158 



CLOUD AND PKECIPITATION. 



{!)) The period during which the sky is too dark for clouds to be seen by the aid of 

 the sunliglit. AVe may take this period as lasting during the time the sun does 

 not rise within 10° of the horizon, i.e., in McMurdo Sound the period from June 

 5 to July 8. Thus the month of June is the only montli which is included 

 in this period. 



(c) The remaining period, during which there is daylight during part of the twenty-four 

 hours and darkness during the remainder, i.e., in McMurdo Sound the periods 

 from March 14 to June 5, and from July 8 to September 30. Thus April and 

 May, July, August and September may be considered as belonging to this period. 



Now during the periods (a) and (b) daylight cannot affect to any appreciable extent the 

 observations of cloud at any time of the day, while during the period (c) it affects the 

 observations to its maximum extent. 



In figure 53 the amplitude of the daily variation of cloud has been plotted for each 

 month, and we see that during the period (((), October to March, the amphtude is very small 



in five of the seven months. 

 Also in the period (6), June, 

 the amplitude is very small. 

 On the other hand each of the 

 two periods into which (c) is 

 divided has a month of maximum 

 amplitude. This alone would make 

 us strongly suspect that the day- 

 light plays a largo part in deter- 

 mining the apparent daily varia- 

 tion of cloud amount ; but when 

 we find that almost exactly the 

 same variation of the amplitude 

 during the year is shown by the 

 observations made on the Fram 

 drift when the variation of day- 

 light were nearly the same, the 

 suspicion becomes almost a cer- 

 tainty. Curve (6) of figure 53 

 shows the amplitude of the daily 

 variation of the cloud for each 

 month as observed on the Fram, 

 the months being arranged so that 

 the same seasons in the north and 

 south correspond. The similarity 

 of the curves for McMurdo Sound 

 and the Fram drift is strilving. In 

 the north the months of constant 

 daylight, August and September, 

 have very small amplitudes and 

 also the winter months November * and December. Between these two minima there are large 

 maxima during the period when each day has a period of daylight and total darkness. 



It is therefore very questionable whether the curves in any month except those of 

 complete daylight and complete darkness give any information about the yearly variation of 



Fig. 53> Amplitude of daily variation of cloud. 



