¹ Ionian Antiquities III 22.

² Newton's excavations at Halicarnassus and Cnidus are published in his ‘Discoveries at Halicarnassus, Cnidus, and Branchidae’ (1862–3: vol. 1, plates, folio: vol. 2, text, cr. octavo), of which ‘Travels and Discoveries in the Levant’ (1865) is a popular abridged account. They are also described in a series of despatches written by Newton on the spot, and addressed to the Foreign Office. These were printed and presented to both Houses of Parliament thus:

1858 ‘Papers respecting the excavations at Budrum.’

1859 ‘Further Papers respecting the excavations at Budrum and Cnidus’.

The despatches sometimes give useful information which does not appear in the Discoveries.

³ Brunn, Gr. Götterideale 42.Google ScholarSmith, A. H.B. M. Catalogue of Sculpture II no. 1300.Google Scholar

⁴ Most opinions have been vague, and have ranged from Scopas on the one side (Gardner, E. A.Six Gr. Sculptors 191Google Scholar) to Praxiteles on the other (Klein, Praxiteles 371Google Scholar). As late as 1938 Süsserott, (Gr. Plastik 178 n. 181)Google Scholar described the statue as the work ‘eines Praxitelikers,’ which is much what Collignon, (Hist. II 362)Google Scholar said in 1897 (‘contemporain de Praxitèle’) and Waldmann, (Gr. Orig. 155) in 1914Google Scholar (‘Kunstschule: Praxitelisch’). These are sure signs that the sculptor has not yet been correctly identified. See also Ruhland, Eleusin, Gòttinen 89Google Scholar: Löwy, Gr. Plastik 84Google Scholar, and—for those who can enjoy the aesthete of 1903—Gurdon, in Weekly Critical Review Aug. 27th, 1903, 136 f.Google Scholar

⁵ Although A. H. Smith (l. c. note 3) rightly described the whole statue as of Parian, and Collignon, (Histoire 362)Google Scholar merely remarked that the head was ‘sculptée dans un bloc de Paros différent du marbre employé pour le reste de la statue’, this did not satisfy Gardner, E. (Handbook (1905) 414)Google Scholar, who said that the body was of inferior local marble, or Lawrence, (Cl. Sculp. 266)Google Scholar, who degraded it to ‘a local stone of poor quality’. I cannot detect any difference in the kind of marble, which is Parian of very fine grain in both head and body. The block used for the head may be of slightly better quality, but the grain is of the same size, and although the colour seems rather warmer, this is due to an accident of preservation, for the protected parts of the statue, e.g. under the sides of the throne, are equally warm in tone. The colder tone is not confined to the clothing, and thus cannot be the remains of a dark paint, appropriate though this would have been. The body has had a severe battering, and has split where the marble was weakest; streaks of schist are almost impossible to find: there is a faint trace of a weak stratum in the middle of the thighs, hardly visible traces of others behind it at intervals along the left edge of the throne, and also a grey streak in the back leg of the throne: but otherwise the block is, to the naked eye, remarkably uniform in structure. Nor is there any need to infer, from the supposed differences of marble and workmanship, that the body is by a different sculptor: it must have been conceived by one man, but the differing states of preservation make it difficult to decide whether it was carved entirely by him. It is perhaps not easy to imagine the creator, if he were present when the statue was assembled, leaving the folds of the himation at the back of the head so inorganically connected with those on the shoulder (pl. IX, b, c). This raises the question whether the statue is, perhaps, a contemporary replica of an original set up somewhere else, presumably in Attica.

⁶ Poulsen, F.From the Collections of Ny. C. Glypt. II 179Google Scholar, ‘sits on a simple backless stool, as does the Knidian Demeter’; so Klein, (Praxit. 368)Google Scholar ‘Marmorsitz ohne Lehne’. See Appendix IB.

⁷ Memoirs XVIII (1941) 71 ‘the Demeter from Knidos in the British Museum must (because of the classicising head and shallow linear drapery) be very late (c. 100 B.C.?)’.

⁸ The material for this essay was mostly collected in 1929: the substance of it was delivered as a lecture to the Hellenic Society in May 1946.

I thank the Trustees of the British Museum for permission to publish pll. I, V, X, e, XIII, XIV: Mr. C. O. Waterhouse for his patience with a difficult subject; pll. III, IV, VII, VIII, IX, a, X, a-d are his photographs: the remain der are mine. Mr. Waterhouse also drew the text-figures 1–4: and Mr. D. W. Akehurst pl. VIII, d.

⁹ The hair is also confined by a ribbon, which shows in front of the fold of the himation on the top of the head (pl. IV, r). In Caylus, Recueil VI pl. XLVI 1, 2Google Scholar is published an engraving of a veiled marble head at that time (1764) in the possession of M. Peyssonel, consul of Candia, and said to have been found shortly before in Rhodes. Though not a replica it bears in the engraving a remarkable though probably accidental resemblance to the head of Demeter. I cannot discover its present whereabouts. Cf. also Einzelaufn. 1190, which looks false.

¹⁰ Throughout this article, even when the throne is being discussed, ‘right’ and ‘left’ refer to Demeter's right and left, not the spectator's.

¹¹ Compare the vase in Leningrad (Schefold, Kertsch. Vas. 19bGoogle Scholar, figure on r.).

¹² Curious that the breaking away of the knees should convey so false an idea of the composition as to deceive F. Poulsen, who (From the Coll. Ny. C. II p. 178) says that the feet are crossed. The relief in Athens (Walter, Rel. kl. Akrop. 68Google Scholar, no. 117) preserves the transverse fold between the legs of a similar statue and explains how the error can arise.

¹³ A startling demonstration of the amount of movement in the whole design is provided by Gardner's, E. A. quaint device (Handbook (1896) fig. 99Google Scholar, and subsequent editions) also adopted by Klein, (Praxit. (1898) fig. 74)Google Scholar of illustrating the statue in looking-glass fashion.

¹⁴ See Appendix I B. Optical corrections, calculated for an oblique view from the right, are apparent in the left cheek, left eye, and in the hair on the left. A further differ ence, which is apt to be obscured because imagination tends, wrongly, to restore the broken edge of the veil on the right level with that on the left, is that the right side of the face was more exposed, and this is a reason to be added to those given above and below, for a three-quarter view of the statue being intended.

Another feature which escapes notice because of the broken veil is the skill with which the carving of the ears and the sides of the neck has been executed from the front, the marble of the veil having prevented access from the sides.

¹⁵ There are some doubtful points about these niches, of which a view is given in Disc. pl. LIV. Newton (id. p. 376) gives the width ofthat on the (spectator's) right as 2 ft. 5⅛ ins. and its height as 4 ft. 4 ins. This would be both too narrow and too low for the statue of Demeter, which is 2 ft. 9¼ ins. wide and 4 ft. 11 ins. high. In the plan on pl. LIII of Disc., the scale shows this niche as 3 ft. 6 ins. wide: the discrepancy may be due to the inclusion of the frame. The left-hand niche is taller, though its height ‘could not be ascertained, as the upper part is broken away’ (Disc. 717), and slightly wider. There is further confusion in the speculations on the possibility of the statue having fallen from one of the niches. It was not the seated Demeter that was found immediately below the cliff-face and almost under the niches (at 7 on Newton's plan in Disc. Pl. LIII), but the standing statue dedicated by Nicocleia (B.M. Cat. Sculpt. II no. 1301), at first identified as Demeter. Base, statue, and fragments of its limbs were found together, so that it would seem that they were near its original position. Moreover, the base is roughly rounded behind, it is dressed only so far back on the sides as a spectator can see from a three-quarter view; even the statue itself—and especially the head—is very roughly finished at sides and back. All the evidence points to its having stood in a niche: yet the left-hand niche seems too low to accommodate both statue and base, which together are just over 8 ft. high.

The seated statue, which we now commonly call the Demeter of Cnidus, was found at 2 on Newton's plan, and if it fell from the left-hand niche, which alone is large enough (though—from a photograph—probably not deep enough) would have had to roll uphill to reach the spot where it was found. Near it was the inscribed base Inscr. in B.M. DCCCXIII. The marble statuette of Kore (see pl. X and Appendix II, p. 25) was also found near (at 1 on Newton's plan), and so were the fourth-century terracottas (at 3 on the plan). Naturally some of the sculptures may have been moved in later times, but to move uphill a statue weighing just a ton is no easy task. Lime-burners commonly prefer to burn sculptures on the spot or break them up into readily portable fragments. The balance of the evidence, then, is against the seated statue having been in the left-hand niche. It may have been on one of the two square bases in the centre of the temenos (Disc. 392, pl. LIII), or, more probably, near where it was found, namely in the building towards its western end. It should be added that, in addition to the objects enumerated above, the thin sheets of lead inscribed with curses were also found at this spot, which would thus seem to be the focus of the sanctuary and a likely position for the cult-statues.

¹⁶ For example, two of the three cult-images of about this date which are reproduced in the relief in Athens (Walter, , Rel. kl. Akrop. 68Google Scholar, no. 117) were apparently turned at an angle to the spectator, and several of the Attic reliefs of the Eleusinian goddesses (e.g. Kern, AM XVII, 125 ff.Google Scholar) for other reff, see n. 4 p. 13) imply groups of which some of the statues were set thus.

¹⁷ The horizontal section is given by Caskey, Boston Cat. 66.Google Scholar

¹⁸ If, as suggested below, he was also a worker in gold and ivory, he may well have needed to study the Parthenos closely.

¹⁹ Casson, Acrop. Cat. II 232 no. 1331Google Scholar: to the reff, there given add Suhr, , Sc. portr. of Greek Statesmen 121 ff.Google Scholar, Gebauer, Alexanderbildnis (AM 63/4 (1938/1939) 101, K.67Google Scholar; Buschor, , Hellenist. Bildnis 9Google Scholar; Bieber, , Proc. Am. Philosoph. Soc. 93, 5 (1949) 380.Google Scholar This head is of course an original. The emphasis is necessary, for opinion has been wavering, confused, and even perverse. There are two copies, one at Erbach (Stark, , Zwei Alexanderköpfe (1897) 12Google Scholar), the other, from Madytus, in Berlin (Blümel Cat. K.203).

It is just possible that two originals of similar type, one in marble and one in bronze, were produced by the fourth-century sculptor, and that the Erbach head, which has more meticulous detail in the hair than the head from the Acropolis, is copied from the lost bronze.

Amelung long ago observed the identity of style between Demeter, and Alexander, (Ausonia III, 1908, 127 ff.).Google Scholar

²⁰ See, for example, the confused discussion in Suhr op. cit. 121 f.

²¹ Cf. Klein, AE 1900 1Google Scholar: Neugebauer, AA 1946–1947 1.Google Scholar

²² Not later than 338, for it was dedicated by Timotheos, who died in that year, as did Isocrates himself (Plut. Vit. X Orat. Isocr. 27).

²³ Paus. V 20, 9. Gebauer's theory (Blümel, , Berlin Cat. V k 203Google Scholar: cf. Buschor, , Mauss. u. Alex. 48Google Scholar) that the Erbach head may be copied from the chryselephantine statue at Olympia, is not irreconcilable with my own. I suspect that the so-called Alcibiades type, identified as Philip by Arndt, (Strena Heibig. 11Google Scholar, Arndt-Br. 467–70), may reproduce the portrait of Philip from this group at Olympia.

²⁴ See Johnson, Lysippos 67Google Scholar (with reff.). Waldhauer, (Über einig. Porträts Alexanders 51)Google Scholar argues that the animals were by Lysippus, the portraits by Leochares.

²⁵ It is customary to assume that Leochares was a native Athenian, but for this there is no direct evidence. Lippold, in RE XII 2, 1994Google Scholars.v. ‘Leochares’, cites one piece of indirect evidence, an inscription of Roman date.

²⁶ Loewy, Inschr. Gr. Bildh. 62Google Scholar no. 80 (‘east of Propylaea’); the head of Alexander was found ‘near the Erechtheum’ (Klein, AE 1900, 1Google Scholar). I regret not having had an opportunity since 1937 of again studying this head at first hand; it is just possible that a fresh examination might give a hint of the kind of body to which it belonged, and thus of the purpose and date of the dedication.

²⁷ This is usually stated as an established fact, e.g. by Geyer, in RE XIX 2Google Scholar, s.v. Philippos 2295, but it is no more than an inference from the authorities there cited, among whom only Pausanias (I ix 4) mentions a statue of Philip, saying that it, and one of Alexander, stood in front of the Odeion, but giving no date for either. I thank Sir Arthur Pickard-Cambridge and Professor A. W. Gomme for help in clearing up this point.

²⁸ Plin. N.H. XXXVI. 30: Scopas habuit aemulos eadem aetate Bryaxim et Timotheum et Leocharen, de quibus simul dicendum est, quoniam pariter cadavere Mausoleum … … … …31: ab oriente caelavit Scopas, a septentrione Bryaxis, a meridie Timotheus, ab occasu Leochares. …

Vitr. VII, praefat. 12: namque singulis frontibus singuli artifices sumserunt certatim partes ad ornandum et probandum Leochares Bryaxis Scopas Praxiteles, nonnulli etiam putant Timotheum.

²⁹ Cf. Amelung, , Aus. III 103.Google ScholarZschictzschmann, Welt als Geschichte I (1935) 435Google Scholar goes to the other extreme, and maintains that there was one designer only for the whole frieze.

³⁰ In situ is a phrase to be used with caution. Newton did not imagine that the slabs were still in position on the building, but wrote (Disc. II 100) ‘They were found lying in a row, and appear to form one continuous composition. They are in very fine condition, and, notwithstanding the extreme salience of the relief, are but little mutilated. It may be inferred from these circumstances that the spot where I discovered them is not far from their original position on the edifice, as, from the great weight of the slabs, one of which is 5 ft. 8 ins. in length, by 1 ft. in thickness, they could hardly have undergone much shifting about without presenting more signs of bad usage on their surface.’ Nothing can be added to this statement (except that Disc. I pl. IV shows that the slabs were not lying parallel with either the eastern or northern wall of the building): each must decide for himself what the probabilities are.

³¹ As did Neugebauer in Studien über Skopas ch. III. See also the reff. in n. 29 above, Lawrence, , Cl. Sculpt. 264Google Scholar, and Wace's article on the whole question of design and execution (Ann. XXIV–VI, 109), which appeared when my own was already at the printer's. The weakness of that article is its failure to suggest how designs were transmitted from the ‘master’ to the carver. Was it by drawing on papyrus, by models in clay or wax, or by some other method? And in what degree of detail? There is of course no ground for assuming that the procedure in decorating a building with sculpture was identical everywhere and at every period, but sculptors commonly find it easier to give expression to their ideas by carving or even by drawing, rather than by modelling, and on full scale rather than in miniature: that is why I think it likely that in the friezes of the Mausoleum the master or masters set out the design full-scale on the surface of the marble, and then supervised the execution, actually taking part in the carving where it was necessary to instruct, correct, or give the final subtlety of emphasis. How otherwise can one account not only for the variety but also for the excellence of the styles? The majority of fourth-century grave-reliefs and votive reliefs are much inferior, both in style and in execution.

³² Repeated attempts, inspired by Pliny's statement, have been made to connect these with Scopas (see Richter, , Sculpture and Sculptors 270)Google Scholar: the latest is by Buschor, (Maus. u. Alex. 46).Google Scholar

³³ Newton, Disc. II 99Google Scholar (cf. 90). He stresses the fact that the find-spots of most of the marbles have little value in determining their original positions.

³⁴ The eyes and lips of Alexander show a slightly freer, more impressionistic treatment than do those of Demeter, and this, unless it be due to the difference of subject, seems to argue a later date.

³⁵ Buschor, in Mauss. u. Alex., passim, argues for two main periods of work on the Mausoleum, the first from a little before 353 to 351, the second round 333.

³⁶ Susserott, , Gr. Plastik 178 f.Google Scholar

³⁷ Before his banishment in 330 B.C. Lippold, (Gr. Porträtst. 95Google Scholar: Kopien 210) puts it at the beginning of the third century, ‘perhaps 280 B.C.’. This seems too late. Among Panathenaic vases one in London (Süsserott op. cit. pl. 6r) dated 332 B.C., with its multiplicity of small folds, has something in common with the statue of Aeschines (though even more with that of Demeter) but looks earlier; whilst one in the Louvre dated 313 (Süsserott pl. 101) looks later.

³⁸ For a summary of the evidence for his date see Richter, Sculpture and Sculptors 282.Google Scholar

³⁹ Richter op. cit. 284 ‘we must admit that Leochares still remains a shadowy figure’.

⁴⁰ Carpenter's, remark in Memoirs XVIII 71Google Scholar, ‘I cannot recall any pertinent instance of an undoubted original so pieced together as early as 300 B.C.’ is baffling, unless the word ‘so’ applies only to the Girl from Anzio. Original statues of the fourth century are not numerous, and few are so complete as to show whether the arms or other parts were made separately. But when a sculptor was making the head separately—and we have a number of fourth-century heads that are so made—there was no law to prevent him dealing similarly with the arms or other projecting parts, if it happened to prove convenient. As early as about 400 B.C., in a statue in Eleusis, (Br.-Br. 536)Google Scholar, one foot was made separately although the head seems not to have been. If there is one undoubted original of the fourth century in existence it is Berlin K.10 (Blümel Cat. III pl. 14: from Athens) which is a head made for in-setting. Then there is the head of Asklepios from Melos (B.M. Cat. Sculpt. no. 550), the head of Zeus in Boston (Caslcey, Cat. 59 no. 25Google Scholar), and the female head there no. 27 (28 and 29—like the patchwork B.M. 1301 from the Cnidian temenos itself—have sometimes been thought later than 300 B.C.). In the statue of Dionysus (B.M. Cat. Sculpt. no. 432), of uncertain date but later set up on the monument of Thrasyllus (Welter, AA 1938, 33 ff.Google Scholar), the head is let into a rounded socket and held at the bottom by a substantial dowel (present hole 8 cm. wide by 5 cm., but this has been enlarged to extract the metal); the pour-channel runs from the front. The left arm, from the shoulder, was also made separately, being fixed into a mortice of which the remains are 15 cm. high by 13 wide. From the Mausoleum there is the colossal statue no. 1000, which, despite contrary opinion (e.g. Buschor, Mauss. u. Alex. 10Google Scholar), should from the coin-evidence be Mausolus (Hill, in Anatolian Studies (Ramsay) p. 207Google Scholar) and must anyhow be of the fourth century: head and arms were all made separately. There is the rider 1045, the upper part of whose body was made separately; the colossal head 1052; the large head of Apollo 1058 (with its fragment of socketed neck 1061); and the smaller heads 1055 (which is life-size, not colossal as stated by Jongkees, (JHS LXVIII 37)Google Scholar), and 1056; finally 1065 (socket to receive inset head) and the necks numbered in the British Museum as MRC 51 A B and C (made for insertion into sockets in statues, and now broken from their heads). Admittedly some of the sculptures found on the site of the Mausoleum may be later than the original building, but scarcely all these. I exclude the head in Berlin from a relief (Blümel Cat. III K.43), the separately-made parts among the sculptures of the Parthenon, and the statue attributed to the Hephaisteion, (Hesp. XVIII pl. 51–2)Google Scholar, as being architectural sculpture; and I omit all archaic examples of the practice.

⁴¹ As in that shown on the relief in Berlin (Blümel, Cat. III K.106 pl. 84Google Scholar), or in the statue in Leningrad (Waldhauer op. cit. III 248 pl. XV (our pl. VIII, f). It is possible that the back of the throne was gable-shaped or rounded at the top, and not horizontal; there is a faint groove on the back of the left shoulder of the statue which may indicate its slope: this is not a convincing explanation, but I am at a loss to account for the groove on any other hypothesis.

⁴² AM XXVI 333, pl. XIV: FR pl. 88.

⁴³ If it was fitted exactly into the slot.

⁴⁴ See Richter, Anc. Furnit. 13 ff.Google Scholar

⁴⁵ B.M. Cat. Sculpt. II no. 1308. Total length 10 cm.: of head only, 9 cm.

For rams in general see Richter, Animals 27Google Scholar, and for late archaic and early classical rams Jacobsthal, Mel. Rel. 135.Google Scholar The features which distinguish our ram's-head from those made earlier, for instance the clay ram's-head cups of Sotades and his followers (Beazley, ARV 451–3Google Scholar) are the differentiation between bone and flesh (which involves deeper and more detailed modelling), increased feeling for the elasticity of the skin, and greater interest in the surroundings of the eye, the shape of the eye-ball and the way it lies in and projects from its socket. But it is sometimes not easy to determine date and style in a subject where Nature has already done so much for the sculptor, whatever his period: some comparisons are possible with fourth-century coins of Salamis in Cyprus, as Mr. E. S. G. Robinson has shown me, and with bracelet-ends, e.g. that in the Ashmolean Museum, from a fourth-century tomb in the Crimea, (JHS V (1884) 68 pl. XLVII—misleading)Google Scholar, but I have found nothing conclusive. One relief almost contemporary with Demeter should be cited, that of Demetria and Pamphile (Conze pl. XL, Diepolder, Att. Grabr. 53 pl. 51Google Scholar), but it is disappointing: the ram's-head is far less vigorous than ours and it lacks the bold aquiline nose (said to be more pronounced in the male animal (Sandars, Beast Book 208Google Scholar)).

⁴⁶ Roman copy of a statue of Cybele of the fifth century B.C. (Waldhauer, Ant. Skulpt. Ermit. III 20 no. 248Google Scholar). Armrails with sphinx-support and ram's-head finial, especially in reliefs and on vases, are too common to enumerate. For a sphinx-support projecting backwards over the cushion of the throne see Walter, Rel. kl. Akrop. 60 no. 104Google Scholar; and for a low arm-rail id. 87 no. 182. The general arrangement is usually the same, though the projection of the ram's-head may vary; for an abnormal arrangement see the rough archaistic statue in Leningrad (Waldhauer op. cit. I no. 9 pl. VII).

⁴⁷ This process consisted in drilling a number of holes with a large drill straight down into the block and then breaking away the walls between them. I previously thought that the hollows visible were the remains of punchmarks, but Mr. J. Brennan, whose help on various technical points I gratefully acknowledge, has convinced me that they are drill-holes.

⁴⁸ As in the statue of Dionysus no. 432 in the British Museum: for its association with the monument of Thrasyllus see Welter, AA 1938 33 ff.Google Scholar

⁴⁹ See pl. VIII, g, and Richter, Anc. Furnit. 20 f.Google Scholar Two archaic panels from Sardis in the British Museum (Pryce, Cat. Sculpt. I 1 (1928) B. 269–70Google Scholar) may have come from reproductions of furniture, but neither is identical in form with the panels missing here.

⁵⁰ It was Mr. J. Brennan who observed that the seat was not level. He points out that there is similar tilting in some of the archaic statues from Branchidae e.g. that of Chares in the British Museum (Pryce, Cat. I 1110, B.278: pl. XIIIGoogle Scholar). Both front and back legs of the throne of Aiakes AM XXXI 151 have a strong tilt backwards. Possibly the tilt was sometimes produced by the insertion of an extra member under the front legs: in the marble throne from Eretria cited in n. 42 above and illustrated on pl. VIII, d, the lowest members are of a different colour (perhaps intended to represent metal): the back legs, unfortunately, were not shown.

The hollowing-out behind (L, N fig. 1) is so different in character—it is not even cut level at the top—that I find it hard to believe it was done by or under the eye of the maker of the statue: more probably by a contractor lightening it for transport. It certainly implies that the back was not intended to be seen.

⁵¹ Newton Disc. pl. LIX nos. 3 and 4: Walters, B.M. Terrac. C.427, 507A.Google Scholar For Newton's no. 4 (my pl. X, d) cf. Neutsch, and others Die Welt der Griechen (Heidelberg 1948) 21, no. 19, fig. 10).Google Scholar

⁵² This type has been frequently studied, notably by Amelung, in Basis des Praxiteles 50Google Scholar: a convenient summary of the literature up to 1932 is given in Rizzo, Prassitele 100, 118.Google Scholar

⁵³ Mr. Reynold Higgins, who has a wide knowledge of terracottas of all periods, has kindly examined these two, and agrees with a date in the late fourth or early third century B.C.

⁵⁴ In Cos, which was as close to Cnidus by race and tradition as it was geographically, have been found three votive statues which reflect, though with variations due to the fancies of Hellenistic sculptors, this same type (Cl. Rhodos V2, 169–85). It was evidently the favourite type of Kore in those parts, and Cnidus has some claim to be the shrine which housed the statue from which all these versions derive.

⁵⁵ B.M. Cat. Sculpt. II no. 1302. Parian marble. Ht. (excl. base) 42 cm.

⁵⁶ In the Vatican and the Albani, Villa. RM IX 142.Google Scholar

⁵⁷ Mingazzini NS 1927 309. In AM LIII 48 Buschor has shown how this relief differs from ordinary votives.

⁵⁸ RM IX 134 ff.

⁵⁹ Disc. 423 ff. Mr. John Cook's current research into the topography of Cnidus will, I believe, illuminate this problem, and may fix the date of the temenos.

⁶⁰ See notes 52 and 61.

⁶¹ B.M. Cat. Sculpt. II no. 1315. Parian marble. Ht. 22 cm.

The date at the end of the fourth century proposed by Süsserott, (Gr. Plastik 194)Google Scholar for the Vienna type is surely too late. I agree with Waldhauer, (Ant. Skulpt. Ermit. III 38)Google Scholar that it is very early (though perhaps not the earliest) in the series. Even the late version on the Mantinean basis preserves the same general arrangement of the hair.

A case has been made out by Arndt, (Festschr. Overbeck 96Google Scholar: cf. also Ruhland and Löwy ll. cc. n. 4 p. 13) for a head in Munich being the original head of Kore from the statue which stood beside Demeter at Cnidus; there is evidence that it may actually have been found at Cnidus. Though recognising the high quality of this head I cannot see that it is of the same style as Demeter; there is no evidence that it came from the temenos. Rizzo, (Prassit. 92 pl. 138)Google Scholar gives Ostia as its provenience (the first publications—for which see Furtw. Beschreib. (1900) 182—describe it as of unknown provenience, bought in Naples) and says that it is of the same type as the ‘maiden from Herculaneum’, which it certainly resembles but does not seem to reproduce exactly.

⁶² The head has the appearance of being on a seated figure. The reverse bears a Victory to 1., r. hand outstretched and a palm-branch in her l.: inscr. ΚΝΙΔ ΑΠΟΛ ΔΩΡ. Not in B.M.C. Caria, but cf. the series p. 96 nos. 90–2.

Some of the copies of a type of Kore related to the types we are discussing (though none of them identical with the Cnidian terracotta pl. IX, d) seem to have had the head veiled. Examples are Einzelaufn. 357, 2284 (head restored but end of veil left), id. 2902, and Waldhauer, Ermitage III no. 279Google Scholar (where there is doubt whether the head belongs). A similar type, so far as can be seen from what remains of the relief, was used on one of the columns of the later Artemision at Ephesus (B.M. Cat. Sculpt. II no. 1211) and also seems to have been veiled. There is no evidence that it was Kore.

⁶³ Newton, Disc. pl. LXXXIX no. 15Google Scholar; Gr. Inscr. in B.M. no. 813.