226 MESSES. R. THRELFALL AND J. A. POLLOCK 



shooting. The best way of carrying out the examination is to draw the thread, 

 comparing it all along its length with a bit broken from one end laid beside it under 

 the cover slip of a microscope slide, the space between the slide and cover slip being 

 filled with stained glycerine. 



When a sufficiently perfect thread is obtained, it is silvered all over, as described in 

 'Laboratory Arts,' p. 222, and the ends are then coppered and soldered up to the 

 supports. The centre of the thread is then coppered for about a centimetre of its 

 length. The lever is adjusted in position, being held tight by a temporary clamp 

 mounted from the back girder. It is then soldered up to the thread by tinman's 

 solder. The excess of silver is removed with nitric acid, and the thread is well washed. 

 It is surprising how difficult it is to do this thoroughly ; we use a brush made of glass, 

 but we are not sure that it is the best way. 



Lever. 



We have made levers of many metals, but at present use one of gilded brass wire 

 of the smallest diameter we could get with our draw-plates, say '005 inch. The 

 conditions to be satisfied are (1) magnetic indifference, (2) undeformability by the 

 arrester, (3) lightness. At first we used to make levers of aluminium foil, shaped like 

 a cross ; we then tried soft annealed copper, in the hope of doing away with secular 

 changes of shape, but it was not stiff enough to resist the arresting. 



Adjustments of the Lever. Having decided to use three turns in each half of the 

 thread, it is necessary to adjust the lever so that the line joining the centre of gravity 

 of the lever to the thread is nearly horizontal when the thread is twisted with this 

 amount of twist. 



It is shown in the part of this paper dealing with the theory of the instrument 

 that with this twist in the thread the lever becomes unstable when the line joining 

 its centre of gravity to the thread rises about three degrees above the horizontal. 

 We have adjusted the lever in the present case so that it upsets when the thread is 

 twisted by about three whole turns. The adjustment is made by weighting the lever 

 with a small drop of fusible metal, rather larger than a pin's head. 



In making a new machine we think it would be worth while to make the lever out 

 of a thickish needle of fused quartz. We have occasionally thought that the thread 

 was slipping in its silvering, but though there is a distinct danger of this happening, 

 we have not had a really clear case. Some attempts to guard against the possibility 

 of this, by grinding the thread flat on one side, failed on account of the way in which 

 the thread was weakened. 



The Spring and its Adjustments. 



It is almost if not quite essential that one end of the quartz fibre shall be carried 

 by a spring, otherwise the difficulty of manipulation becomes intolerable. It is also 



