^{page 262 note 1} We are again indebted to Dr. James Phemister for examining the Maiden Castle material. Dr. Phemister emphasizes the difficulty of precise allocation.

^{page 263 note 1} With this, compare the evidence from the long-barrow, Skendleby, Phillips, Lincolnshire—C. W. in Archaeologia, lxxxv (1935), 53, 79.Google Scholar

^{page 263 note 2} The inadequate data at present available as to climate and water level in Bronze Age Britain will be discussed by Dr. Friedrich Zeuner in the final report.

^{page 264 note 1} Antiq. Journ. xv (1935), 273, and xvi (1936), 268Google Scholar . I have lowered my previous dating in the present report to conform more nearly with the shorter dating urged by Dr. P. Jacobsthal and others, but have throughout used dates rather as relative than as absolute indices. The absolute dating will be discussed in the final report.

^{page 265 note 1} The height is reconstructed by adding to the existing material of the rampart the fallen material from the ditch. The maximum surviving height is 8 ft. (see pl. LX).

^{page 265 note 2} At the Goldberg, , Fund-Berichte aus Schwaben, xx (1912), 27Google Scholar ; and the Lenensburg, , ib. xxi (1913), 36Google Scholar.

^{page 265 note 3} Notably at Hollingbury, , Antiq. Journ. xiii (1933), 162Google Scholar ; and see Curwen, E. Cecil, The Archaeology of Sussex (1937), p. 239Google Scholar.

^{page 269 note 1} Allcroft, A. Hadrian, Earthwork of England (1908), p. 198 and fig. 70.Google Scholar

^{page 269 note 2} Antiq. Journ. xv (1935), 269.Google Scholar

^{page 269 note 3} Bersu, G., ‘Fünf Mittel-la-Tène-Häuser vom Goldberg’, in Schumacher-Festschrift (Mainz, 1930), p. 156 and pl. 15.Google Scholar

^{page 270 note 1} Archaeologia, lxxvi (1926–1927), 26, 34.Google Scholar

^{page 271 note 1} We are greatly indebted to Professor J. Percival for reporting on the cereal grains from Maiden Castle.

^{page 271 note 2} It is difficult to estimate the period taken by a io-in. post to rot, but the conditions under which these revetments were exposed must have conduced to fairly rapid decay. Professor Salisbury, E. J., F.R.S., writes: ‘I believe oak sleepers begin to perish and have to be replaced in 10-16 years. I have known oak posts last some 30 years. Charring would considerably prolong their durability, but how much I do not know, nor do I know if data are available.’ In the present instance, sufficient time had elapsed, not merely to allow for the decay of the posts, but to allow also for a considerable layer of turf to form over the subsided rampart.Google Scholar

^{page 273 note 1} Dr. G. M. Morant has kindly examined this and other human skeletons from Maiden Castle, and will report in detail upon them. The present skeleton was that of a muscular youth, probably between 22 and 30 years old, with a height of 5 ft. 6½ in. and a cephalic index of 72-6.

^{page 274 note 1} The section cut through the main rampart on site E in 1935 showed that the first extension-rampart had there been twice enlarged, but the first enlargement was limited to the summit and would not have shown in our cutting on site H.

^{page 274 note 2} Especially Antiq. Journ. xvi (1936), 269–70.Google Scholar

^{page 275 note 1} In its present state the main rampart is almost everywhere a wreck, due to the systematic plundering of it for the limestones of our ramparts 4 and 5. This plundering has in part been carried out within living memory.

^{page 276 note 1} The average diameter of these pebbles is 1¼-1½ in., and their average weight is just under two ounces. It may here be added that the use of the sling appears to have been greatly extended in Iron Age B, and may well help to explain the multiplication of lines of defence in earthworks of that phase.

^{page 277 note 1} Possibly as a result of the ‘tyranny’ of Cunobelin in eastern Britain at this time. It is here inferred that the Belgic colonization of Wessex, west of the Test, was the result of lateral movement (at a late date) in southern Britain rather than of fresh immigration from Normandy, as has been maintained. Th e matter will be discussed in the final report.

^{page 277 note 2} Mr. Derek Allen, of the British Museum, has very kindly examined the British coins from Maiden Castle, and his notes will be published in the final report.

^{page 277 note 3} Analogous evidence to that here detailed occurred in the inner part of the northern portal in 1935 but was not so clear and consistent, and its significance was not evident at the time of excavation.

^{page 278 note 1} Dr. Pryce has very kindly provided detailed notes which will be published fully in the final report.

^{page 278 note 2} Mr. Christopher Hawkes tells me that the unpublished Roman pottery (now in the British Museum) from this site indicates an occupation under Claudius and Nero. Taken alone this pottery, of which the precise location within the camp is unrecorded, might merely indicate a continued native occupation parallel with that at Maiden Castle; but the quantity of Roman military equipment also found, together with the semi-permanent design of the Roman fort, indicates the latter as the probable main source of the material.

^{page 280 note 1} Antiq. Journ. xv (1935), 271.Google Scholar

^{page 281 note 1} Of 216 Roman coins from site L, 13 were of the second and early third centuries, 23 were of the second half of the third century, and all the remainder of the fourth century, including one of Honorius.

^{page 281 note 2} I am indebted to Dr. G. Carmichael Low for the identification.