682 Colonel Edmond H. Hills [May 19, 



must be centred in prolongation of the vertical axis, otherwise the 

 capillarity between rod and liquid will introduce a force deflecting the 

 telescope from the true vertical. While it would thus appear that in 

 this form of instrument most of the familiar sources of error are 

 minimized, it is interesting to note the introduction of one possible 

 cause of error, quite unfamiliar to astronomers, namely, the deflection 

 that might be due to the attraction of the earth's horizontal magnetic 

 force upon the hanging part. If the telescope-tube were, as is 

 customary, made of iron or steel, this would reach a serious magnitude, 

 and even if a proportion only of the suspended w^eight were of iron a 

 perceptible deviation might result. It would, in fact, not be safe to 

 allow this proportion to exceed one-tenth of the whole weight and it 

 therefore seemed better to exclude the use of iron or steel altogether. 

 There is accordingly none, with the exception of the four thin flat 

 pendulum springs which form the gimbal suspension. 



In detaining you with these short descriptions of recently devised 

 instruments I may appear to have been wandering rather far from 

 my subject, the wanderings of the earth's pole. You will, however, 

 appreciate that in reality they follow very closely from it, being 

 instruments designed with the special object of solving the particular 

 problem we are discussing. 



We will now revert to the diagram of the observed polar motion, 

 and I will indicate how it is possible to analyse this so as to separate 

 the irregular movements from the more orderly fourteen-month pro- 

 cessional rotation. We are justified in assuming that this free preces- 

 sional period is constant in duration and therefore determines the 

 average rate of rotation of the pole of revolution. If, therefore, we take 

 a diagram of the polar movement, which will naturally have its axes of 

 reference fixed in relation to the earth, and convert it into another 

 diagram, showing the same movement, referred to axes rotating in 

 the earth at the average rate of the precessional rotation, we obtain a 

 graph of the irregular part of the polar path. If this irregular part 

 has any well-marked annual period, such period ought to be apparent 

 on inspection of the converted diagram. In the actual diagrams 

 obtained there seems little or no evidence of the existence of a yearly 

 term. 



We now take this second diagram and by the well-known process 

 construct its hodograph, the curve which gives us a measure of the 

 amount and direction of the force which could have caused the 

 movement recorded in diagram number two. This will still be 

 referred to the moving axes, so is not directly available for de- 

 ducing the true direction of these forces in the earth. Before we 

 can do this we must refer the diagram back again to axes fixed in 

 the earth. Thus, finally, we obtain our diagram number four, which 

 may be called the torque diagram, as it represents in direction and 

 relative magnitude the torque or twisting force which has been acting 

 upon the earth to produce the observed movement of the pole. 



