170 REPORT—1904. 
Assumptions of a slightly different kind have been made to account for 
certain phenomena observed in quinquevalent nitrogen compounds ; it has 
been found necessary to suppose that the position occupied by the electro- 
negative radical is not always the same—in fact, that the resultant effect of 
the forces exerted by the nitrogen atom and four electro-positive radicals 
is such as to attract an electro-negative radical which may enter the 
molecule in one of two positions. 
Further, in order to account for the existence of one kind of isomerism 
observed among tervalent nitrogen compounds (p. 172), it has been sug- 
gested that for three groups attached to a nitrogen atom two configurations 
are possible. 
Stress must be laid on a difficulty encountered in the whole of this 
field—namely, the uncertainty as to the cause of certain differences which 
are observed in some compounds, particularly when those differences 
disappear on solution, whether they are to be attributed to isomerism 
of an unstable kind or merely to dimorphism : this difficulty is accentuated 
when the compounds exhibit no distinctive differences in their chemical 
reactions. 
It has been deemed advisable in dealing with this subject to treat the 
tervalent and quinquevalent nitrogen compounds separately, and in the 
latter case to present the facts completely before entering upon any dis- 
cussion of them, because it is necessary to review the whole field in 
attempting to offer a consistent explanation of the various phenomena. 
The time is certainly not yet ripe for drawing any final conclusions about 
this complicated subject, and there is still much work to be done before 
a clear path through the maze-will become evident. 
I. TervaALent NirroGen Compounps. 
(i) Compounds of the type N a be. 
(a) The Problem of Optical Activity.—It might be expected from purely 
dynamical considerations that the most stable configuration for three 
groups attached to one atom (nitrogen) would be that in which all three 
groups were situated in the same plane with that atom ; the distances of 
the groups from the centre of the nitrogen atom would be variable and 
dependent on the nature of the groups, but no displacement out of the 
plane would be expected so long as none of the groups was asymmetric. 
Such a displacement apparently occurs when strain is introduced, as in 
the formation of cyclic compounds or double linkages, and might occur if 
one of the groups became asymmetric. 
This view is supported by the facts, since all attempts to demonstrate 
asymmetry in tervalent nitrogen compounds—other than cyclic compounds 
—have hitherto been abortive. 
Krafft! tried to resolve ethyl-benzylamine and p-tolyl-hydrazine by 
the crystallisation of their neutral tartrates ; Behrend and Kénig ? made 
similar experiments with the tartrates and mandelates of 8 benzyl-hydro- 
xylamine and f nitro-benzyl-benzyl-hydroxylamine, and Ladenburg * on 
the acid tartrates of methylaniline, tetrahydroquinoline, and tetrahydro- 
pyridine. In all cases the salts appeared to be homogeneous, their 
properties remaining unaltered by recrystallisation. 
Regarded in the light of recent experience these experiments are 
1 Ber., 1890, 28, 2780. 2 Ann., 1891, 268, 175. 
3 Ber., 1893, 26, 864. 
