132 ^V1ND. 



found that by accepting it as an indication of a 8.W. wind, tlie number of .S.W. winds 

 became intermediate between the number of south and west -winds, while if it were neglected 

 the number of S.W. winds would be negligible. 



The motion of the clouds also gives valuable indications of the air movements in the 

 upper atmosphere. A Besson nephoscope was taken as part of the meteorological outfit, but 

 after being set up it was not used ; for the very simple reason that it was found prac-tically 

 impossible to stand exposed to the wind, when the temperature was low, sufficiently long to 

 determine with certainty and accuracy the movements of the clouds when they were moving 

 at anything but a rapid rate. It w-as at once realised that for reliable work it was absolutely 

 necessary that the observer should be sheltered when making cloud observations. A long focus 

 lens was therefore fixed iu the roof of the absolute magnetic hut, which was made dark with 

 shutters. This lens cast an image of the clouds near to the zenith on to a horizontal board 

 carried on a vertical axis so that it could turn about its centre and had parallel lines drawn 

 on its surface. The images of the clouds were watched and the board turned until successive 

 clouds appeared to travel along the lines. An index attached to the axis of the board then gave 

 the direction of motion of the clouds. By aid of this apparatus it was jjossible to determine 

 the motion of the higher fine cirrus clouds, the motion of which was exceedingly slow and 

 certainly could not have been measured by an observer exposed to the cold and wind. The 

 importance of some such arrangement is c'ear when it is realised that 40 jJer cent, of the 

 cloud observations were made in \\inds of over 20 miles an hour. As this nephoscope could 

 not be used in the dark very few observations were taken in the winter months, and I 

 regret to say that after I left the ^Viitarctic in March 1912 the cloud motion obser\-ations 

 were discontinued , so that the number of observations is not very large. 



It was practically impossible to determine the motion of the clouds accompanying blizzards 

 owing to the want of detail in the cloud mass (see page 148) and as these formed practi- 

 cally the whole of the low clouds I did not attempt to record the motion of the low clouds 

 at all. The clouds whose motions were measured were practically all either cirrus, alto- 

 stratus or alto-cumulus. 



No direct measurement of the height of the clouds was attempted, but there can be 

 little doubt that the cirrus clouds were well above the summit of Jlount Erebus, while the 

 alto-stratus or alto-cumulus were lower than the summit as they frequently hid from view 

 the smoke cloud. In the following discussion all the observations of cirrus clouds have been 

 grouped together and called high clouds, while the alto-stratus and alto-cumulus have been 

 grouped together and called medium clouds. The cloud and smoke observations thus give 

 information of the air motion at three heights which may ap2>roximately be taken to represent 

 the atmosphere at heights of 10,000—13,000 feet, 13,000—16,000 feet and above 16,000 feet. 



The motion of Erebus smoke and of the clouds could only be observed when they were 

 visible. It often happened that the motion could be recorded every time observations were 

 made on one day and only tnice or twice on another day. If during the whole of the 

 first day the motion remained constantly from say the north and on the second day con- 

 stantly from the south there would be six entries of the iiorth motion and only one or 

 two of south motion, although the south motion continued just as long as the north motion. 

 In order to reduce as much as jjossible this source of error whenever a sequence of the same 

 direction was recorded on any one dav it was only entered once in the reduction of the data. 

 Thus 544 observations of Erebus smoke were reduced to 364, 166 of medium cloud to 133, 

 73 of hi,gh cloud to 67. The same procedure was followed in reducing the observations of 

 cloud motion at Cape Adare ; so that out of 405 observations of the direction of low clouds 

 at Cape Adare only 229 were used and out of 45 of high cloud only 41 were used. 



