Besults of Some Experiments it-^on the Bate of E vapor at ioii. 61 

 4. July 20 to November 22, 1896. 



Colour. 



Mean Temp, at XVII. 



Total Evap. 



Eed 



Yellow 



Green 







72-3 

 72-3 



71-5 

 72-1 

 72-2 



Inches. 

 30-52 

 30-53 



28-49 



Blue 



30-lU 



Purple 



30-95 



During the same period tlie 8-inch copper pan open to the skj^ 

 lost 30-69 inches. 



Considering the roughness of the tools some considerable dis- 

 crepancies in the sequence of the results could only be expected.. 

 Now at 5 p.m. air and water are cooling rapidly, so that the tem- 

 peratures given for that hour are of limited usefulness in comparison 

 with the total evaporation for the day. But in the first set we have- 

 some temperatures at 2 p.m., at which time air and water are both near 

 their maximum. Now the net result of the first set is approximately 

 a loss of about three-quarters of an inch in 41 days for an excess of 

 temperature of about 2^°. Also if we take the temperature 

 variation between the different colours to be about the same 

 throughout the year, we get a loss of about 2^ inches in 125 days for 

 the same excess of temperature. Or, taking ratios, a mean loss of 

 about 2^ inches per annum for each excess of 1° of temperature. 

 As between a copper 5-inch gauge standing in water, and therefore 

 in a relatively humid atmosphere, and another fully exposed to the 

 air, Latham's results give an increase of evaporation which may not 

 differ greatly from 9 inches per annum for each excess of mean 

 temperature of 2° ; or taking only the high temperatures of the 

 middle of the day about 9 inches per annum for an excess of maxi- 

 mum temperature of about b\°. Under all the circumstances this is 

 not greatly different from the numbers derived from the use of 

 coloured glass ; although it must be admitted that the true tempera- 

 ture effect is not obvious from Latham's work. 



During the progress of my experiment it was not expected that it 

 should do more than furnish experience for a better future test. It 

 seemed to be settled, at least, that it was possible to expose any 

 number of gauges in an absolutely identical manner excepting that 

 the one element of temperature could be easily varied at pleasure. 

 Accordingly some better vessels were procured, and some thicker 

 and better glass ; and the subject re-examined under improved con- 

 ditions, during the second half of 1897. In the following series of 



