CHEMICAL COMPOSITION OF FOOl3Cf r fsftES. 
753 
The scales were thoroughly removed from the fresh fish, their heads cut off, and 
the entrails taken out. The meat and skin were then separated as well as possible 
from the backbone, tail, and bony fins. The whole flesh thus obtained was cut fine, 
aud carefully rubbed in a mortar until it formed a homogeneous mass, which was 
used for the analyses. 
The water was determined in the flesh by heating 15 or 20 grammes, 
first at a lower temperature in a water bath and then at 1 10° O., to a con- 
stant weight. As the flesh grew horfiv and hard in drying it was from 
time to time stirred and pulverized with a glass rod or pestle, and thus 
ground fine so as to allow the escape of all of the wate. Alrnen insists 
upon the necessity of thorough and complete drying, and believes that 
it was accomplished by this means. 
The ash was determined by burning the dried substance in air. 
When necessary to avoid fusion of the alkaline salts and consequent 
imperfect ignition, the mass was first charred and then extracted with 
water. Where the ignition was practicable without this previous 
charring and extraction the ash was treated with water, and the solu- 
ble and insoluble salts weighed separately. Where the material was 
first charred the soluble salts were then extracted and weighed and the 
insoluble salts afterwards determined. 
The chlorine was determined in the salts by a titration with a deci- 
normal solution of silver nitrate. Almen remarks that— 
In the fresh fish the chlorine occurs principally as potassium chloride ; in the salt 
fish almost exclusively as sodium chloride. In the former the content of chlorine as 
shown by the analyses must be regarded as correct, while for the latter the amount 
of sodium chloride must be taken into account in order to determine the amount of 
salt used in salting the fish. A small portion of the chlorides may have been vola- 
tilized in the ignition, and the results of the determinations of ash aud of chlorine 
thus rendered somewhat incorrect. Since, however, the method was the same in all 
of the cases, the figures that are given may be accepted as fair indications of the 
relative amounts actually present. 
The amount of fat was determined by extracting the dried aud pulverized flesh 
with ether to which a small quantity of 97 per cent, alcohol had been added. The 
ethereal solution was dried in a water-bath to constant weight. The extraction was 
always complete and the errors can not be important. 
The nitrogen determinations were made by burning w tli soda lime. 
For this purpose 3 or 4 grammes of the finely divided and homogene- 
ously mixed flesh were dried in a water bath. It was u very finely 
pulverized and carefully mixed with the soda lime. 
The combustions were conducted first in glass tubes aud afterwards in porcelain 
tubes, in which latter the operation went on more quietly. The ammonia was 
caught in a standard solution of sulphuric acid of normal strength and titrated 
against soda solution of one-third normal streugth. In the first combustions, colored 
decomposition products appeared, imparting to the sulphuric acid a light-reddish 
color which made the titration somewhat uncertain. The litmus color was, never- 
theless, predominant and the limits of error in the titration could not have exceeded 
0.2 per cent, or at most 0.4 per cent, of the soda solution, which would correspond to 
from 1 to 2 milligrammes of nitrogen, which was of no importance for these investi- 
gations in view of their practical purpose. I succeeded, however, ft, voiding these 
colored combustion products by filling the anterior part of the tube with coarse soda 
H. Mis. 274 48 
