/79*' atitiquitUs {ft StotTund".. 338 



four : One on the top of a hill in the island of Islaj, 

 iitar the sound, called the hill of Lofsit, if my me^- 

 raory serves me well ; another stands on the tof» 

 of a hill in the north east side of the island of Tiree, 

 the name of which I have forgot ; and there are two 

 others near together, in die vicinity of Portree, ia 

 the isle of Skye. These two last are remarkable, 

 among other particulars, for having an area adjoin- 

 ing to them, surrounded with a kind of wall or ram- 

 part, the tower standing in one corner, that seems to 

 have been the most inaccefsible, resembling very 

 much in the plan, a town with its citadel. From 

 these observations, I am inclixned to think that there 

 have been two clafses of circular buildings in Scot- 

 land J one of them for temporary defence, of which 

 the four just mentioned may be accounted examples ; 

 and the other solely for religious ceremonies^ of 

 which the following are well known, and to which 

 all the observations in this paper are strictly appli- 

 cable : One at Dornadilla in the parifli of Hae in Lord. 

 Rae's country ; one at Dunrobin in Sutherland ; ona 

 at Dunagglesgag in Rofsihire ; and its fellow on the 

 opposite side of the frith of Dornoch ; and three ia 

 the valley of Glenelg, in Invernefsfhire, which have 

 been described by Pennant, and which I myself exa- 

 mined with particular attention. I lliauld think it 

 probable tliat this lust clafs of circular buildings may- 

 be the most ancient of the two ; and that the nativeSj 

 in future times, having observed hov/ long the walls, 

 thus built stood firm, have adopted the idea of rear- 

 ing places of defence on the same general plaa. Bufi: 

 this I only offer as a conjecture. 



