is the universal form employed), constitudo Luke xiv. 32, capud Acts xvi. 12 &c.: at is also now and then put for ad, as in Acts xxi. 37: see too aliut ibid. xix. 32: p stands for m, Acts xiii. 34. The diphthong ae is perpetually expressed by the simple e and vice versa: thus with the vocatives plenae and inimicae Acts xiii. 10: praesbyteri Luke xx. 1; Acts xxi. 18 only: raeaedificabo Acts xv. 16: praetium is used eight times, pretiosi only in John xii. 3; we find quaerere Acts xvii. 27, but quero &c. occurs twenty times, questio three; caecidit John xi. 32; saepes Luke xiv. 23; Mark xii. 1; saedeo Matth. xx. 30; Mark x. 46 only saeniorum ibid. vii. 5 only; vadae Matth. xiii. 14: like every other Latin manuscript1 d invariably has caelum: so faenum always (five times): faenus and faenero four times, fenero Luke vi. 34 semel: cena is read 13 times, coena never: jajuno &c. occurs 7 times, jejuno &c. 15, in Matth. ix. 14, 15 varying in consecutive verses: talantum is found in Matth. xxv. twelve times, talentum thrice (both occur v. 28): anticus (like secuntur and locuntur named above, p. xlii) Luke ix. 8 only; thus inicus Luke xvi. 10; 11; xxiii. 41: norus ibid. xii. 53 bis: thensaurus in all eleven places and temptatio in all nine: forsitam Luke xi. 20; xx. 13; Acts xi. 18; xii. 15; xviii. 14 &c.: jenuam Mark xi. 4 only: abraam Matth. iii. 9 bis; John viii. 33; Luke xiii. 28; xvi. 27, but abraham 27 times: istrahel &c. always except in Luke xxiv. 21: patriaarcha Acts ii. 29 only (compare TerрaapXno in Codd. Sinaiticus, Ephraemi and others): santus Acts i. 8 (not vv. 2; 5); iv. 30; vii. 33; xv. 29: passares Luke xii. 6 (not v. 7); carcare &c. xxi. 12 only, but carcere &c. 28 times: clodus in ten places, claudus John v. 3 only: cludo and clusum in all eleven places. Under this head may be brought the familiar practice of writing hii, hiis for hi, his, and the contrary habit of putting fili 44 times for the gen. sing. or nom. pl. of filius (yet not in Mark xiii. 12; Acts iii. 25), and more rarely filis for the dative or ablative plural.

Luke iii. 20;

(7) The abbreviations usual in the Greek text have been enumerated already (see p. xviii): in the Latin the chief are ihs (Cod. Laud. 35 has ihesus, but hiesum Acts xix. 5), xps, ds, dms, sps, and their several cases (ihn John xii. 9; Luke v. 12 for ihm is rare) as is usual in documents of the oldest class. Deus and spiritus (dominus Matth. xiii. 27; Acts xiii. 10) are sometimes written in full, the former often retaining the mark of abridgement (-) notwithstanding, as is likewise the

1 "Coelum is a spurious form, invented about the beginning of the sixteenth century, in conformity with a ridiculous etymology." Munro on Conington's Virgil, Journal of S. and C. Philology, 1860. In Cod. f coelorum, Matth. xx. I must be a mere error either of the editor (Martianay) or of the scribe.

2 On the tomb of that illustrious scholar Isaac Casaubon [d. 1614] in the S. W. aisle of the transept of Westminster Abbey, xpo of the original epitaph has been changed by some ignorant stone-cutter into the barbarous

xto.

case in some parts of Cod. Claromontanus, especially about Romans i. Thus dei is met with 122 times (but never in the Acts), deo 24 times (in the Acts only vii. 40; x. 4), deum only in John vi. 46; x. 33. For the more usual form dms &c., we find dns &c. 16 times in S. Matthew (both occur in xxii. 44), in Luke xxiii. 40, and always (84 times) in the Acts, except dmi iii. 19: dms is the form preferred by Codd. Vercellensis and Claromontanus, dns by Codd. Palatinus, Amiatinus, and Fuldensis. We may possibly think that minute peculiarities of this kind slightly confirm the impression of those who deem the translator of the Acts a different person from him who rendered the Gospels (see p. xxxiii, note 2)1. Add to this that he alone has dum for dm (0), Acts vi. 11; xi. 17; xii. 5; a form also found in Cod. Palatinus: although in S. Luke we see dom i. 16; 46, and in nine other places (besides six in S. John, as also in Cod. Claromontanus); dome (vocative) in Luke x. 40. In Cod. Laud. 35 there are no abridgements in the Latin.

Of compendia scribendi, as distinct from abridged words, the Latin of Cod. Bezae has but few. At the end of a line over the last letter stands for m; in Codd. Palatinus and Claromontanus the line stands for m or n indifferently: a single point (usually the upper point, but sometimes the middle) indicates a termination omitted, e. g. hominib. Matth. x. 33; ossib. ibid. xxiii. 27; sublatisq. Acts xiii. 19. Occasionally an unfinished word has not such point, apparently through oversight; e. g. faciem Matth. xxvii. 22 for faciemus (in later manuscripts like the Cod. Augiensis -us would be indicated by the apostrophus), humiliab Luke iii. 5 (see Adnotationes, Fol. 194 a). The punctuation of the Latin is on the same plan as that of the Greek, described above (p. xviii): in Cod. Palatinus (if we may judge by Tischendorf's facsimile page) the single middle point is rare primâ manu, but more often added by a later pen.

Some grounds for believing that this manuscript was written in the region where it was eventually found in the sixteenth century were stated in the last chapter (p. xxxi): the following philological reasons, so far as they go, would suggest the same conclusion of its Gallic origin.

(1) In addition to the unclassical and indeed ungrammatical use of de to express the genitive in Acts ii. 30; xx. 17 (see p. xl), we find in the style of d distinct traces of the employment of habeo as an auxiliary verb, which is well known to be a notable characteristic of the modern languages of Western Europe (of the French as much as any) as distinguished from the Latin whence most of them sprung. In Mark xiv. 27 σкavdaliσaola (i. e. -0e) is rendered scandalizari habetis by d, but scandalum patiemini by ac, scandalizabimini by f and the Vulgate. Habeo is thus used three times to render peλλw, Luke x. 1; xix. 4; Acts i. 5, although the Greek word is translated by incipio 25 times (sometimes very awkwardly), 15 times by the future participle, three times in other ways (Mark xiii. 4; Acts xvi. 27; xxi. 37).

Two or three peculiar words, which better scholars may perhaps hereafter add to, point to the same conclusion as regards the nationality of the translator. Scholz (N.T. Proleg. p. xxxix)

1 Besides the instances before given we may notice that apxiepeur, which is rendered princeps sacerdotum in all places in S. Matthew (24) and S. Luke (16), in S. John princeps 4 times, princeps sacerdotum twice, in Mark xiv. 47 princeps sacerdos, in the other 19 places of S. Mark summus sacerdos; is in the Acts pontifex iv. 23; v. 17; 21; vii. I; pontefex iv. 6; pontefix (iepevo) v. 27. The reading is ïepewo also in xix. 14, where alone we find sacerdotis.

2 Scholz's examples had occurred to me in complete forgetfulness of what he had written long ago. Like

other bad reasoners, he overlays his really effective arguments by others obviously futile. Thus he urges for the Gallic origin of Cod. D (among others) refectio (kатаλνμа) Mark xiv. 14, though the word is in the Vulgate; sideratus (kullos) Matth. xv. 30; 31; xviii. 8 (but in Mark ix. 43 debilis with the Vulgate), an expressive term found in Pliny (in regard to this word, however, he only follows Mill); and natatoria piscina John v. 2, a mere error of the translator, who unites the two separate words used by the Vulgate for rendering коλuußnepa in the places where it is found (v. 2; 4; 7 piscina; ix. 7; 11 natatoria).

and others have noticed soniis (pepiuvaio) in Luke xxi. 34 only, for which ae have solicitudinibus, bf cogitationibus, c and the Vulgate curis. That sonius, which is not a Latin word at all, is connected with soinus and the French soin is plain enough, and Ducange cites from one Latin and Greek Glossary "somnium ppovтis idurikos," from another "somnior peppvw," whence was corrupted sonius, thence soinus and soin ("Nisi competens soinus eum detineat" Leges Henr. 1. Regis Angliae cap. 29 in Ducange Medii Aevi Latinitas, sub voce Sunnis).

I.

Less certain is the inference drawn from involet as a translation of kλen in John x. 10 only, all the other versions having furetur in that place. Involo is rendered by Ducange per vim auferre, and compared with the French voler, but Servius the Commentator on Virgil, in the 5th century, says "Vola dicitur media pars manûs...unde et involare dicimus, quum aliquid furtim volâ manûs subtrahitur." The best classical example of this use of the word (certainly a very rare one) is Catull. Carm. xxv. "Remitte pallium mihi, meum quod involâsti."

Of applontat (paooe) Mark ix. 18, another of Scholz's examples, I find no notice in Facciolati, Ducange, or other such books. It must be connected with planta, supplanto. Bentley, who read applantat in his hasty fashion, adds "et hoc est allidit humo" (Ellis, Bentl. Crit. Sacra, p. 9). Such forms as sconspectu Acts vii. 46, and yet more scoriscatio Matth. xxiv. 27; scoruscus Luke xvii. 24; scorusco xvii. 24 bis; xxiv. 4 (aσrpay and dorрánтw, but fulgur Matth. xxviii. 3; Luke x. 18) savour more of the initial impure s of the Italian, which plainly sprung from the Latin ex, e. g. sbarcare, scarnare.

CHAPTER IV.

ON THE CHARACTER OF THE GREEK TEXT OF CODEX BEZAE.

Ir results from our investigations respecting the parallel version in this manuscript, that although replete with philological interest as a specimen of vernacular Latin just before it merged into the mediaeval language of the South of France, very little weight can be given to its readings even in those places (comparatively so few) in which it differs from its Greek original. The purpose of the scribe (or at any rate of his immediate predecessor) was simply to copy on the one page of an open leaf and to translate on the other, a very ancient and curious book, arranged to his hand in verses or σríxo, whose present loose and inartificial divisions shew that it must itself have been derived from older documents wherein the orixou had been distributed on an elaborate and regular system, which the carelessness of the writer of the immediate prototype of Codex Bezae has gone far to break up and obscure (see p. xvii). In this the last Chapter of our Introduction we shall aim at proving that the text of Codex Bezae, as it stands at present, is in the main identical with one that was current both in the East and West as early as the second century of our aera. It may very well have been brought into Gaul by Irenaeus and his Asiatic companions about A.D. 170: in some of its most characteristic features it resembles the Syriac versions made at one extremity of Christendom, the citations of the Latin Fathers at the other. Whether Codex Bezae (D) and its allies approach nearer to the verity of the inspired writings than do some of our chief authorities whose extant vellum may be a little older, such as Codd. Sinaiticus (N), Vaticanus (B), Alexandrinus (A) and Ephraemi (C), is too large a question to be entered upon in this place, even if we were in possession of materials for arriving at a definite conclusion, which there is much cause to fear we are not and perhaps never may be. If the high antiquity of the