THE IMPOVERISHMENT OF THE SEA. 19 



means of these averages for the respective periods. They are, therefore, 

 not strictly correct as averages of the entire number of fish taken in 

 the respective months ; but the deviations due to the method are of 

 a minute character and do not affect the general results, especially if 

 a margin of several units in the resultant averages is allowed to cover 

 the errors of method and experiment. As the monthly averages for 

 dabs and long rough dabs were not officially separated during the first 

 two years, I have calculated them anew from the details of the monthly 

 surveys for those years ; and the same alterations have been made in 

 regard to the monthly averages for 1886 as were described in the 

 footnotes to Table I., p. 13. 



I have, moreover, placed in brackets all such averages as are based 

 on less than three surveys in each quinquennial period. This pre- 

 caution shows at a glance which of the averages may be depended 

 upon as accurately representing the general condition of the fauna 

 during the corresponding period. The method falls rather severely 

 upon the averages for St. Andrews Bay, but the natural fluctuations 

 in that area, due to its shallowness and exposed situation, are so great 

 that no less rigorous method could be safely relied upon to eliminate 

 the irregularities due to these circumstances. 



For St. Andrews Bay the two quinquennial averages are seen from 

 the table to be reliable in only two mouths out of the twelve, viz. June 

 and October. They show in each case a fall in the abundance of 

 plaice, correlated with an equality or a marked rise in the number of 

 dabs. The change for June is seen to have been slight ; but for October 

 a great predominance of plaice over dabs in the first period is replaced 

 by a superiority of dabs over plaice in the second period. 



I have not included any statement of the averages for lemon soles 

 and long rough dabs in connection with this area, owing to the great 

 scarcity of these forms in the Bay as shown by Table I. 



For the Firth of Forth seven months out of the tw^elve are seen to 

 afford reliable averages for each period, viz. April to November in- 

 clusive, with the exception of May. The differences between the 

 quinquennial averages are not great, except for August; but it is 

 certainly noteworthy that the general tendency of the change is in the 

 same direction as in the case of St. Andrews Bay. 



The averages for plaice decrease in five months out of the seven, 

 by amounts which vary between 9 per cent, and 33 per cent. The 

 two increased averages show a rise of 3 per cent, and 25 per cent, 

 respectively. 



The averages for lemon soles show a decrease in every one of 

 the seven months ; whereas the averages for dabs and long rough 

 dabs respectively show an increase in every month. 



