no INFRA-RED ABSORPTION SPECTRA. 



showing that the vibration of the original benzene ion has not been 

 destroyed. Such examples as these would indicate that the new bands 

 are due to the groups of atoms substituted. But how are we to estab- 

 lish this with certainty, especially when in the myricyl alcohol the evi- 

 dence is contradictory for the direct effect on the OH-group? 



TOTAL ABSORPTION. 



This is not so well illustrated here as in the work of Friedel and of 

 Zsigmondy (loc. cit.), who used the undispersed radiation. The pres- 

 ent work agrees with theirs in showing that compounds having sulphur 

 and the halogens are more transparent than H, O, OH, or N, which 

 they have replaced. But not all the nitrogen compounds are highly 

 opaque, e. g., nitromethane. The present investigation illustrates best 

 the question of the location of the regions of greatest absorption. Thus 

 in pyridine and picoline a 0.16 mm. layer is almost opaque beyond 6 /x, 

 while methyl cyanide is quite transparent. Methyl iodide is quite trans- 

 parent, since its large absorption bands lie in the region where the radi- 

 ation from a black body is very weak, while the 3.4 /x band is shallow, 

 so that the great transparency of this substance, observed by Friedel, 

 as compared with nitrogen compounds, is apparently due to the lack of 

 absorption bands in the region of intense radiation. As a whole, the 

 work agrees with that of Friedel and of Zsigmondy in showing that the 

 absorption of radiant heat depends upon the manner of the bonding of 

 the atoms in the molecule, as well as upon the kind of compound in 

 which the atoms of an element are united. 



The curves of pyridine and picoline illustrate the meaninglessness 

 here of the application of the law of variation of absorption with thick- 

 ness, which law Angstrom^ found not to hold true in certain cases, and 

 concluded that it is due to the presence of unresolved absorption bands. 

 The pyridine curves show this, especially the one for the thicker film, 

 where the overlapping of the bands has lowered and blotted out the 

 deep depressions in the curves. 



Friedel and Zsigmondy (loc. cit.) found that total absorption does 

 not depend upon the size of the molecule. In the present instance we 

 have noticed that the number and intensity of the absorption bands does 

 not depend upon molecular weight, e. g., in the petroleum distillates. 



^Angstrom : Ofversigt af Kongl. Vetenskaps, Akad. Forhandl., S. 331-352, 1890 ; S, 

 549, 1889. He computed the absorption of a liquid or vapor as a function of the 

 thickness, from the equation (A=Io(i — e'*') and obtained smaller values than those 

 observed. He concludes that the observed bands are complex groups of lines which 

 are unresolved on account of the small dispersion. 



