^{page 2 note 1} Actually, the second period began a little before the end of the second town. In the report for 1929 in J.H.S. xlix. p. 223Google Scholar, the periods, instead of being called first, second and third, are called first, intermediate and second.

^{page 4 note 1} No section of this part of the area is worth giving till it has been cleared to virgin soil.

^{page 6 note 1} See Philippson, Reisen und Forschungen I. geological map.

^{page 6 note 2} Troja und Ilion, p. 47, Fig. 9. It is also found in the new Early Helladic settlement at Hagios Kosmas in Attica: I am indebted to Dr. Mylonas for this information.

^{page 6 note 3} Compare Troja und Ilion, i. p. 36Google Scholar, where it is associated with Troy II (i.e. our Towns IV, V) but not with Troy I.

^{page 7 note 1} Above sea-level. See sections.

^{page 7 note 2} Troja und Ilion, pp. 40, 41.

^{page 7 note 3} The threshold in the side of the ante-room of A. I. 1, a later modification, is exceptional. Where the plans shew gaps in the long side walls of a room, it is due to the destruction of the wall itself and not to the presence of a door.

^{page 8 note 1} See note, p. 7.

^{page 8 note 2} Fimmen, , Kretisch-Mykenische Kultur, p. 45Google Scholar.

^{page 9 note 1} The levels given are those of the rims.

^{page 9 note 2} The shallow bothroi, however, appeared to have been cut short in ancient times by later floors.

^{page 10 note 1} Bulle, , Orchomenos, i. pp. 25 ffGoogle Scholar.

^{page 10 note 2} Blegen, , Zigouriés, pp. 26, 28, 76, 77, 215Google Scholar; Korákou, p. 75; Goldman, , Eutresis, pp. 21, 22Google Scholar.

^{page 10 note 3} Bulle, , Orchomenos, i. pp. 30 ff.Google Scholar; Wace, and Thompson, , Prehistoric Thessaly, p. 195Google Scholar.

^{page 10 note 4} Wace and Thompson, loc. cit.; Heurtley, , B.S.A. xxix. pp. 173–4Google Scholar; Mylonas, , Νεολιθική Ἐποχή, pp. 161–4Google Scholar.

^{page 12 note 1} Troja und Ilion, pp. 90, 91, Fig. 29.

^{page 13 note 1} The walls of houses 3 and 4 are, unfortunately, much broken.

^{page 13 note 2} See Plan, Pl. I. Another ‘flue’ was found in the north-west room of B. V. 1. It is uncertain whether Z. V. 3 and 4 were separate houses: they are numbered separately for convenience.

^{page 13 note 3} One of the hearths is marked with its level 5 m.: the other two adjoin it. The door socket can be distinguished on the plan.

^{page 13 note 4} Mr. Forsdyke suggests that the pithos was for hot water, which could be produced by dropping in stones heated from the neighbouring hearth.

^{page 16 note 1} Mylonas, , Olynthus, i. p. 16Google Scholar.

^{page 18 note 1} Vases in the museum of Mytilene: unpublished.

^{page 18 note 2} Of the tombs dug by Papavasileiou at Manika, groups I, II, IV, V were Cycladic, but III was definitely Troadic: this does not imply an invasion from the actual city of Troy. Papavasileiou, Περι τῶν ὲν Εὺβοὶα ὰρχαὶων τὰφὼν.

^{page 18 note 3} Sveti Kyrillovo may shew Troadic influence, but is probably Thracian.

^{page 18 note 4} See Heurtley, in B.S.A. xxix. p. 180Google Scholar.

^{page 20 note 1} Frankfort, , Studies in Early Pottery, i. pp. 9–11Google Scholar; ii. pp. 63–71 with full references to earlier discussions.

^{page 20 note 2} Similar cooking pots are known from Boz Ëyuk and from the first city of Troy where Schliemann found two examples just above the virgin soil. Both contained human ashes and in the larger were also the bones of a human embryo (Ilios, No. 59, p. 227).

^{page 20 note 3} We know only of Inv. 2468, Berlin Museum für Völkerkunde: thirteen fragments with very coarse fluting.

^{page 20 note 4} Forsdyke, , B.M. Cat. IGoogle Scholar. Part 1, Pl. I.; Ormerod, , B.S.A. xviii. pp. 83, 89 ffGoogle Scholar.

^{page 20 note 5} Demangel, , Protesilaos, p. 20Google Scholar, Fig. 20.

^{page 20 note 6} Occasionally an incised ornament takes its place: see below.

^{page 20 note 7} The first of these corresponds to Schmidt's type a, and the designs of hatched triangles or slanting lines were confined to the flat rim. Cf. S.S. 124–135, 150–4. The second type had a notched rim resembling somewhat S.S. 111 (unique at Troy): the ornament not only covered the rim but was applied both inside and outside in the form of vertical panels of chevrons, or chains of lozenges with a dot in the centre of each.

^{page 22 note 1} B.S.A. xix. p. 56Google Scholar, Fig. 4.

^{page 22 note 2} The shape which we call the horned tubular handle is illustrated on Fig. 6, No. 8. Two only were found in the lowest stratum: they were the most common type in the second period and not rare in the third.

^{page 22 note 3} In all excavated areas except Г where the slope makes itself felt, this change shews above 3·70 m.

^{page 22 note 4} No. 31 on Pl. VII and Fig. 7 is a good example.

^{page 24 note 1} In Fig. 5 a miniature vase, 344, has one handle, but miniature vases appear to retain shapes abandoned by larger vases.

^{page 24 note 2} For the human feet, cf. B.M. A. 55 from Yortan. For the general shape, cf. Berlin Inv. 2314 (119).

^{page 24 note 3} Comparable with the pink-washed wares of Troy III.

^{page 24 note 4} Cf. Berlin Inv. 463.

^{page 24 note 5} Cf. S.S. 111, the unique rim already quoted; S.S. 2465 and 2466 from the second city seem to belong to saucers of this type.

^{page 24 note 6} In all excavated areas except Г it begins at about 4·85 m.

^{page 26 note 1} No. 346 in Fig. 9 and No. 182 in Pl. VII and Fig. 6, 3 shew original patterns; No. 145 is very coarse.

^{page 26 note 2} In 1931, however, we found them in the second period.

^{page 26 note 3} These begin a little earlier, at the end of period II. See Fig. 7, No. 226 and Fig. 9, No. 150.

^{page 26 note 4} S.S. 5863.

^{page 26 note 5} Woolley, , Ur Excavations, i. Pl. XLIX, T.O. 516Google Scholar.

^{page 26 note 6} One of these is a surface find now in the collection of the British School. The other comes from Saratch.

^{page 27 note 1} Another Trojan form which does not occur at Thermi is the hour-glass beaker.

^{page 28 note 1} Fig. 10, a is a polished red ware, very fine; b and c are grey and any polish or coating they may have had has vanished. It is not certain if they were wheel or hand made. For the shapes, cf. S.S. 3491 and 3177.

^{page 28 note 2} Next year's report should include, besides the pottery from the outlying districts, the pithoi and the spindle whorls.

^{page 29 note 1} Measurements of heights include the handle; measurements of diameters do not.

^{page 32 note 1} B.S.A. xix. p. 59Google Scholar (in the British Museum). See also p. 54, Fig. 3, from Thyatira. Liverpool Annals, ii. pp. 145Google Scholar ff. Pl. XXVI.

^{page 32 note 2} Schliemann, , Bericht, 1891, Pl. I. Fig. 2Google Scholar; cf. also Schliemann, Ilios, Fig. 193.

^{page 32 note 3} Another fragment, resembling 29. 3, was found between 5 and 4·50 m. beneath the street which divides Β and Ε.

^{page 33 note 1} See Evans, , Palace of Minos, i. p. 48Google Scholar.

^{page 35 note 1} Evans, , Palace of Minos, i. p. 48Google Scholar, Fig. 13, No. 16.

^{page 35 note 2} The square-headed appear between 3 and 3·5 m., i.e. at a very low level, but below 3 m. only round-headed pins have been found.

^{page 36 note 1} See Blegen, , Zygouriés, p. 183Google Scholar, Pl. XX. 18; Tsoundas, , Ἐϕ. Ἀρχ. 1899Google Scholar, Pl. 10, Nos. 9, 12 (from Chalandriane).

^{page 36 note 2} The metal celts from Troy are of the longer type: cf. S.S. pp. 227–241, 265, 6. Near to ours are the celts from Giöl, Bounarbashi, B.S.A. xviii. p. 93Google Scholar, Fig. 9.

^{page 36 note 3} Such as S.S. 6189–6193, 6202, 6204.

^{page 37 note 1} Δ.Σ., p. 349, Fig. 288.

^{page 37 note 2} Childe, , Most Ancient East, p. 151Google Scholar, Fig. 63. I am indebted to Mr. Hutchinson for the parallel. There is also a resemblance to the Seals, Hogarth, Hittite Seals, Nos. 127–131, in design. As our seal came from the dump it may belong to any period contemporary with or later than the settlement.

^{page 37 note 3} See p. 39 for analyses.

^{page 37 note 4} In the above list, the description does not distinguish between round-headed pins in the strict sense of the word and round-headed pins with pointed tops: the drawing shews which are which. Both types occur at Troy (S.S. pp. 252–3).

^{page 38 note 1} Can the discs have had the same function as those of S.S. 5822 ?

^{page 40 note 1} Contrast the Olynthus celts, where type B was the earliest. Mylonas, , Olynthus, i. pp. 71 ffGoogle Scholar.

^{page 40 note 2} Tsoundas, , Δ.Σ. Pl. 41, 9Google Scholar; Wace and Thompson, op. cit. p. 164, Fig. 111 h (from the VIIIth city at Zerelia which contained elements of several periods); B.S.A. xiv. p. 219Google Scholar, Fig. 19; B.S.A. xxix. pp. 147–9Google Scholar.

^{page 41 note 1} I am indebted to Mr. Hutchinson for this information.

^{page 41 note 2} The implement described on p. 44, besides being of a softer material, belongs to a different class.

^{page 43 note 1} Cf. Kossinna, , Ursprung und Verbreitung der Germanen, p. 234Google Scholar, Fig. 301. In Nos. 29/1 and 30/45 there are slight traces of polish in the hole.

^{page 43 note 2} Hole mainly unpolished.

^{page 46 note 1} Also illustrated as No. 29/8 on Pl. X.

^{page 48 note 1} The functions of the bone tools are not always, of course, certain, but the nomenclature here followed is more or less that used in Wace and Thompson, Prehistoric Thessaly.

^{page 48 note 2} Only the more or less complete examples were counted in the above tests: the statistics of the more fragmentary examples, however, agree.

^{page 48 note 3} S.S. 7928–52. Ath. Mitt. 1899, p. 20Google Scholar, Pl. IV, 6; Wace and Thompson, op. cit. p. 72, Δ.Ε. p. 355; Ἐϕ. Ἀρχ. 1899, Pl. X, 2–5Google Scholar.

^{page 48 note 4} Ἐϕ. Ἀρχ. 1899, p. 104Google Scholar.

^{page 50 note 1} Wace and Thompson, op. cit. p. 72; Δ.Σ. Pl. 47, Nos. 6–11.

^{page 50 note 2} B.S.A. xxix. p. 148Google Scholar, Fig. 28, 10.

^{page 50 note 3} Blegen, , Zigouriés, pp. 196–8Google Scholar.

^{page 50 note 4} From Η, 5–4·5 m., Ζ, 4–3·5 m., Ζ, 4·5–4 m., p. 151.

^{page 50 note 5} S.S. 7650.

^{page 50 note 6} Childe, , Dawn, p. 61Google Scholar.

^{page 50 note 7} B.S.A. xxvii. pp. 32Google Scholar, 3, Fig. 16, No. 13. See also Tallgren, , Pontide Pré-scythique. p. 118Google Scholar, p. 57, Fig. 40.

^{page 50 note 8} The former from Ε, 4·5–4 m., the latter from Θ, surface half-metre.

^{page 50 note 9} Wace and Thompson, op. cit. p. 43, Fig. 19.

^{page 51 note 1} The same race, at the same stage, perhaps, made the vase from Emed in Phrygia and the specimens from between Soma and Panderma in Mysia, J.R.A.I. 1903, p. 373Google Scholar, Fig. 4; Ath. Mitt. 1904, p. 287Google Scholar.

^{page 51 note 2} The grains, etc., will be identified more exactly and published in next year's report.

^{page 51 note 3} Frankfort, pp. 65–7.

^{page 52 note 1} Protesilaos IV is contemporary with the last stage of Troy II: see Demangel, Protesilaos, p. 62Google Scholar. For Ëyuk, Boz, see Ath. Mitt. 1899, pp. 1 ffGoogle Scholar. It is uncertain when this site began but obvious that it continued later than Thermi.