204 
Proceedings of the Royal Society 
probably due to the thermometers being freely exposed underneath 
— evaporation and condensation of moisture from the grass neutral- 
ising the various changes. 
It is intended to compare the dew points given by these screens 
during the hourly observations with those given by a direct hygro- 
meter; the form to be used being that designed by Professor 
Chrystal. This consists of a copper box, nickel-plated, into which 
the bulb of a thermometer is inserted. By means of a double-tap 
arrangement communication can be kept up with two reservoirs, 
one of which is filled with warm water, and the other with water 
surrounded with ice or a mixture of snow and salt. The tempera- 
ture of the water flowing through the box can thus be adjusted 
until a film of moisture has just begun to form on the plated 
surface, and the reading of the thermometer at once gives the dew 
point. With this instrument results differing from each other by 
not more than 0 o, 3 can easily be obtained. 
A few experiments have been made with this instrument, but in 
circumstances under which the dew points given by the wet and dry 
bulb hygrometers approached each other more closely than usual, 
although still maintaining their usual order. The results are some- 
what different from what was expected, but are perfectly consistent 
throughout. The direct hygrometer invariably gives the highest 
dew point, being always a little above the improved Stevenson. 
The mean of fifteen observations shows the dew points given by the 
improved Stevenson to be about 0 o, 7 lower than the direct hygro- 
meter — those by the ordinary Stevenson about 1° — and those by 
the fan apparatus about 1°*5. These show a much closer agreement 
throughout than could have been expected ; and the fact that the 
direct hygrometer is always highest tends to confirm the result 
arrived at by Mr Shaw of Cambridge, that the film of moisture 
begins to deposit on the silvered surface before the actual dew 
point has been reached. 
