of Edinburgh, Session 1885-86. 
953 
portional not to the difference between the temperatures of the wet 
bulb and the dew-point, but to the difference between the vapour 
pressures corresponding to these temperatures. 
The fact that as a whole the lines near each other in the diagram 
are nearly parallel, whatever their length, justifies their being drawn 
straight lines, and in addition there are at least six cases of observa- 
tions with the same temperature of wet bulb where the lines exactly 
coincide, although their length, i.e., the differences between wet and 
dry, are very different. 
It is of course to be observed that a very large number of the 
lines in the diagram are quite irregular in direction. This is 
obviously due to errors of observation, which are necessarily large 
both in the working of the wet and dry, and in the direct hygro- 
meter. On the whole, however, the irregularities are fairly well 
distributed on both sides of the mean direction of the lines, except 
in cases where the air was nearly saturated. Under these circum- 
stances the action of the wet bulb is extremely sluggish, and the 
actual dew-point is below that to be expected from the wet and dry. 
This is further shown by observations taken when, as occasionally 
happens, the wet bulb reads above the dry, as in dry fog. The dry 
bulb thermometer then acts as the more perfect wet bulb, the hygro- 
meter showing that the dew-point temperature is really below that 
of the air. This is of course apart from cases where a sudden 
change of temperature has just taken place, as the wet bulb covering 
prevents its adapting itself to changes as rapidly as the other. 
Sometimes the wet bulb thermometer continues to read above the 
dry for a considerable time, but I have invariably found that in 
these cases the so-called dry bulb was being kept wet, either by 
water dripping from the screen, or by driving fog depositing moisture 
on it, the air being all the time slightly dry. 
Comparing observations extending over a large range of temper- 
ature of wet bulb, we find that the lines in the diagram are not 
really parallel, but that they tend to become slightly steeper as the 
temperature of the wet bulb is lower. From this it appears that the 
factor by which we must multiply the difference of the temperatures 
of the wet and dry to obtain the difference between the vapour 
pressures at the temperatures of the wet bulb and of the dew-point, 
is not a constant, but tends to increase as the temperature falls. 
