The molecules of the compounds placed in column I. contain only tetravalent carbon, and divalent oxygen atoms; those in column II. contain both tetra- and tri-valent carbon, and di- and monovalent oxygen atoms. Brühl also finds' that although the refraction-equivalent of propaldelyde is the same as that of acetone, yet that of the third isomeride, allylic alcohol, is different; thus, A comparison of these formulæ shews, that whereas in isomerides 1 and 2 the trivalent carbon atom is in direct union with an oxygen and a carbon atom, in the third isomeride it is directly bound only to carbon atoms: or we may say that isomerides 1 and 2 contain a divalent group C-O, whereas isomeride 3 contains a tetravalent group C-C, and a trivalent C-O group. If we may draw a general conclusion from the data contained in Brühl's paper, it would seem that the value of the refraction-equivalent of a compound, CO,H,, is independent of the way in which the interatomic reactions are distributed, provided each atom acts on its maximum number of other atoms; but if this is not so, then there is a connection, not only between the valencies of the atoms in the molecule, but also between the distribution of the interatomic reactions, and the refraction-equivalent. The latter part of this statement is illustrated by the fact that the value of (R) found for (CH ̧),CO, and for the acids 1 loc. cit. CH2+COOH, which all contain the divalent group C—O, agrees with that calculated on the assumption that (R) (C-O)"-r1C"+r1O1 but nevertheless the carbon atom in the group C—O, as this occurs in the before-mentioned molecules, is certainly trivalent. On the other hand the actual values found for (R) in the molecules CH2-CH2OH, C2H, &c., where the tetravalent group C-C occurs, agree with those calculated on the assumption that (R) (C-C)=2(C) =11'72. That there is a distinct quantitative connection between the nature of the polyvalent atoms between which direct action and reaction occurs, in unsaturated molecules, and the value of (R) for these molecules, appears certain from the results of researches recently conducted in Landolt's laboratory by Nasini'. Nasini has determined the values of (R) and (RA), for series of carbon compounds containing sulphur, these compounds being divisible into two groups', which may be represented generally as X' He gets the following values for the refraction-equivalent of the sulphur atom according as it acts as a divalent or as a monovalent atom, r1 SH=13'53 r1 S1= 15'09 TA S"=7.65 1 Ber. 15. 2878. 2 As examples of compounds belonging to group (1) may be taken, H,C,-S-H, Wiedemann' has also determined values for rS, from measurements of (R1) of CO(OEt),, CO(OEt)(SEt), CS(OEt),, CS(OEt)(SET), CO(SEt),, and CS(SEt),. The values agree fairly with those found by Nasini; they are as follows, But when it is sought to find values for S in more complex compounds containing oxygen, very different results are obtained according to the structure assigned to the compound molecules in question. Thus, assuming that rO1 = 3'29, and O"= 2'71, the following results are obtained. (1) If SO2=S<↓, r,S"=8′10; but if SO2=S<,,S"=6′94. (2) If SO1=0-S1, r,S1-8'37; but if SO1=S—O, r,STM-6·63. (3) If H2SO, HO-S-O-OH, SI=8'43. = (4) If H2SO, HO-S-OH, r1 S=7.85. I think we are justified in concluding that the refractionequivalents of molecules containing polyvalent atoms, all, or any, of which act directly on less than their maximum number of other atoms, is correlated not only with the actual valencies of these atoms (i.e. the number of atoms between which and each polyvalent atom there is direct mutual action), but also with the distribution of the interatomic actions (i.e. the nature of the atoms between which there is direct mutual action)3. The quantity (R) is conditioned by these factors, refractive index (u), density (d), and molecular weight (M). The values of μ and d vary for each compound when d varies directly as μ, the value of (R) remains constant for 1 Wied. Ann. 17. 577. 2 For data see Nasini loc. cit. 3 This conclusion, if accurate, illustrates the justness of the remark already quoted from L. Meyer, "what we call double or triple linking of atoms does not consist of a repetition of the process which we call single linking." any number of isomerides; but if d varies in some other ratio with μ, the value of (R) is different for each isomeride. The latter condition holds good in such unsaturated molecules as have been just defined. But in many saturated molecules, according to Brühl, the former state of things obtains; hence, although the refraction-equivalent of such molecules is constant, yet their refractive indices vary. This conclusion may be put in general terms thus. Isomeric molecules containing polyvalent atoms, all of which act on their maximum number of other atoms, exhibit equal refractive powers only when their densities are the same1. 141. Values have been assigned to the atomic refractions of the elements, (or the refraction-equivalents of the elementary atoms). Assuming that these values are justified by the experimental data, it is important to mark that in making use of them we do not assert that each atom in a molecule exerts its own refractive power, but rather that a group of certain atoms arranged in this or that manner, (as roughly represented in the structural formula), exerts a definite refractive power, which is increased or diminished by changes in the arrangement, number, or nature of the atoms. It would appear better to assign values to the refractionequivalents of certain groups of atoms, in carbon compounds at least, rather than to each individual atom. The following table contains some of these values, and also recapitulates the atomic refractions' which have already been mentioned. 142. The refraction-equivalents of a few carbon compounds belonging to the benzenoid group have been determined'. The arrangement of carbon atoms in a closed chain does not appear to exert any special influence on the values of the constant in question; thus (R) found for benzene, agrees with that calculated on the assumption that the six atoms of carbon all act as trivalent atoms in the molecule CH. Few, if any, measurements have yet been made of the refraction-equivalents of carbon compounds containing pairs of divalent (trebly-linked) carbon atoms. Brühl2 has determined (R) for a few so-called propargyl compounds, derived from the hydrocarbon C,H,, which possibly has the structure H-C-C-C = H ̧. The refraction-equivalents of some compounds containing much carbon, relatively to the quantities of other elements present, are considerably larger than the values calculated by the use of the numbers given in par. 1413. 1 See especially for data Landolt, Ber. 15. 1038; and Brühl, Ber. 12. 2142. 2 loc. cit. 12. 2146. For data see Gladstone's paper, C. S. Journal, Trans. for 1884. 241. See also ante, par. 138, footnote 3 (p. 308). |