construction, but in plain prose we cannot very easily find a parallel to it. Reference to any commentary will satisfy the reader of the contorted nature of the explanations to which recourse must be had, if the usual method of taking ő, ávτíTUTOV and Báτioμa, as nominatives in apposition with each other, is adopted. Very early this difficulty appears to have been noticed, and ò was altered into in a large number of MSS. But is it necessary to consider ò as the subject of the sentence at all? Is it not at least equally legitimate to take into consideration such passages as Rom. VI. 10, ὃ γὰρ ἀπέθανε, τῇ ἁμαρτίᾳ ἀπέθανεν ἐφάπαξ, ὃ δὲ ζῇ, ζῇ τῷ Θεῷ; Gal. II. 20, ὃ δὲ νῦν ζῶ ἐν σαρκὶ, ἐν πίστει ζῶ; 1 Cor. III. 2, γάλα ὑμᾶς ἐπότισα; and Thucyd. VI. 11, ἅπερ οἱ Εγεσταῖοι ἡμᾶς μάλιστα ἐκφοβοῦσι, and construe ὃ as the cognate object instead of the subject of owe. Thus as ὃ ἀπέθανε = ὃν θάνατον ἀπέθανε, and ὃ ζῶ = ἣν ζωὴν ζῶ, so ὃ καὶ ἡμᾶς ἀντίτυπον νῦν σώζει βάπτισμα = ἣν σωτηρίαν καὶ ἡμᾶς ἀντίτυπον νῦν σώζει βάπτισμα, where it is immaterial whether we connect ȧvTiTUTTOV with the accusative à or the nominative Báπτioμа. The sense thus obtained is particularly simple. The events previously alluded to are stated as having taken place, "while the ark was preparing, by entering into which few, that is eight, souls were brought safe through water; a salvation which, being antitypical, baptism now confers upon us [lit. 'saves us']; (not the putting off of fleshly filth, but an honest [lit. 'good conscience's'] contract with reference to God) by virtue of the resurrection of Jesus Christ." There is thus no necessity of forcing into comparison the destroying water of the flood and the saving water of baptism, no awkwardness of expression or meaning left, and the figurative correspondence between the flood and baptism is complete. As Noah and his family were brought safe through the waters of the flood into a new state of life in the new order of things, so are we brought safe through a figurative death in baptism into a new order of things as Christians, to whom "the old things have passed away, lo! all things have become new."

A. H. WRATISLAW.

EPISTLE TO THE ROMANS.

IN a paper on Two neglected facts bearing on the Ignatian Controversy in the last number of this Journal, I stated (p. 49 sq.) that the heading of the Ignatian Epistle to the Romans in the only extant MS (Colb. 1451) of the Middle Recension is πρὸς Ρωμαίους. This statement was made on the combined authority of Jacobson, Dressel, and Petermann, the three most recent editors. Petermann distinctly says that this is the reading of 'Lat. 1 ut Gr. 1' (i.e. of both the Latin and Greek texts of the Middle Recension); Jacobson and Dressel, if they do not refer to the MS for the title, yet head the epistle πρὸς Ρωμαίους, though it stands between two epistles whose headings represent the other type, Τραλλιάνοις Ἰγνάτιος and Φιλαδελφεῦσιν Ἰγνάτιος. I seemed therefore to have the very best reasons for the statement; but, since the paper was written, I have been at Paris and collated the MS itself. I find that this epistle has no heading at all, the text of the letter being written continuously with the text of the martyrology in which it is inserted, without any title or indication of a break. It is correctly printed in Ruinart's Act. Sinc. Mart. p. 15. We may indeed infer that its title was πρὸς Ρωμαίους from the rendering of the Latin Version (see the last number p. 50), as also from the expression used in the martyrology in introducing it, οἷα πρὸς τὴν ἐκκλησίαν ἐπιστέλλει Ῥωμαίων; but direct evidence fails us. If I mistake not, my argument remains unimpaired notwithstanding this fact: but I take the opportunity of correcting an error into which I had been led by editors of the Ignatian Epistles, and which probably would have been followed by others.

J. B. LIGHTFOOT.

REMARKS ON MR W. G. CLARK'S ARTICLE, ENTITLED "ENGLISH PRONUNCIATION OF GREEK."

THE following brief observations are meant as commentary rather than as reply.

Following the order of Mr Clark's article I would draw attention, in the first place, to a point of some interest with

regard to the pronunciation of x. In Magna Græcia, I learn from Mr Clark, was in some cases pronounced like the soft Italian c. No doubt this would be in the same cases as those in which the Italian c is so pronounced, namely before e and i sounds. I do not however believe that the pronunciation of Magna Græcia arose from a "corrupt following" of the Italians. I should rather be disposed to believe it was the other way. First, because this pronunciation is now prevalent, especially in Crete, a part of Greece perhaps the furthest removed from Italian influence though I dare say it obtains elsewhere. Secondly, because the usual and comparatively hard pronunciation of ê by Greeks before e and i sounds, is already an approach to the Italian soft c, and constitutes one of the main difficulties in the pronunciation of Greek. I never understood till I went to Greece how a sound like chee- could ever come out of what must originally have been kee-, as is apparently the case in Italian. But as soon as I looked at a Greek's mouth when he said καὶ, ἐκεῖνος, κίνδυνος, the mystery was explained. I cannot teach this pronunciation on paper, but I will do my best to describe it. The tongue comes well forward to the palate, so that the tip appears between the teeth. The reason is very clear. The sounds i and e are palatal vowels, as compared with a and c or even ov (Italian u), and the Greek laws of "Sandhi" as the Sanskrit Grammarians would call them require the x to be palatal too. I noticed that whenever this palatal (an idea of the sound of which may be obtained by sounding a k with a t before it as theen), was very forcibly or emphatically pronounced, it had a tendency to become ch, a tendency completely carried out in the Cretan pronunciation. I will not say that the Italians got the c soft from the Greeks in Magna Græcia, but if they developed it themselves they must have developed it, I think, by the same process. It thus becomes plain not only how the ch sound could arise in Italian, but also why e and i are the only vowels before which the change takes place.

As to the theoretical question how did the ancient Greeks pronounce their language? it is hard to answer it without first propounding another; Who were the ancient Greeks? If the Greeks of the Ptolemaic era are meant, then certainly the re

searches of M. Rénan have gone far to shew that the Semitic transcriptions of this period speak with one voice in favour of the modern Greek pronunciation. If the Greeks of the Roman period be allowed the title of ancient, then the bad spelling in inscriptions belonging to that time tell the same tale. Unfortunately I have here room only for assertions and not for proofs.

I quite admit that ɩ, ŋ, v, eɩ, oɩ, and vɩ had at one time different sounds, but I think, to mention nothing else, πîdağ from πηδάω, ἵκω and ἥκω, βύβλος and βίβλος, εἴλη and ἴλη, λέγει aud λέγῃ (2 pass. of λέγω), incontestably prove that at least a strong tendency to Ioticism was prevalent even in the age of Homer.

The very early pronunciation of the diphthongs as simple letters is, I think, conclusively proved by the fact of the diæresis which would otherwise be superfluous, and the fact, be it remembered, is as old as Homer, whenever the sign may have come into use. That a and or were usually not only simple sounds but short too, we know from prosody and scansion.

As to the consonants, B is pronounced in Greece not like our but like the German w, only much more strongly and explosively, if one may use the word. It is not sounded by bringing together the lower lip and the upper teeth, but by compressing the two lips together. Still in the Roman period it was used to transcribe the Roman v, as Þráßios. So too p, and the consonantal sound of u are pure lip letters and very different in point of formation from for v. This is an important distinction, because it is easy to see how à-vτós rapidly pronounced, the v sounded like the German ü, would almost inevitably become aprós; whereas aftós would be, I think, inexplicable. As for and 8 they throw light upon each other. If déλta be pronounced as d in Spanish, and like our own z, I at once understand τωθάσδω becoming τωθάζω; otherwise the change appears mysterious in the extreme. The pronunciation of y as the German g in Tag as early as the Septuagint is placed beyond a doubt by its being used frequently to transcribe the Hebrew y. As to πέντε pende, and ἐμπρὸς embros, who can avoid thinking of 'Aμπрañía and Ambracia, ἐνδελέχεια and ἐντελέχεια ?

With regard to accents I would just remark: first, that having paid the utmost attention to this very point during a three months' residence in Athens I am quite convinced that the accent neither lengthens by its presence nor shortens by its absence any more than, as a rule, it does in English. Aóyos, ovos may appear to be so affected from the fact that the o is followed by but one consonant, forming what in Hebrew would be called an open syllable. But ἐξαίρεσις, πραγματικώς, are pronounced in the most marked manner ἐξέρεσις, πραγ ματικός.

I think Dionysius himself gives us a pretty clear answer to the question what he would have thought of the Accentual modern Heroic measure, when he gives as accentual (πpoowdikovs) the following lines which scan precisely in the same way.

Οὐ βέβηλος ὡς λέγεται τοῦ νέου Διονύσου

Κἀγὼ δ ̓ ἐξεργασίης [reading corrupt] ὠργιασμένος ἥκω. Hephæstion's Enchiridion completes the triplet thus:

Οδεύων Πελουσιακὸν κνεφαῖος παρὰ τέλμα.

What has been called the clashing of the accentual with the quantitative beat constitutes the real beauty of quantitative

measure.

It is this τύπος ἀντίτυπος which makes the charm and melody of the old heroic measure. The accent and quantity of these two words as well as the thought expressed in them seem to me exactly to embody the idea of beauty in quantitative versification, which is, as beauty always is, the harmony of contrasts. Where both coincide, as very rarely in Epic poetry,

(Indignor quandoque bonus dormitat Homerus,)

then the other part of the line (in which, happily for my illustration, this coincidence takes place) is realized:

καὶ πῆμ' ἐπὶ πήματι κεῖται.

E. M. GELDART.