688 REPORT OF COMMISSIONER OF FISH AND FISHERIES. 
that the drying was conducted in the air the treatment of the samp] 
was the same as that described above. 
Water and water-free substance . — For the complete drying, from 
to 2 grammes of the partly dried material were weighed in small dryi 
bottles and dried in hydrogen for two periods of about 2 hours eat 
It is extremely difficult to get an absolutely constant weight, thou 
we find that the object is in most cases approximately attained in t 
above method. The total moisture and water-free substance are co 
puted from the partial drying in hydrogen and from the final dryii 
Examinations of the analytical details will show a number of ca* 
in which the percentage of water-free substance in the portions us . 
for the various determinations differs from those given in the wa 
determination. This is due to the fact that some time interver 
between the determinations, and that our experience shows that thou 
reasonable care was used to procure bottles with well-fitting gh 
stoppers to hold the partly dried material, yet the moisture in 1 
latter was likely to change on standing in the laboratory and tl 
neither the use of rubber stoppers nor that of paraffin on the gl; 
stoppers sufficed to prevent this so completely as to make us feel s 
in using the substance after it had stood for some days, without re 
termination of the water. 
It is only just to say that although no little pains has been taken 
this laboratory to learn how to make accurate determinations of wa 
in animal and vegetable substance, we are far from satisfied with 
success of our efforts. Indeed, I am inclined to regard this as one 
the decidedly difficult determinations, so far as accuracy is concern 
NITROGEN, PROTEIN, ALBUMINOIDS, ETC. 
The nitrogen was determined in the partly dried substance (part 
drying in hydrogen) by the soda-lime method. 
It is customary to compute the albuminoids or protein (nitrogenc 
substances) by multiplying the nitrogen by 6.25. In our analyses tlu 
factor has, in general, appeared to be very nearly correct. I ha 
thought it best, for the present purpose, to state the percentages 
nitrogen and of protein as calculated by multiplying these by 6.25, a 
also to estimate the u albuminoids, etc.” by difference. For the lath 
the remainder left after subtracting the sum of ether extract and a 
from the water-free substance, or the sum of water, ether extract, a 
ash, from the fresh substance, is used. This is not absolutely correct, 
but is more nearly so than the product of nitrogen by 6.25 would be. 
The importance of correct estimation of the nitrogenous matters h < 
us to spend a not inconsiderable amount of time in the study of the 
sources of error and means of avoiding them, especially in the use o i 
the soda-lime method for determining nitrogen in animal tissues. 
Sources of error in the soda-lime method and means for avoiding them. 
As the investigations of these questions have been published in d 
