PRECIPITATION. I59 



cloud, and in these months the amplitudes are so small that no reliance can be placed on 

 them. In fact the summer variation based on the Hut Point observations is almost the 

 exact reverse of that based on the Cape Evans observations, showing that the small 

 amplitude is mainly the result of chance. 



We are therefore reluctantly compelled to conclude that the daily variation of cloud 

 amount teaches nothing reliable as to the true changes in the cloud amount durino- the day 

 except that it is very small, so small in fact that during the gi-eater part of the year it 

 is masked by the subjective effect of the varying amount of daylio-ht. 



Precipitation. 



No rain fell during the whole of our stay at Cape Evans, and up to the present there 

 is no record of rain at either Cape Adare or in McMurdo Sound. On the other hand, a well- 

 developed rainbow was seen to the north-north-east of Cape Evans on February 14 1911. 



Every one who has taken meteorological observations in the Antarctic has been faced 

 with the difficulty of finding some way to record the amount of snowfall. Up to the present 

 this problem has not been solved. The difficulty is that practically all the snowfall occurs 

 during high winds. The snow is carried along in the air and does not settle except in the 

 lee of any obstacle to the winds motion. Any ordinary snowgauge would become full of drift 

 snow in a very short time, while the surrounding ground might not have received any increase 

 to Its snow-covenng. Also it very frequently happens that with a perfectly clear sky from 

 which no snow is falling the wind raises clouds of loose snow from the ground which enters 

 the gauge and so gives a false record. I had many ideas as to measuring the effective 

 snowfall on my arrival in the Antarctic, but none of them proved of any practical use. 



It was no use trying to measure the depth of snow on the ground, nor to keep a space 

 clear and measure the snow on it each morning. For no snow accumulated on ground subject 

 to the full force of the wind, while in sheltered places the snow would accumulate to a depth 

 determined entirely by the size and shape of the obstacle, and be entirely unrelated to the 

 quantity of precipitation. 



My experience is that it is totally impossible to form even an approximate idea of the 

 quantity of snow precipitated from the atmosphere. Even on a perfectly level plain the snow 

 accumulation would be no gauge of the precipitation, for it would receive snow which had 

 fallen miles away and been driven forward by the wind. 



All that is possible in this section is to give a general statement of the snow-covering 

 existing around the station at different times of the year, and to give some idea of the 

 frequence with which snow fell. 



When we arrived at Cape Evans, the small promontory on which the station was built 

 was— as future experience showed— remarkably free from snow. The beach on which the hut 

 was built had no snow-covering at all, so that there were several hundred square yards 

 of black volcanic debris fully exposed to the warm sun. When the expedition left twenty-four 

 months later, the beach was buried in several feet of snow and the hut itself was almost 

 invisible under a great snow-drift which had formed around it. 



After our arrival in January, 1911, there was a slow but steady diminution in the snow 

 and ice accumulations on the Cape, so that the size of the small glacierettes obviously de- 

 creased. This was due mainly to ablation, for it continued throughout the first winter when 

 thawing was quite imposible. A cave had been dug for the magnetic instruments in a small 

 glacierette— in reality the permanent drift which had consolidated into firm ice in the lee of 



