234 On a Mode of erecting light Faults 



advantageous for the uniform closing of all vaults, the inevit- 

 able consequence is, bellying and cracking. If, on the other 

 hand, \ve wish to leave the centering standing till the complete 

 drying of the vault, the wasting of the mortar would cause all 

 the joints to open and crack. But the network formed b'y 

 the mortar in all the joints, gives to a thin vault of heavy stone 

 a peculiar strength ; as the author very clearly ascertained by 

 an experiment for the purpose. There is in this place a kind 

 of stone, which is used very advantageously for the lining of 

 walls, the pannels of ceilings, and the construction of chimneys. 

 It is a conglomerate of loose pumiceous sand, cemented into a 

 coherent mass by a loamy earth ; and is naturally so tender, 

 that it may be rubbed to pieces in the hand ; and, indeed, has 

 hardly more consistence than a swallow's nest. This mass lies 

 in layers some feet under the surface, in the country between 

 Engers and Bendorf : it is worked in stages, cut into loaves of 

 13 in. long, Gin. broad, and 4 or 5 in. thick, "and dried in the 

 air. With these stones set on the long narrow side, he ordered 

 a very flat vault (31 feet span, and 4J feet spring, and about 

 4 feet broad) to be thrown across between two old walls. Its 

 thickness was thus only 6 inches, or the 124th part of the 

 diameter of the semicircle, of which the arch was part ; con- 

 sequently, the stones suffered a pressure equal to that right 

 upon each stone of a semicircular vault of 62 feet diameter, and 

 therefore far greater than that which the stones could have 

 borne on the ground ; and yet this arch, although it could be 

 put in strong vibration with the hand, had so much strength 

 that a man could walk over it. 



That the old vaults were built free-handed, and not upon a 

 boarded centering, no one can doubt. Who would have given 

 himself the trouble, so disproportionate to its object, of making 

 such a boarding vaulted according to each arch of the cen- 

 tering, when he might obtain the same end with one which was 

 quite common ? Besides, the unequal convexity in all such 

 old vaults shews that no gage or model was ever applied, but 

 the observation of the proper form was left to the choice and 

 practice of the mason. We often see, as has already been 

 said, a strong convexity pass into a flat one, or reversely; 

 when probably it had suddenly struck the mason, that he was 



