262 Messrs. Porter and Morris on the Measurement 



Fior. 3. 



soapstone. This substance, curiously enough, was found very 

 much more satisfactory than any we had previously tried. 

 Even the effects of moisture, when no efforts were made to 

 remove it, were comparatively insignificant (though calcium 

 chloride was always used to obtain the best results) ; and it 

 will take a smooth or even a polished surface, which can be 

 renewed at any time if necessary by lightly scraping it. 



(3) The effect of slight tremors on the throws of the needle 



should be minimised. 



In an ordinary suspended needle-galvanometer resting upon 

 a firm support, the only kinds of oscillations 

 which are capable of accumulating to any appre- 

 ciable extent are those produced by a vertical 

 vibration of the suspension. But a simple 

 vertical vibration is unstable in a freely sus- 

 pended body, and tends to resolve itself into 

 an oscillation about the horizontal axis of least 

 moment of inertia (fig. 3), which, in a reflecting 

 galvanometer, gives rise to the up-and-down 

 motion of the spot of light which is so apt to 

 occur. Hence the two points where the needle 

 rests against the stops should be points on the 

 horizontal axis of least moment of inertia. 



This is conveniently accomplished by making 

 the needle symmetrical as far as possible about 

 its longitudinal centre-line, and then placing 

 the stops so as to touch the needle on this line. 

 It was found that the throws from the stops of 

 a needle so constructed were reliable even when 

 it was oscillating so as to produce a violent 

 vertical movement of the image of a scale re- 

 flected from the mirror into a telescope. 



(4) The field at the needle due to given difference of 

 potential between the terminals of the instrument 

 should be as large as is consistent with the high 

 resistance which previous considerations have shown 

 to be necessary. 



There are many advantages in stopping the needle at a 

 long radius from the axis of suspension, in spite of the 

 slightly increased moment of inertia which must result from 

 so doing : perhaps the chief of these is that a comparatively 

 small angular movement will then suffice to get the needle 

 quite clear of its stops. Besides this, however, small to-and-fr o 

 vibrations of either of the stops impart less angular velocity 

 to the needle ; and the coils of the galvanometer can be made 



iii 



