4-NITRO-5-METHYL-2-SULPHOBENZOIC ACID 
209 
dissolved in the crystal. Ordinary recrystallization does not re- 
move the color. The crystals must be ground fine with water and 
allowed to stand with it for a long time in order to extract the 
phthalein. When this is done, they may be recrystallized from 
glacial acetic acid followed by alcohol, and are colorless, with a 
melting point of 206°. They contain no chlorine. These crystals 
were not formed in the Schott en-Baumann reaction. No sul- 
phone chloride was formed by any of these three methods, a fact 
which was due in all probability to the temperature employed. 
The Diphenyl Ester. 
NO. . CHg • CeH, . COOCeHg * SO.OCeH^ 
The product melting at 123° which is formed in all of the 
precedii\g methods has been shown to be the diphenyl ester. The 
melting point of the crude product, is, according to our expe- 
rience, always 115° or 116°, but by recrystallization from alco- 
hol, benzene, alcohol, glacial acetic acid, and alcohol in the 
order named, the melting point becomes constant at 123°. The 
products which were obtained separately from the symmetrical 
and unsymmetrical acid chlorides, are identical in this respect 
as in all others. The pure substance was analyzed for nitrogen 
by the Gunning method, and for sulphur by the Liebig method. 
.2066 g. substance required 4.65 cc .1088N HCl. 
.2059 g. substance required 4.69 cc .1088N HCl- 
.2023 g. substance gave .1144 g. barium sulphate. 
.1631 g. substance gave .0933 g. barium sulphate. 
Theory for C 20 H 15 O 7 NS Pound 
Nitrogen, 3.39 per cent. 3.42 - 3.47 per cent. 
Sulphur, 7.75 per cent. 7.74-7.86 per cent. 
The Product Melting at 206 °. 
The second product which was produced by the action of 
phenol in the presence of pyridine has not been identified pos- 
itively. It gives the odor of pyridine on treatment with strong 
potassium hydroxide, and may be a pyridine salt of the phenyl 
benzoic ester of the sulphonic acid. Nitrogen, where pyridine 
groups are present, cannot be determined by any modification of 
the Kjeldahl method, as the results always are low, but the 
degree of loss in such case is fairly constant. If due allowance is 
made for this factor, the results vary only a few tenths of a 
per cent. 
