^{page 30 note 1} The photograph was taken by chance by Pilot Officer Jonas, R.A.F., and published by MrCrawford, O. G. S., F.S.A., in Antiquity, III, 452Google Scholar. At the time the site was under corn.

^{page 33 note 1} Geologically, ‘The upper soft chalk with flints.’

^{page 33 note 2} Hawkes, C. F. C., ‘Excavations at Quarley Hill, 1938’, Proc. Hants. Arch. F.C., XIV, 1939, p. 152Google Scholar.

^{page 35 note 1} When this crushed chalk is so thin, it is not marked in the section, and is included only when it is thicker than 2 cm.

^{page 39 note 1} Soil such as E is often wrongly interpreted as plough-soil.

^{page 40 note 1} It was easy to walk on this flat bottom, and thus much chalk was deposited there.

^{page 41 note 1} The northern antenna appears to intersect DA, for its head-like thickening extends over the inner margin of the ditch; it has not yet been included in the excavation.

^{page 46 note 1} There is a difference between post-holes and post-sockets. See below, p. 78.

^{page 53 note 1} We have here in part genuine loom-weights, but some of the objects are so crude and formless that they could not have been employed for this purpose. These latter have obviously served to weight and secure roofs.

^{page 53 note 2} ‘Cob’ is the usual local name in Wiltshire for a mixture of ground chalk, clay or loam and chaff, which is still used there to-day for building purposes. The current name is elsewhere ‘chalk and daub.’

^{page 53 note 3} We shall be able to give more exact details about the ovens made of cob only when work has been carried out on the large collection of fragments of their cupolas. All sorts of organic material has been kneaded irito the cupolas, e.g., strings. Imprints of grains of corn and of plants, such as ferns, could also be recognised. Diameter and height of the domes have come to more than a metre. The domes of the ovens made of burnt clay are considerably smaller. Some of the specimens with which we have to do deal are those transportable ovens for holding wood-charcoal, with rectangular openings, like those from Hungary (Ber. Röm-Germ. Komm. 24/25, abb. 10). A specimen of this kind (completion doubtful) belonging to Iron Age A from England (Sutton, Suffolk), in the Ipswich Museum, was recently (Arch. J., XCVI, 1939, p. LII, 2Google Scholar) wrongly interpreted as a hut-urn. As the loom-weight found with it, the goat-skull and ‘bricks’ show, we have not to do with a grave but with a pit filled with refuse.

^{page 53 note 4} Grey and green clay, or that brown loamy clay that fills the fissures of the chalk. Signs of a pottery industry in the settlement were not encountered in the pits.

^{page 54 note 1} What occasionally looked at first sight like a hearth was either coarse material (which, when the pit was being filled, had rolled to the deepest spots of the filling and then appears like a hearth-packing), or those lumps or beds of clay (Lehm) from oven-fragments such as are described above and which can easily be confused with ‘hearth-plate.’ Hearth-plates are smooth, flat clay surfaces marked by fire.

^{page 60 note 1} 27th Ann. Rep. Bureau Am. Ethn., Smithsonian Institution, 1905–1906. Washington, 1911, p. 98Google Scholar, fig. 17.

^{page 61 note 1} Thompson, G. Caton and Gardner, E. W., The Desert Fayum (1934)Google Scholar, chapters IX–XIII.

^{page 61 note 2} Junker, H., Anz. der Akadd. Wiss. Wien. Phil-hist. Kl., 1929Google Scholar, taf. V; ibid., 1930, taf. III.

^{page 61 note 3} For the roasting of corn in ancient times, cf. the note by Crawford, O. G. S. in Antiquity, XII, 1938, p. 289Google Scholar. The agricultural conditions on the islands of Western Scotland (Skye) in 1773 are described by Johnson, (Journey to the Western Islands of Scotland, ed. Chapman, R. W., London, 1934, p. 72)Google Scholar, so clearly that it is worth while to reproduce the passage here for comparison with our statements. ‘In vain they hope for plenty, when a third part of the harvest must be reserved for the seed. When their grain is arrived at a state which they must consider as ripeness they do not cut but pull the barley: to the oats they apply the sickle. … Of that which is obtained with so much difficulty nothing surely ought to be wasted; yet their method of clearing their oats from the husk is by parching them in straw. Thus with the genuine improvidence of savages, they destroy that fodder for want of which their cattle may perish. From this practice they have two petty conveniences. They dry the grain so that is it easily reduced to meal, and they escape the theft of the thresher ! … The oats that are not parched must be dried in a kiln.’

For corn-drying ovens of Roman times cf. the comparison by Corder, Ph., Excavations at Elmswell, East Yorkshire, 1938. Hull, 1940, p. 12–14Google Scholar.

^{page 62 note 1} For the walls of baking ovens and baking domes cf. Germania, 18, 1934, p. 134Google Scholar.

^{page 64 note 1} Eighteen cubic metres represent a storage capacity of fifty-five bushels.

^{page 65 note 1} The same letters have been used to designate the layers as were chosen to explain the layers in the pits.

^{page 82 note 1} The shallow pit, P 124, which may be absolutely contemporary with phase I, was abandoned in Phase II (see above, p. 79).

^{page 85 note 1} If the remains of a dry wall of chalk-blocks between the posts of ring I should have disappeared long ago, or been destroyed by agriculture (supposing they had ever existed) there is another argument for believing that there never was a block-packing between these posts. This argument consists of the position and filling of P 22, 24, 122, 123, 126, 130, 131, which, after the house was abandoned, were cut into its outer wall. Apart from the fact that it is most unlikely that these pits had been dug just at the spot where a block-packing existed before the chalk was reached, there were no signs in the filling of the pits that, when they were again filled-in, such blocks were used in the work.

^{page 90 note 1} A good description with illustrations of a house the size of ours belonging to the Ohamas will be found in the 27th Annual Report of the Bureau of American Ethnology (Smithsonian Institute), Washington 1911, p. 97Google Scholar, plates 22, 23, 43. A ‘dirt lodge’ of the Indian tribes on the Missouri, with four middle posts, appears in Morgan, L. H., ‘Houses and House-life of American Aborigines,’ Contr. to North Am. Ethnol. IV, 1881Google Scholar. Further literature in Safert, E., ‘Haus und Hof bei den Eingeborenen Nordamerikas,’ Archiv. für Anthropologie, N.S. VII, 1909, p. 157Google Scholar.

^{page 90 note 2} Archiv. für Anthropologie, N.S. VII, 1909, p. 44Google Scholar, figs. 7, 10.

^{page 90 note 3} The still unpublished Iron Age C houses at Salmonsbury, (Antiquity, 1931, 489)Google Scholar have, according to the information kindly given by G. C. Dunning, middle posts, and are considerably smaller, and thus belong to another class. The houses of the Glastonbury and Meare lake-villagers are wall-less huts with dome-shaped or conical roofs and vestibules to neutralise the difference in level between inside and outside. Therefore they too belong to quite another group, special, perhaps, to marsh conditions. House DB 2 at Maiden Castle (Dorset), (Ant. J., XVI, 1936Google Scholar, pl. XLVII), has an inner ring of 5 posts but no vestibule and a diameter of 7.20 m.; it is contemporary with our house, but smaller. The so-called hut of site A from Frilford (Berks.) (Oxoniensia, IV, 1938Google Scholar, fig. 4) is apparently only the incompletely excavated part of a round house of similar dimensions to our own. Also, in site C, the double wall of a large round house like ours appears to have been found. As this settlement is contemporary with ours, it would be very desirable if the adjoining area could be uncovered, in spite of the disturbances through older or more recent pits.

^{page 90 note 4} Die Grundformen des Hauses; vol. 1, Haus und Hof im Altertum, Berlin, 1927Google Scholar.

^{page 90 note 5} Oelmann, loc. cit., p. 28, with further literature.

^{page 91 note 1} Loc. cit., p. 28.

^{page 91 note 2} Oelmann, l.c., p. 34.

^{page 92 note 1} Oelmann, l.c., p. 53.

^{page 92 note 2} l.c., p. 35.

^{page 94 note 1} We made some general remarks at the beginning, when describing the houses, (see above p. 78), about the condition of the old surfaces; these remarks hold good for the following discussions as well.

^{page 101 note 1} In ‘Excavations at Quarley Hill, 1938,’ Hants. F.C. and Arch. Soc. XIV, p. 189, Hawkes, ChristopherGoogle Scholar has linked up this period with the Iron B invasion.

^{page 102 note 1} The results have been published in a way deserving of our thanks by MrStevens, Frank, F.S.A., in W.A.M., XLVI, 579Google Scholar.

^{page 103 note 1} W.A.M., LII, 457–96Google Scholar.

^{page 103 note 2} Ibid., pl. II, e.g., pits 21, 25/26, 36/37, 46/47 and 61.

^{page 103 note 3} Ibid., pl. III, e.g., pits 74/75, 83/84 and 86/87.

^{page 103 note 4} Ibid., pl. 1.

^{page 103 note 5} W.A.M., LIII, 59–93Google Scholar.

^{page 103 note 6} Excavations at Cranborne Chase, vol. II, p. 55, 1888Google Scholar.

^{page 103 note 7} Archaeologia, 76 (1927), 1–40CrossRefGoogle Scholar.

^{page 103 note 8} Ibid., fig. P (p. 32) and fig. T (p. 37).

^{page 103 note 9} Proc. Hants. F.C. and Arch. Soc., XII, p. 126Google Scholar; XIII, p. 7. For the air-photograph, see O. G. S. Crawford and A. Keiller, op. cit., pl. XIII b, p. 107.

^{page 103 note 10} See especially Miss Liddell, op. cit., p. 128.

^{page 103 note 11} According to a kind communication from Mr Hawkes, the pottery from the ditch corresponding to our DA is late not earlier than the latest from Woodbury, though still showing that at least 100 years separate the two periods. Thus it is possible, as at Highfield, Salisbury, that there were two periods of fortification. This would be proved if it could be shown that our ditch DA belongs to an early phase of the settlement.

^{page 104 note 1} Miss Liddell, op. cit., XIII, pl. 8, 9.

^{page 104 note 2} Ibid., p. 19.

^{page 104 note 3} Stone, , W.A.M., XLVII, 406–11Google Scholar.

^{page 104 note 4} Crawford, O. G. S. and Keiller, A., Wessex from the Air, pl. XII, p. 102Google Scholar.

^{page 104 note 5} Ibid., pl. XIIIa, p. 106 (with dark patch and hollows ?)

^{page 104 note 6} Ibid., pl. XLI, p. 224.

^{page 104 note 7} Piggott, C. M., W.A.M., XLVIII, 513–22Google Scholar.

^{page 104 note 8} The annual consumption of wheat to-day in the United Kingdom is 4½ bushels per head of the population

^{page 104 note 9} Between 2–3 bushels of wheat or 2–4 bushels of barley are required to-day to sow one acre.

^{page 105 note 1} Proc. Hants. F.C. and Arch. Soc., XVII, 138 ff.Google Scholar

^{page 105 note 2} Ibid., fig. 5.

^{page 105 note 3} See Proc. Hants. F.C. and Arch. Soc., XIV, 189Google Scholar. To these we are able on the kind information of Mr Hawkes to add Bury Hill, Hants.

^{page 106 note 1} Cunnington, M. E., The Early Iron Age Inhabited Site at All Cannings Cross Farm, Wiltshire, Devizes, 1923Google Scholar.

^{page 106 note 2} The map of the settlements in Oxoniensia, IV, 1939Google Scholar, fig. 2 appears to show a number of such settlements situated on low ground, in contrast to the high-level settlements (farms ?) and hill-forts. Frilford itself seems to be such a village-settlement on low ground.

^{page 106 note 3} See especially pl. 5. It may be noted that the first opinion of the excavators was that they were ovens. Cf. also, Corder, P., Excavations at Elmswell, East Yorkshire, 1938. Hull, 1940Google Scholar,

^{page 106 note 4} These relations are clearly shown by the section on pl. 4, 5, in which a filled-up pit (71) has later been built over by an oven (c), having two phases and later still, after further accumulation of the habitation layer, another pit (72) has been sunk and filled up.

^{page 107 note 1} See in comparison against this, Curwen, E. C. in Sussex Arch. Coll., LXXX, p. 215Google Scholar.

^{page 107 note 2} Verification of air-photographs, by excavation, in particular of Major Allen's, should make it easy to find this type of settlement.

^{page 107 note 3} Curwen, E. C., The Archaeology of Sussex, p. 228, 232Google Scholar.

^{page 109 note 1} W.A.M., XLVI, 198–217Google Scholar.