the HIGHLANDS of SCOTLAND. 21 



efficacious than for the former. It is much to be doubted, 

 whether it would be at all pofuble, even in the prefent day, by 

 the utmoft combination of labour and of fkill, to furround a 

 large fpace of ground with a double rampart of ftones, com- 

 pacted by fire, of fuch height and folidity as to ferve any pur- 

 pofe of fecurity, or defence againft a beiieging enemy. Any 

 ftruclure of this kind muft have been irregular, low, fragile, 

 eafily fcaled and quite infecure ; a much weaker rampart, in 

 fhort, than a fimple wall of turf or wooden pallifade. The 

 veftiges yet remaining, as I have already obferved, give no 

 room to fuppofe, that the vitrified mound has ever been much 

 more entire than it is at prefent. The effect of fire upon ftruc- 

 tures reared in the manner I have fuppofed them to have been, 

 will account moft perfectly for their prefent appearance. 



It was from necefTity that the builders of thofe fortifications 

 betook themfelves to a mode of ftruclure fo liable to be de- 

 flroyed by fire. In thofe parts where ftones could be eafily 

 quarried, of fuch fize and form as to rear a rampart by them- 

 felves of fufKcient ftrength and folidity, there was no occafion 

 to employ wood or turf in its conftructon, and it was there- 

 fore proof againft all afTault by fire. Such are the ramparts 

 which appear on the hill of Dun-Jardel, Dun-Evan, and many 

 others, on which there is not the fmalleft appearance of vitrifi- 

 cation. But on Craig- Phad rick, and the other hills above de- 

 fcribed, where, from the nature of the rock, the ftones could, 

 be procured, only in irregular and generally fmall fragments, 

 it was neceffary to employ fome fuch mode of conftruction as 

 I have fuppofed ; and thefe ramparts, though folid and well 

 calculated for defence againft every attack by force or ftratagem, 

 were not proof againft the affault by fire. 



But thofe ancient fortifications prefent a much more curi- 

 ous and more interefting object of fpeculation, than thofe un- 

 certain and indeed fruitlefs conjectures as to the mode in which 

 they have been reared. It is evident, that, were it poffible to 



afcertam 



