23 
The reasons which may be adduced for explaining the discre- 
pancies are — 1 1) The differences in size recorded in the two methods. 
The Tables show, however, that the two classes of experiment 
are in fair agreement with regard to the proportional numbers of 
fishes apart from size. (2) The variability of conditions under 
which successive experiments may be made especially with regard 
to the state of the tide. Our complete experiments extend over 
about one phase of the tide in each case, and are thus more likely to 
afford a more strictly comparable result. Our observations now 
made for seven years on the Northumberland Coast go to show, 
moreover, that, although the results at each haul are tolerably 
uniform, sometimes a gradual and occasionally a sudden change in 
the numbers caught may be seen in a single day's experiment in 
one place. This was illustrated in the report for 1897, by the 
records of successive countings of the smaller iishes. 
All this goes to show that the single haul or experiment, of 
too short duration to cover the diurnal changes to which the fishes 
are liable, is not sufficient to indicate the nature of the bays in a 
comparative sense from year to year. All that it can do well is 
to show the general proportion at a given time of the fishes and 
their sizes, and, therefore, their stages in life, and a comparison 
at different seasons in widely spread stations would show the 
general seasonal migrations. 
We have thus reason to thank Aid. Dent for making the 
experiments which he began in 1892 on the larger scale. It is 
from the results of these complete experiments that we have been 
led to conclude that the in-shore waters are richer to-day in plaice 
and more especially in dabs than they were in 1892, or, in other 
words, that the favourable areas on our coast shown in Chart I. are 
attracting a gradually increasing number of these flat-fishes during 
the two or three months of the summer when the experiments are 
made. As we have pointed out before, the experiments are as free 
as possible from error due to variation, because (1) they have been 
conducted for over ten years, (2i the season during which they are 
made is short— from about the end of June to the beginning of 
September, (8) the experiment is made long enough to cover at 
least one phase of the tide. By the latter statement we mean that 
during the time the experiment is being conducted, the tide more 
or less flows and ebbs. (4) Very little line fishing takes place in 
these bays, and practically none at all during the season of our 
experiments, and trawling is prohibited. 
