Fhotqgraph by V. F. Cook 

 SOME OF THE STAIRCASE FARMS OF THE ANCIENTS 



Each terrace consists of three parts — the wall and the two distinct layers of earth that 

 fill the space behind the wall. All of the ruined terraces show the same inside structure, 

 wherever the walls are removed. The strata that are hidden behind the walls are artificial 

 no less than the stone facing (see also illustrations, pages 508 and 509). 



The underlying stratum, or artificial subsoil; is composed of coarse stones and clay, and 

 is covered by a layer of fine surface soil two. or three feet. thick. The thickness of the 

 subsoil layer depends, of course, upon the height of the terrace. Where clay or other light- 

 colored material is used for the subsoil, the difference between the two layers appears most 

 striking; but the finer texture of the upper layer also renders it very distinct. The lower 

 terraces of this bank are still under cultivation. In the background a part of the megalithic 

 terraces can be seen. A ruined Inca house stands near the base of the precipice at the left. 



tivated lands were widened by building 

 new terraces along the beds of the 

 streams. 



In many cases the work was evidently 

 planned so that large immovable boulders 

 or outcrops of rock could be utilized in 

 the building of the walls instead of being 

 allowed to diminish the area of cultivated 

 land. We may believe that powder or 

 dynamite, to shatter refractory rocks, 

 would have been very highly appreciated 

 among the ancient Peruvians. 



STRAIGHTENING OF RIVER BANKS AND 

 STREAM BEDS 



It would be a mistake to suppose that 

 reclamation work in the bottoms of the 

 valleys was wholly or even principally of 

 the nature of improving irregular land by 

 terracing and leveling behind the walls. 

 A large part of the surface of the valley 



bottoms must have been altogether bare 

 of soil, as the unimproved portions still 

 are — mere wastes of loose stones brought 

 down by the torrential floods. 



The natural behavior of swift moun- 

 tain streams is to cut irregular channels 

 back and forth between the walls of their 

 valleys, but in the terraced valleys of 

 Peru it is the regular condition to find 

 the rivers and smaller streams confined 

 to channels of definite width, and some- 

 times kept in straight courses for several 

 miles at a stretch, as in the case of the 

 Urubamba River near Pisac, and again 

 below Ollantaytambo. In the latter in- 

 stance the river runs for nearly five miles 

 in a straight course, and, although the 

 ancient walls that were built to confine 

 the river have remained intact in only a 

 few places, the artificial nature of the 

 channel is obvious. 



497 



