952 Proceedings of the Royal Society 
caused by incessant small fluctuations of the actual dew-point. If 
the temperature of the hygrometer is kept constant for a time near 
the dew-point, evaporation and condensation take place on the 
surface in quick succession, and the dew-point may vary within a 
considerable range in a very short time. This obviously increases 
the difficulty of making accurate comparative observations. 
The hygrometer, with it reservoirs, tubes, &c., was mounted on a 
stand about six feet away from the thermometer screen, and the 
thermometers were read immediately after checking the rate of 
cooling of the hygrometer, as described. 
Considerable difficulty was at first experienced in finding a suit- 
able method of comparing the observations of dew-points by the 
direct hygrometer with the readings of wet and dry bulbs. After 
several trials, the following graphic method was found to be 
satisfactory. 
Taking rectangular coordinates, temperatures were measured along 
the x axis, and vapour pressures in inches of mercury (taken from 
Regnault’s Tables, as given in “ Glaisher’s Hygrometrical Tables,” 
6th edition, 1876) along the y axis. Each observation with wet and 
dry, and the corresponding hygrometer observation, was then treated 
separately. First a point was taken whose abscissa corresponded 
to the temperature of the wet bulb and ordinate to the vapour 
pressure at that temperature. Another point was then taken with 
abscissa corresponding to the temperature of the dry bulb and 
ordinate to the vapour pressure at the temperature of the dew-point, 
as given by the direct hygrometer. These two points were then 
joined by a straight line. 
Plotting all the observations we get a diagram formed of straight 
lines, and by taking the general trend of these lines we are able to 
see the relations existing between the wet and dry bulb readings 
and the dew-point obtained from direct observation. 
First we note that on the whole the lines near each other are 
almost parallel, from which it at once appears that with a given 
reading of the wet bulb the excess of the air temperature above that 
of the wet bulb is not proportional to the excess of the temperature 
of the wet bulb over that of the dew-point, but to the vapour 
pressure corresponding to those temperatures. That is to say, given 
the temperatures of the dry and wet bulbs, their difference is pro- 
