Meteorological observations. 417 



protection of a celluloid cover. When cooled again, it would resume its 

 movement ; it was the transition from one temperature to another which 

 afïected it. Ultimately, when we had ceased to use it for purposes of 

 observation, Knud Rasmussen carried it as a pocket watch; he had, 

 however, continually occasion to note the same failing. 



After this, we used the better of the two pocket watches which 

 we had wdth us; on the outward journey however, I had the misfortune 

 to break the glass of this. I then carried it in a breast pocket, enclosed 

 in a celluloid case, the front of which, however, would now and again 

 be pressed in on to the hands when I turned over in my sleeping bag, 

 thus stopping the watch for the time being. 



The movement of the other pocket watch was too irregular to permit 

 of its being used for observation purposes. 



Our determinations of longitude on the inland ice are therefore 

 based for the most part on the distances covered, these being determined 

 by means of an odometer, marking tenths of a kilometre. I have both 

 on this and previous trips had occasion to note that the error involved 



— when travelling with sledges drawn by dogs, and according to compass, 



— by taking 2 kilometres' run as = one mile (= 1 minute) in a straight 

 Ипе, is practically imperceptible. The sure and steady driving, due to 

 Knud Rasmussen's and Uvdloriaq's unequalled skill in managing their 

 teams, together with the fact that we were able to check our calculations 

 on arrival at Danmarks Fjord, enabled us, despite the inadequacy of 

 our instruments, to avoid serious error in this respect. 



For survey work, we used both the odometer and also angles taken 

 to and from Hagen's Cairn at Cape Kronborg, Cape Glacier, and the 

 Cairn at Navy Cliff; the two last-named have both been determined 

 by Peary. 



The altitudes taken on the inland ice were reckoned exclusively 

 with the aneroid barometer. The expedition had originally been fur- 

 nished with two compensating holosteric barometers, having different 

 scale intervals for altitudes up to 2500 metres. One of these had, however, 

 previously proved so untrustworthy that we decided to leave it behind. 

 Such measurements are naturally always somewhat uncertain; they suf- 

 fice however, to give some idea of the rise and fall of the country, while 

 the absolute altitudes will as a rule be sufficiently accurate to permit 

 of reliable results being ultimately obtained by comparing with the 

 barometric pressure for the dates in question at nearest meteorological 

 station (Upernivik, North Star Bay). These calculations, however, I am 

 naturally unable to make at present owing to lack of the requisite material. 

 In cases where any alteration in barometric pressure occurred during 

 our stay at any one place, the altitude has been reckoned from the 

 mean. 



These meteorological observations, albeit incomplete, and not made 

 at the times of day agreed upon at the International Meteorolpgical 



