We may now without trouble apply the same line of interpretation to vv. 4958. We may paraphrase again: Dum vitis intacta est, non cara est; sed cum tacta est, cara est. This is balanced by Dum virgo intacta est, non cara est; sed cum tacta est, cara est. V. 56 means simply: "So the maiden, the while she remains intacta, the while she grows old uncared for."

The paper closed with a brief consideration of the question whether dum . . . dum could bear the meaning assigned them throughout the discussion. On this point the author had nothing new to offer, but contented himself with compiling, more completely than has heretofore been done, a list of the authorities by whom this view has been defended and illustrated. Quint. ix. 3. 16 explicitly upholds it, implying that it is an archaism, a very plausible suggestion. The only parallel thus far cited is Pl. Truc. 232, where Lambinus' reading Dum habeat, dum amet is "accepted or repeated by Hand, C. F. W. Müller, Fleckeisen, Schwabe, Schöll, and Key, L. D., s.v.” (Ellis, p. 248, footnote). See also Haupt, Opusc. II., p. 473; Riese and Baehrens in their editions; Schmalz in Müller's Handbuch, II2., p. 509; Hale, Anticipatory Subjunctive, pp. 68, 69; and finally the critical note in the Goetz-Loewe-Schöll edition of the Truculentus. Both Riese and Baehrens cite by way of illustration Verg. Ecl. viii. 42 Ut vidi, ut perii, referring to Savelsberg, Rhein. Mus. XXVI. (1871), p. 135, the latter adding Corssen, De pronunt. II2., p. 856. See, however, Conington ad loc. For similar usages in Greek, see Ellis on v. 45, and Haupt, Opusc. II., pp. 471-473.

This paper was commented on by Professors E. T. Merrill, C. L. Smith, Allen, and by the author.

12. Superstitions and Popular Beliefs in Greek Tragedy, by Dr. Ernst Riess, of Norwalk, Conn.

This paper appears in full in the Transactions.

by Dr. H. W. Magoun, and by the author in reply.

It was discussed

13. Euripides, Hippolytus, 42, by Professor Francis Kingsley Ball, of the University of North Carolina.

δείξω δὲ Θησεῖ πρᾶγμα, κἀκφανήσεται.

The plot of the Hippolytus is set forth in the prologue by Aphrodite : Those who do her homage she puts first in honor; but she throws down all who act presumptuously toward her (5, 6); Hippolytus declares that she is basest of deities, and will pay her no regard (13, 14); she means to be revenged upon him (21, 22).

Her plan of revenge she proceeds to explain. Phaedra, the [second] wife of Theseus, is deeply in love with Hippolytus through Aphrodite's designs (2628). This shall be made the means of accomplishing her purpose (41: ¿XX' OÜTɩ TAÚTη TÓvď ěpwтa Xpǹ πEJEîv). But how? Why, Theseus shall be informed of the affair, and he will curse Hippolytus and put him to death (42-44). But be informed of what affair? Certainly Theseus will not kill his son because

Phaedra is enamored of him.

Wecklein explains that the verse is general in its signification (the matter shall be made public; and I will see that it reaches the ears of Theseus).

Mahaffy and Bury follow Wecklein, and say further, "Euripides does not add that it was to be represented in a false light, for his prologue is only to give a sketch of the plot, not to enter closely into the details." Wilamowitz says that the uninformed hearer can expect nothing from these verses but the guilt of Hippolytus. Wilamowitz, however, is not entirely right. If we join his note with that of Mahaffy and Bury, we shall arrive at what ought to be the conclusion drawn by the uninformed hearer or reader, namely, that either Hippolytus does wrong or his father is misinformed, the latter of which proves to be the truth. Doubt as to the outcome of the play ought to begin, really, with verse 41, where Aphrodite proposes to make use of Phaedra's infatuation as a means of destroying Hippolytus.

Hiller and others object to the verse on the ground that it is contradictory to the issue of the play: Aphrodite does not reveal to Theseus the love of Phaedra for Hippolytus.

To remove the apparent contradiction, Weil suggests the reading deîţaɩ deńσeɩ πράγμα. Other readings proposed are as follows: ἐς φῶς δὲ δείξω πρᾶγμα (Barthold); ἤδη δὲ δείξω πρᾶγμα (Fecht); δείξω δ ̓ ὃ λήθει πρᾶγμα (von Arnim); and, perhaps the most noteworthy, that of Wilamowitz, deięw dè Onoéws waidi, Kȧкpavýσeтαι. Wilamowitz points out that Aphrodite states what takes place immediately and at the end of the play, but leaves out the middle part, namely, the aversion of Hippolytus and the slandering of his character by Phaedra.

On the other hand, we have, in favor of the MSS. reading, the following considerations:

1. The reading del¿w dè Onσeî has the support of all the MSS. without any variant.

2. The meaning is not really inconsistent with the development of the play. It is true, as has been objected, that it is Artemis who reveals to Theseus the love of Phaedra for Hippolytus. But is this what is referred to in verse 42? Not at all. Verses 41-44, though obscure, have but one explanation, as is shown by the development of the plot. "I will use Phaedra's infatuation," says Aphrodite, "to overthrow Hippolytus. Theseus shall hear of the affair (πрâɣμа), and he will curse and destroy my foe." Let us follow, now, the development of her plan : Phaedra at first concealed her love for Hippolytus (394), but finally revealed it to her nurse (350-52); the nurse informs Hippolytus, and is reviled by him (565-90); Phaedra, now that her love has been revealed (596), wishes to die (599), and hangs herself (777); Theseus finds a letter on Phaedra's person (856), in which she accuses Hippolytus (874-86); Theseus bids Hippolytus begone from the country (973); Hippolytus' death is reported (1162); Artemis informs Theseus of the innocence of Hippolytus (1298, 1299) and of the infatuation of Phaedra (1303).

It is clear, then, that the prologue is not really inconsistent with the development of the play, as the information given to Theseus by Artemis is of a sort to clear Hippolytus from censure, while that referred to in the prologue is intended to work his ruin, and reaches Theseus by means of the nurse's revelation to Hippolytus, which causes the writing of the letter and the suicide of Phaedra; — all caused directly or indirectly by Aphrodite, the author of Phaedra's infatuation.

The apparent inconsistency, already referred to, that the prologue does not state exactly what takes place in the play itself, may find a parallel in the Ion. We are informed by Hermes, in the prologue of the Ion, that Xuthus and Creusa, being childless, have gone to consult the god at Delphi (64-67); Apollo is to present his own son by Creusa to Xuthus when the latter enters the temple; Xuthus is to take the boy to Athens, where he is to be made known to Creusa and obtain his rights (69-73). The development in the play, however, is as follows: Creusa is greeted by Ion at the shrine at Delphi (237); Xuthus, on returning from the neighboring oracle of Trophonius (405), meets Ion in the temple and greets him as his son (517); an explanation takes place, and Xuthus bids Ion go to Athens (577); the chorus is commanded to reveal nothing to Creusa on a penalty of death (666, 667), but they tell her everything (761 ff.); Creusa wishes to kill Ion, as she thinks herself wronged by Xuthus (979), but her plan fails (1194 ff.); Creusa recognizes Ion by the garment in which he was exposed when an infant (1395 ff.); — the recognition taking place at Delphi, and not at Athens as told in the prologue.

3. The reading del¿w dè Oŋoeî maintains the line of thought better than any other reading suggested.

Nobody, to my knowledge, disputes the genuineness of verse 44. This verse helps out the MSS. reading in 42. The line of thought is this: "Phaedra loves Hippolytus; this infatuation is good for my purpose; Theseus shall hear of the affair; he will destroy Hippolytus; Phaedra shall die." Theseus is the one to be informed because the one most concerned. When informed, he will kill Hippolytus, — not of course because Phaedra is in love with him, but for one of the reasons given above, namely, that Hippolytus is guilty, or that he believes him guilty. Suppose we read, with Wilamowitz, del¿w dè Onσéws waidi. Then the line of thought is: “Phaedra loves Hippolytus; this is my opportunity; Hippolytus shall know it; his father will kill him; Phaedra shall die." By following the MSS. we are left in doubt as to the real cause of Hippolytus' death; but by following Wilamowitz we can arrive at only one conclusion, and that erroneous, the guilt of Hippolytus. Furthermore, Aphrodite does not give the information to Hippolytus any more than she gives it to Theseus. To both it comes indirectly to the one, through the nurse; to the other, through Phaedra's letter.

To conclude: The objections to the reading del¿w dè Onoeî πрâyμа are not well sustained. I think the reading should be retained, (a) because it has the unanimous support of the MSS.; (b) because it is not really inconsistent with the plot, does not force us to an erroneous conclusion, and may be paralleled in Euripides' own works; (c) because it is suited to the context.

This paper was read, in the author's absence, by Professor Earle, of Bryn Mawr College.

14. One of the Debts of Roman Literature to Early Roman Tragedy, by Professor Karl P. Harrington, of the University of North Carolina.

Although, unhappily, it is quite beyond our power to estimate accurately all the debt of Roman literature to early Roman tragedy, it is clear that the tragedies

of Livius Andronicus, Naevius, Ennius, Pacuvius, and Accius were quite largely responsible for at least the following effects: 1. The awakening of an interest in Greek literature and the cultivation of literary taste. 2. A dissemination of knowledge of Greek mythology. 3. A metrical revolution. 4. The development of a polished literary standard of forms, syntax, and style, which is rendered more noticeable when compared with the colloquialism of comedy. 5. The enrichment of the language by the coinage of new words.

The purpose of this paper is to determine as far as possible to what extent, and along what lines, the language and literature of Rome were enriched by the word-coinage of these early tragic poets.

The period from Livius Andronicus to Accius was the great formative period of Latin. The Romans were suddenly awakened to the fact that their language was a rough and primitive instrument for the expression of the exact, the delicate, the picturesque idea. The early tragic poets must therefore have been constantly hampered, among other difficulties, by the lack of suitable words with which to convey their finer shades of meaning.

We must acknowledge at the start the limitations within which we are compelled to work. We seldom, if ever, can attain mathematical certainty with regard to the author of a word, and the date of its genesis; for some earlier writer, whose works are lost, may have used it. Moreover, a large proportion of the meagre tragic fragments that we do possess have been preserved to us merely because they contain in each instance some unusual word. Again, the condition of the text sometimes leaves us in doubt.

Having frankly admitted, however, that we may look for only a varying degree of probability in our conclusions, we may take courage from the following facts: 1. The old Latin grammarians frequently leave us no reasonable ground of doubt. 2. A comparison of the original Greek in many cases makes it nearly certain that the word arose then and there as a translation. 3. A large number of words bear on their faces the stamp of mere linguistic experiments that were never imitated. 4. A reasonable regard for the laws of word-derivation in Latin and a constant comparison with what seems to have been taking place in that line in each period of the language will enable us to keep our conjectures from going too wide of the mark.

In the preparation of this paper Ribbeck's collection of the fragments of the tragic writers has been used for the text; but, as a rule, conjectural readings have not been taken into account.

LIVIUS ANDRONICUS.

There are but 42 verses or fragments of his tragic writings extant. Of the 10 words which appear to have more or less claim to be considered the creations of Livius Andronicus, 3 are transliterations from the Greek, 4 are new compounds, I is a new form of derivative, and 2 are new adverbs. Not a single one seems to have become very common in the language, and 2 (of the whole number) never occur again. More of the new words are new compounds than are found under any other head. The proportion of new words is nearly one to every four fragments of verses. Two (simus and inhumigo) seem worthy to have endured.

NAEVIUS.

In the case of Naevius more new words occur in his comedies and his Bellum Punicum than in the tragedy fragments; with the former, however, we are not now concerned.

There appear to be about 17 new words in 65 verses, the percentage being but a trifle larger than in Livius Andronicus. Of these, 6 (more than 1) are taken from the Greek, 5 are new compounds (3 of these being of the poetic picturesque type), and the others are new derivatives. Although but one of this list is surely a amağ λeybμevov, most of them seem to have been confined to the ante-classical period, except in so far as they were revived in comparatively late Latin. It is perhaps noteworthy that there are more Greek words in the Campanian Naevius than in the Greek slave Andronicus. It was worth while to use valentia, which might well have held its own against the more cumbrous valetudo; and of course such words as suavisonum and frondifer deserved to be repeated often by succeeding poets.

ENNIUS.

With Ennius the field of our investigation widens. There remain of his tragedy over 400 verses or fragments. Here the proportion of new words appears to be much smaller than in the two preceding writers, there being but 42, or about 1 in 10. A study of these, however, brings out several points of interest. 1. There are only 2 Greek words in the list. Ennius was too great a poet to borrow his diction extensively from the Greeks. 2. On the other hand, there are about 20 new compounds (about of the whole number); there are also several derivative verbs and nouns in -men and -mentum. 3. The äπağ λeybμeva number 7; signitenens and velivolans are typical poetic descriptive words, such as every true Roman poet in the earlier period coined; blandiloquentia and visceratim are strong words and deserved the better fate of being incorporated in the permanent body of the language; obvaro and augifico seem rather needless compounds; but hariolatio is a word that has no good equivalent, and could ill be spared. 4. Derivatives like hostimentum and peniculamentum apparently did not meet with a favorable reception. Compounds in ficus and fico seem to have been overdone, e.g. in the case of augifico and regifice. 5. Quite a number of these words coined by Ennius became the lasting possessions of the language, such as: exalbesco, flammifer, velivolus, deflagro, reciproco, regimen, inauratus, optumates, pervicacia, nearly of the whole number. 6. Several more were used by the early poets after him, influencing the whole language through them, and then, falling out of use for awhile, were taken up again in the post-Augustan or postclassical period, such as: propitiabilis, derepente, nitido, elimino, flacceo, evisceratus. 7. Ennius was not conservative in the form of words, as is evidenced by his variant forms in the case of common words. So, for instance, he uses caementa (f.), but caementum (n.) was the form that endured; similarly sanguen for sanguis, veter for vetus; further we find velivolans and velivolus side by side, the latter of which endured; and tabum was added to tabes, and both endured. 8. It is not easy to see why such words as expectoro, altisonus, scrupeus, convestio, and conglomero should not have cut a more important figure in the later language. 9. If, on the whole, Ennius coined fewer words than might have