Marine Barometer fo'r Errors due to Swinging. 



171 



curve on the same diagram; the value of d was obtained by 

 direct measurement. The agreement is such as to provide 

 confirmation of our theoretical conclusions. 



Fig*. 2. — Variation of error with time of oscillation. 

 Full line is theoretical curve from equation (5). 



v\ 





Observed vgfues © 



1-7 1-8 1-9 2-0 2-1 



Time of oscillation in seconds. 



(V) A fitting was made for the barometer casing, which 

 enabled the instrument to be swung about various axes 

 parallel to one another along the tube, and method (ii.) was 

 tested by noting the corresponding displacements of the 

 mercury when the apparatus was swung precisely as in the 

 previous test. 



Fig. 3 gives in open circles the results of observation, and 

 it again appears that compensation is capable of realization, 

 the distance of the correct position of the axis from that for 

 which the instrument was designed being 21*5 cm. 



These observations also provided a check upon equation (1\ 

 which has been plotted in the same figure. For evaluating 

 this expression, d was determined for one position of the 

 axis ; if the axis was displaced x cm., d became d + a : 

 I was given by the usual equation for a compound pendulum 



P = 



+ P 



where 5 is the distance of the centre of mass ot the 



whole suspended system below the point of support and k is 



