SEA-FISHERIES LABORATORY. Lae 
Such are, indeed, described by Miss Williams,* but the results 
obviously require confirmation in view of the general statements 
that carbohydrates are absent, or are present in negligible quan- 
tities only, in fish flesh. 
It is also probable that the apparent deficiency in the 
analyses is to be traced to the employment of the particular 
factor, 6-25, by which the nitrogen actually found analytically 
is assumed to be converted into proteid. There has been 
much discussion as to the value of the factor which is appropri- 
ate to fish proteids, and there is a good resumé in Atwater’s 
paper} already noticed here. This investigator gives a list 
of analyses in which the propriety of using the factor 6-25 is 
doubtful, and he discusses whether or not the percentage of 
proteid present in fish flesh ought to be estimated by difference, 
or actually by determination of the nitrogen and multiplication 
of the latter by 6:25. It seems to me that the method of estim- 
mating proteid by difference—that is assuming that whatever 
is not water, nor ether-extract, nor non-volatile mineral matter 
—is “ proteid ’’—is good enough for dietetic purposes, but it 
is obvious, from a consideration of Atwater’s results, that it 
is not accurate enough for metabolic studies. 
Thus some of the percentages of water, fat-extract, proteid 
(that is, N x 6-25) and mineral matter in the Manx Summer 
herrings is only about 98 °%4. Winter herrings give more satis- 
factory results, the sum of the above percentages being 99-7 %. 
On the other hand analyses of red herrings gave an error in 
_ excess, the water, fat, proteid and ash adding up to 102-5 %. 
It was thought that the analysis of“ proteid ’’ was inaccurate 
here, and the estimation was repeated with the result that 
26-52% of (N x 6-25) (the first analysis), 26-45 and 26-52 were 
obtained for the percentages of (N x 6-25) in the dried residue. 
We have, therefore, a total range of about 24 % above and below 
* Loc. cit. Reducing substances reckoned as glucose vary between about 
1*5 and 9°5 of the net weight of cooked flesh. 
t Loc. cit. 
