I start with two contradictory statements: (1) Jaeger, Aristotle (Eng. trans.), p. 330: ‘Thus all indications point to a late date for the origin of the philosopher's zoological works.’ (2) D'Arcy Thompson, Historia Animalium (Oxford Translation), Prefatory Note: ‘It can be shown that Aristotle's natural history studies were carried on, or mainly carried on, in his middle age, between his two periods of residence at Athens.‘

It is well known that the trisyllabic ending of the pentameter, which, with the Augustan elegists, declines in popularity until it is practically rejected by Ovid, meets with a revival in Martial. His pentameters show a considerably higher proportion of trisyllabic endings than those of either Tibullus or Propertius.1 Is it possible to dis-cover any reasons, conscious or instinctive, that conditioned the revival of this type of ending?

Was Isocrates influenced by Thudydides? Wilamowitz at first suspended judgement and later decided he was not, but he did not go into the question. Attempts have since been made to prove close and direct influence. The question assumes greater interest and importance because of the immense influence of rhetoric on the writing of history in the fourth century and of the generally accepted tradition that Isocrates’ pupils included well-known historians like Ephorus, Theopompus, and the ‘Atthis’ writer Androtion.

In A.D. 60 Nero instituted at Rome a quinquennale certamen which he called Neronia. It was in three sections—athletics, chariot-racing, and music (which included singing, and recitations in prose and verse)—after a Greek model. This model would be that of the Pythian rather than of the Olympic Games, for at the latter there was no regular musical contest; though perhaps Nero got his immediate inspiration from the Augustalia at Naples, which was an athletic and musical festival of a religious nature.

Every reader of Roman poetry must be struck by the fact that atque is so much more frequently elided than left unelided; and that the rarity of unelided atque is not—a matter of chance may be seen from a comparison between the poets' treatment of this word and that of others of a similar metrical structure: i.e. disyllables beginning with an open long vowel and terminating with an open short one. Such words ending in -que or -ě (utque, inque, ipse, ante, ire, ecce, and the like) are common enough in Roman poetry and are, particularly in elegiac verse, more often found unelided than elided. It would be a waste of time to give figures for them all, but those for ille may be cited as an example. In Virgil, Aen. 1 and 2 the proportion of unelided to elided ille is 57 per cent, to 43 per cent.; in Catullus' hexameters 50 per cent, to 50 per cent.; in his lyrics 75 per cent, to 25 per cent. In his elegiacs there are 3 unelided to none elided. In Ovid's Met. the proportion is 83 per cent, unelided to 17 per cent, elided; in hisArs Am. 1 and 2 84–6 per cent, to 15–4 per cent.; in his Fasti 1 and 2 97–5 per cent, to 2–5 per cent. In Tibullus 1 and 2 the proportion is 87 per cent, unelided to 13 per cent, elided; in Propertius 1 and 4 73–7 per cent, unelided to 26–3 elided.

I present first what I take to be a more interesting item than the others.

O.C 716–19. δ' εὐρετμος ἔκπαγλ' λα χερ

σ ✝παραπτομναπλτα

θρῷσκει, τν κατομπδων

νηρῄδων κλουθος.

The above is Pearson's text, except that I have transferred the last syllable of his 716 to the beginning of my 717. Careful consideration of the metre of this stasimon has convinced me that 716 is rightly regarded by Schroeder as an ionic trimeter (what Pearson's line would be I have no idea, and his in 703 looks oddly detached); further, that 717 is what (I and) most people call a glyconic, as is also 718. Naturally therefore I follow Hermann and Jebb as regards the text in 704, and have thus no use for the three conjectures mentioned by Pearson as substitutes for our admittedly corrupt παραπτομνα.

Professor R. J. Getty has drawn attention (C.Q. xxvii, 1933, p. 139) to a tenth- or early eleventh-century manuscript of Statius’ Thebaid, hitherto examined only in Book I, namely Turonensis (T: codex Parisinus nouv. acqu. lat. 1627). Dr. Klotz, in his Teubner edition of 1908, gave citations from Book I, and wrote (Praefatio, p. xiv), ‘dolendum est sane de hoc codice primum tantum librum innotuisse, sed cum Roffensis libri maxime affinis accuratiorem notitiam haberemus, collatione quamvis -aegre careri posse nobis visum est.’ I have collated both T and Roffensis (r: codex Regius Mus. Brit. 15.C.X) in full, and find firstly that the citations of r by Klotz are far from accurate, especially in the last six books, and secondly that while it is indeed true thatT and r are very closely related, T is a much better representative than r of their common source.