3?8 PROCESS FOR 



ceflary to (IK the 'turfs together. Plates of any 

 metal are both too expenfive, and require to 

 be covered with wood, and yet are no fccurity 

 again ft fire. The ufe of flate for roofing houfes, 

 although it may be found in Sweden, is almoll 

 entirely unknown to my countrymen. Tiles there- 

 fore are the only proper materials that remain 

 for us to ufe. And it is cafy to fee ot what con- 

 fequence it is to have them compact and folid. 



in. The ordinary Faults of Tiles. 



SOFTNESS is the greateft fault of tiles. I have 

 fcen tiles burnt in the bell brick kilns in Sweden, 

 which on a roof of 30 degrees of inclination ab- 

 forlxed water like a fponge. The water retained in 

 the poies of tiles is congealed in winter, and the 

 expanlion of the ice fplits and (hatters the tiles, 

 fo as to render them in a fliort time abfolutely 

 ufelefs. 'I'o keep out the \U|ter, tiles are in fume 

 places incrullcd with u thin covering of vitreous 

 matter; which adds corilidenibly to the ex- 

 pcncu. But, if tiles were more thoroughly burnt 

 on the furfacc, fo as to be nearly reduced tofu- 

 Jion, they might, in my opinion, be rendered 

 fo hard as to abibrb very little moifture, and to 

 be almoft entirely proof aguinll the influence of 

 froft. JJut, before faying more on this head, 

 I mull enter into a more accurate examination 

 wf the nature both of pure and common clay. 



S iv. 



