ON THE STEREOCHEMISTRY OF NITROGEN. 171 
inconclusive, since similar attempts failed to resolve quinquevalent nitrogen 
compounds. The failure might have been due either to hydrolytic dissocia- 
tion of these salts of weak acids with weak bases, or to partial racemism, 
which the recent work of Professor Kipping has shown to occur so 
frequently. 
Reychler ! eliminated the first-mentioned possible cause of failure in 
his attempts to resolve methyl-ethyl @ naphthylamine by fractional 
erystallisation of its dextro-camphorsulphonate from non-hydroxylic 
solvents ; the writer also made a similar attempt with methyl-benzy, 
aniline, but no resolution was effected in either case. 
The evidence derived from all the foregoing experiments (with the 
exception of that in which p-tolyl-hydrazine was used) is useless if the 
change of valency direction established in the case of sulphur and selenium 
also occurs in nitrogen during the change from ter- to quinquevalency. 
There is, however, no evidence of this, and the ease with which cyclic 
compounds, such as pyridine and piperidine, form quaternary ammonium 
derivatives is not in favour of this view. 
On this account some experiments were made by Mr. J. P. Millington 
and the writer? with the object of obtaining evidence free from this 
objection, by avoiding a change in the valency of the nitrogen during the 
process of resolution. Benzyl-phenyl-hydrazine d-camphorsulphonate and 
the brucine salt of methyl-ethyl aniline sulphonic acid were submitted 
to fractional crystallisation, but without effecting any separation into 
fractions of different rotatory power ; in the first case it is probable that 
no change of valency of the nitrogen atom in question takes place, and in 
the second it is certain. 
Evidence of quite a different kind has been adduced by Messrs. 
Kipping and Salway.* It was proved that when an externally compen- 
sated acid chloride reacted with an externally compensated primary amine 
(each containing an asymmetric carbon atom) to form a substituted amide, 
this consisted of four compounds, which being enantiomorphously related 
in pairs give rise to two externally compensated compounds easily separable 
by fractional crystallisation. This, therefore, forms a method of testing 
for asymmetry in carbon compounds, and by analogy it should be capable 
of detecting asymmetry in nitrogen compounds. The products of the 
interaction of dl-benzyl-methyl-acetyl chloride with methyl aniline, 
p-toluidine, pheny]-hydrazine and benzy] aniline were examined and found 
to be homogeneous. The active (d) acid chloride was also allowed to react 
with p-toluidine and with benzyl aniline ; but, as before, the product 
appeared to be a chemical individual. Finally, the active chloride was 
treated with an active amine ; if the nitrogen atom form a centre of 
asymmetry, in addition to the two already present, then two compounds 
should be formed ; but again the product was homogeneous. 
Clearly, therefore, all the evidence is in favour of a plane configuration 
for tervalent nitrogen compounds. 
(b) Isomeric Tervalent Nitrogen Derivatives N a 6 c.—Although the 
tervalent nitrogen atom is incapable of giving rise to optical activity, it 
seems capable of giving rise to another kind of isomerism of which there 
are two distinct examples. 
' Bull. Soc. Chem., 1902[3], 27, 979. 
2 Proc. Camb. Phil. Soc., 1904, xii. 489. 
3 Trans. Chem. Soc., 1904, 85, 438. 
