30 ON THE OLDER FORMS OF 



sent an ao^a'roiration similar to that in a mnseum of all the 

 constrnctions cnstomaiy in Greece, Lower Italy and Sicily, 

 and this snpposition has ventied itself, for further studies 

 showed that at the places in Greece and Italy, which we 

 visited, certain particular kinds of construction are used 

 almost exclusively, and that the variety and midtiplicity 

 of forms found at Olympia occurred nowhere else to the 

 same extent." 



Speaking of the marble roof, Graeber says : "The gen- 

 eral system and scheme of the antique marble roof is well 

 known through many publications. This system, however, 

 has not been invented for the marble roof, but had its 

 prototype in the clay-tile roof. The antique roof had to 

 pass through centuries of evolution till it attained that per- 

 fection which we admire in the Parthenon of Athens, and 

 the Zeus Temple of Olympia and many other ediHces. As 

 regards elegance, one may even say subtility of perfection, 

 the Greek tile roof ranks even above the marble roof." 



I cannot forbear quoting further from this valuable me- 

 moir of Graeber's. He says in regard to the attachment 

 of tiles on the roof: "A securing of the tiles on the raft- 

 ers by means of nails did not take place ; onl}^ the lowest 

 tile, next to the gutter, was always secured by iron or 

 bronze nails to the rafter. All tiles with nail-holes, there- 

 fore, belong, without exception, to the gutter, and just so 

 little did the tiles have projections for hanging them to 

 the laths as it is assumed erroneously of the marl)le roofs, 

 but they rested directly on the rafters, and maintained 

 themselves in their position in part by their weight, in 

 part by supporting themselves throuirh the next lower tile 

 by means of the cutting on their lower surface. This may 

 have occasioned, under certain circumstances, a heavy 

 strain; for instance, a sliding down, involving even the 

 lowest gutter or moulding tiles." And he refers to the 



