256 Mr. H. Darwin. On the Small Vertical Movements 



of the decimals was not reliable. This error existed because the 

 horizontal movement of the trunnions along its bearings was not 

 strictty parallel to the surface of the top of the rods from which the 

 measurement was taken. As the readings from one rod only were used, 

 it would have been better if this lateral displacement had been impos- 

 sible. With care, however, consecutive measurements agreed within 

 0*01 mm., showing that the method was capable of far greater accuracy 

 than was required. 



During the experiment the stone sank more than the range of the 

 micrometer screw. The arm was unsoldered, moved upwards suffi- 

 ciently far to allow the screw to be used again, and was then re- 

 soldered. This operation, no doubt, introduced a small error. 



The curve marked " Movement of Stone " in fig. 2 represents the up 



Fig. 2. 



Feb. Mar Apl. May June July Aug Sept Oct. 



1380 



and down movements of the stone from February 19 to October 9, 

 1880, due to the varying dampness of the ground. 



The points corresponding to each observation are surrounded by a 

 small circle ; their vertical distance apart is the movement of the stone 

 magnified 8 times, each division of the scale representing l'mm. ; the 

 horizontal distance apart is proportional time. 



The following are the observations from which the curve is con- 

 structed. The numbers in the second column give the distance 

 moved downward by the stone from its position on February 19, 

 1880:— 



