HYDROLOGY OF NEW YORK 



335 



AT hen the measurements were first begun, it was considered 

 that the formula Q=1142 was best suited to the form of the 

 dam, but after more careful consideration it was apparent that 

 the results given by this formula were somewhat in excess of 

 the actual discharge, specially for the low-water flows. The 

 computed discharges, as shown by columns (3) and (4) of the pre- 

 ceding tabulation, are somewhat irregular. This result is due to 

 the disturbing effect of the irregular sections of the crest, the 

 highest point of which was 2 feet above the lowest. 



Column (5) shows the percentage variations between the dis- 

 charges as determined by a sharp-crested weir, up to 5200 cubic 

 feet per second, and the discharges computed by the formula. 

 For discharges beyond 5000 cubic feet per second the original 

 determination has been used. An extension of the plotted curves 

 shows that some little distance above 5000 cubic feet per second 

 discharge, the results of the two methods are substantially the 

 same. The two curves crossed at the point of about 6000 cubic 

 feet per second discharge. For discharges above 10,000 or 

 15,000 cubic feet per second there is probably an error in the 

 results of from 5 to 10 per cent. Below 5000 cubic feet per 

 second it is believed that the results are now accurate within a 

 few per cent. Francis's formula, Q=3.33 L H% has been used 

 for the weir computations. 



The measurements taken previously to the construction of the 

 weir and the rating of the dam, have all been corrected to con- 

 form to the new determinations; hence all the data of the 

 Genesee measurements of this table may be considered as 

 accurate within the limits stated. 



The original Genesee river data show for a portion of the 

 range more error than is consistent with good work, and which 

 remained inexplicable until the experiments at Cornell Univer- 

 sity were carried out. These experiments showed that for high 

 heads the tendency was to neutralize the differences at lower 

 heads. The flow of all the weirs, in short, became sensibly uni- 

 form at from 5 to 6 feet depth and at 10 feet depth, there will 

 be very little difference. 1 



lTen feet depth on the crest is not unusual in flood flows. In order to 

 measure such flows, a weir formula should be worked out to at least 10 feet 

 depth on the crest Such a formula will npply without more than 15 per 

 cent to 20 per cent error to almost any form of crest. 



