Wave-length of Light. 257 



coefficient of expansion of these bars was very thoroughly 

 investigated by Prof. Rogers, and was found to be 



17*946 /i per metre per degree C. 



The absolute lengths of S a j and S a 2 depend on long series 

 of comparisons with Prof. Rogers's bronze yard and metre E 2 

 and steel copies thereof. Upon the relation existing between 

 R 2 aQ d the Metre des Archives depends, then, the absolute 

 value assigned to the wave-length of light, since the close 

 agreement of the various series of comparisons executed by 

 Prof. Rogers between R 2 and the speculum-metal standards 

 show that no sensible uncertainty exists in the relations 

 between them. 



The yard and metre R 2 is of the alloy known as Bailey's 

 metal, this being the material of the Imperial Yard and many 

 other standards. The graduations are upon platinum-iridium 

 plugs, the polished faces of which are in the plane of one sur- 

 face of the bar when supported at its neutral points. The 

 relation of the metre R 2 to the Metres des Archives rests on 

 a very large number of comparisons made with two entirely 

 independent secondary standards; the copper metre desig- 

 nated T, and the brass yard and metre designated C.S. A 

 full account of these comparisons is contained in vol. xviii. of 

 the Proceedings of the American Academy of Arts and 

 Sciences. 



The metre T is on platinum plugs in a pure copper bar 

 and was traced and compared by M. Tresca in 1880, from the 

 Conservatoire line-metre No. 19, the relation of which to the 

 Metre des Archives was very exactly known. 



The yard and metre C.S. has its graduations on silver 

 plugs in a brass bar. The yard was compared directly with 

 the Imperial Yard in 1880, and the standard was then sent 

 to Breteuil, where it was compared with the International 

 Metre by Dr. Pernet. 



There were thus two completely independent sources from 

 which the relation of R 2 to the Metre des Archives could be 

 obtained. The results derived by very elaborate comparisons 

 with each of these were as follows : — 



where A is the Metre des Archives. In addition to the 

 very close agreement of the above, further evidence was ob- 

 tained by deriving the relation between the yard and metre 

 from R 2 , the yard R 2 having been exactly determined by 



