SECTION IIL., 1895. [19] Trans. R. 8. C. 
[T—On the Estimation of Starch. 
By THomas MACFARLANE. 
~~ 
(Read May 15, 1895.) 
To determine the quantity of starch contained in cereals, food pro- 
ducts or feeding stuffs has always been regarded as a tedious and uncertain 
operation. Indeed, so much has this been the case that, in by far the 
greater number of analyses of these substances, the starch has not been 
estimated but is included in the percentage of non-nitrogenous extractive 
or carbo-hydrates. This figure is obtained by calculating the difference 
between 100 and thesum of the constituents directly determined ; namely, 
water, proteids, fat, fibre and ash. This remark applies to the completest 
of the analyses collected together in Kénig’s “ Chemische Zusammenset- 
zung der menschlichen Nahrungs und Genussmittel,” and to the work of 
Grandeau, Maercker, Clifford Richardson and many other investigators. 
These analyses embrace all cereals, leguminous fruits, oleaginous seeds, in 
~hort every grain and seed species which is used for food ; flour and 
bread of every description, prepared flours, infants and invalids foods, 
tubers, roots, vegetables, ete. 
The substances thus included in the term non-nitrogenous extractive 
or carbo-hydrates are quite considerable in number, but not by any means 
similar in properties. ‘The terms mentioned may, but do not always, 
include starch, gums or pentosans, dextrine, cane sugar, grape sugar, 
pectin, lichenin, bitter principles and colouring matters. In order to get a 
correct estimation of the starch, most of the last named substances require 
to be previously removed which is always tedious and not always possible. 
The difficulties of the problem have been set forth by Professor Stone, in 
an article contributed to the Journal of the American Chemical Society 
(1894, p. 726) who, in introducing it, says: “ There is no longer much 
“doubt among chemists that, in food analysis, the present practice of 
“ classifying a large number of widely varying substances under the head 
“ of non-nitrogenous extract matter, as a homogeneous material, is wholly 
‘erroneous and misleading.” 
In the same article Professor Stone gives the experimental results of 
estimating the starch in a mixture of starch, sugar and dextrine, in cotton 
seed meal, wheat middlings, hay, wheat, bran, cornmeal, wheatflour, dried 
potato and pure potato starch. There were five different methods em- 
ployed, all of which gave results closely agreeing in the case of the pure 
potato starch, but more or less discordant in the flours and feeding stuffs 
on account of the presence in these of some of the substances other than 
starch above mentioned. Except in the case of asbestus method the pro- 
cesses followed were based upon inversion by acids; the starch is con- 
