Dr Fenton, A Delicate Reaction for Carbohydrates. 
25 
ting the crystals of bromo-methylfurfural and estimating their 
amount. In this way it was shewn that ketohexoses and cellulose 
yield about 20 or 30 per cent, of the crystalline bromo-compound, 
whereas the aldose sugars give only about 2 or 3 per cent, of a 
residue which is uncrystallisable. It was previously stated [Trans. 
Chem. Soc., 1901, 363] that if any of the bromo-compound is 
formed by aldoses, its quantity is too small to be detected. By 
the method described in the present communication however, it 
can be shewn that bromo-methylfurfural is actually formed, 
although in small quantity, even in the case of aldose sugars. 
It is not improbable that this small quantity results from a partial 
transformation of aldose to ketose under the influence of the 
strong acid employed. [Compare Ost. Zeit. angetu. Chem., 1905 (18), 
1170.] 
By taking advantage of the highly sensitive reaction with 
malonic ester, which was mentioned above, one may detect the 
merest trace of bromo-methylfurfural and this test may conse- 
quently serve for the identification of all hexoses, whether aldose 
or ketose. 
The mode of operating is as follows : A minute quantity of the 
solid substance to be examined is slightly moistened with water, 
mixed with a drop or two of phosphorus tribromide * gradually 
heated on a water-bath to 90 — 100°, and kept at this temperature 
until the mixture has turned dark coloured. It is then cooled, 
stirred with a little alcohol and a few drops of malonic ester ; 
alcoholic potash is then added until the solution is alkaline. On 
now diluting the resulting mixture with a large volume of water, 
or alcohol, the fluorescence is at once apparent. 
The reaction is strongly marked even with the aldoses, but is, 
as might be expected, more intense with ketose sugars and 
with cellulose. In comparison with the other pure hexoses 
examined, galactose appears to give the weakest indications ; 
this fact is perhaps connected with the peculiar configuration 
of galactose. [Compare Morrell and Crofts, Trans. Chem. Soc., 
1902, 6G7, and Fischer, Ber., 1891?, 2031.] 
A saturated aqueous solution of hydrogen bromide may be 
employed in this test instead of phosphorus tri bromide ; the 
latter is, however, preferable in the case of substances which 
are less readily attacked, such as cellulose. The chlorides of 
phosphorus may also be used, but their action is perhaps less 
certain. 
Positive results were obtained with dextrose, la 3 vulose, ga- 
lactose, sorbose, cane sugar, maltose, lactose, rafAnose, starches, 
* It is best to dissolve the tribromide in some inert solvent of sufficiently high 
boiling point, such as toluene. 
