498 



THE NATIONAL GEOGRAPHIC MAGAZINE 



A road that rims along the river has 

 utilized the top of a buried wall as a 

 pavement. These buried walls, which 

 occur also in other places, make it evident 

 that narrowing of the channel of the river 

 was accomplished by gradual stages. In 

 this case the area of cultivated land was 

 widened for about 12 feet toward the 

 river by building a new wall closer to the 

 river and filling in behind it. The old 

 wall was left in place, but buried in the 

 mass of the terrace and covered with 

 earth so that it did not interfere with the 

 cultivation of the land. Thus_ the land- 

 starved people gained an additional strip 

 of land, only a few feet wide, but several 

 miles long. The river may have furnished 

 the stones for the new wall, but the layer 

 of surface soil must have been brought 

 from a distance. 



THESE ANCIENT PERUVIANS BUILT AQUE- 

 DUCTS THAT ARE UNEQUALED 

 ELSEWHERE 



The ancient aqueducts of Peru have 

 challenged the most attention from for- 

 mer travelers, and they do not become 

 less worthy of admiration because they 

 are now seen to be only one feature of a 

 highly specialized agricultural system. 

 The construction of the irrigation chan- 

 nels was an enormous undertaking, per- 

 haps not equaled in any other part of the 

 world. From the aqueducts alone it is 

 evident that agriculture must have at- 

 tained a high development, which may 

 explain why the olhcr branches of the art 

 have received less consideration. 



Plow numerous and extensive the an- 

 cient aqueducts really were may never be 

 known, but the subject is worthy of much 

 more extensive study than it has received 

 thus far. If the accounts of earl)- Span- 

 ish writers are to be accepted, some of 

 these aqueducts were very long. Garci- 

 lasso wrote of one of them as 55 leagues 

 long and another 120 leagues, with a 

 depth of 12 feet. 



Where the soil was loose the channels 

 were paved with stones for many miles. 

 ( )n rocky slopes or precipices channels 

 were cut into the cliffs, and in some cases 

 tunnels of considerable length are said to 

 have been drilled. 



Where channels were being carried 

 along the sides of steep slopes, the usual 

 method of passing vertical surfaces or 

 overhanging rocks was to build up a wall 

 from below to the height of the channel. 

 In sheltered angles such walls remain in 

 place after the channels that ran along 

 the exposed slopes have entirely disap- 

 peared. 



WHERE SQUASHES ARE TIED 



The favorite courses for the ancient 

 irrigation channels, and by far the best 

 from an engineering' point of view, were 

 along the very crests of the ridges and 

 spurs of the mountains. In such places 

 the water-courses were cut, and now have 

 worn deep grooves. Thus there is no 

 danger of the channels washing away or 

 being filled up by drainage from above, 

 as with channels that run along the 

 slopes (see page 517). 



Cultivation was by no means confined 

 to the walled terraces that usually follow 

 the lower slopes of the mountains, but 

 was carried all the way up, on any slopes 

 that were not too steep to permit the ac- 

 cumulation of soil. It was not necessary 

 to build terraces to get rid of rocks on 

 slopes that are so steep that rocks roll 

 off. Slopes are even now cultivated 

 where squashes have to be staked or tied 

 to keep them from rolling down the 

 mountain, and where potatoes must be 

 picked instead of being shaken from the 

 vines. As these higher slopes are cool 

 and cloudy, there is much less need of 

 irrigation than in the valleys below. ' 



Though stone terraces were seldom 

 built on the high slopes, a system of nar- 

 row earth terraces or transverse ridges, 

 somewhat analogous to the contour farm- 

 ing in our Southeastern , States, was in 

 general use. Remains of such ridges 

 cover large areas of the higher slopes. 

 Usually there are a few large ridges at 

 intervals, with numerous smaller ridges 

 between. Many of these smaller ridges 

 can be seen from the valleys below only 

 when the light comes from a particular 

 angle, so as to cast shadows across them. 

 When lighted from in front the inequal- 

 ities are not shown, and nobody would 

 suspect that such steep slopes, now pro- 

 ducing only a sparse and scattered 



