170 Prof. Duffield and Mr. Littlewood on Correction of a 



appears possible to obtain compensation by arranging 

 that l = 2d, and it is satisfactory .to find that an ordinary 

 marine barometer lends itself to this adjustment. A feature 

 which is also favourable is the independence, as regards first 

 order quantities, of this adjustment upon the value of « ; if 

 once made, it should hold good for all small amplitudes of 

 swing. 



In practice, there are two means whereby the necessary 

 adjustment may be arrived at : — 



(i.) By keeping the point of suspension fixed, when d will 

 be unchanged, and altering I, the length of the equivalent 

 simple pendulum, until I — 2d. This may be effected by 

 loading the barometer to alter its period of oscillation. 



(ii.) By altering the position of the point of suspension, 

 moving it up or down the barometer tube. Both I and d arc 

 thus varied until the required relationship is attained. 



(6) Marine Barometer No. 1408, kindly lent by the 

 Meteorological Office, was submitted to a series of tests. 



(a) When swung about its original axis for 10 minutes 

 through an angle of about 16° (« = 8°) the mercury rose 

 through 2'6 mb. divisions, indicating that the deviation from 

 the vertical introduced a + ve displacement larger than the 

 — Te displacement due to the centrifugal force occasioned by 

 the swinging. With other amplitudes of swing the displace- 

 ments were found to be proportional to their square in 

 accordance with equations (4) and (5). 



(6) Method (i.) for effecting the adjustment was tested. 

 For convenience in obtaining a suitable range of periods of 

 vibration, the axis was raised 23*7 cm. above its usual 

 position and kept there during the series of experiments ; a 

 board to carry weights was fixed to the top of the instrument 

 and the displacement of mercury noted for various periods 

 of vibration. The amplitude, u , was maintained for ten 

 minutes at approximately 8° by gently pushing the tube 

 with the finger as it passed through the vertical position ; 

 at the end of this time the barometer was brought to rest 

 and the displacement noted. 



The experimental results are shown by open circles in 

 fig. 2, the ordinates being the rise or fall of the mercury 

 in mb. divisions and abscissae the corresponding periods 

 of vibration. It appears at once that there is some period 

 for which there will be no displacement. The periods plotted 

 are those of the forced vibration. 



The observed values maybe compared with the values of y 

 given by equation (5), which have been plotted in the full 



