has a pleasant odour, and is hydrolysed almost quantitatively by boiling concentrated aqueous potash to the corresponding tetramethyleneCH2 CH2

monocarboxylic acid,

CH2CH COOH'

4.3 grams of cyanotetramethylene were boiled with 6 grams of potash dissolved in 20 c.c. of water for 6-7 hours in a reflux apparatus. When the evolution of ammonia had ceased, the liquid was allowed to cool, and then acidified with dilute sulphuric acid; the colourless oil which separated was extracted with ether in the usual way, and, after drying over anhydrous sodium sulphate and distilling off the ether, the residual liquid, on fractionation, passed over between 190° and 193°. This was again distilled, and the fraction 191-192° analysed.

0.2842 gave 0.6234 CO, and 0.2060 H2O. C-59·82; H=8·05. CHO, requires C=60·00; H=8.00 per cent.

The characteristic calcium salt of this acid was prepared; it is very soluble in water, and crystallises from very concentrated solutions in silky needles.

Preparation of Pimelic Acid from the Bye-product of the Reaction between Trimethylene Dibromide and Ethylic Sodiocyanacetate.

The thick oil, which is left in the steam distillation flask, after distilling off the ethylic cyanacetate and cyanotetramethylenecarboxylate, and does not solidify on cooling, probably consists of diethylic aa1-dicyanopimelate,

COOC,H•CH(CN)•CH, CH, CH, CH(CN)-COOC,H.

2

It was extracted with ether, the ethereal solution dried over calcium chloride, and the ether distilled off; the residual oil was then boiled for about 20 hours in a reflux apparatus with concentrated aqueous potash. At the end of this time, complete hydrolysis had taken place, and the evolution of ammonia ceased. The liquid was acidified with sulphuric acid, and, after saturating with ammonium sulphate, extracted with ether; the ethereal solution was dried over anhydrous sodium sulphate and the ether distilled off, when an oil was left which was heated in a bath of fusible metal to 210° until the evolution of carbon dioxide had ceased. The dark coloured residue, which showed no signs of crystallising, was etherified by heating with 5 times its volume of absolute alcohol and one-fifth its volume of concentrated sulphuric acid for 2 hours in a reflux apparatus. The product was poured into water, extracted with ether, washed with sodium carbonate solution, dried, and the ether distilled off. On fractionating the residue at 40 mm. pressure, almost the

whole quantity distilled between 160° and 180°. This fraction was hydrolysed by boiling with methyl alcoholic potash, and after complete removal of the alcohol by evaporation with water, the residue was acidified and extracted with ether in the usual way. On allowing the ethereal solution to evaporate in a vacuum, an acid crystallised out which melted between 97° and 102°, and after recrystallising from benzene was finally obtained in a pure condition; it melted at 105°, and consisted of pure pimelic acid, as the following analysis shows:

2

0.1429 gave 0.2775 CO2 and 0·1004 H2O. C-52-33; H=7·80. C-H1204 requires C=52.50; H=7.50 per cent.

This acid gave a calcium salt, which separated from its cold, saturated solution on heating, a behaviour which is also shown by the calcium salt of normal pimelic acid.

The authors wish to reserve the further examination of the sub stances described in this paper.

OWENS COLLEGE,

MANCHESTER.

XCI.-Influence of Substitution on Specific Rotation in the Bornylamine Series.

By MARTIN ONSLOW FORSTER, Ph.D., D.Sc.

ALTHOUGH recent years have witnessed the rapid multiplication of optically active substances, much remains to be learned respecting the influence of substitution on specific rotation before the guiding principles of the subject can be properly established. The theory propounded in 1890 almost simultaneously by Guye and Crum Brown, attracted considerable attention at the time, and has exerted a marked influence on the character of subsequent investigations. Some among the latter have seemed to afford confirmation of the theory, and more particularly of that deduction from it which predicts the occurrence of a point of maximum rotation in a homologous series (compare Frankland and MacGregor, Trans., 1893, 63, 1417, and 1894, 65, 756; Guye and Chavanne, Compt. rend., 1893, 116, 1454; 1894, 119, 906, and 1895, 120, 452). The general tendency of recent work, however, has been unfavourable to the original conception. Numerous cases of disagreement between results predicted and those found by experiment have led inevitably to a fuller recognition of influences exerted by the qualitative nature of substituent groups, influences foreshadowed by Crum Brown, and latterly

verified experimentally by other workers (compare Frankland and MacGregor, Trans., 1896, 69, 119, and Frankland, this vol., 347).

In the hope of gaining some information on this subject, I have prepared several alkyl derivatives of bornylamine, and determined their rotatory power in the liquid and dissolved states. Bornylamine was chosen because, in the first place, although numerous homologous series of optically active ethereal salts and alkyl oxides have been examined, the effect of replacing the aminic hydrogen of an optically active base by alkyl groups does not appear to have been studied. Secondly, it was recognised that a primary base, in which the aminogroup is attached to asymmetric carbon, offers two points of attack in one of the four groups causing asymmetry, instead of the one occurring in hydroxylic and carboxylic groups. An additional advantage which bornylamine offers may be found in the fact that it should give rise to one series of derivatives in which asymmetric nitrogen occurs in the trivalent condition, and a second containing quinquevalent asymmetric nitrogen.

Two objections to the choice of this substance, however, present themselves. In the first place, it contains two asymmetric carbon atoms besides that to which the amino-group is directly attached, and, secondly, this carbon atom forms part of a ring system, and consequently the "product of asymmetry" cannot be calculated by Guye's formula. On the other hand, active primary bases are not very numerous, and, in general, are not very readily obtained, whereas bornylamine can now be prepared without much difficulty (Forster, Trans., 1898, 73, 390).

The

At the outset of this investigation, much time was occupied in attempting to prepare methylbornylamine, and a suitable process for obtaining the compound was not discovered until the study of the other derivatives enumerated in this paper had been concluded. cause of the difficulty lies in the fact that the action which methylic iodide exerts on the primary base is sufficiently vigorous to give rise to the dimethyl derivative in a cold ethereal solution, a large proportion of bornylamine being precipitated from the liquid in the form of hydriodide. Even when the more usual method of heating the base with methylic iodide and alcoholic potash in a reflux apparatus is adopted, it is found that the methyl and dimethyl derivatives are produced in approximately equal quantities, along with a considerable amount of trimethylbornylammonium iodide. A fresh obstacle is encountered on isolating methylbornylamine from the basic mixture by conversion into the nitrosamine, inasmuch as this method cannot be relied upon to yield a base having the maximum rotation. So misleading, in fact, is the result of using this process, that at one

time I regarded ethylbornylamine as having greater specific rotatory power than the lower homologue (Proc., 1899, 15, 71).

By having recourse to a method which avoids both direct methylation and treatment with nitrous acid, it is possible to obtain highly purified methylbornylamine without much loss of material. The procedure is as follows. Bornylamine is first condensed with benzaldehyde, yielding benzylidenebornylamine, which melts at 57°; this compound is then heated with methylic iodide, which converts it into an unstable methiodide, melting at 215°. On attempting to recrystallise this derivative from a solvent containing water, it is resolved almost quantitatively into benzaldehyde and methylbornylamine hydriodide, from which the base may be liberated in the usual way.

The remaining alkyl derivatives of bornylamine are readily obtained. The nitrosamine method of separating the secondary from the tertiary bases may be applied in the case of ethylbornylamine without impairing the specific rotatory power of the product. At this point the action of alkyl iodides sustains a sharp check; monalkyl bases are still obtained without difficulty, but no production of the corresponding dialkyl derivative occurs under ordinary conditions. The following table summarises the information which has been gained respecting the rotatory power of bornylamine and its alkyl derivatives.

From this it appears that the introduction of a single methyl group produces a striking increase of rotatory power. The second column of figures will show that a maximum in the molecular rotation occurs at the third term of the homologous alkylbornylamines, which form a series in this respect quite similar to those examined by Guye and Chavanne, Frankland and MacGregor, Tschúgaeff, and others. The first column indicates, however, that the specific rotatory power of these compounds undergoes a regular decline from the first member of the series; groups such as this, in which the maximum occurs at the first term, are not common, but the phenomenon has been observed

by Guye and Chavanne in the valeric series (Compt. rend., 1893, 116, 1454), by Purdie and Williamson in studying the ethereal salts of methoxy- and ethoxy-succinic acids (Trans., 1895, 67, 957), and by Frankland and MacGregor among the dibenzoylglycerates (Trans., 1896, 69, 118).

In view of the suggestion that the true optical behaviour of initial members of such series may be possibly masked by molecular association (Frankland, this vol., 347), the molecular volumes of the secondary bases, experimentally determined, are compared in the following table with the values calculated by Traube's method (Ber., 1895, 28, 2724).

Although the higher homologues are not associated, these data suggest the possibility that molecular association may occur to a slight extent in the case of the methyl derivative.

A more novel point presents itself on comparing the specific rotatory power of bornylamine and its monalkyl derivatives with that of the corresponding tertiary bases. Reverting to the table on p. 936, it will be noticed that the effect produced on the rotation of bornylamine by introducing a single alkyl radicle into the amino-group, compared with the result of replacing both atoms of hydrogen, is very considerable. If it can be shown that other groups of alkylated amines exhibit the same characteristic, the phenomenon should have some bearing on a discussion of the influence exerted by substitution upon rotatory power. For it will be recognised that, in the series under consideration, it is destruction of molecular symmetry, and not increase of mass, which produces the more marked effects on the optical activity. Thus, on comparing the specific rotatory power of ethylbornylamine and dimethylbornylamine with that of the primary base in benzene, it is found that although the radicle attached to asymmetric carbon undergoes in each instance a change in mass from 16 units to 44, the specific rotatory power is raised through only 2·5° in the one case, and through 33.2° in the other.