204 Proceedings of the Eoyal Society of Edinburgh. [Sess. 
5 : 
With a different wind direction the 
smoke from the forge was rising rapidly 
out of the valley and drifting over a hill. 
At the top of the hill the smoke was still 
mostly going right over our heads,** and 
the readings in forty minutes were from 
140 to 220, while 50 yards down the slope 
during thirty minutes the variations were 
from 75 to 130. 
Curve B was taken near Garforth 
colliery, seven miles east of Leeds. It 
shows readings taken at distances of 100, 
350, and 880 yards from a tall chimney 
when the wind was blowing smoke towards 
the instruments. About two hours after 
the last reading was made the instruments 
were taken to a position about half a mile 
to the windward of the chimney, and 
during twenty minutes the reading was 
never above 135. 
It will be seen that fresh smoke reach- 
ing the collector caused an increase in the 
positive potential gradient. We commonly 
got readings of over 800 volts at distances 
over a mile from large chimneys. The 
interpretation of these measurements is 
slightly complicated owing to the ordinary 
variations in the potential gradient due to 
other causes. It is consequently more con- 
venient to study the effect of smoke by 
means of passing trains, as in that case the 
smoke effect is limited to a definite interval 
of time, as will be seen from the following 
curve. 
Curve C . — A slight wind was blowing 
from a railway 300 to 400 yards away 
(wind direction roughly at right angles 
to the railway). The ground level was 
below the level of the railway, and smoke 
from trains was wafted slowly down to 
