( 444 ) 



spectroscope. When llie sljt is opciKMl widely cacli spectral line is 

 seen as a leclaii^lo wilh l)riiilit i-ini;s or parts of rings as (lie case 

 may be. TIh^ pari of (he spec(rinu in (he (ignres I'efers to (he (wo 

 yellow and (he lireen mercury lines. In (ig. J (he (wo rectangles 

 corresponding to (he two yellow mercury lines arc super[)Osed. The 

 greeii mercury line is largely ov^orexposed. I iiave reproduced it also 

 in order to give an idea of the dispersion used. The intensity of the 

 magnetic Held in ligures J ajid 2 was about 5()()() (niuss. 



1( is a very beautiful sigld (o \Aatch the moving system of rijigs, 

 while the magnetic force is slowly inci'eased. The I'ings ).,■ and X„ 

 are lirst seen approaching, (hen coinciding, separadng, coinciding 

 for a value of (he lield of aboul loOOO dianss wi(li (he n(;\t ring 

 ).Q, |)assing oxer this ring, etc. 



For measui'emen(s i( is necessary (o reduce the width of (he sli(, 

 as in Fig. 2. Owing (o rise of temperature the rings have somewhat 

 expanded. 



6. For measurements, which T hope to communicate in a future 

 paper, I ha\e used not oidy the method of diameters resumed al)ove 

 (§ 3) l)ut also till' nu'thtxl of the coincidences ') for the distinct values 

 of the magnetic force, which bring to coincidence X,. and ).^., or ).,. 

 and Xo with /„. 



Concerning the difticulties attendijig the use of the method of 

 coincidences P\\bry and Pkrot') remark: 



"Même avec ce perfectionnement, ia médiode présentait des incon- 

 \énients assez graves : 



1. La necessité d'éclairer simultanément Tappareil par les deux 

 sources entraine des pertes de lumière assez iinportantes ; 



2. Les coincidences ne sont bien observal)les ((ue lorsque les deux 

 systèmes d'anneaux ont des éclats comparables, et cette condition 

 n'est pas toujours facile a réaliser ; 



3. La recherche de la coincidence entraine toujours des tatonne- 

 ments et l'on n'est Jamais sur (lorsque la periode est courte) d'eji 

 rencontrer une qui soit exacte." 



The drawbacks (o (he method, mentioned sul) 1. and 2. are 

 eliminated in the ap[)licatiGn to radiation in a magnetic lield, now 

 under consideration. V>\ variation of the current in the electromagnet 

 the coincidence can be attained with the desired degree of accuracy 

 and hence also the third objection is obviated. 



J) Fabry et Perut, Ann. de Chim. et de Phys. p. 12, T. 25, Janvier 1902 



