[993] Lewis’s Topographical Dictionary.

[994] The passage of the river Lune in Lancashire is similarly defended by the mottes of Melling and Arkholme.

[995] The dates given are those of the Brut, and probably two years too early.

[996] Meyrick’s History of Cardigan, p. 146.

[997] Meyrick’s History of Cardigan, p. 146.

[998] Lewis’s Topographical Dictionary.

[999] We do not include the castles which the Welsh rebuilt. Thus in 1194 we are told that Rhys built the castle of Kidwelly, which he certainly only rebuilt.

[1000] Malcolm Canmore himself had passed nearly fourteen years in England. Fordun, iv., 45.

[1001] Burton remarks: “To the Lowland Scot, as well as to the Saxon, the Norman was what a clever man, highly educated and trained in the great world of politics, is to the same man who has spent his days in a village.” History of Scotland, i., 353.

[1002] Dr Round has brought to light the significant fact that King David took his chancellor straight from the English chancery, where he had been a clerk. This first chancellor of Scotland was the founder of the great Comyn family. The Ancestor, 10, 108.

[1003] Fordun, Annalia, vol. iv.

[1004] It is tempting to connect the extraordinary preponderance of mottes, as shown by Dr Christison’s map, in the shires which made up ancient Galloway (Wigton, Kirkcudbright, and Dumfries) with the savage resistance offered by Galloway, which may have made it necessary for all the Norman under-tenants to fortify themselves, each in his own motte-castle. It is wiser, however, to delay such speculations until we have the more exact information as to the number of mottes in Scotland, which it is hoped will be furnished when the Royal Commission on Historical Monuments has finished its work. But this work will not be complete unless special attention is paid to the earthworks which now form part of stone castles, and which are too often overlooked, even by antiquaries. The New Statistical Account certainly raises the suspicion that there are many more mottes north of the Forth than are recognised in the map alluded to. In one district we are told that “almost every farm had its knap.” “Forfarshire,” p. 326.

[1005] Cited by Fordun, v., 43.

[1006] Benedict of Peterborough, i., 68, R. S.

[1007] Fordun, v., 26. Bower in one of his interpolations to Fordun’s Annals, tells how a Highlander named Gillescop burnt certain wooden castles (quasdam munitiones ligneas) in Moray. Skene’s Fordun, ii., 435.

[1008] That Fordun should speak of the castra and municipia of Macduff is not surprising, seeing that he wrote in the 14th century, when a noble without a castle was a thing unthinkable.

[1009] Burton actually thought that the Normans built no castles in Scotland in the 12th century. Messrs MacGibbon and Ross remark that there is not one example of civil or military architecture of the 12th century, while there are so many fine specimens of ecclesiastical. Castellated Architecture of Scotland, i., 63. It is just to add that when speaking of the castles of William the Lion, they say: “It is highly probable that these and other castles of the 13th century were of the primeval kind, consisting of palisaded earthen mounds and ditches.” Ibid., iii. 6.

[1010] Mote is the word used in Scotland, as in the north of England, Pembrokeshire, and Ireland, for the Norman motte. As the word is still a living word in Scotland, its original sense has been partly lost, and it seems to be now applied to some defensive works which are not mottes at all. But the true motes of Scotland entirely resemble the mottes of France and England.

[1011] Scottish Review, xxxii., 232.

[1012] Scottish Review, xxxii., 232.

[1013] Ibid., p. 236.

[1014] This list is mainly compiled from Chalmers’ Caledonia, vol. i., book iv., ch. i. The letter C. refers to Dr Christison’s Early Fortifications in Scotland; N., to Mr Neilson’s paper in the Scottish Review, 1898; O.M., to the 25-inch Ordnance Map; G., to the Gazetteer of Scotland. It is a matter of great regret to the writer that she has been unable to do any personal visitation of the Scottish castles, except in the cases of Roxburgh and Jedburgh. It is therefore impossible to be absolutely certain that all the hillocks mentioned in this list are true mottes, or whether all of them still exist.

[1015] Registrum Magni Sigilli, quoted by Christison, p. 19.

[1016] A plan is given by Mr Coles in “the Motes, Forts, and Doons of Kirkcudbright.” Soc. Ant. Scot., 1891-1892.

[1017] M‘Ferlie, Lands and Their Owners in Galloway, ii., 47.

[1018] This description, taken from the Gazetteer, seems clear, but Mr Neilson tells me the site is more probably Woody Castle, which is styled a manor in the 15th century. The N. S. A. says: “There is the site of an ancient castle close to the town, on a mound of considerable height, called the Castle Hill, which is surrounded by a deep moat.” “Dumfries,” p. 383.

[1019] Annals, ii., 196, cited in Douglas’s History of the Border Counties, 173.

[1020] Round, in The Ancestor, 10, 108.

[1021] Dr Christison distinctly marks one on his map, but Mr Coles says there is no trace of one, though the name Marl Mount is preserved. Soc. Ant. Scot., 1892, p. 108.

[1022] See the Aberdeen volume, p. 1092.

[1023] See Grose’s picture, which is confirmed by Dr Ross.

[1024] The name Tom-a-mhoid is derived by some writers from the Gaelic Tom, a tumulus (Welsh Tomen) and moid, a meeting. Is there such a word for a meeting in Gaelic? If there is, it must be derived from Anglo-Saxon mot or gemot. But there is no need to go to Gaelic for this word, as it is clear from the Registrum Magni Sigilli that moit was a common version of mote, and meant a castle hill, the mota or mons castri, as it is often called.

[1025] Chalmers, Caledonia, iii., 864. Sir Archibald Lawrie, however, regards it as doubtful whether Arkel was the ancestor of the earls of Lennox. Early Scottish Charters, p. 327.

[1026] M‘Ferlie, Lands and Their Owners in Galloway, ii., 140-141.

[1027] See plan in MacGibbon and Ross, Castellated Architecture, iv., 341.

[1028] The name Maccus is undoubtedly the same as Magnus, a Latin adjective much affected as a proper name by the Norwegians of the 11th and 12th centuries.

[1029] Lawrie, Early Scottish Charters, p. 273.

[1030] MacGibbon and Ross, i., 279.

[1031] Proceedings of Soc. Ant. Scotland, xxxi., and N. S. A.

[1032] See Armstrong’s History of Liddesdale, cited by MacGibbon and Ross, i., 523.

[1033] Round, The Ancestor, No. 11, 130.

[1034] Benedict of Peterborough, i., 67. See Mr Neilson’s papers in the Dumfries Standard, June 28, 1899. Mr Neilson remarks: “It may well be that the original castle of Dumfries was one of Malcolm IV.’s forts, and that the mote of Troqueer, at the other side of a ford of the river, was the first little strength of the series by which the Norman grip of the province was sought to be maintained.”

[1035] “Mottes, Forts, and Doons of Kirkcudbright,” Soc. Ant. Scot., xxv., 1890.

[1036] The Annals of the Four Masters mention the building of three castles (caisteol) in Connaught in 1125, and the Annals of Ulster say that Tirlagh O’Connor built a castle (caislen) at Athlone in 1129. What the nature of these castles was it is now impossible to say, but there are no mottes at the three places mentioned in Connaught (Dunlo, Galway, and Coloony). The caislen at Athlone was not recognised by the Normans as a castle of their sort, as John built his castle on a new site, on land obtained from the church. Sweetman’s Cal., p. 80.

[1037] The meagre entries in the various Irish Annals may often come from contemporary sources, but as none of their MSS. are older than the 14th century, they do not stand on the same level as the two authorities above mentioned.

[1038] “Hibernicus enim populus castella non curat; silvis namque pro castris, paludibus utitur pro fossatis.” Top. Hib., 182, R. S., vol. v. In the same passage he speaks of the “fossa infinita, alta nimis, rotunda quoque, et pleraque triplicia; castella etiam murata, et adhuc integra, vacua tarnen et deserta,” which he ascribes to the Northmen. This passage has been gravely adduced as an argument in favour of the prehistoric existence of mottes! as though a round ditch necessarily implied a round hill within it! Giraldus was probably alluding to the round embankments or raths, of which such immense numbers are still to be found in Ireland. By the “walled castles” he probably meant the stone enclosures or cashels which are also so numerous in Ireland. In the time of Giraldus the word castellum, though it had become the proper word for a private castle, had not quite lost its original sense of a fortified enclosure of any kind, as we know from the phrases “the castle and tower” or “the castle and motte” not infrequent in documents of the 12th century (see Round’s Geoffrey de Mandeville, Appendix O, p. 328). We may add that Giraldus’ attribution of these prehistoric remains to Thorgils, the Norwegian, only shows that their origin was unknown in his day.

[1039] See Expug. Hib., 383, 397, 398.

[1040] I am informed that the “Crith Gablach,” which gives a minute description of one of these halls, is a very late document, and by no means to be trusted.

[1041] Vide the Irish Annals, passim.

[1042] There is another story, preserved in Hanmer’s Chronicle, that the Irish chief Mac Mahon levelled two castles given to him by John de Courcy, saying he had promised to hold not stones but land.

[1043] Joyce’s Irish Names of Places, p. 290.

[1044] See J. E. Lloyd, Cymmrodor, xi., 24; Skeat’s English Dictionary, “town.” In the “Dindsenchas of Erin,” edited by O’Beirne Crowe, Journ. R. S. A. I., 1872-1873, phrases occur, such as “the dun was open,” “she went back into the dun,” which show clearly that the dun was an enclosure. In several passages dun and cathair are interchanged.

[1045] Joyce, Irish Names of Places, p. 273.

[1046] Annals of the Four Masters, 1166.

[1047] See Orpen, “Motes and Norman Castles in Ireland,” in Journ. R. S. A. I., xxxvii., 143-147.

[1048] Sweetman’s Calendar of Documents relating to Ireland, i., 412.

[1049] That a motte-castle of earth and wood seemed to Giraldus quite an adequate castle is proved by the fact that numbers of the castles which he mentions have never had any stone defences. It may be a mere coincidence, but it is worth noting, that there are no mottes now at any of the places which Giraldus mentions as exilia municipia, Pembroke, Dundunnolf, Down City, and Carrick.

[1050] This word must not be understood to mean that this new type of castle was Edward’s invention, nor even that he was the first to introduce it into Europe from Palestine; it was used by the Hohenstauffen emperors as early as 1224. See Köhler, Entwickelung des Kriegswesen, iii., 475.

[1051] Newcastle, Worcester, Gloucester, and Bristol are instances.

[1052] Rhuddlan is an instance of this.

[1053] Book of Rights, p. 203.

[1054] It must be admitted that in the most recent and most learned edition of the Anglo-Saxon Chronicle the topographical identifications are quite on a level with O’Donovan’s.

[1055] The Annals have not been used, partly because in their present form they are not contemporary, and partly because the difficulties of identifying many of the castles they mention appeared insuperable.

[1056] See especially two papers on “Motes and Norman Castles in Ireland,” in English Historical Review, vol. xxii., pp. 228, 240. Mr Orpen has further enriched this subject by a number of papers in the Journ. R. S. A. I., to which reference will be made subsequently.

[1057] The only castles still unidentified are Aq’i, Kilmehal, Rokerel, and Inchleder.

[1058] It should be stated that the great majority of the castles in this list have been visited for the writer by Mr Basil T. Stallybrass, who has a large acquaintance with English earthworks, as well as a competent knowledge of the history of architecture. The rest have been visited by the writer herself, except in a few cases where the information given in Lewis’s Topographical Dictionary or other sources was sufficient. The castles personally visited are initialled.

[1059] Annals of Loch Cè.

[1060] Orpen, Eng. Hist. Rev., xxii., 249.

[1061] Orpen, Eng. Hist. Rev., xxii., 450, citing from MS. Annals of Innisfallen.

[1062] The poetical list enumerates the places which were “of the right of Cashel in its power.” The prose version, which may be assumed to be later, is entitled “Do phortaibh righ Caisil,” which O’Donovan translates “of the seats of the king of Cashel.” But can one small king have had sixty-one different abodes? Professor Bury says “The Book of Rights still awaits a critical investigation.” Life of St Patrick, p. 69.

[1063] Ibid., p. 449. See Westropp, Trans. R. I. A., xxvi. (c), p. 146. Mr Orpen informs me that the Black Book of Limerick contains a charter of William de Burgo which mentions “Ecclesia de Escluana alias Kilkyde.” No. cxxxv.

[1064] Journ. R. S. A. I., 1898, 155; and 1904, 354.

[1065] Eng. Hist. Rev., xxii., 452.

[1066] Butler’s Notices of the Castle of Trim, p. 13.

[1067] Eng. Hist. Rev., xxii., 458.

[1068] Eng. Hist. Rev., xxii., 441.

[1069] “Exile municipium,” Giraldus, 345. See Eng. Hist. Rev., xx., 717.

[1070] Annals of Ulster, 1177.

[1071] See Orpen, “Motes and Castles in County Louth,” Journ. R. S. A. I., xxxviii., 249. The town walls are later than the castle, and were built up to it.

[1072] Cited by Westropp, Journ. R. S. A. I., 1904, paper on “Irish Motes and Early Norman Castles.”

[1073] Annals of Ulster, 1186.

[1074] Round, Cal. of Doc. preserved in France, i., 105, 107.

[1075] “On the Ancient Forts of Ireland,” Trans. R. I. A., 1902.

[1076] Orpen, “The Castle of Raymond le Gros at Fodredunolan,” Journ. R. S. A. I., 1906.

[1077] Annals of Innisfallen.

[1078] Orpen, Eng. Hist. Rev., xxii., 449.

[1079] “On some Caves in the Slieve na Cailliagh District,” by E. C. Rotheram, Proc. R. I. A., 3rd ser., vol. iii. Mr Rotheram remarks that the passages in the motte of Killallon, and that of Moat near Oldcastle, seem as if they were not built by the same people as those who constructed the passages at Slieve na Cailliagh.

[1080] Annals of Ulster.

[1081] Annals of Loch Cè.

[1082] Orpen, Eng. Hist. Rev., xxii., 448.

[1083] Ibid., p. 242.

[1084] Annals of Ulster. See Orpen, Eng. Hist. Rev., xxii., 443.

[1085] Annals of Ulster.

[1086] Annals of the Four Masters, vol. iii. See Orpen, Journ. R. S. A. I., vol. xxxix., 1909.

[1087] Orpen, Eng. Hist. Rev., xxii., 448. A place called Graffan is mentioned in the Book of Rights, and on the strength of this mere mention it has been argued that the motte is a prehistoric work. Trans. R. I. A., vol. xxxi., 1902.

[1088] Mr Orpen.

[1089] Giraldus’ words are: “Castrum Lechliniæ, super nobilem Beruæ fluvium, a latere Ossiriæ, trans Odronam in loco natura munito.” V., 352. See Eng. Hist. Rev., xxii., 245.

[1090] See Orpen, Eng. Hist. Rev., xxii., 456, and Journ. R. S. A. I., xxxvii., 140.

[1091] Orpen, “Motes and Norman Castles in County Louth,” Journ. R. S. A. I., xxxviii., 241, from which paper the notice above is largely taken.

[1092] Eng. Hist. Rev., xxii., 242.

[1093] The castle is casually mentioned by Giraldus, v., 100, and the date of its erection is not given.

[1094] As far as the writer’s experience goes, terraces are only found on mottes which have at some time been incorporated in private gardens or grounds.

[1095] Journ. R. S. A. I., vol. xxxix., 1909.

[1096] Piers, Collect. de Rebus Hib., cited by Orpen.

[1097] Mr Orpen says: “The castle was ‘constructed anew’ in the sixth and seventh years of Edward I., when £700 was expended.” Irish Pipe Rolls, 8 Edward I., cited in Eng. Hist. Rev., xxii., 454.

[1098] Line 3178.

[1099] The annular bailey, with the motte in the centre, is a most unusual arrangement, and certainly suggests the idea that the motte was placed in an existing Irish rath.

[1100] See Appendix M.

[1101] Annals of Loch Cè.

[1102] Giraldus, v., 313.

[1103] This keep has a square turret on each of its faces instead of at the angles. A similar plan is found at Warkworth, and Castle Rushen, Isle of Man.

[1104] Orpen, Eng. Hist. Rev., xxii., 248.

[1105] Figured in The Tomb of Ollamh Fodhla, by E. A. Conwell, 1873.

[1106] Gir., i., 255, 277.

[1107] Eng. Hist. Rev., xxii., 457.

[1108] In five cases the mottes are now destroyed.

[1109] The dates of the building of numbers of these castles are given in the Annals of Ulster and the Annals of Loch Cè.

[1110] Cal. of Pat. Rolls, 1232-1247.

[1111] The tower at Malling was supposed to be an early Norman keep by Mr G. T. Clark (M. M. A., ii., 251), but it has recently been shown that it is purely an ecclesiastical building.

[1112] The only stone castles of early date in France which the writer has been able to visit are those of Langeais, Plessis Grimoult, Breteuil, and Le Mans. The two latter are too ruinous to furnish data.

[1113] Given in D’Achery’s Spicilegium, iii., 232.

[1114] This can be positively stated of Baugé, Montrichard, Montboyau, St Florent-le-Vieil, Chateaufort, and Chérament. M. de Salies thinks the motte of Bazonneau, about 500 metres from the ruins of the castle of Montbazon, is the original castle of Fulk Nerra. Histoire de Fulk Nerra, 57. About the other castles the writer has not been able to obtain any information.

[1115] See Halphen, Comté d’Anjou au xiième Siècle, 153.

[1116] The building of Langeais was begun in 994. Chron. St Florent, and Richerius, 274.

[1117] It somewhat shakes one’s confidence in De Caumont’s accuracy that in the sketch which he gives of this keep (Abécédaire, ii., 409) he altogether omits this doorway.

[1118] Measurements were impossible without a ladder.

[1119] It is well known that William the Conqueror left large treasures at his death.

[1120] The keep of Colchester is immensely larger than any keep in existence. Mr Round thinks it was probably built to defend the eastern counties against Danish invasions. Hist. of Colchester Castle, p. 32. Its immense size seems to show that it was intended for a large garrison.

[1121] Cours d’Antiquités Monumentales, v., 152, and Abécédaire, ii., 413-431. De Caumont says of the keep of Colchester, “il me parait d’une antiquité moins certaine que celui de Guildford, et on pourrait le croire du douzième siècle” (p. 205), a remark which considerably shakes one’s confidence in his architectural judgment.

[1122] As only the foundations of Pevensey are left, it gives little help in determining the character of early keeps. It had no basement entrance, and the forebuilding is evidently later than the keep.