— 192 — 
»4) Such a survey will nevertheless be subject to certain errors, which may in some 
cases be of the same numerical order as the figures for the variations noted from year 
to year. : 
»b) The relation which HeLLAnD-HAnsen and NANSEN believe to have discovered 
between the fluctuations in hydrographical and biological conditions may in certain 
cases be merely fictitious, and due to errors in the observations. 
»6) The alleged relation between the extent of the coastal water area and the yield 
of the sprat fishery is not confirmed by the investigations of recent years.« 
Biological investigations as to the fluctuations in quantity of cod liver 
and roe. 
The hydrographical investigations are thus still far from being able to furnish any 
full and simple explanation of the fluctuation m the quality of the fish. It will also be 
evident, on closer consideration, that no satisfactory result can be arrived at in this 
respect until the biological phenomena which it is desired to explam are far more 
accurately and distinctly defined than has been the case in the course of the investiga- 
tions made by HELLAND-HANSEN and Nansen. The present writer has long entertained 
the project of carrying out such preliminary biological investigations; not until 1913, 
however, was it possible to realise this plan, by personal observation of the Lofoten 
and Finmark fisheries. The data here obtained will be briefly described in the following. 
First of all, it appeared to be of importance to study the variation (in weight 
and volume) of the liver in fish of different size and different condition as regards 
quality. With this end in view, several hundreds of cod were measured (as to length) 
and weighed, the liver and roe of each bemg then weighed, for purposes of comparison 
with the length and weight of the fish. It was immediately apparent that the liver, 
as every fisherman and dealer knows, exhibits very great variation in size and weight, 
even in fish of the same size. On examining a large number of fish, however, and calcu- 
lating the average weight for the different lengths, it is noticeable that these average weights 
increase at a high rate of progression with the size of the fish. Fig. 121 illustrates the method 
of investigation here employed. For each centimetre group of cod (e. g. for fish from 
50—59 cm. in length) the single liver weights are grouped according to size. We thus 
obtain a view both of the variation in weight, and the average weight of the group. This 
is indicated in the figure by the fully drawn curve, which shows an increase in the average 
weight from each size group to that next following. In like manner, I examined a great 
number of samples, arranging the individuals, as a rule, in 5 cm. groups, as in the size 
investigations described in Chap. III. The results of all these investigations tend to 
show an increase in the average weight of the liver with increasing length or weight 
of the fish. 
Comparison between size of fish and weight of liver. 
In view of the importance of more closely considering these features of the question, 
I have drawn up the table on page 194, showing the results obtained by investigation of 
two samples, one consisting of spent Lofoten skrei (from Sorvaagen) taken in the latter 
part of April 1913, and the other of Finmark fish, taken on the 6th of May in the same 
