Mr Fenton , Oxidation in Presence of Iron. 371 
and a little caustic alkali a beautiful but transient violet colour. 
Ferric chloride gives a deep red colour, and cupric acetate an 
olive-green precipitate. 
This compound is identical in every respect with the ‘secondary’ 
dioximido-propionic acid CH (NOH) . C (NOH) . COOH which 
Soderbaum obtained from dibromopyruvic acid. 
Oxidation with cupric hydroxide. 
The oxidation -product mentioned above when further oxidized 
with cupric hydroxide, in presence of alkalies, gives a large yield 
of mesoxalic acid. 
Taking this result together with those above described con- 
clusive evidence is afforded as to the constitution of the product — 
that it is the semi-aldehyde of mesoxalic acid. 
The formation of this substance from tartaric acid under the 
conditions mentioned might be explained in several ways. At 
first it was thought probable that dioxytartaric acid was the 
initial product, and that this gave rise to mesoxalic semi-aldehyde 
with loss of carbon dioxide and water. But all attempts to 
prepare the substance from pure dioxytartaric acid, by heat, 
action of acids or of iron salts, were unsuccessful or unsatisfactory. 
It was observed however that the product of the action of chlorine 
on tartaric acid in presence of iron gives, in the first instance, 
a notable reaction of dioxymaleic acid with iron salts, and it 
seemed probable that the latter acid might be the initial product 
in formation of the semi-aldehyde. Starting upon this assumption 
attempts were made to prepare the semi-aldehyde directly from 
pure dioxymaleic acid, and these eventually were completely 
successful. 
When a solution of dioxymaleic acid is treated with a ferric 
salt, such as ferric sulphate, it is almost quantitatively converted 
into mesoxalic semi-aldehyde and carbon dioxide, 
C 4 H 4 0 6 + Fe 2 (S0 4 ) 3 = C 3 H 2 0 4 + 2FeS0 4 + H 2 S0 4 + C0 2 . 
After removal of the iron salt and free numeral acid the 
aldehyde remains as a thick syrup. It has not yet been induced 
to crystallize but experiments are being made in this direction. 
The typical oxidation-products of glycerol which are theoreti- 
cally possible are eleven in number, namely — glyceric, hydroxy- 
pyruvic, tartronic and mesoxalic acids, glyceraldehyde, dihy- 
droxy-acetone, tartronic semi- and di-aldehydes, hydroxy-pyruvic 
aldehyde, and mesoxalic semi- and di-aldehydes. Of these the 
first six are well known ; mesoxalic di-aldehyde is known only in 
form of oxim, and the remaining four were hitherto unknown. 
The isolation of the semi-aldehyde of mesoxalic acid therefore 
brings the series one stage nearer completion. 
VOL. XI. PT. V. 
27 
