137 



That is, there were 13 hauls in this year that contained U or 

 less than 51 plaice, 6 that contained 51, but less than 100, and 

 so on. The lower line is to be read as follows : there were 

 37 hauls in the year, and in all of these either no plaice or some 

 plaice were caught ; there were 24 hauls in which either 51 or 

 more than 51 were caught ; there were 18 hauls in which 

 either 101 or more than 101 were caught, and so on. When 

 we plot the figures of the last line we get a J-shaped curve 



^oo- 



5oo 



Boo- 



/oo- 



r^4U)>or~-^'^5 



0\ On 



Fig. 16. The continuous, angular line represents the average catch 

 per hour's fishing with a trawl-net of 6-inch mesh. The individual annual 

 averages are "smoothed" by taking three-yearly averages yearly. The 

 catch for the period 1895-7 rises to 895. The series of columns represents 

 the results of the other method described in the text. Several years' records 

 are lost. 



(or exponential). Then we take half the area of this curve 

 on the steeper side and note what is the range of frequencies 

 covered by the half area. 8uch ranges are represented by the 

 columns of Fig. 16. 



This kind of average gives us the nearest approximation 

 to the experience of the fishermen. The latter would say 

 " we usually got about 100 fish per haul," or so. This kind 



