^{page 345 note 1} Rahtz, P. A. and ApSimon, A. M., PPS, 28 (1962), 289–328Google Scholar, and especially fig. 5 (p. 295).

^{page 345 note 2} Definite in S.1 and S.6; questionable in S.2 and S.3; probably removed by later ditch in S.4.

^{page 345 note 3} In S.7, F.2, the daub from Houses A, underlies the West Bank buildup, i.e. phase 3.

^{page 345 note 4} Either this bank skirts House B, in which case House B was built before the bank; or House B cuts into the bank, in which case the bank was built before House B. In either case, the idea was to build a new house with bank and ditch around it.

^{page 345 note 5} We are grateful to Peter Fowler for discussion of this point.

^{page 345 note 6} These, together with some of the shallow gullies in this area, form a rather unclear pattern of two parallel rows.

^{page 347 note 1} Ibid., 325: of the charcoal samples used, nos. 2 and 7 are of phase 2; no. 3 probably of phase 2 (in the area of House A, though not assigned to A1 or A2 in our plan, fig. 1); nos. 4 and 5, from hollow F.18, are probably of phase 1, and certainly the section of F.18 (ibid., fig. 15) shows this hollow is sealed by the bank of phase 3.

^{page 347 note 2} Hodson, Pace, PPS, 30 (1964), 104Google Scholar; and see the excavator's remarks, p. 303, penultimate paragraph. We are grateful to Philip Rahtz for reading this note in typescript, and commenting on it.

^{page 347 note 3} Only postholes of the excavator's qualities 1 and 2 have been planned (PPS, 28 (1962), fig. 5, p. 295Google Scholar) and p. 301, para. 3, for explanation). The numbers of the postholes in each phase are: House A.1: 5, 6, 10, 11, 13c, 15, 20, 24, 25, 26, 27, 30, 31, 35, 36, 39, 43, 47, 50, 52, 55, 58 and F.23; House A.2: 3, 9, 13a, 17, 19, 21, 24 again, 28, 29, 32, 34, 40, 46, 48, 50 (different socket), 53, 54, 56, 57, F.5 and F.6.

^{page 347 note 4} PPS, 28 (1962), 304Google Scholar, final paragraph.

^{page 347 note 5} For these, see ibid., 303, and also (for section drawing, S.7) fig. 6 (p. 297) and photos on pl. 44, upper and lower. The hachures in the centre and at the east of the house are re-interpreted as marking the scarped edges of an oval artificial hollow in bedrock, whose date must remain uncertain: S.7 (fig. 6) shows the small sunken ‘plat-form’ between these two scarps (one scarp is marked ‘cutaway for House A’; the other is directly under the gap between ‘HOUSE’ and ‘A’).

^{page 347 note 6} Pace the excavator, who suggested tentatively that F.21 was a storm-water gully. Since it is unlikely that eaves would project 6 or 7 feet, F.1 and F.21 are too far out to be drip gullies for a single-ring house. A possible alternative is that they are part of a fence enclosing an unroofed area round the exterior wall of the house: this seems impossible to disprove strictly, but the regularity with which the grooves encircle the postholes seems a point against it.

^{page 347 note 7} See section S.7 (ibid., fig. 6, p. 297) for the position of F.2, described (p. 303) as ‘dark and yellow clayey soil with signs of reddening and hardening at the base’; and compare the daub scattered near a bedding trench at Craigs Quarry (Piggott, S., PSAS, 91 (1957–1958), fig. 4, p. 68Google Scholar.)

^{page 347 note 8} PPS, 28, 1962, p. 305Google Scholar.

^{page 347 note 9} In the following paragraphs of discussion of the structure, our views have profited greatly from discussions with Miss Sheila Gibson, with Mr. Peter Winchester, and with Brian Scott of the Queen's University Department of Archaeology.

^{page 347 note 10} Clark, Grahame, Prehistoric England (1962), 67Google Scholar.

^{page 347 note 11} The structure is so much larger than the roundhouse at Veii (Close-Brooks, and Gibson, , PPS, 32 (1966), 349–52Google Scholar) that the woven withy ring-beam found there might seem too weak. For while withies might well be strong enough when in tension to withstand the outward thrust of the roof, they would probably fail to resist the compression caused by winds; whereas rigid timbers would maintain the stability of the structure, albeit, one may imagine, with an eerie creaking and groaning on stormy nights.

^{page 347 note 12} Radial collar-beams at the level of the top of the stand seem to require an impossibly cumbersome central crossing.

^{page 348 note 1} It may be noted that, from the plan (fig. 2) and photographs (pl. XLIII), the Shearplace houses lie slightly off the crest of their hill, on ground sloping down to the east, and seem thus to have some shelter from west and southwest winds.

^{page 348 note 2} The area, and therefore the weight, of the roof is much greater lower down than higher up: the skirt here would in fact weigh about 3/2 times the cone.

^{page 348 note 3} Oelmann, F., Haus und Hof in Altertum (Berlin, 1927), 29Google Scholar, a Northwest Brasilian (Maloka Indian) house; and Erixon, S., in Folkliv, 1937, 124 ff.Google Scholar, figs. 1 and 2 (roundhouses in the Roman Campagna); and Erixon, S. in Arkeol. Stud. tillagnade H.K.H. Kronprinz Gustaf Adolf (1932), 249 ffGoogle Scholar. These parallels also support our reconstruction of the roof with few, if any, cross-ties or collars in its upper part.

^{page 349 note 1} To avoid confusing the drawings, these horizontal purlins have been omitted from the roof skeleton on figs. 2 and 3; they are visible under the thatch in the background of the house reconstructions, and are shown in section on fig. 1.

^{page 349 note 2} Burstow, G. P. and Holleyman, G. A., PPS, 23, 1957, 167–212Google Scholar; Stone, J. F. S., PPS, 1941, 114–33Google Scholar, especially pl. IV; Wilts. Arch. Mag., 47 (1937), 640–59Google Scholar.

^{page 350 note 1} New Barn Down: Curwen, E. C., Sussex Arch. Collns., 75 (1934), 137–70Google Scholar. Plumpton Plain: Holleyman, G. A., Curwen, E. C. and Hawkes, C. F. C., PPS, 1 (1935), 16–59Google Scholar. Park Brow: Wolseley, G. R., Smith, R. A. and Hawley, W., Archaeologia, 76 (1927), 1–6, 14–16CrossRefGoogle Scholar; especially fig. C, p. 4. Kingley Vale: Curwen, E. C., Sussex Arch. Collns., 75 (1934), 209–15Google Scholar: apparently overlain and partly destroyed by ‘Celtic Field’ lynchets. Preshute Down: Piggott, C. M., PPS, 8 (1942), 48–61Google Scholar. Encombe, Dorset, hut 6: Cunliffe, B. and Phillipson, D. W., PPS, 34 (1968), 200Google Scholar.

^{page 350 note 2} Feachem, R. W., PSAS, 94 (1960–1961), 79–85Google Scholar: an inner ring of stout posts, with projecting porch on the southwest, and stakeholes presumably marking inner and outer faces of a wall 0·80 metres thick, with its outer face (running round significantly just inside the scarped platform) aligned on the projecting porch holes. A thick wall might eliminate the need for ‘slots’ set at the entrance between inner post ring and outer stake ring.

^{page 350 note 3} See section in PPS, 23 (1957), fig. 18 (p. 189)Google Scholar; and plan, pl. XVI.

^{page 350 note 4} The ‘banks’ marked with hachures on the plan are in fact very slight (see photo, ibid., pl. XXa; and statement, p. 168, para. 4) and it cannot be assumed that they would have protected any gully from erosion.

^{page 350 note 5} See Piggott, , Ancient Europe (1965)Google Scholar, fig. 86.

^{page 350 note 6} Pace Prof. Piggott (ibid., fig. 87, at p. 153), hut V might still be interpreted as a roundhouse, with the porch (pointing south-east) formed of posts in the holes marked −12, −14, and also (rather far away) the pair −10, −6; and the main internal ring in the holes (running clockwise from the porch) marked −12 (inner post of porch), −10, −15, −10, −9, −11, −9, −12 (?), −10 (porch post); this inner ring is then 21 or 22 feet in diameter, the whole house 40 feet in diameter, and one pair of holes is interposed at the entrance (−12 and −14) as at Shearplace Hill. These are only slightly elongated and suggest more accuracy in planning than at Shearplace Hill.

^{page 350 note 7} The internal ring and entrance of house I is again apparently of 2 phases: structure 1 (there is no evidence as to which is the earlier) made use of the elongated slots (−14 and −18) and of the shallower postholes ring (running clockwise from the porch: −9, −8, −7, −6, −6, −7, −10, (?), −4) and may also have made use of the outer “funnel” of 4 posts (−9, −9, −12, −15); while structure 2, of deeper postholes (running clockwise, these are: the unnumbered post to the right of ‘quern’, −14, −9, −10, −9, −6 (possibly ? if so, re-used), −10, −10 (re-used), −11) may also have made re-use of the slots −18 and −14, or else only of the outer posts (−9, and −9; and perhaps also −12 and −15).

^{page 350 note 8} Little Woodbury: Bersu, G., PPS, 6 (1940), 30–111Google Scholar. Pimperne: Harding, and Blake, , Antiquity, 37 (1963), 63–4Google Scholar. West Brandon: Jobey, G., Arch. Aeliana, 40 (1962), 1–34Google Scholar. Scotstarvit: Bersu, G., PSAS, 82 (1947–8), 241–63Google Scholar. West Plean: Steer, K. A., PSAS, 89 (1955–1956), 227–51Google Scholar.

^{page 351 note 1} Piggott, S., Ancient Europe (1965), 236Google Scholar.

^{page 351 note 2} PSAS, 89 (1955–1956), 249Google Scholar; PPS, 20 (1954), 96–7Google Scholar; Arch. Aeliana, 40 (1962), 30Google Scholar.

^{page 351 note 3} And the fact that the single-ring, central-posthole design is not (pace Feachem, , PSAS, 92 (1958–1959), 23Google Scholar) limited to the Bronze Age in date is shown by the Iron Age one at Maiden Castle, Dorset, hut DA (Wheeler, , Maiden Castle (1943)Google Scholar, pl. VIII, and fig. 17): this, then, is a different tradition, in use contemporary with the multiple-ring design for perhaps a millennium.

^{page 351 note 4} Shearplace Hill, House A, will be 35 and 38 feet in external diameter, instead of 27 feet; Itford Hill, hut B, will be 25 feet, not 20; hut D, 33 feet, not 22; hut M, 23 feet, not 14; hut N, 22 feet, not 12; Thorny Down house I (structure 2) will be 32 feet instead of 18. These compare with Little Woodbury (45 feet), Pimperne (45 to 50 feet) and West Plean house 2 (38 feet).

^{page 351 note 5} Notably that of supporting the roof at the entrance. Iron Age roundhouse plans lack the postholes or slots clumsily interposed in Bronze Age houses between the outer and of the porch and the main ring of posts.