that Matter takes up no Room in the ^jEther. 845 



passed a little platinum hook serving for the attachment of a 

 quartz fibre. 



Fig. 1. Fig. 2. 









o 



1 

 1 







thickness 

 * '0444cm. 



1 

 1 



S 



» 

 1 



thickness 



IS 



=-023cm. 





<'342-> 

 cm. 



K--38: 

 cm 



74- -> 





1 



1 

 i 



V 



G. The attachment of the fibres called for special care, 

 because a gradual drift of " zero," which would be of little 

 consequence in many instruments, might here be a serious 

 trouble. Various forms of soldering were tried, but on the 

 whole the best results were got with comparatively soft 

 cements, such as flexible shellac and a hardish form of 

 adhesive wax. (The quartz fibres were of course allowed 

 to untwist before their ends were cemented.) Though the 

 rate of creep immediately after suspension might then be 

 fairly rapid, in a few days the soaking out of the torsional 

 strains was so far complete that observations could be made 

 and the still continuing creep allowed for. A persistent drift 

 in azimuth taking place in a smooth manner was naturally 

 less objectionable than far smaller changes occurring irre- 

 gularly ; but it was essential that the progressive displacement 

 should not be so rapid as to pass beyond the range of the 

 adjustments in the course of a few days. 



7. It was evident from the first that convection currents 

 within the suspension-chamber were likely to be a source of 

 trouble ; but it was not practicable to surround the suspended 

 plate by massive metal walls, leaving only a minimum of 

 clearance for free libration ; for the gravitational forces 

 exerted on the plate by the walls would then have been 

 relatively large, and minute tilts of the apparatus as a whole, 

 by shifting the position of the plate within the enclosure, 

 would have caused these forces to vary. 



8. At the outset it had been intended to exhaust the 



