150 GEOLOGICAL SURVEY OF NEW JERSEY. 



FORESTED CATCHMENTS. 



Merrimac, 40 19.4 20.3 0.9 5 



Connecticut, 53 20.9 21.5 0.6 3 



Upper Hudson, 80 21.2 19.1 +2.1 +10 



Hemlock Lake, Large. 14.7 17.4 2.7 18 



Hackensack, 60 246 25.2 0.6 2 



Passaic, 58 23.4 24.7 0.7 3 



Pequannock, ......,. 78 241 24.2 o.i o 



Ramapo 75 23.9 24.3 0.4 2 



Batsto, c ... 88 27.9 28.0 o.i o 



Great Egg Harbor 88 28.9 29.1 0.2 i 



Potomac, 52 24.8 289 4. i 17 



Savannah . 65 31.8 35.7 3.9 13 



Clear Creek, 129 13.1 02 2 



Excepting for the Savannah, the mean temperature is used in 

 all cases, but for that stream the summer temperature is 

 decidedly lower, as compared with the mean for the year, than 

 for other streams, consequently as the summer is the season of 

 maximum evaporation, the evaporation is estimated on the basis 

 of the summer temperature. 



The fifth column of this table gives, in inches, the amount by 

 which the measured evaporation is greater or less than the 

 evaporation computed from rain-fall and temperature, and the 

 last column shows the percentage of the measured evaporation 

 which this discrepancy represents. It is not usually to be 

 expected that the measurements are correct within five per 

 cent., considering possible errors in both rain-fall and stream 

 measurements ; therefore, it is significant that 17 streams having 

 an average length of record of 12 years, differ from the results 

 given by the formula less than 5 per cent., the mean being but 

 \y 2 per cent., while three more, with average records of n 

 years, vary from 6 to 8 per cent., the algebraic mean being 3 

 per cent., and the remaining 14 streams, which differ more 

 widely, have an average length of record of 6.5 years only, 

 the average difference for these being 5 per cent. As has been 

 previously intimated, these larger differences would, in many 

 cases, be reduced by a more detailed computation, which would 

 take into account the distribution of rain-fall by seasons. The 

 burden of evidence of these stream gaugings, therefore, is 



