296 REPORTS ON THE STATE OF SCIENCE, ETC. 



tight compartment, ignoring the absorption bands exhibited by their compounds 

 in the spectral regions on either side of that which they arbitrarily select. 



There are, however, certain observed facts in connection with absorption 

 spectra which have now been establislaed beyond all possible question, and 

 these cannot be ignored in any discussion of the problem. In the first place, 

 it may be categorically stated that every known substance, wnether elementary 

 or compound, possesses characteristic frequencies not only in the visible or 

 ultra-violet, but also in the infra-red, and these frequencies, of course, are 

 exhibited as absorption bands. In discussing these characteristic frequencies or 

 absorption bands it is convenient to subdivide the whole spectrum into four 

 sections, the visible and ultra-violet from X = O.S fx to 0.1 /i, the short wave 

 infra-red from X = 20 n to 0.8 u, the long wave infra-red from /\ = 4!00ju to 20//, 

 and the very long wave infra-red from A = 3,000,a to 400 u. 



The first real advance was made by Coblentz {Carnegie Irut. Pull., No. 35, 

 1905), who examined the absorption spectra of a large number of substances in 

 the short wave infra-red, and found that definite atomic groups, such as 

 CH3, OH, NH2, etc., exhibit characteristic absorption bands in that region, 

 of whatever compounds these groups may form a part. This is the first 

 observation which definitely connotes constitution (not structure) with 

 absorption. 



The next advance was made by Bjerrum (Nernst Festschrift, p. 90, 1912), 

 who applied Lord Rayleigh's principle to infra-red spectra and showed that 

 the breadth of an absorption band is due to the summation of frequencies 

 characteristic of a substance. If F be the characteristic frequency of a mole- 

 cule in the short wave infra-red, then associated with F are the subsidiary 

 frequencies F + nR, where E is a characteristic frequency of that molecule in 

 the long wave infra-red and ?! = 1, 2, 3, etc. The absorptive power of these 

 subsidiary frequencies decreases as the value of n increases, with the result 

 that an absorption band is produced consisting of a series of equidistant sub- 

 groups which progressively decrease in intensity the farther they are from the 

 centre. It was shown later (Baly, Phil. Mag., 29, 223 (1915)) that this structure 

 is not confined to the infra-red bands, but is also shown by visible and ultra- 

 violet bands. 



A third phenomenon, which has considerable importance, is that the central 

 frequency of every absorption band shown by a substance in the visible or ultra- 

 violet is alwaj's an exact integral multiple of the central frequency of a very 

 pronounced band shown by that compound in the infra-red (Baly, Phil. Mag., 

 27, 632 (1914)). Reference has been made to the fact that a given substance 

 can exhibit different absorption bands in the visible and ultra-violet according 

 to the conditions of solvent, etc., and it is these bands which form the basis 

 of the structure-absorption theory. These bands, however, only differ in the 

 fact that they are different integral multiples of the same infra-red band. 



Further, the Bjerrum conception of the structure of absorption bands has 

 been extended. It is well known that many bands have not only been resolved 

 into sub-groups, but also that these sub-groups have been resolved into fine 

 absorption lines. These absorption lines are symmetrically distributed in an 

 analogous v\'ay to the sub-groups — that is to say, if S be the frequency of the 

 principal line of a sub-group, the remaining lines in that subgroup are given 

 Ijy S'±nA, where A is a constant and 71 — I, 2, 3, etc. Just as in the case of the 

 principal lines of the sub-groups, the frequencies of which are given by F + wR, 

 where R is the frequency of an absorption line shown by the substance in the 

 long wave infra-red, so the constant A is the frequency of an absorption line 

 sho^vn by the substance in the very long wave infra-red. 



Lastly, the frequencies A, R, and F appear to be related together, as shown 

 by the cases of sulphur dioxide (Baly and Garrett, Phil. Mag., 31, 512 (1916)) and 

 water (Baly, Phil. Mag., 39, 566 (1920)). These two compounds are the simplest 

 yet dealt with, and in both of them there are found evidences of the existence 

 of three constants, Ai, A3, and A3, of the same order of magnitude, and of two 

 larger constants, Ri and R;, these also being of like order of magnitude. The 

 relation between these various frequencies is as follows : The frequency Ri is 

 the least common integral multiple of one pair of the three frequencies Ai, Aj. 

 and A3, whilst the frequency R2 is th:^ least common integral multiple of another 

 pair of Ai. Aj, and A3. Again, the frequency F, which is the central frequency 



