The effect of the substitution of an alkyl group for hydrogen varies, as we have seen, with the nature of the acid. Such a substitution causes always a decrease of the first dissociation constant in acids of the acetic series and in malonic acid, but an increase in the case of monoalkylated malonic acids and in succinic acid. We must either assume that no specific effect can be assigned to an alkyl group, or we must attribute these results to a secondary action. Such a secondary action is the probable alteration in the relative positions of the two carboxyl groups, caused by the substitution of alkyl groups. Of two dibasic acids in which the two carboxyl groups occupy different relative positions, we should expect that one to give an anhydride the more readily in which the carboxyls are the nearer together; but from their experiments on anhydride formation, Bone and Sprankling1 concluded that the carboxyl groups of all dialkylated succinic acids are in the same relative position. It must, however, be remembered that such a rough method of measurement as this cannot be expected to give information as to the small changes of distance which, since the power of forming ions is such a highly constitutive property, would be sufficient to cause the variations observed.

It appears probable that the real explanation of the apparently irregular action of the alkyl groups is, that the substitution of these groups for hydrogen produces changes -perhaps small-in the configuration of the molecule of the acid, the nature of which depends on the alkyl groups and on the nature of the acid. This effect is superimposed on the specific effect of the alkyl groups, which would be observed if the substitution caused no change in configuration, and which varies with the nature of the group, and its distance from the carboxyl group. There is, of course, no reason to suppose that the action of an alkyl group is essentially

1 Loc. cit. See also a later paper by Bone, Sudborough, and Sprankling (J.C.S., 85. p. 534 (1904), in which the conclusions arrived at are different from those described in the text.

different from that of any other substituent; but owing to the comparatively large specific volumes of these groups, the changes in configuration they cause are probably greater than those caused by such substituents as Cl or OH.

Halogen and other substitution-products of dibasic acids. Monochloromalonic acid HOOC.CHCI.COOH Monochlorosuccinic acid HOOC.CH.CHCI.COOH

K1 X 10"

4'0

Walden.

0.28

Racemic acid

Mesotartaric acid

Tetrahydroxysuccinic acid HOOC.C(OH)2.C(OH), COOH 124 Skinner.

The effect of substitution of chlorine and of hydroxyl is seen to be similar to their effects in the monobasic acids. Chlorine has a smaller effect in malonic than in succinic acid, and the effect of chlorine in acetic acid is greater than in either.

As is required from the theory of stereo-isomerism, the constants of lævo- and dextro-tartaric acids are the same. Since the constant of racemic acid is equal to that for 1- and d-tartaric acids, the solution of racemic acid must contain a mixture, and not a compound of these two acids.

The constants are seen to be greater than those of the saturated acids to which these acids correspond.

Isomers of this type show plainly the connection between the distance between the carboxyl groups and the values for the ionisation - constants. Comparing analogous compounds, K1 is the greater, and K, the smaller, the nearer the

two carboxyls in the molecule. The introduction of methyl diminishes both K, and K2, but the decrease of K, is smaller for maleïc than for fumaric, while the decrease of K, is smaller for fumaric. This may be explained by assuming that in addition to its specific effect, the methyl group causes the distance between the carboxyl groups to increase, and that the increase is greater for maleïc than for fumaric acid.

As in other cases, the effect is greatest when the substitution takes place in the ortho position. It is seen that the usual

relation between K, and K, again appears.

It has been proved (p. 114) that the ionisation-constant K1 of a dibasic acid is the sum of the ionisation-constants corresponding to the two carboxyl groups. If, now, on esterifying one of the carboxyl groups, the effect is simply to take away the power of giving ions from one carboxyl group, and no change is produced in the ionising power of the other carboxyl group, the ionisation-constant of the ester - acid should be one-half the value of K, for the acid itself, if the acid is symmetrical. And the ionisation-constant (K,) of an unsymmetrical dibasic acid should be the sum of the ionisation-constants of the two ester-acids derived from it.

The following numbers are from Walker's measurements:

1

Ratio of diss. constant of acid to that

of ester-acid.

3'6

2.2

1

Leaving out the acids with large ionisation - constants (K1 × 105 greater than o‘1), we find that the ratio given in the third column varies between 2 and 2'5. If the substitution of ethyl for hydrogen had no effect except that of taking the power of forming ions away from the COOH group, the ratio would be in each case 2, since all these acids are symmetrical. The following numbers for unsymmetrical acids are from the measurements of Wegscheider :—1

1

The mean of the ratios obtained by dividing the sum of the ionisation-constants of the ester-acids by the value of K1 for the acid is, for methyl acid-esters, 1'07, and for propyl acidesters (one example) o'97. This ratio would be unity, if the alkyl groups themselves had no effect on the ionisation

constants.

Organic Tribasic Acids.-Three stages of ionisation are possible, e.g.

It has been found that the ionisation-constants calculated on the assumption that the ionisation takes place only according to the first equation, show no irregularity until large dilutions are reached. Thus, the second and third ionisation-constants must be small, compared with the first ionisation-constant.

For citric acid the three constants have been approximately determined :— 2

1 Monatshefte für Chemie, 23. 316 (1902).

2 W. H. Smith, loc. cit.

[It is to be noted that the mode of ionisation of tribasic acids is probably very complicated. An acid such as citric acid can undergo the first stage of ionisation in two ways:—

The anion produced according to the first of these equations could undergo the second ionisation in two ways, giving as anions,

The anion produced according to the second equation could undergo the second stage of ionisation in only one way, giving an anion

CH2.COO'

CH.COO'.

CH.COOH

Then there would be two kinds of anion in solution capable of undergoing the third stage of ionisation. The case would be still more complicated with an unsymmetrical acid.]

The following rules have been found to apply to the