are satisfied and give judgment in her favour and the reward her life has earned. grata humus, the earth as including the 'underworld,' oi xeóviol. So Tellus is appealed to in 1. 19. 16. Compare Eur. Herc. Fur. 45 xlovòs μéλaivav opøvnv eloéßaivev, and Prop. 11. 6. 31 a gemat in terris, &c.

101. moribus et caelum patuit. Ennius makes Scipio say ap. Senec. Ep. 108. 34 si fas endo plagas caelestum adscendere cuiquam, mi soli caeli maxuma porta patet. Compare for the sentiment Hor. Od. 3. 2. 21 sqq. and in a different connexion 4. 2. 23, 24 (of Pindar) uires animumque moresque aureos educit in astra nigroque inuidet Orco and compare iv. 18. 34 n. sim digna. Mr Pretor has suggested to me that sum would be more in keeping with Cornelia's self-asserting character, and this may be granted. But sim is probably genuine, and means 'may my deserts make good my title to be rewarded like my ancestors.' We have the subj. where the ind. might be expected in the very similar passage iv. 17 (18). 31 sqq. at tibi nauta pias hominum qui traicit umbras huc animae portet corpus inane tuae qua Siculae uictor telluris Claudius et qua Caesar ab humana cessit ad astra uia. merendo, by my deserts': absolutely, as in Virg. Aen. 6. 664 quique sui memores alios fecere merendo.

102. honoratis auis, the dat. of motion towards, to the abode of my distinguished ancestors.' Cf. 1. 20. 32 ibat Hamadryasin n. The мss. read equis or aquis. But the dative auis is justified by the imitations in the Consol. Liv. 125 tumulo portaris et igni, 188 nec poscunt tura ferenda rogo and v. 330 ille pio (si non temere haec creduntur) in aruo inter honoratos excipietur auos. Compare v. 162 ad ueteres conditus ibis auos. ossa, as in v. 20. uehantur, by the boat which conveys the good el. 7. 56 sqq. and 1.c. in last note. This led to the alteration aquis.

ADDENDUM IV. 7. 49. p. 165.

Professor Gildersleeve (American Journal of Philology, IV. p. 210) points out that to suppose that the passage refers to a real chamber or bed belonging to Paetus is inconsistent with pauper v. 48. "Non tulit, he writes, is oùK ễTλŋ=non is fuit qui ferret, from which we get for the contrast sed is fuit qui mallet. This Paetus was not the man to bear the sound of the piping storm, but he was the man (to have) his head propped on feather pillows of shot colours in a chamber (i.e. stateroom) of thyian wood or (of) Orician terebinth." For this use of eáλauos cf. Athenaeus 5, p. 207, who tells us the ship which Archimedes built for Hiero had a áλapos with doors of ivory and thyon wood.

APPENDIX A.

We have no existing мs. of Propertius that can be traced back earlier than the 14th century: and it is probable that all go back to a single archetype.

The Codex Neapolitanus (N.) has of late generally been considered to be the best. But its claims have been lately attacked by Prof. Baehrens (in his Preface, p. vii.), who has been replied to by Herr Leo (Rh. Mus. 31, p. 431), by Prof. Ellis (American Journal of Philology 1. p. 389), and Prof. A. Palmer (Hermathena, Vol. IV. 40-72). It is beyond my limits to enter upon this controversy; so I content myself with observing that it seems to me that Baehrens' views require very considerable qualifications. It is probably to be assigned to the 14th century.

Other мss. are the Codex Vossianus (about 1360, imperfect), the Florentinus (end of 14th or beginning of 15th century), Ottoboniano-Vaticanus (end of 14th century), Dauentrianus (1410-1420). All of these have been recently collated by Baehrens; and from them he reconstructs the archetype which he calls O.

Besides these may be mentioned the Perusinus (1467), which once belonged to Scaliger, and of which there is a collation in the edition of Mr Palmer, its rediscoverer, and the Groninganus (G.) which was formerly held in high esteem, but has lately fallen into disfavour, owing to its numerous interpolations.

That the text of Propertius as preserved to us in these мss. is far from perfect, every scholar will concede. It is none the less clear, however, that in many quarters there is a disposition to make it out to be more corrupt than it is, and to apply alteration in place of interpretation. I will enumerate some of the chief sources of its mistakes.

(i) Clerical errors. These are chiefly due to confusions of the cursive character, and shew that the lost archetype was written in cursive, e. g. stemus for flemus, 11. 1. 2.

But some of the corruptions go back to an earlier stage, and shew uncial corruptions, e. g. II 1. 5 COGIS for coccis; v. 11. 30 AERA for AFRA; Iv. 18 (17). 38 FVNDET for TVNDET.

(ii) Transpositions.

(iii) Lacunae.

Of these I have spoken, Introduction (pp. xlviii. sqq.).

(iv) Interpolations. There are very few of these. I am convinced of the spuriousness of only two passages, Iv. 3. 55, 56 (from 1. 2) and v. 9. 41, where a line has been lost. In certain cases injudicious editing has caused lines to appear in unsuitable positions.

Thus in III. 32 (26). 61 sqq. we have two versions of the compliment to Virgil, as Mr Munro has acutely pointed out; II. 1. 37, 38 is a genuine Propertian distich, out of place in its present position. Perhaps it was written in the margin of the poet's Ms. This may have been the case too with Iv. 6 (7). 23, 24 and note.

I have added a comparison of the readings in this edition with those of Baehrens and Palmer.

ADDENDUM.

Since the first edition appeared the comparative merits of the Mss. have been the subject of a careful examination by Solbisky (Comm. Ienenses II. (1883)). The division into books has been discussed from a new standpoint by Th. Birt (das antike Buchwesen, pp. 413 sqq.), who makes it probable that the poems consist of a single book published separately (the Cynthia monobiblos) and four books published together (tetrabiblos syntaxis). For a fuller account see the article Propertius in the Encyclopædia Britannica (new ed.).