marked on the last syllable, and the grave which is understood in other syllables are the same; and that, being the same, they must produce the same effect. But the mark (') over the last syllable is the sign of the acute accent, and raises that syllable: whereas the grave accent, which is understood to apply to a syllable over which there is no mark, depresses that syllable. The correct way of stating the rule would be thus: "Gravis nunquam signatur, sed in omni syllaba, cui deest signum, intelligitur." It has been supposed by some, that syllables so marked, though depressed in the middle of a sentence, were raised at the end of a sentence, to prevent the sound being lost by the sinking of the voice. But if such were the rule, why confine it to these particular words? why not apply it to every final syllable? That the mark (') must stand for an acute may be further shown from the testimony of Cicero and Dionysius, which will be cited hereafter, that every word has an acute accent. Where then is the acute in the word feòc in the middle of a sentence? Not on the first syllable, because that, not being marked, must be understood to be grave; not on the second, if we consider the () to stand for a grave; so that the word would be without an acute, which is impossible.

The effect of the circumflex is not so easily defined. To say that it sustains or makes long the syllable affected by it, is to give a very im

perfect account of it; because, though all circumflexed syllables are long, all long syllables are not circumflexed. Quinctilian says expressly, that a circumflexed syllable and an acute are the same : "Præterea nunquam in eadem flexa et acuta, quoniam eadem flexa et acuta." (i. 5, 31.) Apollonius, with greater precision, says that they are the same in power though not in nature: Προηυθέτισται γὰρ τὰ ἐγκλιτικὰ μόρια ἐπὶ τέλους ἔχειν τὴν ὀξείαν, ἢ φύσει, ἢ δυνάμει λέγω δὲ, δυνάμει, διὰ τὰ περισπώμενα. (Syntax, ii. 18. p. 138.)

There is another passage of the same author, which at first sight strikes us as being in direct contradiction to this. In speaking of the accentuation of interrogative adverbs he says: Ta πύσματα ἢ φύσει θέλει βαρύνεσθαι, ἢ δυνάμει τὰ γοῦν ὑπὲρ μίαν συλλαβὴν, ἔχοντα τόπον τῆς βαρείας, πάντα βαρύνεται τὰ δὲ μονοσύλλαβα, οὐ δυνάμενα ἐκτὸς τῆς ὀξείας γένεσθαι, δυνάμει ἐβαρύνθη περισπασθέντα. (De Adverb., Bekker, Anecdot. Græc. p. 584.)

The solution of the apparent inconsistency must be sought in the peculiar nature of the circumflex accent, which was compounded of an acute and a grave: Αἱ μὲν κατὰ μίαν συλλαβὴν συνεφθαρμένον ἔχουσι τῷ ὀξεῖ τὸ βαρὺ, ἃς δὴ καὶ περισπωμένας καλοῦμεν. (Dionys. xi. 76.)

It appears that the voice was first raised to the pitch of an acute, and then, before passing on to the next syllable, was dropped to a grave. This is sometimes explained by saying, that the sound was repeated twice, first with an acute and then

with a grave, oŵua for instance was pronounced σόομα. oooua. This explanation is probably very near the mark, but still we must remember that these two sounds were blended into one syllable: owμa, σώματος, and ῥύος were doubtless all differently pronounced. At any rate, however, it is clear that the syllables having the circumflex mark were raised, and we may therefore apply the term accent to them, as well as to those having the mark of the acute.

INSTANCE SELECTED AS A GUIDE FOR THE VOICE IN

READING.

5. Having thus pointed out the distinction between accents and accentual marks, and having endeavoured to give an accurate definition of each, I now proceed to the question whether our pronunciation ought to be guided by the marks; that is, whether we ought, in speaking, to raise those syllables, and those only, over which we observe a mark. It may be asked, to what manuscript or to what book I would refer as the standard of the accuracy of the marks? and in answer to that question, the reader's attention may be called to the striking agreement in all the manuscripts and books in their general manner of placing the marks. It would indeed be absurd to contend that particular exceptions do not occur; copyists must have varied in carefulness as well as in knowledge; blunders must have been made in the accentual marks as well as in the order

and orthography of the words; but the general manner of placing the marks shows such an agreement between the different writers as could not possibly have resulted from accident. The deviations from this general manner have not been sufficiently numerous to throw any doubt or uncertainty over the system; but every critic who has studied it feels himself justified in saying that such and such words in a manuscript are wrongly marked, with as much confidence as he would say, that such and such words are wrongly spelt; appealing for the correctness of his criticism to an overwhelming majority of other manuscripts. And accordingly those who rely on the accuracy of the marks have usually contented themselves with contending in general, that our pronunciation ought to be guided by them wherever we find them. It has however occurred to me, that it may be better to refer to a particular passage of a given manuscript or book; first, for the sake of greater precision as to the very words to which our rules are to be applied; and secondly, because, by counting the number of marks in a passage of limited extent, we are enabled to show exactly the proportions in which the rules which we assume are observed or violated, and to reduce both rules and exceptions to arithmetical statement. I have selected for this purpose three of Dr. Charles Burney's manuscripts of the New Testament, in the British Museum. The passage fixed on was the first

chapter of St. Luke's Gospel to the end of the twentieth verse. The manuscripts selected are the following: No. 20 of the Burneian collection in the British Museum, written by Theophilus in 1285; No. 21 of the same collection, written in the year 1292 by Theodorus; No. 18 of the same, written in 1366 by Joasaph. I have chosen these manuscripts on account of their bearing the exact dates at which they were written. The marks in them have the appearance of being written at the same time with the words, and I think it may be fairly assumed that they were so. With respect to some of the more ancient manuscripts, this point might not perhaps be so easily conceded. When we find only one manuscript marked out of many, we may doubt whether it may not owe its marks to the accident of its having fallen into the hands of some one, who marked it in a later age: but as it is universally agreed that the use of the marks had become very general before the thirteenth century, there is little probability that a copier would allow a manuscript unmarked to go out of his hands; and I should think, in general, a writer who had made up his mind that the manuscript should be marked, would mark the words as fast as he wrote them.

The following is a copy of the first twenty verses of the first chapter of Saint Luke's Gospel, as they stand in the manuscript of Theophilus ; I have divided the text into verses for the con