FUTURE WATER SUPPLY OF THE ADIRONDACK MOUNTAIN REGION. 467 



of crest of experimental weir, of 0.66 feet ; Lebros, witli limiting head of 0.80 feet 

 and length of crest, 0.66 feet ; Francis, limiting head, 1.62 feet and length of crest, 

 10 feet; Hamilton Smith, limiting head, 1. 73 feet and length of crest, 2.59 feet; 

 Fteley and Stearns, limiting head, 1.60 feet and length of crest, 19 feet; Bazin, 

 limiting head, 1.77 feet, length of crest, 6.56 feet; and the Cornell experiments, 

 made by the present writer, limiting liead, mostly somewhat over 5 feet* and length 

 of crest, 6.56 feet. It is thus seen that the quantity of water flowing in the Cornell 

 experiments was much larger than in any of the previous ones. For the higher 

 heads if was about 260 cubic feet per second. 



The writer wishes to be understood as not in any way criticisirig the work of pre- 

 vious experimenters with the weir, but he may properly point out that a very large 

 proportion of the previous work was designed on what may be termed a Liliputian 

 basis. Very great ingenuity has been displayed in measuring minute differences of 

 head, length of weir, etc., with no other result than in confusing the subject. 

 Indeed, at the present time, engineers, when speaking of weir measurements, appar- 

 ently measure water with a teacup. As a matter of universal practice this would 

 be perhaps allowable in a country where the streams are nearly all small ; but for 

 the great bulk of weir measurements to be made in the United States, where rivers 

 with catchment areas of several thousand square miles are common, water must be 

 rather measured by the bucket-full, and the work, instead of being Liliputian 

 becomes Titanic. Without pursuing the matter of flow over dams further we may 

 saj- that as a broad proposition, in the light of present information, where there is 

 a dam on a stream with tolerably uniform crest, the most accurate method of meas- 

 urement will be by gagings over such dam.f 



Gaging by current meter may also be referred to. This method has some advo- 

 cates who consider it superior to any other. Probably the reason for such an 

 opinion is that tests of a current meter in uniform masonrj' or concrete channels, 

 where the conditions are the most favorable possible to obtain, have shown fairly 

 accurate results; but in a shallow stream, flowing over a boulder bed, the conditions 

 are so different as to make rational comparison impossible. The current meter is, 

 indeed, only really useful when the following conditions obtain : 



1. A smooth, uniform channel for a considerable distance on either side of the 

 point selected. 



2. Considerable depth at the point selected and for several hundred feet to 

 either side. 



* The head varied in different experiments. I'or detail sec pa]>er On tlie Flow of Water Over Dams, 

 f On the point, see Water Supply and Irrigation papers of the U. S. Oeological Survey, No. 5O. 

 Methods of .Stream Measurement, p. 14. 



