39. That Koipavos is a proper name as indicated by κρέκει. Κοίρανος of Miletus or Paros is said to have been saved from drowning by a dolphin, and the similarity of this story to that of Arion suggests music. Пάoσ (« embroider»), in 1. 41, is associated with the other sense of xρéxεw (weave), and the primary meaning of wάoσε leads not inaply to σπόδον.

II, 5-8. The cohesion of this group depends on the alternative identification of Gyges as (1) the Lydian king, (2) the brother of Kottus and Briareus; cf. Hes., Theog., 149, Κάττος τε Βριάρεώς τε Γύης (v. 1. Γύγης) θ', ὑπερήφανα τέκνα.

11. L. Koo.

15. Θρύπτεται is a probable reading and is supported by Kаnπаdoxía, though why that province in particular was chosen as the representative of Asiatic effeminacy is not clear.

19-20. The Capitol is associated with Zeus rather than Cronus, who however as identified with Saturn was especially honoured in Rome, and according to Suidas the Capitol Σατούρνιον πρότερον ἐκαλεῖτο.

21. τε[μ]πλον : πέπλον is hardly possible, whereas the Latin word is commended by its two neighbours.

23. The surface of the papyrus is damaged in the middle of the symbol below this line, and the formation of the symbol remains in doubt; the two small half-circles may well have been connected by a longer curve, through which the horizontal stroke passed.

24-27. Here the connexion preceeds by way of contrast enuía characterises the guardian of holy places, while its opposite is typical of the gambler.

32 sqq. Λέανδρος appropriately follows διαπερᾷ, and κρηπίς, if that is the right reading (Κρήτης is less suit

able), is not unintelligible (= quay?), though not quite the word that would be expected.

37-39. τρίπους looks obvious between Πυθία οι Πυθώ and e. g. iɛpetov, but cannot be read. The π is followed by a vertical stroke which would siut p, v or, but if it is v, there is no room for an o.

40-41. For the Dioscuri and Samothrace, cf. PLUTARCH, Aemil. Paul., 23, διέπλευσεν εἰς Σαμοθράκην καὶ διαφεύ γων ἐπὶ τοὺς Διοσκούρους ἱκέτευεν; PHILO BYBL., Fr. 2, 11, Διόσκουροι ἢ Καβειροὶ ἢ Κορύβαντες ἢ Σαμοθρᾷκες.

III, 1. The letter after Zapd was apparently x or v, and Σαρδικός or Σαρδίνι[ο]s could be read. Σικυών is presumably misplaced; cf. I, 16, n.

3-5. Teyens is a false form and Teye άrns is probably to be read; cf. PAUSAN., VIII, 53. In any case the prominence of Apollo and Artemis at Tegea makes the transition to Cynthus easy, and this suggests the name of another hill also ending in -v0os. The correction in 1. 5 may be by the original hand.

10-12. Only three lines instead of the normal minimum of four have been given to this group, and perhaps two words were mistakenly written in one line, as in III, 1 above. In l. 12, Uτvov is clear, and the preceding vestiges rather suggest a, but this gives no word; the frepetition of the syllabe -vov is noticeable.

22. L. ép. This word is followed by a high stop, in addition to the usual double dots.

Oxford.

SUR LE (N) ÉGYPTIEN

DEVENANT P (R) EN COPTE,

PAR

M. PIERRE LACAU.

Le rapprochement établi par Spiegelberg (1) entre le mote nmhw et le copte p2€ (sah.) PEM2E (boh.) libre est extrêmement intéressant pour le sens. Nous avons là une étymologie qui éclaire d'un jour tout nouveau le terme égyptien si discuté. Mais au point de vue phonétique le rapprochement semble d'abord faire difficulté. En effet, un initial ne devient jamais p en copte il correspond dans cette position soit à xe =AAC «langue, soit à N:NAY «heure. Il faut expliquer cette anomalie.

:

Spiegelberg, qui a bien vu la difficulté, propose la solution suivante. Il y aurait eu assimilation du n initial de nmḥw au r final du verbe 'r dans l'expression 'r-nmḥw qu'on rencontre en démotique (2) et qui se retrouve en copte sous la forme ppм2Є: Єppeм2e «devenir libre ». Mais, en réalité, nous n'avons pas d'autre exemple d'une pareille assimilation régressive. Nous connaissons seulement le phénomène inverse: le N (article pluriel ou préposition) s'assimilant au p initial du mot qui le suit, exemple : Pwмe devenant PPOмe (3). L'explication proposée demeure donc une simple hypothèse.

Sethe suggère une autre possibilité le mot nmhw aurait été décomposé, par étymologie populaire, en p

(1) Ä. Z., 53, 116.

(2) GRIFFITH, Rylands, III, 2, V, 2 VI, 2, VII, 3 — Tous ces exemples dans des contrats de servage".

Corpus pap. n° 7.

(3) STEINDORFF, Kopt. Gram.', $ 50-51. Le phénomène est d'ailleurs restreint à certains textes anciens du dialecte sabidique.

RECUEIL CHAMPOLLION.

46