Dr. Ketteler on the Dispersion of Light in Gases. 337 



The complete determination of the optical properties of the 

 gases required a connected and uniform measurement of all the 

 elements concerned. In the first place it was needful to measure 

 more accurately than had hitherto been done the wave-lengths 

 of the three sources of homogeneous light which I employed, 

 namely the lithium-, sodium-, and thallium-flames. For this 

 purpose the method recommended by Fizeau* was followed. 

 This consists in finding the coincidences which occur when an 

 appropriate system of Newton's rings is illuminated by each pair 

 of the above-mentioned flames simultaneously. By help of an 

 apparatus in which the plates needed to produce these rings were 

 contained — the upper one being fixed but so arranged as to be 

 capable of assuming any required inclination to the lower one, 

 while the latter could be moved vertically up and down, without 

 the slightest twisting, by means of a lever worked bya rnicrometric 

 screw — the number of rings intervening between two coincidences 

 could be observed : the following, for instance, are examples of 

 numbers actually thus obtained: — 



/x Na _1156 fi Tl _ 770 /L6 T1 1537 



~~ ~ 1015' ~~. 7 699' y^T 7 1225* 

 The delicacy of the apparatus was such that, in moving the 

 lever, the finger had to pass over a length of more than 10 milli- 

 metres in order to cause a displacement to the extent of a single 

 fringe. As the mean of twenty-three observations of this kind, 

 the following ratios were obtained : — 



^- = 1-138953 ; ^- a = 1 '101568 ; ^ = 1*254638. 



The product of the first two numbers gives 1 "254634. 



In order to obtain the absolute values of the wave-lengths, we 

 must adopt for / Na Fraunhofer's value — 0*0005888 millim., 

 inasmuch as Fizeau's contemplated repetition of this measure- 

 ment has not yet been published, and to determine it myself lay 

 beside the scope of my investigation. We thus obtain, in hun- 

 dred-millionths of a millimetre, 



/ Li = 67061-6, / Na -58880-0, Z T1 =53451'0, 

 numbers which may be regarded as relatively tolerably certain, 

 as far as the fifth place. 



It was now possible to proceed to the determination of the 

 dispersive properties of gases. The gases examined were the 

 five following : — air, carbonic acid, hydrogen, cyanogen, and 

 sulphurous acid. 



The gases were placed in tubes between thick interference- 

 plates, such as those introduced and several times described by 

 * Annates de Chimie et de Physique, 3 e ser. vol. lxvi. p. 429. 



Phil Mag. S. 4. Vol. 32. No. 217. Nov. 1866. Z 



