84 Part III.—Twenty-fourth Annual Report 
in the U. S. Commissioner’s Report on Fish and Fisheries. 1888, Part xvi., 
1892). This report deals with the chemical composition and nutritive 
values of food fishes and aquatic invertebrates. Analyses were made of 
the flesh of 123 specimens of American fishes belonging to 52 species, and 
among these are given analyses of fresh, salted, and pickled herrings. He 
also collects all the analytical literature and gives a table of the maximal, 
minimal, and average amounts of the food principles present in the 
herring as given by different authorities. 
I give it here in order to show that there are marked variations in the 
analytical results, especially as regards the fat. ; 
| : Protein 
Water. eee i.e. Fats. Ash. 
: b u Ss ance. N x 6:25. 
| 
Maxima. 7611 | 30:97 19°12 11-01 1:9 
| 
Minima. 69°03 | 23°89 15°31 4°89 1:5 
Average. | 7210 | 27-90 17-75 8-02 1-69 
| | | 
Atwater adopted all the precautions which he thought necessary to 
render his analyses accurate, and devotes some part of his paper to a criti- 
cism of the methods of analyses employed by his predecessors, explaining 
in some cases the differences in the results by probable analytical errors. A 
very important and most probable cause of divergence in the analytical 
results was a difference in the condition of the herring at the time of the 
analysis. As we shall see, there are great differences in composition of the 
muscle of the herring at different periods, and these differences are to be 
observed in herrings obtained from the same waters, when they are 
examined at different months of the year. 
For purposes of comparison it is extremely desirable that the herrings 
should be obtained from the same district at regular intervals for at least 
two years, and that analyses should be made of muscles and genitalia. 
This would be a work of no great difficulty if three or four persons 
were occupied with the investigation, but it could not be covered by a 
single worker unless he devoted his whole time to the work. I have been 
unable to devote more than a comparatively small portion of my time 
during the last two years to the work, and it is ouly during the last year 
that I obtained satisfactory material for the research. 
The work, therefore, will deal with herrings mainly obtained from the 
Loch Fyne district during 1905 and 1906. 
Only these fish were employed for analysis which had arrived at the 
laboratory in absolutely fresh condition. During the summer months 
they were sent in a double walled japanned tin carrier, an ice-salt mixture 
filling the space between the two walls. At other times they were sent 
with salt sprinkled over them or without any preservative. 
The fish were measured and weighed, and the ovaries or milt removed, 
weighed, and examined microscopically, the ova being measured by means 
of an ocular micrometer. The measurement for length was made from 
end of snout to end of tail fin; for girth, in front of the dorsal fin around 
the thickest part of the fish. The measurements are given in centimetres, 
The weights of the fish and reproductive organs are given in grammes. 
