^{page 133 note 1} Mem. Soc. Préh. Francaise, IV, 1 (1919), pp. 1–51Google Scholar.

^{page 133 note 2} Kendrick, T. D., The Axe Age, 1925Google Scholar.

^{page 133 note 3} Archaeologia, LXXXV (1936), 253 ff.Google Scholar

^{page 133 note 4} Keiller, and Piggott, , P.P.S., 1938, 122 ff.Google Scholar

^{page 133 note 5} Ausgemeitselte Tiiren in Megalithgräbern der Pyrenäenhalbinsel, Marburger Studien, 147–55Google Scholar.

^{page 133 note 6} Lysons figures the porthole to the north chamber (Proc. Soc. Ant., 2nd ser., II, 277Google Scholar, fig. 6) and Davis and Thurnam in Crania Britannica publish a picture of the north porthole looking inwards over the porthole into the chamber (fasc. 59, fig. 3), and one of the south chamber looking outwards through the porthole, (fasc. 59 fig. 2). The latter drawing is, at least in part, imaginary, since it shows no dry walling closing the south porthole (see p. 140 infra.)

^{page 133 note 7} Antiquity, I, 230 (1927)Google Scholar.

^{page 133 note 8} Op. cit., 40–1.

^{page 134 note 1} Lewis, A. L. in J.R.A.I., 1910, 342Google Scholar. The distinction must not be over-emphasised, as Kendrick (op. cit. 40–1) points out, since some of the Asian portholes (e.g. in Syria) are as much as 2 ft. in diameter: but these are exceptional examples.

^{page 134 note 2} Sometimes both the ghost holes and the larger holes are referred to as portholes. Thurnam appears to have been one of the first to apply the term porthole for holes in megaliths and he clearly used it for the larger holes ‘through which the tomb might be entered in the creeping posture.’ (Archaeologia, XLII, 216Google Scholar)

^{page 134 note 3} Op. cit., 40–2.

^{page 134 note 4} P.P.S., 1938, 128Google Scholar.

^{page 135 note 1} They have been said to be soul-holes or to have a sexual or solar symbolism. Baudouin advocates this solar symbolism (op. cit., 40–2) and claims that the two-piece portholes represent the rising and setting sun placed together.

^{page 135 note 2} Thurman says of portholes that they were ‘naturally hollowed’ (Archaeologia, XLII, 216)Google Scholar, but this is not so.

^{page 135 note 3} Coutil and Baudouin adopt this classification and define the two classes thus ‘1. seule pierre avec perforation circulaire; 2. deux pierres échrancrées et ŕeunies.’

^{page 135 note 4} According to Ekwall, (Concise Oxford Dictionary of English Place Names, 372Google Scholar) Rodmarton (Redmertone D.B., Rodmarton 1220 Fees) derives from ‘Tūun by a reedy lake.’ (O.E. hrēodmere).

^{page 135 note 5} Proc. Soc. Ant., 2nd ser., II, 2, 275Google Scholar; Our British Ancestors, 1865. No. 27 in Witt's Archaeological Handbook of Gloucestershire and No. 56 in Crawford's Long Barrows of the Cotswolds.

^{page 135 note 6} Crania Britannica, vol. II, fasc. 59.

^{page 136 note 1} First identified by Mr T. D. Kendrick in op. cit., 15.

^{page 136 note 2} Long Barrows of the Cotswolds, pp. 144–5.

^{page 137 note 1} If war had not come, the writer intended to verify the strong impression formed during the excavation that the chambers themselves were first erected before the mound was built over them. This method was abundantly clear at Nympsfield, and if it should later prove to have been the general practice, the ‘dolmen’ problem would be nearer to a solution. Crawford & Daniel have argued that Dolmens are merely the remains of long barrows, where the rest of the mound has been carted away. This argument has always raised the question as to why a great mound should be moved in districts where land has practically no value. The suggestion is now made that ‘dolmens’ in most cases represent unfinished long barrows and not denuded ones.

^{page 139 note 1} Op. cit., vol. II, fasc. 59, fig. 2.

^{page 139 note 2} Op. cit., 59.

^{page 140 note 1} Op. cit., II, fasc. 59, fig. 3.

^{page 140 note 2} According to Thurnam (op. cit.), this slab was 3 ft. long.

^{page 140 note 3} Archaeologia, LXXXVI, 128Google Scholar.

^{page 140 note 4} Wilts. Arch. Mag., XLVI, 383Google Scholar.

^{page 140 note 5} P.P.S., 1938, 198Google Scholar.

^{page 141 note 1} P.P.S., 1939, 125Google Scholar.

^{page 141 note 2} P.P.S., 1938, 201Google Scholar. Also at Creevykeel, , J.R.S.A.I., LXIX, 64Google Scholar, and Skendleby, , Archaeologia, LXXXV, 87–8Google Scholar. In the report of the latter excavation several examples are quoted.

^{page 141 note 3} P.P.S., 1938, 202Google Scholar.

^{page 142 note 1} At Notgrove this distance was found to be 31 ft.

^{page 142 note 2} If this suggestion is proved correct by future work, the west wall will cross the barrow at the same point in proportion to the difference in length as at Nympsfield.

^{page 142 note 3} J.R.A.I., vol. V, 155Google Scholar.

^{page 142 note 4} Op. cit.

^{page 143 note 1} This was also noted at Nympsfield, Uley and Winterbourne Stoke long barrows by Thurnam, op. cit.

^{page 143 note 2} Archaeologia LXXXVI, p. 155Google Scholar.

^{page 143 note 3} It is highly probable that the unornamented fragment published by Kendrick, , Axe Age, p. 15Google Scholar, as Neolithic B ware (on account of its coarseness) is really coarse gritty Neolithic A. As Mr Hawkes has pointed out to the writer, before the Abingdon excavation Neolithic A ware was expected to be smooth and fine, like the remaining pieces here; actually this piece is no coarser than a great deal of the Neolithic A ware now known.

^{page 143 note 4} It will be recalled that this was first suggested by Thurnam, (Proc. Soc. Ant., 2nd series, III, 168Google Scholar) and that MrSmith, Regd. A. (Archaeologia, 1926, 81Google Scholar) added some examples to his list.

^{page 143 note 5} P.P.S., 1938, 204Google Scholar.

^{page 143 note 6} Two brown and one white pebbles were found at Notgrove, and at Nympsfield long barrows, Archaeologia, 86; P.P.S., 1938, 204Google Scholar.

^{page 143 note 7} Proc. Soc. Ant., 2nd ser., II, 278Google Scholar.

^{page 145 note 1} Crawford, , Ord. Surv. Prof. Papers, No. 6 p. 4Google Scholar, and No. 4, p. 4; Daniel, , P.P.S., 1939, 165Google Scholar.

^{page 146 note 1} Archaeologia, XVI, 362Google Scholar. These letters are said to come from the Rev. W. H. Thornbury, but this is an error.

^{page 146 note 2} Witts, G. B., Archaeological Handbook of the County of Gloucester, 73–4. Cheltenham, 1883Google Scholar.

^{page 146 note 3} Not two as stated in Archaeologia, XVI, 363Google Scholar. Crawford also suggests that there were two chambers (Long Barrows of the Cotswolds, 115).

^{page 146 note 4} Op.cit.

^{page 146 note 5} Crawford states (op. cit., 115), that Burden's etching is printed in the second edition of Fosbrooke's Encyclopedia (op. cit. infra), but this is not so.

^{page 146 note 6} The only existing drawings of the Avening porthole known to us are that of Burden's already cited and reproduced here, and a drawing preserved in the Gloucestershire Red portfolios (A–S, pl. 1) of the Society of Antiquaries of London, and presented to the Society at the time of the Avening excavation. Burden's drawing has been reproduced many times.

^{page 146 note 7} The height of this stone could not be ascertained with safety.

^{page 146 note 8} Op. cit., fasc. 59, pp.2, 3.

^{page 147 note 1} Archaeologia, XVI, 362Google Scholar.

^{page 147 note 2} Proc. Cotteswold Nat. F.C., V, pl. 1, opp. p. 278Google Scholar.

^{page 147 note 3} Op. cit., 73.

^{page 147 note 4} Op. cit., 115.

^{page 147 note 5} Ib., 115–6.

^{page 148 note 1} Fosbrooke, T. D., Encyclopedia of Antiquities and Elements of Archaeology, Classical and Mediaeval, 1840 (2nd edition), 547Google Scholar.

^{page 148 note 2} Barrow Tump is just over the Avening parish boundary at the present day.

^{page 148 note 3} Witts, (op. cit., 73–4), seems to suggest, however, that the barrow existed in his day.

^{page 148 note 4} Archaeologia, XVI, 362Google Scholar.

^{page 148 note 5} Archaeologia, XVI, plate, LVII.

^{page 148 note 6} Gloucestershire red portfolios (A–S, pl. 1). We are indebted to Mr. H. S. Kingsford for drawing our attention to these drawings.

^{page 148 note 7} Subscription to his etching.

^{page 148 note 8} Proc. Cotteswold Nat. F.C., 1872, 277Google Scholar.

^{page 149 note 1} Gentleman's Magazine, 1789, 393Google Scholar.

^{page 149 note 2} Long Barrows of the Cotswolds, 73.

^{page 149 note 3} Ib.

^{page 149 note 4} Bristol and Glou. Arch. Soc., 1930, 142Google Scholar.

^{page 149 note 5} P.P.S., 1938, 122 ff.Google Scholar

^{page 149 note 6} Ib., 125.

^{page 149 note 7} P.P.S., 1938, 124Google Scholar.

^{page 150 note 1} Op. cit., 166–8.

^{page 150 note 2} Ib., 113.

^{page 150 note 3} Ib., 165–6.

^{page 150 note 4} Ravenhill, T. H., The Rollright Stones, 17Google Scholar.

^{page 150 note 5} Op. cit., 113. The Long Stone is made of Dagham Stone (called locally ‘holey stone’) which weathers in this extraordinary way—hence the name.

^{page 150 note 6} Archaeclogia, 42, 243Google Scholar.

^{page 151 note 1} In a letter to Henry Rowlands printed in the latter's Mona Antiqua Restaurata (2nd edition of 1766), p. 320Google Scholar.

^{page 151 note 2} Clinch, George, in V.C.H. Kent, 1 (1908), 320Google Scholar, refers to two transverse stones dividing the chamber into two and Playne's sketch refers to two stones. Jessup, (Arch. of Kent, 79Google Scholar) suggests that one may be a fallen capstone and Playne thinks that both may be fragments of a capstone (Coll. Cant., 139).

^{page 151 note 3} Playne, George, Coll. Cant. (1893), pl. 25, opp. p. 139Google Scholar.

^{page 151 note 4} Arch. Kent, 79.

^{page 151 note 5} Arch. Camb., 1864, 292Google Scholar.

^{page 151 note 6} Ant. Cornwall, 177–8.

^{page 151 note 7} Made on the margin of an old estate map of the Manor of Lanyon in 1778. See Hencken, , Bronze and Iron Age in Cornwall and Devon, Ph. D. thesis, Cambridge University LibraryGoogle Scholar. My thanks are due to Dr Hencken for allowing me to consult this valuable work.

^{page 151 note 8} J.R.A.I., 1910, 342Google Scholar. Blight thinks the two outlying stones were the remains of a stone circle and perhaps the two 8 ft. stones are also.

^{page 151 note 9} Arch, of Cornwall and Scilly, 46–7; BICD, 12–14.

^{page 151 note 10} For an account of the Tolven see Henderson, Charles (ed. Doble, Canon G. H.), A History of the Parish of Constantine in Cornwall (1939), pp. 8–9Google Scholar.

^{page 151 note 11} Blight, in Arch. Camb., 1864, 293Google Scholar.

^{page 152 note 1} Op. cit., 9.

^{page 152 note 2} Arch, of Cornwall and Scilly, 47.

^{page 152 note 3} Ib., 295.

^{page 152 note 4} Antiquity, 1927, 230Google Scholar. For a view of this site see ib., plate V, opp. p. 229. See also Trans. North Staffs. F.C., XLIII (1908–1909), 133 and 195–6Google Scholar.

^{page 152 note 5} Arch. Camb., 1864, 292–9Google Scholar.

^{page 152 note 6} Blight lists two stones here but Brash (Gents. Mag., 1864, 693Google Scholar) makes these two into four by listing them twice, once under Rosemodress and once as Bolleit.

^{page 152 note 7} Simpson, (Proc. Soc. Ant. Scot., VI, Appendix, 24Google Scholar) refers to Kits Coty House in Kent as having a ‘holed’ capstone and compares it with Trethevy, but neither of these sites are connected with the portholes we are here discussing.

^{page 152 note 8} Op. cit., 217.

^{page 152 note 9} Ant.J., 1924, 406Google Scholar.

^{page 152 note 10} Mr Crawford now concurs in this view. In a letter to me dated 20th May, 1938, he says, ‘I agree with you that it is not likely to be an antiquity at all.’

^{page 153 note 1} For his full report see Archaeologia, 85, 253 ff.Google Scholar

^{page 153 note 2} Arch., 85, 258Google Scholar.

^{page 153 note 3} Fergusson, , Rude Stone Monuments, 472–3Google Scholar.

^{page 153 note 4} In a letter to me dated February 17th, 1937.

^{page 153 note 5} Archaeologia, 85, 258Google Scholar.

^{page 153 note 6} The shape of the stone at present is certainly remarkable and unusual as Hemp says, but then he argues that stone 3 was cut in a remarkable way to allow entrance to a side-chamber, so that stone 6 might also have been cut into its present shape.

^{page 153 note 7} Using the phrase in the sense we have defined it above, p. 1.

^{page 153 note 8} I am indebted to Mr Kinvig who first drew my attention to this feature.

^{page 153 note 9} Ant. J., 1936, 376Google Scholar.

^{page 153 note 10} The alleged porthole stone, sometimes quoted in this connection, at Brimham Rocks in Yorkshire, is an entirely natural site (see Rooke, , Archaeologia, VIII, 209)Google Scholar.

^{page 153 note 11} I am very indebted to Mr and Mrs Hawkes for putting at my disposal their unpublished notes on their discoveries in Ireland.

^{page 154 note 1} This site is no. 7 on Wood-Martin, 's plan (Rude Stone Monuments of Ireland, fig. 93, p. 112)Google Scholar.

^{page 154 note 2} Described by Borlase, , Dolmens of Ireland, 245–6Google Scholar, and illustrated there, fig. 231, p. 246.

^{page 154 note 3} The Axe Age, 43.

^{page 154 note 4} The Dolmens of Ireland, 203–6.

^{page 154 note 5} J.R.A.S.I., 67, 170–1Google Scholar, no. XV in their list.

^{page 154 note 6} Ib., 171.

^{page 154 note 7} Dolmens of Ireland, 127.

^{page 154 note 8} Of the partition slab if F be the end of the monument. For a plan see Borlase, op. cit., 70.

^{page 154 note 9} Ib., 70–1.

^{page 154 note 10} Leask and Price, , P.R.I.A., 43 (1935–1937), C. 79Google Scholar.

^{page 154 note 11} J.R.S.A.I., 67, 303Google Scholar and plate opposite.

^{page 155 note 1} On the Boyne culture in Ireland see Powell, , P.P.S., 1938, 239 ff.Google Scholar

^{page 154 note 2} Childe, (Prehistory of Scotland, 25 ff.Google Scholar) calls the south-west Scottish groups of the Carlingford, the Clyde and Solway groups. I have elsewhere (P.P.S., 1937, 80 ff.Google Scholar, and Antiquity, 1937, 183 ff.Google Scholar) for the sake of convenience called the regional groups of the Carlingford in southern Britain, the Manx, south Peak, Gwynedd, Dyfed and Penwith groups.

^{page 157 note 1} On these Co. Waterford sites and on the distinction of the Tramore group see Powell, , P.P.S., 1940Google Scholar, pt. 2.

^{page 157 note 2} P.P.S., 1937, 71 ff.Google Scholar, and ib., 1939, 143 ff.

^{page 157 note 3} Op. cit., supra.

^{page 158 note 1} The Passage-Graves of Iberia from which the Boyne tombs are derived admittedly have portholes; but only some of the Iberian Passage-Graves possess these: see infra.

^{page 158 note 2} P.P.S., 1939, 143 ff.Google Scholar

^{page 158 note 3} What we are here calling ‘kennel-holes’ do occur in Sardinia. See infra.

^{page 158 note 4} It is based on Leisner, Baudouin, Coutil, Montelius and Sprockhoff and not on a field-survey of all the sites concerned.

^{page 158 note 5} We are not concerned here with the alleged analagous portholed monuments in Algeria, Cyprus, Palestine and the Caucasus nor with the ghost-holed monuments of India.

^{page 158 note 6} Marburger Studien, 147 ff.

^{page 159 note 1} Leisner, op. cit., plate 63, fig.

^{page 159 note 2} Ib., plates 3, 58 and 59. On kennel-holes in general see infra.

^{page 160 note 1} On the Marne grottoes see the Baron de Baye, L' Archéologie préhistorique; Roland, in Bull.Soc.Préh. Francaise, 1910, 521 ffGoogle Scholar. and ib., 1911, 669; and the Favret, Abbé, Revue Arch., 1923, 198 ff.Google Scholar

^{page 160 note 2} For plans of some of these see Lewis, A. L., J.R.A.I., 1910, 337 and 339Google Scholar, and Baudouin and Coutil, op. cit.

^{page 160 note 3} Naturally if the Gallery–Graves are to be open they must be built on a hill-side so that one end is free and open. It may well be that the closed Gallery–Grave evolved in the Paris Basin out of the open Gallery–Grave.

^{page 160 note 4} Manuel d'Archéologie, I, 400Google Scholar.

^{page 160 note 5} Forde's statement that ‘over half the galleries reported have holed stones’ in his discussion on the Paris Basin tombs (Am. Anthrop., 1930, 63Google Scholar) is not correct.

^{page 160 note 6} Axe Age, 27.

^{page 160 note 7} Plan by Lukis, W. C. in Chambered Barrows (MS. in the Lukis and Island Museum, Guernsey), no. CX, pp. 82Google Scholar and opposite.

^{page 160 note 8} Journ. Brit. Arch. Assoc., 24 (1868), p. 40CrossRefGoogle Scholar. Forde, (Man, 1929Google Scholar, no. 80) refers to portholes in ‘PassageDolmens’ at Garen-dol, Kerlouan, in the Morbihan and the laterally chambered Parc-ar-dolmen, St. Pol de Léon, north Finistére. Lukis, if he refers here to the Parc-ar-dolmen site, certainly suggests it is a Gallery-Grave and the Lukis plan in Chambered Barrows of Garren-dol is unquestionably of a Gallery-Grave.

^{page 160 note 9} The Archaeology of Jersey, 254–8 and plate 11, opp. p. 8.

^{page 161 note 1} For plans and diagrams see Montelius, , in Antiq. Tidskr. Sverige, 13, pp. 206–7Google Scholar.

^{page 161 note 2} Sprockhoff, Die Nordische Megalithkultur, plate 16 and fig. 77.

^{page 161 note 3} For a distribution map of the Swedish portholed megaliths see Montelius, , Medd, från Ostergotlands Forntninnesförening, 1907Google Scholar, reproduced in Nordman, , ‘Megalithic Culture of Northern Europe’, Finska Fornminnesföreningens Tidskr., 39, 53Google Scholar.

^{page 161 note 4} Nordman (op. cit., 54–5) mentions in this connection the site at Gravlev, Aalborg in north Jutland. This is, as he points out, not a porthole but an analagous device (see infra).

^{page 161 note 5} See especially Kendrick, The Axe Age, passim.

^{page 161 note 6} Montelius, (Antiq. Tidskr. Sverige, 13, 198–200)Google Scholar, and Kendrick, (Axe Age, 28 and 43Google Scholar) compare King Orry's Grave with Kerlescant. Kerlescant is a closed Gallery and King Orry's Grave an open one.

^{page 162 note 1} Childe, , (Trans. Glasgow Arch. Soc., new ser., VIII, III, 136Google Scholar) suggests an interesting example of this personal factor in the diffusion of tomb types from Brittany to Ireland.

^{page 162 note 2} P.P.S., 1935, 14–3Google Scholar.

^{page 162 note 3} See Childe, , Prehistory of Scotland, 27 ff.Google Scholar

^{page 162 note 4} J.R.S.A.I., 1938, 260 ff.Google Scholar

^{page 164 note 1} For a recent account of this site see the Hebrard, Abbé, Cahiers d'Histoire et d'Archeologie, 1935, 445Google Scholar. Fondouce, Cazalis de (L'Hérault préhistorique. Montpellier, 1922Google Scholar) describes another and similar monument near the farm of Moustachon in the Commune of Causse-de-la-Selle.

^{page 164 note 2} Am. Anthrop., 1930, 63Google Scholar.

^{page 165 note 1} The origin of the Scandinavian portholed megaliths probably forms an exception to what we have been saying here about the parallel development of the porthole in Iberia, north France and Britain. The south Swedish portholed megaliths reveal a number of complete and striking resemblances with the French portholed megaliths; they are open or closed Gallery-Graves, usually sunk in the ground to the level of the capstones, with central portholed partitions. The porthole in this region is not a single cultural trait appearing in a variety of tombs as in Britain; it is part of a linked complex and was probably diffused as a whole from the Paris Basin or south Germany, though how, is not clear at the moment.