378 



NATURE 



[August i8, 1904 



organic compounds, although he dealt fully with the question 

 of molecular volumes in his final communication in 1889. 



As a pioneer, Kopp had very great difficulties to contend 

 against when he began his researches ; data were scanty and 

 far from accurate, and the substances which could be most 

 easily obtained and, it was thought, readily purified were, 

 unfortunately, those which were the least likely to lead to 

 normal generalisations. Water, the alcohols, and the 

 organic acids all contain a hydroxyl group, and we now 

 know that the physical properties of these substances are 

 abnormal in nearly all respects, owing, probably, to the fact 

 that their molecules tend to associate together ; moreover, 

 the esters, which are formed by the interaction of acids and 

 alcohols, do not behave quite normally, and there is prob- 

 ably molecular association, though to a much smaller extent 

 than with the hydroxyl compounds. 



There can be little doubt that if Kopp had been able, in 

 the first place, to obtain a considerable number of pure 

 substances of normal behaviour, such as the paraffins or 

 their halogen derivatives, he would not have been led to the 

 erroneous conclusions which he defended with such vigour 

 for so many years. If we take the normal paraffins as the 

 simplest class of organic compounds, we find that, instead 

 of the boiling-points rising by equal intervals as the series 

 is ascended, the rise, which is very large for the lowest 

 numbers, becomes smaller and smaller as the molecular 

 weight increases. This fact is, of course, now well known, 

 and various formulae have been suggested to reproduce these 

 boiling-points. Thus Walker has proposed the formula 

 T^aM', where T is the boiling-point on the absolute scale 

 of temperature, M is the molecular weight, and a and h 

 are constants. Ramage has this year suggested that this 

 formula applies only to the CH, chain linkage, and that 

 the influence of the terminal hydrogen atoms is considerable 

 in the case of the lowest members, but diminishes as the 

 chain lengthens, and becomes eventually either constant or 

 negligible. In other words, the lower members of the series 

 cannot be regarded as truly homologous, and that is a point 

 which is, I think, important to bear in mind. Ramage 

 suggests a new formula, T = ii[M(i -2"")]*. where a is 

 Walker's constant, 37-377.^, and n is the number of carbon 

 atoms in the molecule. He assumes, however, a constant 

 difference for CH, in the case of the alcohols, the aldehvdes, 

 and the ketones, but I doubt whether the boiling-points of 

 the last two classes of compounds are yet sufficiently well 

 established to allow of any certain conclusions being drawn 

 from them. 



I am inclined to think that it may be useful to regard 

 the value of A (the rise of B.P. for an increment of CH,) 

 as being mainly a function of the absolute temperature, 

 and I would provisionally suggest the formula A = '^.'Jn^^ ,' 

 where A is the difference between the boiling-point, T, of 

 any paraffin and that of its next higher homologue. 

 Taking the boiling-point of methane as io6°.75 ^bs., the 

 values for the higher members agree better with the 

 observed temperatures than those given by Ramage 's 

 formula, as w-ill be seen by the first table on the next column. 



I do not wish, however, to lay much stress on the actual 

 form of the equation, or on the particular values of the 

 constants ; the chief point I wish to direct attention to is 

 that A may be regarded as a function of the temperature. 



Suppose that we replace a terminal atom of hydrogen 

 in each normal paraffin by chlorine, so as to form the 

 homologous series of primary alkyl chlorides. The boiling- 

 points of these chlorides are much higher, and the differ- 

 ences, A, are much smaller than for the corresponding 

 paraffins, but the gradual fall in the values of A as the 

 series is ascended is unmistakable. The same remarks 

 apply to the bromides and iodides, the boiling-points being 

 still higher and the values of A smaller. 



But the point of chief interest appears to me to be this : 

 if the values of A for the halogen derivatives are plotted 

 against the absolute temperatures, the points for the most 

 part fall near the curve constructed for the paraffins, and 



represented by the formula A =wLn8v=^ The first value of 

 A is decidedly low in each case (average deviation from 

 curve 2°. 7) ; the later ones are rather high in nearly every 

 case (average deviation o°-86). Similar results are in 



NO. 1816, VOL. 70] 



Boiling-point (abs. temp.) 



general obtained with other homologous series of com- 

 pounds in which molecular association is not believed to 

 occur, but, as will be seen from the following table, the 

 deviations from the normal paraffin curve are greater in 

 the case of those series the lower members of which, 

 according to Ramsay and Shields, are characterised by 

 molecular association. 



Associating Substances. 



4-2-9 



+3-85 

 4-2-85 

 + 1-58 

 4-5-24 



In the great majority of cases the deviations are greatest 

 for the lowest members of a series, the calculated values 

 of A being almost invariably higher than the observed, and 

 this may perhaps be explained in the manner suggested by 

 Ramage. I have, therefore, divided each series into two 

 groups, the first ending and the second beginning with the 

 lowest member of the series which contains a CH, group 

 linked to two carbon atoms. Thus, of the alkyl chlorides, 

 the first group contains CH3CI, CH3CH.CI, and 

 CH3 — CH, — CH,C1, and the second group begins with 



