44 



They might have gone a little farther and cut off the angles at the 

 top as well as at the bottom : 



FIG. D. TILES SCALLOPED OR ANGULATED TOP AXD BOTTOM. 1 



This latter seems to be a further development of Figure C, but 

 none is so perfect, or so economic as regards weight, as Figure B. 



The tiles, then, so formed are in some instances thickly strewn 

 upon the floors of Roman dwellings. Many of them have the nails 

 still left in the hole made for that purpose, and it seems, from the 

 manner in which they are disposed, that, the Roman dwelling being 

 abandoned, the timbers of the roof gave way, and the tiles thus fell 

 upon the floors, where they became intermixed with whatever 

 articles might have been left in the rooms, some of which we would 

 notice here. 



WALL PAINTING. 



Among the tiles, and still adhering to the portion of the wall left 

 intact, will usually be found and it was at Thornford in a great 

 degree of profusion a quantity of wall decorations, and, as the 

 plaster and colouring are still very fresh and perfect, we ventured to 

 ask Professor Church how he considered these colourings to have 

 been formed; and we have much pleasure in transcribing the follow- 

 ing interesting letter : 



* We beg to offer our thanks to the Cotteswold Field Club for the loan of the cuts 

 p., c, and I), which were made by us to illustrate some roof tiles dug up at Bradford 

 Abl.a.-. Ki>. 



