^{page 41 note *} The Survey was made possible by a generous grant from the Inter-University Council for Higher Education in the Colonies to the Royal University of Malta for this purpose.

^{page 41 note 1} Mr Zammit also prepared the drawings for this paper with the exception of figs. 3 and 4, which are my own.

^{page 43 note 1} See diagram, fig. 2.

^{page 43 note 2} The term ‘apse’ is used because it has been consecrated by previous usage and for want of a more precise term. They are called ‘hemicycles’ by some of the old writers, but this is clumsy and not specially accurate. These features are not, of course, apses in the normal architectural sense of the term, and must be regarded as having developed from separate circular or oval chambers opening on opposite sides of a central court.

^{page 44 note 1} Despott, G., J.R.A.I., LIII, 1923, pp. 18–34Google Scholar and pls. I-II; Murray, M. A. and Thompson, G. Caton, Excavations in Malta, Pt. I, pp. 10–12Google Scholar and pl. 1. pt. II, pl. V.

^{page 45 note 1} Brea, L. Bernabò, ‘The Prehistoric Culture-Sequence in Sicily’, Sixth Ann. Rep. Univ. of London Institute of Archaeology, 1950Google Scholar, pl. 1, 3.

^{page 46 note 1} Published (very inadequately) in the Bulletin of the Valletta Museum, No. 1, 1929. Also short notes in the Annual Report of the Valletta Museum for 1923–4 and 1925–6.

^{page 48 note 1} Notably Bur Mghez (The Cave of the Goats), published by Tagliaferro, in Man, October, 1911Google Scholar.

^{page 48 note 2} A note on these appeared in the Annual Report of the Valletta Museum (M.A.R.) for 1947–1948.

^{page 48 note 3} The example illustrated in pl. II, 2, which is in the collections of the Valletta Museum, has previously appeared in Mayr, , Die Vorgeschichtliche Denkmäler in Malta, München, 1901Google Scholar, pl. XII, 2, and another (pl. II, 4) by Manneville, Mile. E. de in Commission Internationale pour la Préhistoire de la Méditerranée Occidentale, Conference de Barcelone, 1935Google Scholar, pl. XXXIX, 3.

^{page 49 note 1} This seems to be the remote ancestor of the ‘amphorae’ with twin tunnel-handles of Period I C. What is probably the intermediate form in Period I B is represented only by fragments of the lugs and shoulder. See fig. 8, Shape 30, and p. 55.

^{page 50 note 1} Cp. Brea, op. cit., pl. III, 3, showing a bowl from Predio Iozza, near Gela. There is a large fragment of a similar bowl from Kordin III, the decoration of which matches that of the Sicilian example very closely. (See pl. X, 2).

^{page 51 note 1} A short note on this site, with a plan of the cave, but no illustrations, was published by Orsi, , B.P. XLVIII, pp. 58–61Google Scholar.

^{page 53 note 1} Ant. J., vol. VIII, No. 4, p. 481Google Scholar and pl. LXXVI, figs. 1 and 2. The sherds shown in fig. 1 are decorated in the Zebbug style, those in fig. 2 in the Ggantija style.

^{page 57 note 1} Tagliaferro, , ‘The Prehistoric Pottery found in the Hypogeum at Hal Saflieni, Casal Paula, Malta’, Liverpool Annals, vol. III, June 1910Google Scholar.

^{page 57 note 2} For Gheizu, Ta see M.A.R. 1933–1934Google Scholar, Appendix A, p. VI; Cave, North, M.A.R. 1949–1950Google Scholar.

^{page 57 note 3} The earlier, or South Building of the Ggantija is unique in that the inner pair of apses is larger than the outer one, which seems to support the theory advanced here. None of the plans so far published (they are all based on the one made for La Marmora in 1834) adequately convey the ‘clover-leaf’ pattern of the original nucleus.

^{page 58 note 1} Excavation reports in Archaeohgia, LXVII, pp. 127 ff.Google Scholar, LXVIII, pp. 263 ff., and LXX, pp. 179 ff. Full publication in Prehistoric Malta, 1930.

^{page 58 note 2} The precision and accuracy of the carving on the stone altars, etc., strongly suggest the use of metal tools in fashioning them.

^{page 60 note 1} The volute-patterns are often employed in pairs to form oculus motifs, especially on the offering-bowls, which are most usually decorated with this motif. This fact seems to have escaped the notice of previous students, and was only brought to mine by the accident of a photograph being taken from the correct angle to show it. This happened after the completion of the present paper.

^{page 61 note 1} One is illustrated in Ugolini, Malta, Origine della Civiltà Mediterranea, fig. 43.

^{page 61 note 2} Mayr, V.D., fig. 4.

^{page 61 note 3} M.A.R., 1928–9.

^{page 61 note 4} Zammit, T., Peet, T. Eric and Bradley, R. N., ‘The Small Objects and Human Skulls found in the Hal Saflieni Prehistoric Hypogeum’, Malta, 1912Google Scholar, pl. VIII, no. 4.

^{page 61 note 5} P.M., fig. 17.

^{page 61 note 6} No illustration of this interesting fragment has yet been published.

^{page 62 note 1} Archaeologia, LXVIII, pl. XL and pl. XLI, nos. i and 2. No. 3, wrongly included in the caption with the inlaid wares, is actually a sherd of Period I B (comet-style) scratched ware with red incrustation.

^{page 62 note 2} J.R.A.I., LIV, p. 84 and pl. vn, 9. Best illustrations in Ugolini, op. cit., figs. 27, 28 and 29.

^{page 62 note 3} e.g. that illustrated in Ugolini, op. cit., fig. 23 and pl. III. See also J.R.A.I., LIV, pl. V.

^{page 62 note 4} Ugolini, op. cit., pl. XI, and J.R.A.I., Lrv, p. 90 and pl. LX.

^{page 63 note 1} P.M., p. 14 and fig. 1.

^{page 64 note 1} Zammit, Peet and Bradley, op. cit., pl. X, 1, 3, 10, 12. No. 10 seems to be the head of a snake.

^{page 64 note 2} Zammit, Peet and Bradley, op. cit., pl. VI, 4 and 5.

^{page 64 note 3} Name suggested by Zammit, , P.M., p. 82Google Scholar.

^{page 64 note 4} P.M., p. 89 and fig. 16. They are also known from other Period I sites.

^{page 64 note 5} Zammit, Peet and Bradley, op. cit., p. 13 and pl. XII, 1–6 and 8. Also P.M., p. 83 and pl. XXIII, 3.

^{page 64 note 6} P.M., pl. XXV, 1; Ugolini, op. cit., fig. 36. The two portions were found together and almost certainly belong to the same facade. Ugolini's illustration shows the lower left-hand corner, Zammit's the upper righthand part.

^{page 64 note 7} P.M., pl. XXV, 2.

^{page 65 note 1} Zammit, Peet and Bradley, op. cit., pl. XIII, 5.

^{page 65 note 2} Published at same time as Tarxien Period I material.

^{page 68 note 1} P.M., pl. XVII; M. A. Murray, Corpus of the Bronze Age Pottery of Malta, pls. V and VI.

^{page 68 note 2} Corpus, pl. III—three together to right of jug.

^{page 68 note 3} P.M., p. 55.

^{page 68 note 4} P.M., pl. XV; Corpus, pls. VII–IX.

^{page 69 note 1} Cp. Schliemann, Tiryns, figs. 159, 160.

^{page 69 note 2} Ant. J., VIII, p. 481Google Scholar and pl. LXXVI, 3.

^{page 69 note 3} M. A. Murray, Excavations in Malta, pts. I–III.

^{page 69 note 4} This wall was partially cleared of later débris in 1881, but no publication was ever made. Mayr, in 1901, came to the conclusion that it was of a defensive nature and enclosed a settlement. See V.D., pp. 687–93 for full description and bibliography of earlier references. Orsi, , B.P., XXIIIGoogle Scholar, pl. VIII shows a similarly planned wall built with an almost identical technique, at Monte Finocchito, near Noto. The date is uncertain; Orsi would make it 700–670 B.C., but there were much earlier remains at Finocchito. In any case, it probably represents a traditional style of fortification.

^{page 71 note 1} Prof. Stuart Piggott has suggested to me that these painted patterns might be crude imitations of those on Mycenaean wares. The use of red paint on a light buff or cream background is certainly suggestive of this, as are some of the patterns themselves.

^{page 72 note 1} The opinion of Dr P. Dikaios, who examined the fragment in the Valletta Museum, shared by all others to whom I have shown it.

^{page 72 note 2} E.M., pt. II, pl. XVII, 11, and Peet, , ‘The Study of the Prehistoric period in Malta’, P.B.S.R., V, pp. 141–63Google Scholar and pl. XV, 52, 61, and 69.

^{page 72 note 3} E.M., pt. III, pl. VIII, 3.

^{page 73 note 1} P.B.S.R., V, pp. 149–63 and pls. XIII–XV.

^{page 73 note 2} This can be seen in the change of colouring in the ware from red to black, the introduction of the maeander and excised zig-zag motifs in the decoration, and in the shapes of some vases, e.g. fig. 12, shapes 109–12.

^{page 73 note 3} A silo-grave, 14 ft. deep, with pottery of Bahria type and a Punic lamp, and nearby a second tomb of similar type, but altered and re-used in Punic times, were found by Mr J. B. Ward Perkins at Mtarfa in 1939. See Perkins, Ward, ‘Problems of Maltese Prehistory’, Antiquity XVI (1942), p. 34Google Scholar, also M.A.R., 1938–9, Appendix B, p. XII.

^{page 75 note 1} P.B.S.R., V, pl. XV, 59, 60. I have since traced one of these in a collection of sherds from Malta in the Museum of Archaeology and Ethnology at Cambridge. It is definitely a handle of the type described and is made of red-slipped buff ware of the Borg in-Nadur kind. Cp. B.P., XXIX, pl. XI. 1.

^{page 75 note 2} P.B.S.R., V, pl. XIV, 34–8.

^{page 75 note 3} P.B.S.R., V, pl. XIV, 41–3.

^{page 76 note 1} See note I, p. 90 infra.

^{page 76 note 2} Brea, op. cit., p. 19 mentions it. I do not know who first made the observation.

^{page 77 note 1} Brea, op. cit., p. 22.

^{page 77 note 2} e.g. B.P. XVI, pl. VII, 7, 9, 14, and pl. VIII, 17, 18.

^{page 77 note 3} Brea, op. cit., p. 19.

^{page 77 note 4} At Mgarr and Kordin III. Much pottery of this type was also found on Sites I and II at Kordin, but here the remains were not so well preserved and the plans are not clear.

^{page 78 note 1} Brea, op. cit., p. 22 and pl. III, 1 and 2. At some sites near Etna this pottery seems to supersede that of Stentinello, and so might be contemporary with early San Cono in the rest of Sicily and with Zebbug in Malta.

^{page 78 note 2} See Brea, op. cit., p. 23 for description and bibliography of sites.

^{page 78 note 3} Brea, op. cit., pl. III, 3.

^{page 79 note 1} Observation made by Prof. R. Battaglia in a letter to Dr J. G. Baldacchino with reference to this find.

^{page 79 note 2} For plank-idols see The Swedish Cyprus Expedition, vol. I (plates), pls. XVIII, XIX, XX, XXV, XXVI, and Myres, , B.S.A., XLI, pp. 78 ffGoogle Scholar. and figs. 4 and 5. The resemblance to the schist plaques of Spain and Portugal is obvious and has been remarked on by Martín, B. Sàez in Actas y Memorias de la Soc. Esp. de A. E. P., XIX, 1944, pp. 134 ffGoogle Scholar.

^{page 80 note 1} Stevenson, R. B. K., ‘The Neolithic Cultures of South-East Italy’, P.P.S., 1947, pp. 85–100Google Scholar.

^{page 81 note 1} Childe, , The Dawn of European Civilisation, 4th ed., p. 250 and p. 294Google Scholar.

^{page 81 note 2} The multi-pierced lug is identical with Déchelette, , Manuel, vol. 1Google Scholar, fig. 207, I, and with examples from Arene Candide. The doubly pierced lug, on the other hand, is very common at Ruju, Anghelu. See Not. Scav., 1904, p. 314Google Scholar, fig. 12, nos. 1, 2, 4, 6, 13.

^{page 82 note 1} Ugolini, op. cit., fig. 60.

^{page 83 note 1} The bead is illustrated in Zammit, Peet and Bradley, op. cit., pl. X, 14. Savory, in Ant. J., XXXIII, p. 227Google Scholar, mentions a vase in the form of a dove found by the Leisners in the tholos of La Zarcita, which is possibly of Cycladic inspiration, and Dr G. E. Daniel has drawn my attention to the decorated bone tube from the ‘dolmen’ of Cabut (Gironde) (Déchelette, , Manuel, vol. 11Google Scholar, fig. 145) which greatly resembles those found in Early Cycladic tomb-groups.

^{page 83 note 2} See Jacobsthal, P. and Neuffer, E., ‘Gallia Graeca’, in Préhistoire, II, Fasc. 1, p. 37Google Scholar for bibliography.

^{page 83 note 3} Olalla, J. M. Santa, ‘Jarro picudo de Melos hallado en Menorca (Baleares)’ in Cuadernos de Historia Primitiva, Año III No. 1, pp. 37–42Google Scholar.

^{page 83 note 4} A rather poor example is illustrated by Mayr, , Die Insel Malta im Altertum, München, 1909Google Scholar.

^{page 83 note 5} This fact is stated in Sir Them. Zammit's notebook of the Tarxien excavations for 1918.

^{page 84 note 1} I hope to study these objects in more detail elsewhere.

^{page 84 note 2} This was pointed out to me by Prof. Stuart Piggott.

^{page 84 note 3} See note 2, page 81, supra.

^{page 84 note 4} This refers to the sequence defined by Hawkes, J. in Arch. J., LXXXVIII, 67–159Google Scholar from evidence found by P. Helena at the Grotte de Bize and elsewhere, and used by Childe in the Dawn.

^{page 86 note 1} Bovio, Marconi, ‘La Cultura tipo Conca d'Oro’, Mon. Ant. 1943Google Scholar.

^{page 86 note 2} e.g. B.P. XVIII, pl. II, 1; cp. Murray, Corpus, pl. XXIII, 2a, pl. XXIV, 3a, and pl. XXVII, 17. Cp. also B.P. XXIV, pl. XX, 9 and pl. XXI, 2 and 6 with Corpus, pl. XXXI, 12.

^{page 86 note 3} Murray, , Corpus, p. 2Google Scholar and pl. XXXVIII.

^{page 86 note 4} However, whilst on a visit to the excavations at present in progress at Leontinoi in June of last year I was shown a newly discovered cremation burial in an urn with decorated Thapsos grey pottery, so that a complete revision of ideas in this respect may yet become necessary.

^{page 86 note 5} This influence could have come directly from the West coast of Greece and the islands which lie off it. Some of the pottery illustrated by Dörpfeld in Altithaka is faintly suggestive of such a connection.

^{page 87 note 1} Brea, , ‘Civiltà preistoriche delle isole eolie’, Archivo de Prehistoria Levantina, III, pp. 69–93Google Scholar.

^{page 87 note 2} Cp. B.P. XXIX, pl. X, 3 and 5; and pl. XI, 6; B.P., XXX, p. 299, fig. 95.

^{page 89 note 1} For Montechiaro, Palma see B.P. XLVIII, pp. 44–62Google Scholar. So far as I know, however, no illustration of the vase mentioned here has been published.

^{page 89 note 2} In Syracuse Museum. So far as I know unpublished.

^{page 89 note 3} See Atti d. R. Ace. di Palermo for S. Angelo Muxaro.

^{page 90 note 1} With material from Lipari excavations in the Syracuse Museum. Unpublished.

^{page 91 note 1} A decoration very like that of Bahria appears on some clay objects from the necropolis of Canale Ianchino in Calabria. See section on Period II c.

^{page 91 note 2} Brea, Isole eolie, pl. XI, 1 and XII, 5. The handles in these instances do not correspond, but a large number of the bowls have high strap handles of the Bahria type.

^{page 91 note 3} Cp. askos from Pantalica in Peet, Stone and Bronze Ages in Italy, fig. 275, after Orsi, in Mon.Ant. IXGoogle Scholar.

^{page 91 note 4} See also note 3, page 73, supra. An inquiry to the Museo dell' Etruria at Florence has since revealed that the tomb-group in question is synthetic. The Bahria bowl and the Punic urn which it covers are numbered in different series and are certainly not from the same grave. The material from Ward Perkins' Grave 2 at Mtarfa, which was mislaid, has now been traced, and it proves to be entirely of the Borg in-Nadur type. Period II c wares are absent. It should be remembered that pottery of the Period IIB type was also found at Bahria, and was, in fact, known to Peet and others as ‘Bahria red’ ware. We have, therefore, to face the fact that there is still no grave known of Period II c, and no real link between this phase and Period III (Punic). Despite this double disillusionment, however, there is no reason to alter the proposed dating of this phase, but we shall have to wait for future research to provide us with the much-desired link between it and the succeeding period.

^{page 92 note 1} P.M., pl. XXIV, 4.

^{page 92 note 2} M.A.R., 1910, p. 7Google Scholar.