greater than that of any other previously known member of the series, but also that it was over twenty times as great as that of the isomeric trans-acid. The actual numbers obtained were:

In the first place, it will be observed that the value of k for our cisacid approximates to that given by Auwers for the acid he obtained from ethyl a-bromoisovalerate (0.2938), but that, whereas his values diminished considerably on dilution (see p. 655), ours remain nearly constant. The chemical properties of the two acids (his and ours) are so much alike that we are not disposed to doubt their identity. But since the trans-acid is now shown to have a constant less than onetwentieth that of the cis-isomeride, Auwers' conductivity results can hardly be explained on the supposition that he was dealing with a mixture of the two in which the cis-form predominated. As a matter of fact, we made a mixture of 9 parts of the cis-acid with 1 part of the trans-acid, and although the mixture melted in the neighbourhood of 178° (the temperature given by Auwers as the melting point of his acid), its "dissociation constant" (if such a term can be strictly applied to a variable quantity) was always much lower than that of the pure cis-acid, thus:

Mixture of 90 per cent. cis- with 10 per cent. trans-s-Diisopropylsuccinic Acids (temp. 25°).

The chief interest, however, connected with these results is the indication they afford that the constants of two stereisomeric saturated acids may differ very greatly. Professor J. Walker has in a private communication drawn our attention to the fact that the constant for maleic acid (k = 1·17) is thirteen times as great as that of fumaric acid (k=0·093). But the two cases are in no wise parallel, for the stereoisomerism of two 8-dialkylsuccinic acids is of a totally different character from that of maleic and fumaric acids, although it must be confessed that the conventional use of the terms trans- and cis- in the case of the former is misleading.

Our experiments on the conductivities of the two s-dipropylsuccinic acids yielded the following numbers:

There is nothing extraordinary about these results, for although there is a marked difference between the constants of the two stereoisomerides, it is not abnormal. Thinking that the exceptional behaviour of the s-diisopropyl acids might be connected with the isopropyl grouping, we determined to examine aa1-methylisopropyl- and aa-propylisopropyl-succinic acids. Professor Perkin was kind enough to furnish us with specimens of the two inactive isomerides of the former, whilst those of the latter were of our own preparation. Conductivity determinations gave the following results:

The aa,-methylisopropylsuccinic acids were described in a paper by W. H. Bentley, W. H. Perkin, jun., and J. F. Thorpe (Trans., 1896, 69, 270).

trans-aa-Methylisopropylsuccinic Acid (m. p. 174-175°).

cis-aa-Methylisopropylsuccinic Acid (m. p. 125–126°).

trans-aa1-Propylisopropylsuccinic Acid (temp. 25°).

cis-aa-Propylisopropylsuccinic Acid (temp. 25°).

Thus, we find in the case of the aa,-methylisopropyl acids that the constant of the cis- is four times that of the trans-isomeride, whereas there is no greater difference between the two stereoisomeric aa,-propylisopropyl acids than there is between the corresponding twos-dipropyl acids. Hence it would seem that the peculiar behaviour of the two diisopropyl acids is in some measure due to the additive effect of the two isopropyl groups. In order to see whether this was wholly so or not, we prepared and determined the constants of the methyl hydrogen salts of the two s-diisopropyl acids. We were astonished to find that the masking of one of the carboxyl groups, whilst it only lowered the

constant of the trans-acid by about 40 per cent., reduced that of the cis-acid to one-twentieth of its original value.

Thus :

The foregoing results, which for the sake of reference we have tabulated at the end of this paper, indicate that the connection between the dissociation constant and constitution of an organic acid is not so direct as many would have us believe it to be. Professor Walker has recently stated (Trans., 1900, 77, 398) that the increase in the constant of succinic (or glutaric) acid when the hydrogens are successively replaced by alkyl groups, is not due to the primary effect of the substitution, but rather to an increase in the proximity of the two carboxyl groups. Adopting the Perkin-Bouveault formula for camphoric acid, which represents it as a succinic derivative, as follows:

he accounts for its small dissociation constant (0.0025) by saying that the rigidity of the ring formation prevents the approximation of the carboxyl radicles, and thus allows the primary effect of substitution to become manifest. The view that the proximity of the carboxyl radicles in succinic acid increases as the hydrogen is successively replaced by alkyl groups, is supported by the fact that the tendency to anhydride formation is more marked as the number of alkyl-substituted groups increases, but our results show that such considerations do not wholly explain the facts. Since our experience is that cis-s-di

propylsuccinic acid yields an anhydride with acetyl chloride if anything more readily than does the cis-s-diisopropyl acid, it can hardly be contended that the two carboxyl radicles in the latter are nearer each other than they are in the former, and yet the constant of the latter is nearly five times as great as that of the former. We hope shortly to publish, in conjunction with Dr. Sudborough, important evidence on these matters derived from a study of the methyl hydrogen salts of substituted succinic acids, but the facts recorded in our present communication are sufficient to show that the subject calls for much further investigation before any general principle can legitimately be deduced.

We desire to state that the expenses incurred in this research have been largely defrayed by a grant from the Research Fund of the Chemical Society.

THE OWENS COLLEGE,

MANCHESTER.

LVI.-The Interaction between Sulphites and Nitrites. By EDWARD DIVERS and TAMEMASA HAGA.

THE present paper gives an account of a series of experiments the results of which seem to leave no room for doubt as to the truth of the following propositions respecting the sulphonation of nitrites : (1) normal sulphites have no action on nitrites; (2) pyrosulphites react but partially with nitrites; (3) pyrosulphites react completely with nitrous acid or its equivalent of nitric oxide and nitric peroxide (nitrous fumes); and (4) sulphurous acid and nitrous acid, or the oxides and water equivalent to them, interact in such a way that the base of the sulphite, that may be used in place of the sulphurous acid, is needed only to preserve from hydrolysis the products of the interaction. Concerning these assertions, we would point out that

VOL. LXXVII.

3 A