SEA-FISHERIES LABORATORY. 133 © 
others. Since the abundance and average sizes of the plaice 
caught vary greatly from month to month, this unequal 
fishing disturbs the averages. Also one very large catch of 
plaice among a number of very ordinary catches raises the 
average unduly and gives us a distorted idea of the variation. 
So, mstead of average catches per hour’s fishing, we 
try to find some other form of average. I have arranged the 
catches made by the fish-trawl and shrimp-trawl in groups 
of 3 years, which overlap—thus, 1892 to 1894, 1893 to 1895, 
1894 to 1896, etc., and then the numbers of plaice caught 
per haul are arranged in groups of 0 to 50 fish, 51 to 100, 
101 to 150, and so on. Thus we consider all the separate hauls 
made during each of the overlapping groups of three years, 
as follows :— | 
Nos. of plaice 
caught per haul ... 0-50 51-100 101-150 151-200, ete. 
Nos. of times each 
of these results 
was obtained ...... 24 29 5 7 ete. 
We see that 24 of the hauls made during the years 1912-14, 
for instance, contained from 0 to 50 plaice, 29 contamed 
from 51 to 100, 5 contained from 101 to 150, and so on. 
This has been done for each of the groups of years 1892-4, 
1893-5, 1894-6, ete. The results are, in general, similar to 
those obtained simply by tabulating the average numbers 
of plaice caught per haul (or, what is very much the same 
thing, per hour’s fishing), but the irregularities due to accident, 
which bulk so largely in the application of the latter method, 
are avoided, and we are prevented from making erroneous 
conclusions. 
The result is—the smaller the catch of plaice the more 
often it is made, and wice versa. If we make graphs for each 
of the three-years periods, as is suggested by the above incomplete 
table, we get J-shaped curves, the tail of the J being drawn 
