( 604 ) 



border of the band (if there i« no magnetic field) | has the value 

 V t}, the coefficient v being some moderate number, say between 3 

 and 6 (§ 9), and for one of the components of Zk.eman's doublet we 

 have i = ?. In the magnetic field used by Hallo the distance of 

 the components from the middle of the original line amounted to 

 0,15.1.^., half the breadlli of the line being 0,5 J. £•., as has 

 already been said. 



We have therefore the following relation between ij and 5 : 

 g:jjij= 0,15:0,5 



3,3 

 'i = — S (32) 



V 



On the other hand, a point in the spectrum, at which the angle 



of rotation per unit length was approximately equal to nnity, was 



^ , /35 

 situated at a distance of 1,6 A. h.\ — — ot the mutual distance of the 



J 30 



two D-lines) from the middle of the original line. Tins being 10 times 

 the distance from this line to one of the components, we have 

 approximately 



§ = 10$. 



On substituting this value and (32) in the formula (31), it appears 

 that the terms tj" may be omitted. Hence, if (13) is taken into account, 



n Ne 



*ƒ) = 0,005 — = 0,005 ~- , (33) 



c s H 



or since cp = 1 is, 



A^e = 200H. 



The strength of the magnetic field in these experiments was 9000 



in ordinary units, or 



9000 

 H = = 2600 



in those used in our equations. Taking for e the value (30), I finally find 

 A^=i4 X 10'*- 



• ^ 12. The value of i] may likewise be calculated, both for the 



carbonic dioxyde and for the sodium flame. In the first case we can 



avail ourselves of the formula (27), in which /•„ is now known ; 



the result is 



)( jt 



)i = — =z = 2,5 X 10"- 



2,'/-„ Xk, 



Foi' the sodium llamc we first draw from (33) 



