CHEMICAL COMPOSITION OF FOOD-FISH 
) 
in the discussion of the constitution* of the flesh of fi r 
place, as will Christison’s statement regarding the com 1 
and foul salmon.* 
Various other chemists have studied the composition of tin flesh and 
other organs of fishes, but the purposes and methods have been such 
that the results need not be referred to in detail here. I should, how- 
ever, call attention to an article by Dr. Davy,t in which the attempt is 
made to compare the nutritive values of different fishes by their per- 
centages of dry substance. 
PROPORTIONS OF PROTEIN IN THE FLESH OF FISHES. 
In the recapitulations of European analyses of the flesh of fishes the 
percentage of protein as estimated by multiplying the percentages of 
nitrogen by the factors 6.25, the percentages of u albuminoids” as esti- 
mated by difference, and the footings, water-f protein + fats -fash, are 
given where the data permit. One object in this has been to obtain a 
crude test of the correctness of the analyses, it being assumed that a 
wide divergence between the figures for nitrogenous material by the 
two methods or, what is the same thing, a wide divergence between the 
sums named and 100, would indicate error of analysis. Doubtless some 
of the wider divergences in the analyses here cited are due to typo- 
graphical errors. But considering the fact that even the directions 
for determining nitrogen by the soda-lime method as found in standard 
works on quantitative analysis are such as to involve danger of serious 
error, and that numerous analysts in following them have been unable to 
obtain reliable results, it would not be strange if errors in the nitrogen 
determinations had crept into some of the earlier analyses. My experi- 
ence has also warned me that the dangers of error in the determinations 
of water and fats are greater than is sometimes supposed. 
The object of these remarks is not to criticise the most valuable 
work of the investigators whose results are quoted, but rather to 
enforce the need of improvement of the analytical methods. Meanwhile 
the figures for albuminoids by difference perhaps represent the actual 
quantities of nitrogenous material more nearly than those obtained by 
multiplying the nitrogen by 6.25. They are used in th^ f ables beyond, 
in which the analyses are summarized. 
* Cited by Smith, Foods, p. 108. 
t Edinburgh New Philosophical Journal, Oct. 1853, p, 225 ; Dingle 
131, 1854, p. 390. 
Mis. 274 49 
