210 Transactions of the Boyal Society of South Africa. 
The solution of the chromate in acid also turns green on dilution, but this- 
is due to simple re-precipitation of the blue substance in fine division. 
The dirty-violet solution, obtained on allowing the diimine salt solu- 
tions to stand, gives with alkali a dark violet base which is an aminoazo 
dye. The substance has not yet been properly investigated, but in all 
probabihty it is the dimolecular polymer of the diimine suspected by 
Willstatter, viz., azodixenylamine or diamino-azobisdiphenyl : — 
It can be made direct from benzidine acetate and ammonium persulphate 
or by insolating a mixture of benzidine and cupric acetates. Its ferro- 
cyanide is crystalline, almost black with bronze lustre, 0-3349 gave 
0-0286 Fe,03 - Fe = 5-98 . IC^.ll^o'^^ + H.Fcy requires 5-92 per cent. 
The HCl-salt of this dye appears to be the product of the action of 
aqueous FeCl^ on neutral benzidine solution : very dark violet substance 
sparingly soluble in water. Found 7-1 per cent. CI. The dye is quite 
different from Willstatter's diimine, and gives with H2SO4 a very dark 
olive coloration. 
One remarkable observation has recently been made, which throws, 
doubt on the possibility of the existence of the free diimine. When a 
freshly-prepared solution of the diimine nitrate is made faintly alkaline 
with ammonia, the brown precipitate obtained gives, immediately after 
washing, a yellowish olive coloration with II2SO4 ; but on treatment with 
boiling water it deepens in colour and then gives the normal crimson 
shade. Whilst it is possible that the fresh substance is merely a highly 
hydrated diimine, it is more likely that it is the true diimine and that 
the substance giving the crimson colour (Willstatter's compound) is an 
isomeric or polymeric variety. It might be a simpler explanation of the 
whole subject to suppose that the yellow nitrate belongs to the true diimine 
series, whereas the blue solution formed by diluting it, and also the blue 
chromate and ferricyanide, are quinhydronic in character, intermediate 
between benzidine and the diimine. The analytical figures agree with 
either assumption. 
Two other substances of the class have been made by adding CrO^ 
to solutions of salts of 5-5' dibromo-3-3' tolidine and 3-3' dibromobenzidine 
respectively. Both resemble indigo but are amorphous, and their ana- 
lytical figures were unsatisfactory. The first gave Br = 23*3, Cr = 17-05 
per cent., and the second, Br = 37*2, Cr = 5-2 per cent. 
NH, 
N : N — < 
0-0 
— NH^. 
