240 Proceedings of the Royal Irish Academy. 



by our ignorance of the meaning of the technical term tuarad, on which all 

 the MSS. are agreed, with slight differences of spelling. 1 It probably denotes 

 some part of the roof. But. leaving to future discoverers to find out what 

 the " very low tuarad" of the house may have been, we note that the build- 

 ing had "a lofty floor" {lar ard . We can guess why the floor was raised 

 aloft with the help of the description in YD. According to this account 

 there was a height in the west, and a great height in the north, but a low 

 •! to the east. To explain this we must remember that the house was 

 built over or against a tumulus. The top of Llie mound (the " great height '") 

 was to the north, the the mound to the ist of the house was 



natural level of the ground. A- the -lope of the mound would encroach 

 ,,,, £] n ad-level of tin- house built againsi it, the floor would have t> 



if sufficient breadth. The diagram (fig. 1) 



■. attempt to represent th( - a building that would answer 



to the conditions. A structure, thus cleverly adapted to an awkward Bite, 



might well »n*s triumph," if that could be the meaning 



.nd a person building his own house after the model of the 



|. M risiu " might be supposed to be placing his household undei 



her protection. Whatever i bought of these conjectures, it can at 



claimed that they afford a not unreasonable explanation of certain 



remarkable thai -one tion ia urgently called for. 



:'.. /, /. Xri/I 



This structure, whii I irily identified, had four doors 



in I. (J, and II >. It d not appear whether 



il,,.. surrounding ramparl or in the enclosed buildii s 



probably the much ruined to allow us to deter- 



m ii I nature of on the 



aim - A' wiped oul 



\ • againsi the outer south- 



raith " I 



ide the rampart, and on the 



I i in confirmation 



• . ly not :: !'!> : the two 



passages seem t' . VD gives a different 



situation for the grave. J of the monument are there enume- 



The w..rd seems to be a ami Kryiptwr. 

 • - .. VD in LOO. 



1 Tara, pp. 160, LI 



