546 MR A. J. HERBERTSON ON 



5 - G to 31*6 grammes in weight, and the increase during an experiment varied from 

 G"0316 to 0*2026 gramme for the first tube. The average rate of aspirations was 

 O'J to 3*3 litres per minute. Here the average increase in weight of tube II., omitting 

 cases where the gain was obviously due to the exhaustion of tube I., is 4*3 milli- 

 grammes. At Fort- William this was still lower than at Ben Nevis. At Montpellier 

 the gain was much greater in tube II. ; the tubes not retaining their drying properties 

 nearly so long. 



In the Ben Nevis experiments an error of 1 milligramme in the weighing would 

 make an error of from 0'4 to 15*6 per cent, in the mass of water vapour measured, 

 in Fort- William experiments one from 0'4 to 0*9 per cent., and in those at Montpellier 

 an error of from 0'5 to 3-2 per cent. At Ben Nevis, however, the very dry conditions 

 which made the first tube gain so little that a milligramme would have produced a 

 large percentage error, were not common. 



In the final results at Ben Nevis and Fort- William, when expressed in grammes 

 of vapour per cubic metre of air aspirated, the percentage of error will be only about 

 one-fiftieth of the values given above, supposing the column of air aspirated is correctly 

 determined. 



The various measurements of the capacity of the aspirators did not differ by 

 0"01 litre, giving a possible error of 0'04 per cent, in the measurement of the 

 volume of water run out at Fort- William, and 0*03 per cent, at Ben Nevis. An 

 error of 1° C. in estimating the temperature of the air inside the aspirator at Fort- 

 William would lead to an error of less than 0"02 per cent, in the calculation of 

 amount of air. 



To study the results of the later determinations — 314 in number — a diagram was 

 constructed, the air temperatures being drawn along one axis, and the departure of the 

 wet bulb from the dry along the perpendicular one and the mass of vapour in each 

 cubic metre of the atmosphere written at each point thus fixed. The results are not 

 numerous enough to permit curves connecting points of similar value being drawn 

 with absolute precision. 



On the same diagram there was drawn a series of lines derived from Sprung's 

 formula : — 



reckoning p = 7 55 mm . The connection between the vapour pressures and the mass of 

 vapour in a cubic metre of the atmosphere was derived from the latest Smithsonian 

 Tables by finding the temperatures at which saturated vapour had a mass of 2, 4, 6 . . 

 grammes per cubic metre, and then the pressures of saturated water vapour at these 

 temperatures. 



The experimental results, whether they were obtained at high or at low levels, agree 

 with the general trend of these lines. 



