94 
tected, was, taking the mean of 16 days, 1*28 Fahr. warmer 
than the lower one, sometimes the difference being very 
considerable. The lower thermometer with the bulb below 
the level of the screen gave mean readings 04 Fahr. colder 
than my screen, while the one in the middle of the metal 
screen gave 0 9 Fahr. warmer than in my wooden one. 
I also for comparison frequently hung my alcohol mini- 
mum thermometer on the north side of my wooden screen 
so that it was always in the shade, and this gave on an 
average for 25 days a temperature of 1*7 Fahr. warmer than 
inside the box. I consider that much of the difference must 
be attributed to heat reflected from the snow. On one 
occasion when it was 5*7 Fahr. higher than inside I swung 
it for a few minutes, by which it was reduced until the two 
closely approximated. 
The screen which I used was large and specially adapted 
for the evaporation experiments I proposed making. The 
inner one was the box I used last year, 90 cm. wide by 75 
cm. by 60 cm. high (see loc. cit. p. 162), but as I unfortu- 
nately had no pavilion shade this year, I had to cover it 
with another louvre screen and left a space of about 25 
centimetres all round between these two louvre boxes. I 
still think that this construction is not unsuitable for such 
a climate as this, although with so much wood the thermo- 
meters wi]l not indicate changes quite as rapidly as should 
be the case, and further, 1 am very doubtful about the 
Stevensen screen being suitable, but as the question of 
screening is so important and as the conditions are so 
different from those which obtain in England, I purpose 
making some exact comparisons of different methods, as 
perhaps some of the figures showing most variation may 
have arisen from the screening being only of a provisional 
character, and I do not think that the Swiss metal screen 
ought to have a board at such a short distance below the 
