466 RESTORATION OF THE WATER. 
erounds, and thence in part escapes again into the bed of the 
river, by similar conduits, or in the form of springs and rivulets. 
Interesting observations have lately been made on this subject 
in France and important practical results arrived at. It was 
maintained that mountain irrigation is not ultimately injurious 
to that of the plains below, because lands liberally watered in the 
spring, when the supply is abundant, act as reservoirs, storing 
up by absorption water which afterwards filters down to lower 
grounds or escapes into the channel of the river and keeps up 
its current in the dry summer months, so as to compensate for 
what, during those months, is withdrawn from it for upland 
irrigation. Careful investigation showed that though this pro- 
position is not universally true, it is so in many cases, and there 
can be no doubt that the loss in the volume of rivers by the ab- 
straction of water for irrigation is very considerably less than 
the measure of the quantity withdrawn.* 
* See VIGAN, Etude sur les Irrigations, Paris, 1867 ; and Scorr MoNCRIEFF, 
Irrigation in Southern Hurope, pp. 89, 90. 
he brook Ain Musa, which runs through the ruined city of Petra and final- 
ly disappears in the sands of Wadi el Araba, is a considerable stream in winter, 
and the inhabitants of that town were obliged to excavate a tunnel through 
the rock near the right bank, just above the upper entrance of the narrow Sik, 
to discharge a part of its swollen current. The sagacity of Dr. Robinson 
detected the necessity of this measure, though the tunnel, the mouth of 
which was hidden by brushwood, was not discovered till some time after his 
visit. I even noticed, near the arch that crosses the Sik, unequivocal remains 
‘of a sluice by which the water was diverted to the tunnel. Immense labor 
was also expended in widening the natural channel at several points below the 
town, to prevent the damming up and setting back of the water—a fact I be- 
lieve not hitherto noticed by travellers. 
The Fellahheen above Petra still employ the waters of Ain Musa for irriga- 
4ion, and in summer the superficial current is wholly diverted from its natural 
channel for that purpose. At this season, the bed of the brook, which is com- 
posed of pebbles, gravel, and sand, is dry in the Sik and through the town ; 
but the infiltration is such that water is generally found by digging to a small 
depth in the channel. Observing these facts ina visit to Petra in the sum- 
mer, I was curious to know whether the subterranean waters escaped again 
to daylight, and I followed the ravine below the town for a long distance. Not 
very far from the upper entrance of the ravine, arborescent vegetation ap- 
peared upon its bottom, and as soon as the ground was well shaded, a thread 
