The statement that masks were not introduced on the Roman stage until after the time of Terence is still repeated by editors (e.g. Menaechmi, Thoresby Jones, 1918, p. 19; Menaechmi, Moseley aria Hammond, 1933, p. 19) and has the support of Pauly Wissowa (Bieber, s.v. ‘Maske’, 1930) as well as Daremberg and Saglio (Navarre, s.v. histrio); it may, in fact, be regarded as generally accepted. Yet so long ago as 1912 A. S. F. Gow (J.R.S. ii. 65–77) put forward strong arguments on the opposite side; his article, though mentioned with respect in Bursian (1936, pp. 51 ff.) and referred to by Schanz-Hosius (i. 149), has not yet been satisfactorily answered, so far as I am aware. Gow did not claim that a final solution of this problem could be attained on our present evidence, but he did show that the orthodox position is open to attack. I hope to prove that the arguments in favour of the early use of masks are even stronger than he claimed.

As Headlam remarked more than thirty years ago, ‘It is a strange fact that the order of words in a Greek sentence has never been clearly appreciated.’ The emphatic word is placed at the beginning of the sentence, or as near it as the structure of the language permits. That is the general law, which I have discussed at length in my edition of the Oresteia. My object in this article is to make an exhaustive study of its operation in a limited field.

Neglect of the ‘codices deteriores’ has caused the ascription of a considerable number of readings in Cic. De Diuin. to the conjectures of scholars of the fifteenth, sixteenth, and later centuries, though manuscript evidence, in some cases of a much earlier date, is to be found. Even if the presence of such readings in the manuscripts is due to conjecture and to no other cause, credit for priority should be given to the manuscripts. The following notes are restricted to the evidence of the London MSS.

Herodotus tells us in book vii, ch. 220, that the Pythian priestess gave the Spartans a warning couched in hexameters, of which the second line begins ἢ μ⋯γα ἄστυ ⋯ρικυδ⋯ς. To this text the admirable commentary of How and Wells takes exception in the following note: ‘The synizesis ἄστυ ⋯ρικυδ⋯ς is intolerable. Read δ⋯μ' ⋯ρικυδ⋯ς, ἄστυ being a gloss, H. Richards, Cl. Rev. xix. 345.’ Doubtless this union of vowels is harder than that of υω in Ἐριν⋯ων (Eur. Iph. Taur. 1456) or in γεν⋯ων and δυωδεκ⋯μηνον, accepted by Christ in his edition of Pindar (Pyth. iv. 225; [Nem.] xi. 10). Let us grant that from the point of view of elegance it is even intolerable; the question still remains, should the line be altered?

Greek art and literature follow parallel courses through the long period from Homer to Euripides. Homer and Euripides, Dipylon vases and the latest white lekythoi are as far apart from each other as it is possible for works in the same medium to be. The distance can only be explained by a similar change in the views of artists, writers, and their public.

Peter Elmsley once said ‘difficillimum opus esse accuratam librorum collationem’. He was pointing out some mistakes in Porson's collation of the editio princeps of the four plays of Euripides, which Porson had said that he himself had collated ‘summa cum religione, ne dicam superstitione’. These were both men who knew Greek and who could collate manuscripts. So we ⋯ν;θρωπ⋯σκοι who would follow in their footsteps in this most difficult task cannot expect to do more than increase in some small measure the accuracy and completeness of existing collations. It is to be hoped that others in their turn may in due course improve on our work also, so that finally, with sufficient workers in the field, a close approximation to truth may be achieved.

The first study of Plutarch's prose-rhythm was made by Dr. A. W. de Groot, whose results were published in certain preliminary articles and in his Handbook of Greek Prose Rhythm, a work which is one of the landmarks in the history of its subject. In it he insisted that to discover which forms of clausula were favoured or avoided by any author it was not sufficient to make a count and discover which were frequent, which infrequent; for a form may be frequent not because an author feels it suitable for the end of a sentence, but because he likes it at any point in his sentence, or even because we should find it frequent if we picked out words by chance from the dictionary. To discover which rhythms are specially sought or avoided at the end of a sentence we must compare the ends with the sentence as a whole. This Dr. de Groot did for a number of texts, included in which were selections from Plutarch's Lives.

The construction assumed for the first sentence in this passage is that adopted by Verrall and Headlam, apart from some differences in detail. It seems unlikely that ἰδών can refer to what precedes, as some have thought; for it can hardly be supposed that the poet, who is using speed and economy (a point often neglected by critics), would pause to tell us that the great Seer merely ‘sees’ two distinctively coloured birds which ‘have appeared near by in a conspicuous station’ compare Homer's method (B321). Also it would involve us in the doubtful construction Ατρεϊδας έδάη λαγοδαίτας for ‘he knew the Ἀ. to be the λ.’ (or vice versa); the nearest parallel to this would appear to be Xen. Anab. 1. 7. 4 οἴους γνώσεσθε τοùς ἄνθρὡπονς, where the adjective makes some difference.

The following remarks on the aviary simile have been prompted by Professor Hackforth's article in C.Q. January 1938, pp. 27 ff., in which he in turn comments on certain points in Professor Cornford's treatment in his Plato's Theory of Knowledge.

Commenting on 199c–d C. (137) suggests that P.'s criticism in that passage might be met by the inclusion in the aviary of ‘complex objects such as the “sum of 7 and 5”.… It is this object (sc. the complex object “the sum of 7 and 5”) that I identify with 11 when I make my false judgement.… False judgement can be explained as the wrong putting together of two pieces of knowledge’—i.e. 11 and ‘the sum of 7 and 5’. C. (138) further suggests that P. has overlooked this possibility because he persists in speaking as if we judged not ‘that the sum of 7 and 5 is 11’ but that ‘12 (the number we are seeking) is 11 (the number we lay hold of)’. H. replies that this criticism of C.'s is not valid because P. understands the knowledge of 11 and 12 to include a knowledge of the various sums that make up these numbers. There could, therefore, be no ‘question of “putting together” such a piece of knowledge as 7 + 5 = 12 with a different piece of knowledge, viz. the knowledge of 12. They are not different pieces; they are the same piece.’

The new edition of Chariton's novel Chaireas and Callirhoe by Mr. Warren E. Blake (Oxford, 1938) has a twofold merit: it is founded on a new and accurate collation of the only extant manuscript, a Florentinus, and it gives a full collection of the conjectures in and since the editio princeps by D'Orville (Amsterdam, 1750). The new text has already provoked some suggestions by Prof. H. J. Rose in C.Q. xxxiii, 1939, 30, but I should like to propose alterations in some more passages