120 Dr. L. Silberstein on Molecular 



these circumstances the above result seems to acquire an 

 additional interest, inasmuch as it would account at least 

 for the breadth of the bands, if not for the large total 

 amount of absorption. In fact, according to (36) we can 

 have almost any breadth due to BR alone, i. e. without 

 taking account at all of the collisions of molecules. Thus, if, 

 in round figures, \ = 1000 and V= 2000 A.U. (that is, by the 

 little table on p. 117, if R/R c == 1*10), then we should have, 

 for SB=-^qR c , 6V = 1200 A.U., L e. what would be called 

 a " general absorption " from 2000 to 3200 A.U., and even 

 for 8R=— yooojRc a rather broad band, stretching over 

 12 A.U. The "general absorption^ in the former case 

 would, of course, be a "weak" one, such as are observed 

 in the fatty alcohols. 



5. Triatomic Molecules, 



Let the three centres 0\ } 2 , 0%, i. e. the equilibrium 

 positions of the dispersive particles within their atoms, form 

 any (fixed) triangle. Let R 1} R 2 , ^3 be the sides of this 

 triangle, 2 O z , O z 1 , 0\0 2 , respectively, and lt 2 , 6% its 

 angles (fig. 2). It will be most convenient to write the 



Fis. 2. 



equations of motion in terms of the components of the dis- 

 placements r t - taken normally to the molecular plane Oi0 2 3 

 and along the two sides meeting in 0%. The normal com- 

 ponents will be free from interatomic action, and need, 

 therefore, no further attention. If i\ r b A r 2 , N z are the 

 three atomic refractivities, and if the planes of the mole- 

 cules are all parallel to one another, then the normal 



