¹ See Welcker, , Die griechischen Tragoedien, ii., p 644 i. p. 349Google Scholar; Ribbeck, , Die römische Tragoedie, pp. 162 and 561Google Scholar; Robert, , Arch. Zeit., 1878, p. 16Google Scholar; Wecklein, , Münchner Sitzungsberichte 1888, i., p.87Google Scholar; Wernicke, Pauly-Wissowa i., 2156; Engelmann, , Archaeologische Studien zu den Tragikern; cf. also below, p. 111Google Scholar.

² Robert, loc. cit. p. 18, tries to prove, on very insufficient evidence, the existence of a prologue in the customary form, and, strangely enough, supposes it spoken by Echo.

The scholion which refers to the beginning of Andromeda's lament, Fr. 114, as tells against and not in favour of Robert's theory, as Wecklein demonstrates, p. 87, ff. Engelmann's use of the British Museum Hydria, E 169, as evidence for a prologue of this kind, which S. Reinach, (Revue Critique, 1900, p. 109Google Scholar) thought ingenious but Bethe, (Deutsche Literaturzeitung, 1900, p. 2082)Google Scholar rejected, falls to the ground along with the accompanying interpretation of the vase.

³ Wecklein very rightly separated them, but gives the second verse to Cepheus.

⁴ In support of this view note that in Ennius (assuming that Ribbeek, p. 167, was right when he gave Fr. xvi of the incerta to Andromeda) Perseus addresses his question as to who she is not to the mother herself but to the daughter.

⁵ Robert considers such a promise inconsistent with the later refusal by Cepheus, and he does not believe that Perseus spoke with Cepheus at all, previous to the death of the monster. This view is refuted by the vases, the fragments, and the action of the piece.

⁶ Several fragments of Ennius belonging to this part are preserved, v. Ribbeck, p. 168 f.

⁷ Laertius Diogenes, iv. 29, quotes these words as ἐχόμενα therefore immediately following Fr. 129. In the story told there, certainly the one quotation follows as an answer to the other, but this was scarcely possible in Euripides, as the question seems to require stichomythy, while these verses evidently form the close of a longer speech. Indeed, the verses themselves give no answer to Perseus' question: ὦ παρθρ´ν᾿ εὶ σώσαιμί σ᾿ εἴσῃ μοι χἀριν and in the earlier situation they are scarcely suited to the virgin modesty of Andromeda.

⁸ In Wecklein's attempt to assign the character of Phineus to Euripides' drama, we look in vain for any argument even partially satisfactory. His theory which gives Fr. 149 to Phineus (in whose case we should expect to find the ἔρως emphasised), 141 to Phineus, spoken to Cepheus, 142, and then later 143, to Phineus, similarly addressed to Cepheus, is to me inconceivable. Indeed, even though Wecklein finds in Fr. 144 “the most convincing proof of a dispute of the kind between Phineus and Cepheus,” there are obviously no proofs of it.

⁹ I have marked the first four vases with the same numbers as Trendelenburg, , Annali, 1872, p. 169Google Scholar, thus:

A. Mon. d. I. viiii. 38, Engelmann, p. 72.

B. Memorie delľ accademia ercolanense, 9. Pl. to p. 197; cf. Watzinger, de vasculis pictis Tarentinis, p. 39Google Scholar.

R. Rochette Mon. ďAnt. fig. Pl. 41; Engelmann, p. 73.

D. B.M., Vases, F 185Google Scholar, iv. Pl. vii. Engelmann from ditto, p. 8.

E. Krater, Berlin, Arch. Anz. 1893, p. 91, 50Google Scholar; Jahrb. 1896, Pl. 2; Engelmann, p. 69.

F. Hydria, Berlin, Arch. Anz. 1893, p. 93Google Scholar, 57.

G. Vase in Bari, Engelmann, p. 6, perhaps the one briefly described by Barnabei, in Bull. d. I. 1885, p. 50Google Scholar, which Helbig, p. 52, compared with a Capuan vase not represented in any of the drawings belonging to the Institute.

H. S. Angelo vase. See below. All these vases were found in Lower Italy (F ?). E is assigned by Furtwängler as Attic to the end of the 5th century B.C., so that it might have been painted soon after the performance of Euripides' Andromeda.

¹⁰ Cf. Trendelenburg, p. 118. But the special form to which Milchhöfer and particularly Wolters, (Ath. Mitth. 1891, xvi. 391Google Scholar ff.) drew attention can perhaps only be traced on H and, imperfectly, on C.

¹¹ It is impossible to guess what the unskilful craftsman who painted C meant by the two “conversing” figures to the right. Their gestures make them appear deaf and dumb. But cf. Trendelenburg, , Annali, 1872, p. 114 f.Google Scholar, on the restoration of the picture.

^lla Engelmann, loc. cit. p. 76, prefers to assign this part to Aphrodite.

¹² Hence the absence of the mask of Phineus would be no proof of the absence of that character in Euripides' play. Robert, p. 15–20, evidently considers these to be the masks of the whole drama, not of one scene. Similarly Wernicke.

¹³ Mon. d. I. vi. 40.

¹⁴ This is expressed so clearly that one hardly understands how Koerte, , Urne Etrusche, ii. p. 103Google Scholar, could dispute it solely on account of trifling and unimportant inconsistencies.

¹⁵ Patroni, G., Atti della r. accademia di Napoli 1894, vol. xvii, ii., Pl. 5Google Scholar, referred it to an adventure of Paris. The correct interpretation was touched on in Röm. Mitth. 1895, p. 95, and more exhaustively treated by Trendelenburg, , Arch. Anz. 1896, xi. 204Google Scholar.

¹⁶ Koerte loc. cit. p. 104, 3, assumes a change of scene on grounds which are unintelligible to me, like most of his remarks in this connection.

¹⁷ If this, as we may assume, was referred to in the narration of the fight, then certainly Perseus himself could not have been the narrator.

^17a Loc. cit. especially pp. 163, 169 and 564. For the opposite view v. Robert, loc. cit. 5, 17, 12.

¹⁸ Birch, , Archaeologia, 1855, 36, 1 p. 53, 68, Pl. 6CrossRefGoogle Scholar; Minervini, , Memorie d. accad. ercolan., 9, p. 221Google Scholar; Trendelenburg loc. cit., p. 111; Furtwängler, , Arch. Anz. 1893, p. 91Google Scholar; Knatz, , Quomodo Persei fabulam artifices graeci et romani tractaverint, Bonn diss. 1893, p. 34, 2, and 53Google Scholar; Bethe, Jahrb. 1896, p. 299Google Scholar; Wernieke, loc. cit., p. 215, f.; Smith, C., B.M. Vases, iii. p. 152Google Scholar; Engelmann, loc. cit., pp. 10 and 66.

¹⁹ Arch. Anz. 1855, p. 65. His own explanation (Tithonos, Eos, Memnon), was as foolish as possible.

²⁰ Probably the painter did not intend to make one post shorter than the other. It happened while the background was being blacked in round the head of the kneeling Ethiopian.

²¹ Fr. 125. The words παρθένου τ᾿ εἰκώ τινα ἐξ αὐτομόρθων λαΐνων τειχισμἀτων σοφῆς ἄγαλμα χειρὁς have been misunderstood by those ancient writers who talk of an αὐτοφυὲς ἄγαλμα ‘Automorph’ i.e. “natural formation” means the stone structure in front of which, like an image in high relief, made by an artist σοφῆς χειρός) the maiden stands. In C, where one is almost reminded of skulls (cf. Accius, Fr. x. Immane se habet templum obvallatum ossibus), and on Etruscan urns it resembles a niche in the rock. Cf. Bethe, , Jahrb. 1896, p. 296Google Scholar. From this is taken Antiphilus' epigram in Ovid iv. 672, Apollodor. ii. 43.

²² Thus Birch, Minervini; Bethe, p. 298 f., even goes so far as to suppose that the Ethiopian seated to the left on E is female and the representative of the Chorus, although the figure is exactly the same as the one in the Vulci hydria.

²³ With this we may compare the relief of Aphrodite raised from the sea, for it is strikingly similar and older than Sophocles, . Ant. Denkm. ii. Pl. vi. Röm. Mitth. 1892, p. 71, Pl. iiGoogle Scholar.

²⁴ This flabby figure reminded Engelmann (loc. cit., p. 66) of a puppet, or at least he assumed that this idea had occurred to others. Further, in the representations of the Andromeda of Euripides, A, B, Cepheus is supported by an Ethiop. Whether Euripides transferred this attendant figure from Phineus to Cepheus, regarding it merely as a support of old age as in the case of Hecuba, (Hec. 594)Google Scholar, or whether the vase-painter made the change, seems impossible to decide.

²⁵ On this account Birch and Trendelenburg supposed the figure to be female, for the offerings were such as could be intended only for a woman.

²⁶ Robert, p. 17, with whom Bethe agrees, conceives of Sophocles' Andromeda as played in front of the palace. He appears to have no definite grounds for this view except his own opinion that Perseus (whom he quite rightly thinks of in Sophocles' play as walking, not flying on at his first appearance) could arrive on foot only near the palace and not on the shore. There seems to be no difficulty here except in Robert's own preconceptions. For the Vulci vase-painting refutes his view that “Perseus' first appearance is from the shore after slaying the monster.” Tümpel, Die Aethiopenländer des Andromedamythus, p. 132Google Scholar, would like to shift the scene to Persia because the σάρητον mentioned in Fr. 131 is a Persian garment! In p. 177 he tries to exclude Phineus from the Andromeda of Sophocles! Unfortunately the vase shows, not only Phineus, but also his negro attendants.

²⁷ Did old Pheres, like Phineus, walk with some support ? Certainly it was in accord with the situation to represent him as feeble as possible.

^27a Compare the Cepheus of Euripides on the vases ABE.

²⁸ In Apollodorus, also, Phineus is the only adversary. Hyginus gives Agenor as the betrothed of Andromeda, and makes him an ally of Cepheus.