WATER AND VOLCANIC ACTIVITY — DAY AND SHEPHERD. 



281 



ing SO2 and SO3 may be a factor of considerable significance. The 

 first reading of the instrument might well be approximately correct, 

 but subsequent readings would surely all be subject to the effect of 

 uncertain amounts of SO^ and SO3 carried by the instrument in con- 

 sequence of the first exposure. This would have the eifect of render- 

 ing all the subsequent readings of the series quite valueless as a meas- 

 ure of the water content either of the air or of the cloud. 



In order to support the view that the atmosphere within the cloud 

 containing SO2 and SO3 is necessarily drier than air which does not 

 contain these substances, several measurements of dew points were 

 made b}^ us in an appropriate laboratory apparatus, of which the 

 results will be found in the table below. The first column contains 

 the dew point of air at varying degrees of saturation; the second 

 column the dew point of the same air to which 1 per cent of SO^ 

 (air -f- 1 per cent SO, is still respirable) has been added. All obser- 

 vations are in duplicate. 



Observations of dew points. 



Dew-point observations in the nature of the case can make no pre- 

 tensions to high accuracy, but the effect of charging the air with a 

 very small quantity of SO2 is shown most convincingly. The effect 

 of the addition of SO3 would have been still greater than that of 

 SO2, since it forms H2SO4, a notable dehydrating agent; but this 

 effect is somewhat more difficult to examine experimentally and so 

 Avas not undertaken; indeed, it was unnecessary, in view of the fact 

 that the point at issue is abundantly proved by the observations con- 

 tained in the table above. Brun has therefore proved no more with 

 his hygrometer measurements than that the great white cloud does 

 not consist entirely of water vapor, but it is not possible to estimate 

 the percentage of water contained in it from any figures based on 

 dew-point determinations under the conditions which he describes. 



From this evidence it appears clear that the observations of fact 

 noted by both Green and Brun may for the most part be precisely as 



