NO. 8.] SUPPLEMENT. 89 



so that the plane of oscillation was parallel with the longitudinal direction of 

 the beam, and further, the solid construction of the ship, I take it that the 

 simultaneous movement of the substratum may be assumed to have been 



trifling, and mainly confined to oscillations of the pendulum apparatus itself. 

 It is true that Scott-Hansen remarks that the level moved when he ap- 

 proached the apparatus from various sides, but only very slightly; and he 

 also states that from the coincidence apparatus, he could approach sufficiently 

 near to observe the level, and see that it did not move at his approach. 1 



It seems to me that the correctness of the above assumption is corrobo- 

 rated by the observation made on the ice on June 8th, 1895, for this, as the 

 table on page 60 shows, leads to a result exactly similar to that of the 

 observations made in the saloon, if we except those of January 16th, 1894 

 and 1896. It is true that on the day in question (June 8th, 1895) only one 

 observation was made with the one pendulum 34; but nevertheless it seems 

 unlikely that the accordance should be due to chance. It may also be men- 

 tioned that among the observations made on the 10th June following, out 

 on the ice, there is one with pendulum 33 2 , that gives far too large a period 



1 On page 41 an error has slipped in, where, instead of "two men . . walked . . across 

 the floor, causing a movement in the level as they passed near the apparatus", it should 

 read " . . across the floor. Scott-Hansen states that movement is perceptible in the 

 level when he crosses the floor in the neighbourhood of the apparatus". 



" On page 54, the period is wrongly given as 0'5056139. It should be 05056189, as 

 calculated on page 35. 



