168 THE ANTARCTIC MANUAL. 
the level of the surface of the ice, and its temperature only went 
down to 7°5° C. 
“ All the temperatures were taken with a mercurial thermometer, 
which was whirled at the end of a string so that its velocity was 
about 6m. per second. It was not protected in any way, so that 
the temperatures observed with it are not free from a certain error 
due to radiation and reflection, although it was always shaded from 
the direct sun. These errors are not usually great with a whirled 
instrument, and most of my observations have to do with differences 
of temperatures observed with the same instrument and under 
similar circumstances. On the glacier the thermometer, when 
whirled, was not apparently affected by radiation or reflection from 
the ice, and only very slightly by that from the sun. On land I 
reinarked that the greatest disturbing effect is produced by sunlight 
reflected from grass. If the thermometer was whirled in the shade 
of a north wall with a grass field or hill-side close by, the thermo- 
meter would be immediately affected to the extent of one or two 
degrees, according as the sun shone on the grass or was obscured by 
acloud. The effect was immediate the moment the sun came out; 
sunlight reflected from rocks and light-coloured surfaces did not 
produce the same effect. 
“On the 19th August I returned to the glacier. At 11am. in the 
valley below the glacier I found the temperature of the air 22°C, 
and the wet bulb 12°°5, whence the vapour tension is 5:0 mm., and 
the relative humidity 26. In determining the temperature of the 
air by whirling the thermometer I found variations of as much as 2°. 
The hot puffs of air made themselves felt most markedly, and showed 
that the real variations of the temperature of the air were much 
greater than the thermometer showed. At 1 p.m., on the hill-side, 
to the west of the tongue of the glacier, and at a height of about 
2100 m. above the sea, four good observations of the temperature 
were made, giving 17°°5, 18°°0, 19°°5, and 19°-0; they are all 
equally trustworthy, and represent the average temperatures of the 
air during the minute, or minute and a half, that the thermometer 
was whirled. The mean of these values, 18°:5 is taken as the tem- 
perature of the air. For determining the temperature of the wet 
bulb, the bulb of the thermometer was wrapped round with one 
thickness of Swedish filtering paper thoroughly moistened, and the 
thermometer was whirled as before and until the temperature ceased 
to fall; it then stood at 9°°5. Still higher up the hill at an altitude 
of 2250 m., the temperature of the air at 2 p.m. was 18°°5C. Having 
returned to the spot where the observations had been made at 
