178 THE RATE OF GROWTH [ch. 



11cm., 26, 44 and 60 cm., as the mean or modal sizes of four 

 successive broods. The dwindhng heights of the successive cusps 

 are a first approximation to a "curve of mortahty," shewing how 

 the young are many and the old are few. Again, plotting the several 

 sizes against time, we should get our curve of growth for four years, 

 or a first rough approximation to it. Thus we learn from a random 

 sample, caught in a single haul, the mean (or modal) sizes of a fish 

 at several epochs of its fife, say at two, three or even more successive 

 intervals of a year ; and we learn (to a first approximation) its rate 

 of growth and its actual age, for the slope of the growth-curve, 

 drawing to the base-fine, points to the time when growth began. 



^, Cod. Nov. 1906. F. of F 



20 JO 40 



Length, in centimetres 

 Fig. 44. A catch of cod, shewing a multimodal curve of frequency. 



Another haul, soon after, will add new points to the curve, and 

 confirm our first rough approximation. 



An experiment in the Moray Firth, a month or two later, shewed 

 the first three annual groups in much the same way; but it also 

 shewed another group, of about 90 cm. long, and others larger still. 

 At first sight these did not seem to fit on to our four successive 

 year-groups, of 11, 26, 44 and 60 .cm.; but they did so after all, 

 only with a gap between. They were older fish, six and seven years 

 old, which had come back to the Moray Firth to breed after spending 

 a couple of years elsewhere. 



It was thought at first that every such experiment should tally 

 with another, and bring us to a more and more accurate knowledge 

 of the growth-rate of this fish or that; but there were continual 

 discrepancies, and it was soon found that the rate varied from place 



