272 UPPER AIR OBSERVATIONS. 



thread it was found to be broken and the instrument could not be found. The cause of 

 the breakin" of the silk after the instrument had been deit.ached from the balloon long 

 remained a mystery, and numerous expedients were employed to find either the cause or 

 a remedy. The opportunities for doing balloon work were very few, nothing could be done 

 in the dark months, and the intervals between the psriods of wind were so short and 

 occurred so irregularly that by taking advantage of every opportunity the problem was not 

 solved until th3 work was near its end. The cans 3 for the breaking of the silk was really 

 very simple. So long as the balloon was rising and drawing out the thread there was no 

 strain to break it. As soon however as the instrument fell it left a great bight of silk 

 in the air, the instrument after it had fallen on the ground anchored one end of this 

 bi^ht and soon there was enough silk lying on the ground at the other end to anchor 

 that end also. Thus while a great length of silk was still in the air and falling only 

 slowly both ends were fastened. If there was any wind at all the long length of thread 

 which remained in the air was subject to a considerable pressure which it was unable to 

 withstand, and it therefore broke. 



The msthod devised for minimising this difficulty was as follows: The instrument was 

 attached as before by a fuse, timed to release the instrument in about fifteen minutes. After 

 the silk thread had run out for about ten minutes it was broken near to the reel 

 and a small rubbsr balloon filled with air rapidly attached to the end. This was then 

 allowed to go free, and it was raised on the end of the thread until the instrument was 

 released about five minutes later. The length of thread used was reduced in this way by 

 one-third. After the instrument had fallen, for some time the balloon remained in the 

 air for it fell quite slowly, during this period it was free to move under the tension 

 of the thread and so allowed the thread to lay itself in the direction of the wind. After 

 the red balloon had been released its course was watched through field glasses, and as 

 the time was only short the chance of the glasses becoming fogged was much reduced. 

 The position in which the small balloon fell was noticed, and on going out to it the 

 thread was found and followed up to the instrument. When this method was used 

 the thread was never found broken, but it had only just been successfully developed 

 when the weather imiproved so much that the summer method described above could be 

 used. 



Tlie first method — theodolite used for watching the ascent and descent of the instrument 

 which had a small red balloon attached to mark its position — was used for five ascents, and 

 on each occasion the instrument was recovered. The second method — simple thread — was used 

 on twelve occasions. Five times it was successful, and the instrument was recovered 

 by following the thread ; six times the thread was found broken ; and on the remaining 

 occasion the thread led on to new thin ice and so could not be followed. The third 

 method— short thread with small red balloon used to mark the end — was employed four 

 times, on three occasions the instrument was recovered, while on the fourth occasion the 

 balloon moved in the direction of the sun so that it was lost sight of and not found 

 again. 



Thus out of twenty-one ascents, thirteen instruments were recovered at once, one instru- 

 ment was found later on the floe giving fourteen records in all. Two records were not 

 satisfactory, so that there remain twelve for discussion.* 



as 



e 



* The resiilts of tho observations .ire plotted on figure 13, page 42. 



