34 



«j!ai constat^ avee surjirise que la rotation pour eette menie lon- 

 «gueur d'oude t^tait bien. comme I'indique le tableau, — 1*^21' efr 

 «qu'il n'y a pas Ik une faute d'iuipression. mais une faute dans le 

 «trac6 de la courbe dent Torigine est ^vidente: le dessinateur a 

 «plac^ le point exactement un deg^re trop haut. Vous verrez vous- 

 *meme que cela change un peu, non pas la forme de la courbe xjjj 

 «mais la place du second maximum », 



The, observed values of ip as now corrected are plotted aoainst 

 those of I in the accompanying diagram; they are shown thus, 0. 

 The wave-length corresponding to t^ = appears to be 5,72; the 

 wave-length A„, (in previous notation) is situated at 5,31 or nearly 

 80. Assuming 



(1) Xf, = 5.72 . 10-5 cm ; /^^^ = 5.31 . iQ-^ cm 



and applying the equation 



(2) ±/,Q = /.,) ^mt 



(see § 9 of the paper quoted) we obtain 



(3) r= 0,782. 10-5 cm. 

 Again, from i^, = A'^-\- F Xq we deduce 



(4) X„, ^ 640 . 10-5 cm 



and this agrees fairly well with the general march of the observed 

 values of ip. 



From the observed rotations, the following values of the con- 

 stant D have been calculated by equation (2) Art. 8 of the paper 

 quoted, T and Xq having the values assigned to them in (3) and 

 (1) above. 



X \tp\ D 



5,22 . 10-5 cm 20 50' 4,21 . lO"" cm^ 



5,62 .... 1021' 4,28 



5,81 .... 1046' 6,45 



5,89 .... 2° 30' 5,55 



Assuming 



(5) D = 4,28. 10-11 cm2 



equation (2) Art. 8 of the paper quoted leads to the following results: 



