¹ His name was naturally omitted from the report, but is stated by residents of Siphnos to be Lorenzo Martiglio.

² Délos XII, 8 ff., figs. 8–9, where other parallels are quoted.

³ The plan shows some disconnected foundations in this area which probably date from the late sixth or early fifth century.

⁴ Measurements in this section are given in metres, levek in height above sea-level.

⁵ Pottery and other objects such as terracotta. figurines which must have come from this deposit were abo found farmer down the hill in the trenches YA, YB, YC.

⁶ The almost total absence of votives after 550 can be explained by the supposition that temples were spring-cleaned at comparatively long intervals. The votives of the living generation would thus be still housed in the temple; only occasionally would a broken object be added to the rubbish heap.

⁷ It might of course have extended farther than the marble wall, assuming the latter to be later in date; but in that case the marble wall was designed to include a smaller area than existed before. One would naturally expect expansion rather than duction.

⁸ Scranton, R. L., Greek Walls (Harvard U.P., 1941).Google Scholar The surface of the two sections of our wall are what Scranton would, I think, describe as hammer face.

⁹ Levels: I, 87; II, 867; III–V, 865. IV and V are visible in PI. 3. 4; the other three were found in the test trench in 1935.

¹⁰ It made an angle of 100° with the fifth-century wall. For the sake of clarity it has been omitted from the plan.

¹¹ Pl. 3. 4 shows areas 4 and 5 after this wall had been removed.

¹² The sides vary between 8·5 and 10·5 m. in length.

¹³ The bottom of the hole is 84·73 m.; the grooved channel outside 84·79 and the top of the schist slab 85·06.

¹⁴ The greater part of the water supply of the modern village is rain water collected in cisterns from the roofs. The nearest spring in the valley is a weary trapse 300 ft. down and up again.

¹⁵ The floor stratum was unfortunately not clearly defined, but could be inferred at two points as will be explained later. It would of course be unreasonable to expect the floor of a Geometric house to be nearer level than the pavements in modern Athens.

¹⁶ The possibility that the hole could be used for the passage of liquids in the other direction, e.g., for emptying slops, is ruled out by the levels. The slab outside would also be pointless.

¹⁷ It is of course just possible, though very unlikely, that the house was entered by a wooden ladder from a door more than half a metre above the floor.

¹⁸ Marble may have been purposely avoided in order to render the village less conspicuous to sea raiders. One can imagine that these houses, built of the same material as the cliffs under which they nestle, would have been completely invisible from seaward except to a vessel navigating close inshore. It is worth noting that the wall in question is a party wall between two houses, and would therefore not be visible from outside. But this may be an over-ingenious explanation of what was probably a fortuitous choice of materials.

¹⁹ The level of the stone base is 86·13 m. and the slab below 86·04 m.

²⁰ E.g. in the Agora at Athens: Hesperia II, 547, where prehistoric parallels are cited.

²¹ The area outside the wall was badly disturbed by the intrusion of a cistern and two wells of medieval date.

²² In Pl. 4. 5 contrast the Geometric wall in the foreground with the motley medieval wall behind it.

²³ The following examples are published in the catalogue: oinochoe i (Pl. 12. 20); lid 5 (Pl. 13. 24); crater 9 (Pl. 14. 29); skyphos 13 (Pl. 14. 16). The Protocorinthian aryballos 2 almost certainly belongs to the votive deposit; it was found about 0·3 above the estimated floor level.

²⁴ Upside down in the photograph.

²⁵ Délos XII, fig. 17.

²⁶ A Delt 1923, 7.

²⁷ Typen der archaischen Architektur, 103. Doric temples in Delos from 540 to 450 B.C. are listed by R. Vallois, L'architecture hellénique et hellénistique à Délos, 128 f. T. J. Dunbabin kindly drew my attention to this book.

²⁸ It is hard to say whether, both here and elsewhere, such projecting spurs are a natural formation or whether they may have been formed artificially by quarrying the stone of which the houses are built.

²⁹ E.g., the amphora fragments 1, 9, in the catalogue below. The former was found under the schist slab on the hearth.

³⁰ Greek and Roman Architecture, 41, n. 5.

³¹ The property belongs to the Rapheletos family of Kastro.

³² One Mycenaean sherd was also found.

³³ Measurements in this and the following section are in millimetres.

³⁴ For the latest discussion on Boeotian bell-shaped figurines see F. R. Grace, Archaic Sculpture in Boeotia, ch. 1.

³⁵ A helmeted Arterais occurs, however, among the lead figurines at Sparta (AO, 274, pl. CXCVI).

³⁶ The peak in front makes it unlikely that long hair and not a helmet is meant.

³⁷ AJA 1934, 128; cf. Hampe, Sagenbilder, 67.

³⁸ Centaurs with conical helmets occur in oriental art, e.g., Perrot-Chipiez III, 604, fig. 412; cf. the curious conical headgear of the centaur in the New York bronze group (AM 1930, Beil. 38; Hampe, op. cit., pl. 30) which Buschor identifies with Zeus and Typhon.

³⁹ I owe this reference to Mrs. Joan Stubbings, for whose kind assistance in this section I am most grateful.

⁴⁰ Evans in JHS 1894, 1897. It is interesting to note in passing that a pale-green steatite seal with linear script comes from Siphnos (JHS 1894, 287, fig. 19).

⁴¹ A bronze ring from Vrokastro (Hall, Vrokastro, 120, fig. 70 E) also has a bezel which appears to be circular, but one cannot be certain of the shape from the drawing.

⁴² Later examples of this attribute : Cook, , Zeus I 291.Google Scholar

⁴³ Mycenaean examples discussed by Furtwängler, Ant. Gem. III, 35 ff.; Minoan, : CAH Plates I, 200dGoogle Scholar; Frankfort, loc. cit., 301, fig. 103 (Matz, pl. XIII, 6); cf. also the Cycladic ‘harp players,’ Bossert, Altkreta, 16 ff.

⁴⁴ Hesperia, Supp. II, 54. Another in New York illustrated by Valentin Müller in an interesting article on early Greek sculpture in Met. Mus. Studies V, 157 ff. T. J. Dunbabin calls my attention to another, Oxford, 1934. 320, said to be from Attica.

⁴⁵ For a recent discussion on the subject see Hesperia Supp. VII, section VIII.

⁴⁶ The following special abbreviations are used in this section:

Buschor=Buschor, E., ‘Kykladisches,’ AM 1929, 142 ff.Google Scholar

Eilmann = Eilmann, R., ‘Frühe griechische Keramik im Samischen Heraion,’ AM 1933, 47 ff.Google Scholar

Pfuhl = Pfuhl, E., ‘Der archaische Friedhof am Stadtberge von Thera,’ AM 1903, 1 ff.Google Scholar

Rubensohn = Rubensohn, O., ‘Die prähistorischen und frühgeschichtlichen Funde auf dem Burghügel von Paros,’ AM 1917, 1 ff.Google Scholar

Technau = Technau, W., ‘Griechische Keramik im Samischen Heraion,’ AM 1929, 6 ff.Google Scholar

⁴⁷ The summary accounts in JHS 1935, 163, and AA 1935, 237, which mention bucchero among the finds are therefore to be corrected. No archaic bucchero was in fact found.

⁴⁸ Judging by the description given by Rubensohn, in AM 1917, pp. 54, 67Google Scholar, the Siphnos sherds are imports from Melos rather than of local Parian manufacture.

⁴⁹ See BSA XXII, 187; BM Catalogue of Vase I, 1 p. xxx.

⁵⁰ Several other examples are quoted by Miss Lorimer, BSA XLII, 76 ff.

⁵¹ Upside down in the illustration.

⁵² On its side in the illustration.

⁵³ Upside down in the illustration.

⁵⁴ Some of the plain black glaze sherds may date from the second half of the sixth century, but they are in any case not numerous.

^54a J. M. Cook informs me that the Acropolis fragments seen by Brückner and Pernice (AM 1893, he. cit.) connect the nought and cross pattern with spiral and cable bands and a scrap of a figured band. These fragments were recently rediscovered under the floor of the Acropolis Museum—hence Graef's denial of their existence. Cook has also found a fragment with this pattern (three O's in the vertical column) on the site of ancient Kythnos.

^54b I have not had an opportunity of studying in detail Professor Picard's learned article in Mon. Piot 1941, which contains much relevant material.

^54c J. M. Cook has called my attention to two Attic representations of armed horsemen, dating from the beginning of the orientalising period: Berlin 31006 (Neugebauer, Führer, p. 8) and Karlsruhe (Welter, Bausteine, pl. i; Jdi 1907, 99, fig. 12).

⁵⁵ A fragmentary r.f. pelike was found in Grave 12 and is noted on p. 88.

⁵⁶ As these fragments are glazed on the inside I thought at first that they must come from a kantharos. I think it possible, however, having regard to their thickness, that they come from an amphora, since I now observe that W. Slope amphorae may be glazed inside.

⁵⁷ See also Attic fifth-fourth century, no. 1 (p. 58).

⁵⁸ The Cycladic material from Thera is restricted, with very few exceptions, to Theran and Parian. The most that can be deduced from it is the relation ship of Theran and early Parian, but in many cases the evidence is doubtful as the graves were almost always used for several burials as in Crete.

⁵⁹ A sherd on the Acropolis: Graef 312. Possibly the amphora fragment no. 19 above. A few other exceptions are given by Payne in his review of Kunze's Kretische Bronzereliefs, JHS 1933, 123.

⁶⁰ JdI 1937. I should also like to pay a tribute to the painstaking publication of the pottery from Rheneia and Delos by Dugas and Rhomaios in Délos XV, XVII. Even if one may quarrel with the classification, their systematic presentation of the material, coupled with the excellent illustrations, makes it a really valuable contribution to the study of Cycladic pottery.

⁶¹ Délos XV.

⁶² Sieveking Hackl p. 44, no. 456.

⁶³ Ac 10, 11 appear to be Naxian. Karusos (JdI 1937, 190, n. 3) believes the whole group to be Naxian. See my remarks on the Ad group below.

⁶⁴ The transition can best be followed in Dugas' group Ab.

⁶⁵ Although I have not had an opportunity of comparing linear Parian from Siphnos with that in Paros, I do not think, judging from memory, that there is any distinction of fabric between the two.

⁶⁶ Cf. AA 1900, 20; ibid. 1923–24, 118.

⁶⁷ The lion and the stag are certainly very close to Pfuhl J 7 but I do not think anything else about the vase is strikingly Parian. Payne's arguments (JHS 1926, 205) have, however, convinced Dugas (Délos XVII, 7, n. 3). The style is certainly quite unlike Naxian, although certain features of the decoration can be paralleled there-no doubt because they are part of a Cycladic koine. For instance, the same type of maeander occurs on Délos XV Bc 6, which is ap parently under Cretan influence, and perhaps the fragment Buschor, Beil. LIII, 3; cf. Délos XV, Ac 10, f. For the sudden change of pattern in one frieze, cf. the Naxian amphora Délos XVII, Ba 8 (lozenge to maeander). The find of a griffon's head in Naxos (AM 1929, 156) also shows that the type may not have been confined to Paros.

⁶⁸ This dating does not account, if Johansen's dating is followed, for the striking resemblance of the birds on these vases with those on Protocorinthian globular aryballoi from Cumae, a connection which seems to be confirmed by one of the grave groups in Thera (Thera II, tomb 90), where two globular aryballoi were found with a developed orientalising Parian amphora. The same difficulty is encountered in the dating of Protoattic. See BSA XXXV, 203, where J. M. Cook proposes 700 as a more likely date for the transition from globular to ovoid aryballoi. This date would seem to be in line with the chronology advanced by Payne in PV.

⁶⁹ Cf. the Protocorinthian plastic vases Payne PV Pl. 25, dated to the second quarter. Karusos assigns the griffon jug to the last quarter of the eighth century on the strength of its being contemporary with the Cumae jug. See below.

⁷⁰ Discussed by Payne JHS 1926, 205 ff.

⁷¹ Hampe (p. 56) places it in the first quarter. Against this, Cook, op. cit., 207.

⁷² My list of Parian in the Delos publication agrees with Karusos' except that he includes Délos XVII, C and excludes Délos XV, Ac. I have omitted the group of skyphoi Ae, most of which are of course Parian.

⁷³ I had already reached the same conclusions as Karusos (with the exceptions noted below) about the Naxian group before reading his article, thanks to Buschor's illuminating publication of the sherds in Naxos. I note this for what it may be worth as a confirmation of Karusos' grouping.

⁷⁴ The connection, first noted by Dugas in Céramique des Cyclades, was forcefully refuted by Payne, (JHS 1926, 203 ff.CrossRefGoogle Scholar), who pointed out the very close similari ties with the ‘linear island’ (Parian) style. Payne's view was adopted by Buschor and Karusos and also by myself at the time when I reviewed Délos XVII (JHS 1935, 248). Dugas remained unconvinced (Délos XVII, 21), and I now believe him to be right.

⁷⁵ But cf. the lion on the heraldic amphora from Thera.

⁷⁶ The kotyle Délos X, 44 belongs to the last quarter of the century together with most of the polychrome plates. The crater ibid. Pl. 9, 39 is hardly earlier than the second quarter of the seventh century. The lozenge ornament occurs on the protome amphorae. The handles are decorated like those of Melian vases. For the vertical maeander (not a Geometric type) cf. Bc 21 and the centre stripe on the Siphnos goddess.

⁷⁷ Closely related to Melian vases decorated with animals, mostly horses, in outline style. The Athens amphora with riders (MuZ, 105) belongs to this period. Be 19 is contemporary with the ‘storm spirit’ amphora mentioned by Payne JVC, 78. I withdraw the suggestion I made in JHS 1935, 248 that these vases are actually Melian.

⁷⁸ Note the now familiar Naxian floral. Dugas has already pointed out the close parallels between this vase and the protome skyphos illustrated on the same plate.

⁷⁹ The left-hand winged horse on the skirt of the Siphnos goddess shows an embryonic version of the down-swung jaw. One of the lions in the bottom panel has a flame-shaped tongue exactly like those on the C group amphora and skyphos (Délos XVII, pl. 15 a, 20. 9a), but unfortunately this panel is only just visible to the naked eye, let alone in a reproduction.

⁸⁰ Other arguments, including a strong one—that of fabric—are advanced by Dugas in Délos XVII, 21; cf. ibid., 8 where the amphora Bc 1 is noted as another link between the two groups.

⁸¹ Horses are rare in Parian (Thera II, fig. 406; griffon jug). Karusos contrasts the horses on the heraldic amphorae with those on the protome amphorae, but it is evident that the distinction is one of date rather than style.

⁸² Cf. the large cable on the back of Thera II, fig. 403 with Ba 3, 6. Pfuhl, J 7 is clearly the Parian retort to the heraldic amphorae.

⁸³ JdI 1937, 192, n. 3. Be it said that Karusos' version is a considerable modification of Kunze's still earlier dating.

⁸⁴ The running dog ko tyle found with the Naxian amphora in Thera can hardly belong to the last quarter of the eighth century as Karusos suggests. This would be unlikely, even if we keep Johansen's date for the transition period.

⁸⁵ Cycladic connection noted by Watzinger (Gnomon 1934, 573).

⁸⁶ Payne seems to have considered this view but abandoned it on grounds of technique (JHS 1926, 205, n. 10). The feasting lions are a favourite western motive, although also popular on Boeotian fibulae. Grazing horses are rare in Cycladic (Subgeometric: Délos XV, Bb 8, Bc 4) before late Melian. They are, of course, common in Attic from Geometric times.

⁸⁷ This does not prove that these vases were made in Siphnos, since there are several examples in Paros from the Delion, including skyphoi and some unusual shapes, which unfortunately still remain unpublished. J. M. Cook informs me that vases of this fabric are also reported to have been found in Kimolos.

⁸⁸ The fabric is often indistinguishable.

⁸⁹ The relation with Ac is not one of fabric alone. Ac 6, 7, 9 have a strong resemblance to Ad. Our amphora no. 10 perhaps belongs to a transitional stage. The horse's mane is very like that on Ac 9. The fragment with a bird is closer to Pfuhl, J 14.

⁹⁰ I was at first inclined to think that the ‘besondere Gruppe’ was the Geometric stage of Ad and therefore Parian and that the home of the Ab group and ‘Euboic’ amphorae in Thera was still to seek. But it seems impossible to draw a hard and fast line between the ‘besondere Gruppe’ and the early Ab amphorae.

⁹¹ The early appearance of rays in an otherwise undistinguished fabric is something of a puzzle.

⁹² The ‘bowl’ CVA Providence, Pl. 4 is early Melian, not Naxian.

⁹³ I am greatly indebted to Professor Bernard Ashmole, Mr. D. B. Harden and Mr. Denys Haynes for help in the preparation of this part of the report, and to Professor Ashmole for permission abo to publish the objects from the, British Museum illustrated in Figures 18, 19 and 24.

⁹⁴ Mr. C. M. Kraay has identified this coin from a description and rubbings, and has kindly confirmed the identification of the other coins.