NO. 8.] PENDULUM OBSERVATIONS. 57 



substratum will always cause a diminution in the pendulum's period of oscillation. 

 It is there stated that I have observed this actually to be the case in 

 experiments carried out in the Christiania Observatory 1 ; and I have also 

 since had an opportunity of making a similar observation in Trondhjem. 

 From these it is clear that the experiments made on the Fram when she 

 was drifting with the ice, can never give too low values for the acceleration; 

 it would be more reasonable to expect the values to be too high, if the ice- 

 covering was in irregular, trembling motion during the observations. 



In order to obtain some idea of the motion of the ice, I have made an 

 examination as to the condition of the wind on the days of observation, and 

 immediately before them, as the motion will depend chiefly upon the wind. I 

 find the following statements: 



January, 1894, slight wind on the 15th and 16th. 



March, 1894, calm on the 15th, but rather windy on the 16th, 



continuing so during the night. 



June, 1895, rather strong wind on the 7th and the morning of 



the 8th, the rest of the latter day being calm; strong 



ESE wind on the 9th; blowing fresh on the 10th, 



but less on the llth. 



November, 1895, no wind from the 14th to the 24th; had been a 



breeze on the llth. 



January, 1896, calm all the month, no screwing and little wind. 

 April, 1896, little wind from the 26th to the 30th, no screwing. 



It will been seen from this that the masses of ice ought to have been 

 in the greatest state of disturbance in June, 1895. If we examine the drift, 

 we find that the ship has drifted as much as 19'9 km. in the two civil days 

 from the 8th to the 10th June, and that the distance drifted diminishes to 

 4'8 km. in the following 24 hours (10th and llth June). If we now consider 

 the above given values for the periods of oscillation, we find that the lowest 

 values are found just on the 10th and llth June, 1895, although the latitude 

 is more than 1 lower than that reached on the 14th and 23rd November, 

 1895. That this cannot be due to a greater local value of gravity is ap- 

 parent from the fact that the mean period of oscillation observed on the llth 



i 1. c. p. 8. 



