upsets the theory held by Metcalfe, Henry, and Kappes that Vergil follows the plan of a Roman house in this place and refers to the cavaedium by cavae—a - aedes in line 487, to suppose that he uses the word atria in these two passages, not in its ordinary sense, but as an equivalent for the Homeric uéyapov, which appears to have also had an open roof.

The remaining citation has reference to the feast in the palace of Dido, of which he says, I. 638: mediisque parant convivia tectis; cf. the Homeric use οἱ δῶμα for μέγαρον, II. VI. 316, etc. Servius regards both this passage and the citation as having reference to the early customs of the Romans, and Metcalfe, Beck. Gal. p. 250, cites his quotation from Cato on line 726 (730) as an evidence that in the early days the atrium was the dining-room of the house. The same authority, however, on the following page cites the passage itself as an evidence that the atria of Vergil's time had become "very magnificent." This comes painfully near convicting Vergil of an anachronism; but is it necessary? We have no means of knowing what idea Vergil had of a Carthaginian house; but as Carthage was destroyed more than seventy years before he was born, and the scene is laid in Homeric times, it seems very unlikely that any factor of that kind entered into the question. On the other hand (there is hardly a page of the Aeneid which does not contain some item suggestive of Homer), the palace of Dido is represented as magnificent, I. 637, cf. Od. VII. 81 ff., the city excites the admiration of Aeneas, I. 421 f., cf. Od. VII. 43 ff., and the whole setting of the passage seems to be Homeric. Again, the μéyapov, or dŵμa, was the dininghall in which feasts were held, cf. Od. XX. 248 ff. etc., and finally, Dido, while upbraiding her faithless lover, laments that she has no little Aeneas to play in her aula, IV. 328. It does not help matters to suppose that aula here stands for atrium, as is commonly done in the case of III. 354 (Servius and Heyne seem to regard it as an avλ), although Vergil expressly says, 295, that Helenus is reigning over Grecian cities. It is much simpler to believe that Vergil is consistent (it is like him), and as he must have been acquainted with Greek houses, cf. Hor. Car. I. 3, 1-8, it seems quite probable that he conceives of Helenus as receiving his guests in an avλý such as he himself had seen. The other three passages in which aula occurs all point to an intentional use of the word as appropriate to the passage, not as a poetic makeshift — Aen. I. 140, the palace of Aeolus, cf. Od. X. 10; G. II. 504, of foreign conquests, cf. lines 487-97, which tend to give a Greek tone to the passage and suggest that the conquests are in the East; and G. IV. 202, where the bees in question are of the Greek variety, cf. 177, and therefore make an avλý, figuratively speaking, even if they do elect Quirites, 201; for the poet is not to be held to too strict an account. Again, there is abundant evidence that Vergil was a careful user of words. The huts of Carthage are called magalia, I. 421 and IV. 259, and those of Libya mapalia, G. III. 340; a Trojan in prayer speaks of the tholos of a temple, IX. 406; and even his use of thalamus (the metre forbids cubiculum and dormitorium is late) goes to show how careful he was to be consistent. In three passages it refers to a room in Dido's palace, IV. 133, 392, 495; but in fourteen (?) others, — G. IV. 189, 333, 373: Aen. II. 503; VI. 280, 397, 521, 528, 623; VII. 97; VIII. 372; X. 497: Ciris (?) 217, 512, — with hardly an exception, the color or setting is so distinctly Greek or Trojan (Homeric) that the word seems not merely appropriate, but technically correct. In the Latin sense, 'marriage-bed,' 'marriage,'

he uses it eight times, Aen. IV. 18, 550; VII. 253, 388; IX. 591; X. 388, 648; yet all but two of these, IX. 591 and X. 388, seem to be in keeping with the other passages and both cases occur in the later books, which show other evidences of increasing freedom in dealing with his subject. It is very probable, then, that Vergil means an auλý when he uses aula; and, if he does, it seems clear that here again, in the two passages relating to Dido's palace, he has used the word atrium as an equivalent for the Homeric μéyapov. Such an explanation relieves all four of the passages from difficulty, is in strict keeping with Vergil's methods and character, and is far more natural than the supposition that he pictures the houses of his own day and then attaches usages and customs of his remote ancestors to give the whole an antique flavor. He may indeed have been influenced somewhat in his conception of ancient palaces by what he saw about him in Rome; but that is quite a different matter from supposing that he took these things as a basis rather than his Homeric sources.

I have thus far been able to find nothing in positive support of the above views. Heyne compares I. 725 with Od. I. 365, and adds below: "Non lucernas vel candelabra posuit sed lychnos, funalia . . . cf. Odyss. ", de regia Alcinoi, 100 sqq.," and he says of porticibus longis, II. 528: “Si Homerica et non sua potius tempora sequutus est, at@ovσav expressit, quae aʊλǹv ab utraque parte ornabat"; but his note on II. 512 and the excursus on the passage make it clear that he holds practically the common view. He says: "Graecis poetis erat ara Iovis Hercei (Διὸς Ερκείου) in atrio aedium Priami, ἐν αὐλῇ· eam aram Virgilius in impluvium, si interiora domus ita appellare licet, transtulit . . . ut Penatium ara esset; propius hoc ad Romanum morem. v. Excurs." In the excursus he makes interiora domus refer to the peristylium of a Roman house. He does, however, recognize that there are difficulties in the passage. If Vergil means the uéyapov, all these difficulties disappear, and that he does seems to be the only logical conclusion; for, as was suggested at the beginning, he could not use a Latinized form of μéyapov. In the sense of the main hall of the dvopúv, the word is cited only in Homer. In Herodotus it is used of sacred edifices alone, and in later times it seems to have been confined entirely to underground caves sacred to Demeter and Persephone, in which sense it would probably have been understood by his readers, if Vergil had been bold enough to turn it into a Latin word. The best thing that he could do was to use the word atrium in its place, very much as we should use the word hall or halls to-day if writing a poem in English under similar circumstances; for it is probable that every foreign word which he used was familiar to his readers in the sense in which he used it. He accordingly used the technical term where he could do so, and translated elsewhere. Finally, the common view, that atria in the passages corresponds to avλý, loses sight of the fact that the Homeric avλý was not a room at all, but an open, unpaved court.

25. On the Accent of certain Enclitic Combinations in Greek, by Professor Francis G. Allinson, of Brown University.

This paper appears in full in the Transactions.

Professor J. Irving Manatt, of Brown University, then made some remarks on recent progress in Mycenaean archaeology.

26. Notes on the Hippolytus of Euripides, by Professor J. E. Harry, of Georgetown College.

woman,

I. THE CHARACTER OF PHAEDRA.

Down to the time of Wilamowitz-Möllendorff it was generally believed that Euripides represented Phaedra as being what she pretended to be - a virtuous who really tried to remain true to her husband, and earnestly desired that her passion for Hippolytus should not be revealed to the young man. Wilamowitz says this view is not the correct one, that Phaedra is only playing a part in her dialogue with the nurse (516-524), that she really sees through the design of her servant, and hopes that she will approach the youth for whom she has conceived such a strong passion. Kalkman follows W.-M.'s lead, and others

subscribe to the same opinion.

But, if we should adopt this view, would we not impute excessive subtlety to Euripides, as well as to Phaedra? Would we not demand too much of the audience? The average Athenian was not dull, but could he (in the brisk dialogue of Euripides) have taken all these subtle points which have escaped the scrutinizing glance of all the painstaking students, only to be discovered after the lapse of twenty-three hundred years? If this was the poet's design, and he has lacked an interpreter from that day to this, was it not too deep for even the quick-witted Athenian? Could he understand the real significance of Phaedra's words when she declared over and over again what her feelings are and what she has determined to do?

Phaedra was not spotless any more than Hippolytus, but she could not be called unchaste.1 Every utterance of hers shows how she tried to stem the tide and die evkλeńs. Her great misfortune is not to have hearkened to the voice of reason in time. Whither her passion might have carried her (even the strongest have succumbed) we can only conjecture, for the nurse precipitates matters by revealing to Hippolytus the whole situation. That this was done without the queen's knowledge and consent is clear. That she, perhaps, intuitively divines somewhat of her servant's purpose without knowing to what lengths she would go, does not affect the question. In her present state of mind she is easily led on by the nurse. She is not able to take the lead herself until she is roused by the terrible reality, until her worst fears have been realized. Then she summons up all her strength and carries out her previous resolution, viz., to take her own life. But now her reputation is at stake, and another shall suffer as well as she — κακόν γε χατέρῳ γενήσομαι θανοῦσ ̓ (728).

This conception of Phaedra harmonizes with the circumstances attending the production of the two dramas. Euripides must have intended that his second play should be entirely changed: he could not have retained the old Phaedra without deceiving his audience, and this he would not wish his players to do; for, as Hamlet says, their business is to tell all.

1 So Puntoni De Phaedrae indole et moribus in Euripidis Hippolyto Stephanephoro, Pisa, 1884.

II. THE GREEK STAGE AGAIN.

In view of the stand taken in some quarters recently against the 'no-stage' theory, I merely wish to emphasize what Pickard says in the American Journal of Philology, Vol. XIV., p. 83: “The height of this ‘stage,' the lack of means of communication with the orchestra, its slight depth, its distance from the cavea, the doors leading out on the level of the orchestra, the arrangement of the seats themselves, all unite to prove that this structure could never have been used as a stage."

At line 58 of the Hippolytus a chorus of attendants enter, remain for some time, and depart with their master (l. 112). For this scene the broad level of the orchestra is better suited than the narrow platform of the stage. Indeed, to one who has had something to do with the management of a troop of young men on a much wider platform under not entirely dissimilar circumstances (at least so nearly alike that any difficulties of representation in the one would obtain for the other) it seems almost impossible to put this scene on a stage eight feet deep. So in the scene where Hippolytus and the nurse are within the palace. Phaedra is farther away from the audience than the chorus, and consequently nearer the palace. Suddenly she hears a noise within and commands the chorus to keep quiet that she may hear. It is Hippolytus upbraiding the nurse, and, as soon as she discovers the real state of affairs, she breaks out with: ἰώ μοι, αἰαὶ αἰαῖ· [ ὦ δυστάλαινα τῶν ἐμῶν παθημάτων (569-70). The chorus does not understand, and Phaedra bids the choreutae step up to the door and listen for themselves (raîod' émiorâσai Ħúλais). She does not tell them to mount an elevated stage. True, they do not move, but this is clearly a device of the poet to convey to the audience an account of what is going on in the palace (577-80).

III. MISCELLANEOUS.

1. 32. xônμov is read by MAC2P and the scholiast, ěkdŋλov by VC1N. Editors vacillate between the two. Nauck reads ἔκδηλον and says: “ ἔκδημον deteriores libri" (which is not true). Wilamowitz has exônμov and translates ihr fernes Lieben. Wilhelm Pecz (in his study of the tropes in the three tragic poets, Berliner Studien für class. Phil., Vol. III.) considers ěkdŋuov as referring to Hippolytus by metonymy (love for the lover). Weil proposes to alter these three lines, and omits épŵσ'ěpwr' ěkdnμov. Bury (Class. Review, III. 220) feels sure that ëŋλov must be the true reading, but his arguments are not convincing. ěkdŋλov, conspicuous (cf. II. V. 2), makes sense - such as it is; the opposite adnλov, secret (Thuc. VIII. 108), would be just as appropriate. Cf. 40, 42, 139 (KρUTTY TÉVOEL). The passion was secret in both places (in Athens and at Troezen); it only became more violent when Phaedra came to the latter place. But many conjectures might be made, e.g. ¿\yeɩvóv (cp. 775), and yet none of them could be received as certain emendations. Blomfield, as well as Hartung, excises 32 and 33, but if these verses are removed, a sufficient explanation is wanting for 30 and 31: these two in turn are made necessary by 29. Moreover, ἐπεὶ δὲ in 34 corresponds to πρὶν μὲν in 29.

1. 33. wvbμašev is the reading of the MSS. and the scholiast. Several changes have been suggested. Meineke and Wilamowitz read wvouacov. Kirch

hoff has ὀνομάσουσι. The future is certainly to be desired, but the change is unnecessary: the tense can be explained on the basis of the praesens propheticum.

1. 42. πрâyμа: MSS. Wilamowitz changes to raidi, unnecessarily. His reviewer in the Classical Review says that it is a harmless reading, if it had MS. authority, and harmonizes with 520, as the traditional reading does with 690. Neither line affects the question particularly. Line 41 gives the negative, 42 the positive side. In ll. 43-46 mention is made of the death of Hippolytus, in 47 of Phaedra; so we have a regular gradation in time: Theseus, Hippolytus, Phaedra, the events referred to occurring in the reverse order — a sort of v☛tepov πρότερον. The chorus swears μηδὲν κακῶν σῶν ἐς φάος δείξειν ποτέ (714); nevertheless, the whole matter does come to the light. Cf. the exclamation of the chorus in 367 ὅλωλας, ἐξέφηνας ἐς φάος κακά.

1. 79. 8ools: MSS. Porson changed orois to oσris, a reading which is not objectionable, to be sure; but it is not so natural as doois. Nauck says: mit Porson's Aenderung ist dem Sinn der Stelle wenig gedient. In spite of the fact that most editors have followed Porson (Monk, Weil, Wilamowitz, Nauck in the third edition) the MSS. reading seems to me to be preferable. Cf. 3, 6, 442, 451, 1015, 1302. Consider the sentence ¿XX' . . . duws parenthetical, understanding aurous with elλnxev, and it becomes clear. It is not necessary to make elλnxev neuter, as in Hom. Od. IX. 160, for the verb means here took them for her own.'

11. 168-169. Translate Much-revered by me she always comes to the rescue,' not as Mahaffy and Bury (after Weil) explain: 'she walks in the number of the gods,' nor as Paley takes it, thanks to the gods.' ovv beoîo is a stereotyped phrase meaning 'with the blessing of heaven.' pora is not used absolutely (the meaning is 'she comes to me') and this verb is purposely selected. μerà eŵv occurs only in H. F. 180, where it is a matter of gods among gods. It is strange that this line has been so often misunderstood. Tycho Mommsen explained it correctly. Herwerden in Revue de Philol. for 1878, p. 19, says: requiro év beoîoi poira, showing that he has the same conception of the passage as Weil, but is dissatisfied with the preposition σύν. Hadley changes φοιτᾷ to ἐφοίτα, unnecessarily.

1. 277. Daveîv. This is the reading of the MSS. The only possible interpretation is that the nurse repeats the eaveîv of the chorus by way of reply, and then adds ἀσιτεῖ δ' εἰς ἀπόστασιν βίον to explain the means Phaedra chose: she desires to die, and to accomplish this, you see, she is starving herself.' The nurse knows her mistress wishes to die (248-9, 305, 314, and especially 322), but the cause which lies back of this resolution (τὸ δεινὸν τοῦθ ̓ ὅ σ ̓ ἐξαίρει θανείν) is what she has been trying so hard to discover (39-40, 271, 273 wávтa yàp σiyâ Táde, 279, 283, 284, 297, 303) and the very thing which Phaedra takes such pains to conceal. Consequently, no valid objection can be raised to the reading of the MSS. Wilamowitz feels sure that eaveîv has crept in from the preceding line; so he removes the word and fills up the gap with ouk old', a harmless reading, but no better than many others which suggest themselves, e.g. oiya (cf. 273, 279, 297): the nurse reiterates that all her efforts have been in vain, and this word would be very appropriate here, in fact, more appropriate than our old', for the nurse does know the answer to the latter of the two questions (Daveîv weipwμévŋ), as is shown by her interrogation in 322.