SECTIONAL TRANSACTIONS.— B, 359 



polarities the point of constraint must be known. This may be determined by 

 steric considerations, and also possibly by the nature of the reacting compounds. 

 It would appear to follow that (1) the current practice of labelling certain 

 atoms negative or positive is only correct in so far as it indicates the condition 

 they lend to assume ; (2) the key atom of a molecule is the point at which 

 constraint originates, and hence that one which, in the initial stages of a reaction 

 with a second molecule, is first associated with this. Conceivably, therefore, 

 it may vary from case to case. Attention is drawn to the incomplete nature 

 of explanations wliich simply refer these to the polarity of a given group. The 

 course of Lapworth's argument explains why considerations of reactions based 

 en Werner and Fiiirscheim's views have led to correct prognostications. Never- 

 theless, it would seem in some respects inconsistent with more purely electronic 

 conceptions. In connection with these, it is pointed out that physicists are 

 by no means in agreement as to the details of atomic structure, and that 

 hence it seems premature to associate a discussion of the valency of organic 

 compounds with any one theory. Rather it seems advisable at present to limit 

 discussion to the valency electrons. It is further suggested that a more com- 

 plete conception of chemical reactions may be gained by considerations of the 

 lines of force associated with electrons. Exception is taken to the mode of 

 derivation of the property of induced polarity employed by Kermack and 

 Robinson, and an alternative is suggested. 



It is emphasised that the intermediate production of induced alternate 

 polarities only represents one course by which stable equilibrium is attained. 

 Other factors may contribute to this end, and for this reason a reaction may 

 take a course which would hardly be anticipated from the ordinary way of 

 applying the plus-minus notation. 



8. Mr. E. D. WiLLi.\MSON. — The Determination of Compressibilities 

 up to High Pressures and Applications to High-Pressure 

 Chemistry. 



A number of compressibilities have been determined at the Geophysical 

 Laboratory, Washington, in recent years. The method previously used, how- 

 ever, is not sufficiently accurate at the lower end of the pressure range, and so, 

 especially in the case of liquids, some supplementary method is required. A 

 new form of pyknometer has recently been developed and has proved satisfactory 

 in filling the gap. A feature of the instrument is that continuous readings by 

 means of a movable electrical contact may be taken without removal from the 

 pressure chamber. 



In the study of the chemical effects of pressure on systems of more than one 

 component it is necessary to know the compressibility of each solution in order 

 to compute the volume changes on which these effects depend. The volume 

 changes can be readily calculated from the slopes of the density-composition 

 curves. Even in the case of a simple system, such as a salt and water, it is 

 necessary to make a number of other measurements iu addition to those of 

 compressibility. For instance, good density-composition data must be obtained 

 at atmcspheric pressure. Also, some form of equilibriimi determinations, such as 

 those of vapour pressure or E.M.F., must be made in order to calculate the 

 initial differences in ' free energy ' {used in the same sense as by G. N. Lewis) 

 between the solid salt and salt in solution. For the case of H2O — NaCl almost 

 all the necessary data for the complete elucidation of the system under pressure 

 have been obtained, and a beginning has been made with some others. 



9. Dr. E. F. Armstrong, F.E.S. — The Hydrogenation of Fats. 

 (See below, No. 11.) 



10. Prof. A. F. HoLLEMAN. — The Rule of the Conservation of Sub- 

 stitution-Type of the Benzene Nucleus. 



1. The rule can only be applied without restrictions to introduction of a 

 second substituent in CeHjX. 



2. Pure p-o-substitutions as well as pure /n-substitutions must be regarded 

 as limits; all substitutions really observed are more or less mixed. 



3. The two types of substitution are not so sliarply different as is generally 

 believed. 



