No. 5« 1915- ROALD Amundsen's ANTARCTIC expedition: METEOROLOGY. 57 



The observations gave 



+ II m. + 28 m. 



DifF. o m. 5 m. 



The differences between the judged and computed values of ))i are 

 presumably caused by the following circumstances. 



The pairs of observations from which the differences of height are compu- 

 ted are not synchronous. During the march from one station to the next 

 the pressure varies at both stations. 



The observations are made at two different places not in the same 

 vertical. 



These causes seem to eliminate each other perfectly on the Barrier, 

 but not on the Plateau where the road was steeper and more rugged. 



The probable error attaching to a computed height increases with the 

 distances from the original station, the initial errors being multiplied in 

 the process of calculation. Each step carries with it the error of the fore- 

 going step, and adds to this its own error. Taking in this way the square 

 root of the sum of the squares of the two we have, when dh is the error 

 of one step, dH the error for the station and H the height. 



The preliminary heights which Capt. Amundsen gives in his book »The 

 South Pole« are greater than mine. They were computed from uncorrec- 

 ted observations of pressure and a higher pressure at sea-level than that 

 generally existing in the Antarctic. 



