524 



On the whole, then, it is probable that the apparent raising of the potential at a given height in the air 

 by ilriftins; snow is real, and not a spurious effect due to the presence of the observer and his instrument. 

 An additional argument in favour of this conclusion is afforded by the fact that in many cases where 

 high potentials were observed the drift is recorded as being along the ground, in some cases entirely below 

 the level of the instrument. Again, there is at least one case of abnormally high potential where the 

 snow was observed to be drifting, but not in the immediate neighbourhood of the instrument (August 

 19, 1902, 4.30 p.m.). It is of course possible that even in these cases fine snow particles were being 

 driven against the observer and instrument. It is more probable, however, that the high air potentials 

 recorded are, in these cases at least, genuine, and the probability of the still higher potentials, observed 

 when the drift was more pronounced, really indicating high air potentials is increased. 



From December 3, 1902, till March 14, 1903, observations were taken with the collector at 15 feet 

 above the ground as well as at the usual height, and after that date at 12 feet. With the collector at 

 such heights the possibilitity of any charge on the observer's clothing influencing the potential taken up 

 by the collector may be considered as excluded ; the chance of electrification of the wire connecting the 

 collector to the electrometer is, on the other hand, increased. While these high-level collectors were in use 

 the number of occasions of drift at the time of the electrical observations was small (only six), but on all 

 except one of these abnormally high values were observed. It is probable, then, that the high readings 

 obtained with drifting snow really indicate high air potentials, but it has to be remembered that it is not 

 altogether impossible that they may be due to electrification of the collector or its connecting wire by 

 friction. 



Assuming the high readings accompanying drifting snow to indicate abnormally high air potentials, we 

 may perhaps most naturally conclude that the high potential is due to a positive charge acquired by the 

 ice particles as a consequence of their friction against the ground. In many cases in which this abnormally 

 high potential was observed the drift is described as being along the ground ; sometimes it is definitely 

 described as entirely below the level of the instrument. If the increased potential is entirely due to the 

 charge carried by the ice particles, then the increase in the vertical electrical field or potential gradient will 

 not extend above the level of the drifting snow, the vertical field within the drift-filled air (assumed to 

 form a uniform layer in contact with the ground) diminishing from the ground upwards to the upper 

 surface of the drift at a rate proportional to the density of the charge, the field above this level being 

 unaffected by the drift. Under the simple conditions assumed (which may not differ much from the 



d' 2 V 

 actual) we should everywhere have -rr^ = - iirp, V being the potential in electrostatic measure, and p the 



charge per cubic centimetre. 



One might hope to get some light thrown upon the nature and distribution of the charge carried by the 

 lower layers by a comparison of the results obtained at 15 feet with those of the ordinary 5-feet observations. 

 If the increased potential accompanying drifting snow were entirely due to charges below the 5-feet level, 

 then the electric force above this level would be unaffected, or, in other words, the difference of potential 

 between points at the 15- and 5-feet levels would be the same as under normal conditions. If, on the other 

 hand, the charge causing the disturbance was entirely above the higher level, the increase in the electric 

 force would be the same throughout the 15 feet (a horizontally stratified distribution being assumed) ; the 

 ratio of the potential observed at the two points would remain constant if the charge in the air below the 

 higher one were always negligible. 



The data for such a comparison are, unfortunately, very scanty, but, so far as they go, the results show 

 that with snow in the air accompanied by wind the difference of potential between the two levels was 

 increased, the relative increase being, however, not so great as that of the potential at the lower level, ?>., 

 the ratio of the potential at 15 feet to that at 5 feet was on the whole diminished. This would indicate a 

 positive charge neither entirely above the 1 5-feet level nor entirely below the 5-feet level. There is, 

 unfortunately, no case (where the data for both levels are available) in which the drifting snow is recorded 

 as being entirely below the 5-feet level. The results are, then, consistent with the view that the disturbance 

 is entirely due to a positive charge carried by the drifting snow particles, but the date are insufficient to 

 justify any more definite conclusion. 



