NATURAL BRIDGES BY STREAM EROSION 323 



"The explanation of the formation of the bridge is as follows : The stream 

 which flows underneath the bridge has been able to excavate, owing to a for- 

 mer waterfall and the peculiar platy structure of the rhyolite, in which curved 

 layers of extremely different physical texture and friability offered a favor- 

 able site for attack by frost and water."* 



The excavation was accomplished by the water which flowed over the 

 fall (now the top of the bridge) excavating with its sediment the rock at 

 its base, assisted by the separation of the thin plates of lava by the ex- 

 pansion of the water on freezing. This excavating continued until a 

 layer of more porous lava (represented by the space between the slabs 

 composing the bridge) was found, which permitted the water to flow 

 downward into the cave. This process was repeated, thus producing a 

 bridge made of two vertical layers. 



h. Remnants of lava stream roofs not included. — The remnants of lava 

 stream roofs are not included as natural bridges, since they are to be 

 classed as arches. It should be stated, however, that arches of this sort 

 perform the functions of bridges, since they permit men and animals to 

 pass over a cavity which otherwise it would be necessary to bridge. 



4. TRAVERTINE-CEMENTED STREAM DEPOSITS UNDERCUT BY STREAM 



EROSION 



The writer has neither seen nor been able to learn of any bridges of 

 this origin in North America. Professor Friih cites several examples 

 from Europe and South America which should be mentioned here, in 

 order that our classification may be as complete as possible. 



a. Bridges in Arcadia and Messina, Greece. — In northern Arcadia, 

 in Greece, a brook disappears under a tunnel 100 meters long whose roof 

 consists of earth and debris cemented by travertine. It supports fields, 

 a few trees, and a roadway.® 



Another in northern Messina, in Greece, is described as being a tunnel 

 of mountain debris cemented by travertine. 



h. Puenta del Incas, South America. — Another cited by the same 

 writer,^^ from Gussfeldt, is near the Banas del Inca. It is 40 meters 

 long and 40 meters broad and consists of valley debris cemented by traver- 

 tine over which a compact layer of travertine has been formed. 



5. BY THE UNDERCUTTING OF A PETRIFIED LOG 



An unusual natural bridge (plate 23, figure 1) formed by stream ero- 

 sion is that in the petrified forest of Adamana, Arizona, where a petrified 



8 H. F. Cleland : American Journal of Science, vol. 20, August, 1905, p. 123. 

 » Friih : Naturbriicken, ibid., p. 368. 

 10 Ibid., p. 363. 



