1^ Proceedings cf the Koj/al Irish Academij. 



and oue or more crescent walls, we find Caherlismacsheedy,' a single crescent 

 on a clift' in Burren. The type is very common in England and Scotland, the 

 finest and most complex specimen being the two conjoined forts of Coldingham, 

 each with three walls. We note in England the forts of Embury Beacon, 

 Devon, and Blackers Hill, other typical examples being the inland Scottish 

 "Doon" of Xunmill and Errickstane." In France we have many fine 

 examples. To select a few : There is a double crescent earthwork on a spur 

 at Caudebec, and near it a promontory fort. The camp of Bois de Kouret 

 has two stone walls built each in two sections, as in the Irish forts. The 

 masonry, too, is identical with that in the forts of western Ireland. At 

 Sanau, in Cantal, a single crescent wall encloses a garth 170 feet long and 

 105 feet deep on the edge of an inland cliff. The fort of St. Maurice at 

 Beaulieu (E.M.) has two crescent rings enclosing many house-sites, some of 

 the Eomau period. The ground slopes back from the inland clilV as at Dun 

 Aengiisa. Mont Milan (Cote d'Or) has also two walls; but, unlike the Irish 

 forts, it has flanking towere believed to be of the same period as the fort. It 

 is noteworthy that the Irish, no less in their forts than in many of their 

 later castle courts, were entirely indifferent to the advantages of flanking 

 defences.' In Hungary lies the great ring of Bdny, an earthwork of three 

 crescent mounds, extending for 1700 feet along a steep bluff. In Switzerland 

 and Perm we have crescent works fencing mountain spurs ; and crescent ring- 

 walls are found in Sweden.' The type also occurs in America. 



Whether the walls of Dun Aengusa were all crescents from the first we 

 can never know^ ; for they, their foundations, and the rock for over 300 feet 

 beneath are devoured by the Atlantic and " their memorial has perished with 

 them." We have no means of calculating^the advance of this destruction ; 

 and it may have been very slow for centuries. 



The fall of cliffs is nowhere uniform ; even at the same locality all depends 

 on the jointing, the currents, and the prevaQing winds. A fort like Doon, 

 in Iraghticonor,* Kerry, was evidently a crescent from the first ; the sea has 



' Journal B.S.A.I., vol. xxxi., pp. 275, 276. 



'Si* Dr. Christison's " Early Kortificaiiong in Scotland," pp. 130, 131, Coldingham ; p. 13J, 

 Arbory ; Upper Cademiir;p. 201, Raecleuchead ; p. 206, Krriclutane ; and Mr. A. R. AUcrofl's 

 "Earthwork of Kngland," pp. 53, 64, 69, 63, and 113. 



3 " Socielc Prcbiaioriquc de France," Bulletin. Tome iv., p. 311 ; vi., pp. 231, 41.i. Rapport, 

 Tome v., p. 76. Compteu Rendua (1906), pp. 27, 30, 36, 52; and Dr. Adiien Gucbhard's 

 "Enceintes Prchiatoriques " (1907), pp. 11, 12. 



♦Tlie notee on many o£ theae continental forts aro brought together in "Ancient Forts of 

 Ireland," sections 8, 11, 20, with plans, &c., figs. 2, 3. For a Scandinavian " ring mur," crescent- 

 shaped in plan, see Borlase, " Dolmens of Ireland," vol. iii., p. 1133. For Russian forts in Perm, 

 see "Camps Retrancbcs," Dr. A. Gucbbard in " L' Association Fran^aise pour ravanccment des 

 Sciences " (36th Congress, 1902), p. 3. From plans by M. Vkdimir Tolmatcheff. 



' Described in u|>upvr iea<l before R.S..\.I. in July, 1909, now printed, but as yit unpublished. 



