56 0. E. SCHI0TZ. [NORW. POL. EXP. 



At Khabarova, where the observations as a whole were not so successful, 

 the period for pendulum 34, as has been already mentioned, was found too 

 small, on account of the rapid rise in the temperature during the experiments. 

 Even with the highest value, which is the only one retained above, the dif- 

 ference in question is only 2356 X 10 ~ 7 . For this reason, I have thought it 

 best to give the period of oscillation for pendulum 34, half the weight of that 

 for pendulum 33. If we suppose, moreover, that the difference between the 

 periods of oscillation is 2447 X 10 ~ ', we find the mean period of oscillation 

 of the two pendulums to be 0'5058721, the value given in the table. 



What distinguishes the above observations of the Fram expedition from 

 others that have hitherto been made, is that they have been made upon the 

 open sea, over depths of water of more than 2000 metres. They were rendered 

 possible by the fact that the vessel was frozen into the sea-ice, and drifted 

 with it. The great pressure, however, to which the ship was exposed, even 

 in the middle of winter, shows that this mass of ice cannot be regarded 

 at all times as one coherent layer, drifting along with one motion for the 

 whole. The various parts of the ice-covering may be moving at variance 

 with one another, and this movement may be carried so far as to cause the 

 ice-covering to burst at such places where the compression or distention be- 

 comes too great. If this be the case, it is to be feared that even if the ice 

 is apparently motionless, there may be imperceptible movements and trem- 

 blings in the covering that may affect the oscillations of the pendulum. 

 The influence that the motion of the ice may have upon the pendulum's 

 period of oscillation ought therefore to be more carefully investigated. It is 

 easy to show that a simple motion of translation, even if not uniform, but 

 uniformly accelerated or retarded, will produce no change in the period of 

 oscillation; the motion need not, I presume, be imagined to be more com- 

 plicated than this in the comparatively short time that each observation lasts, if 

 it is possible at all to consider the motion of the ice to be regular. The case 

 would be different with the irregular movements and tremors of the ice. As 

 I have attempted to demonstrate in a previous work 1 , these tremors of the 



1 0. E. SCHIOTZ, Resultate der im Sommer 1893 in dent ndrdlichsten Theile Norweyens 

 ausgefMirten PendelbeobacMungen nebst einer Untersuchung fiber den Einfluss von 

 Bodenerschutterungen auf die Schwingunyszeit (Kristiania, J. Dybwad 1894.) p. 15 

 et seq. 



