176 THE RATE OF GROWTH [ch. 



more may be kept in a tank or pond and watched as they grow. 

 Both ways are slow and difficult. The advantage of large numbers 

 is not obtained ; and it is needed all the more because the rate of 

 growth turns out to be very variable in fishes, as it doubtless is 

 in all cold-blooded or " poecilothermic " animals: changing and 

 fluctuating not only with age and season, but with food-supply, 

 temperature and other known and unknown conditions. Trout in 

 a chalk-stream so differ from those in the peaty water of a highland 

 burn that the former may grow to three pounds weight while the 

 latter only reach four ounces, at three years old or four*. 



It is found (and easily verified) that shells on the seashore, kind 

 for kind, do not follow normal curves of frequency in respect of 

 magnitude, but fall into size-groups with intervals between, so 

 constituting a multimodal curve. The reason is that they are not 

 born all the year round, as we are, but each at a certain annual 

 breeding-season ; so that the whole population consists of so many 

 "groups," each one year older, and bigger in proportion, than 

 another. In short we find size-groups, and recognise them as age- 

 groups. Each group has its own spread or scatter, which increases 

 with ^ize and age; even from the first one group tends to overlap 

 another, but the older groups do so more and more, for they have 

 had more time and chance to vary. Hence this way of determining 

 age gets harder and less certain as the years go by; but it is a safe 

 and useful method for short-lived animals, or in the early lifetime 

 of the rest. Aristotle's fishermen used it when they recognised 

 three sorts or sizes of tunnies, the auxids, pelamyds and full-grown 

 fi^h; and when they found a scarcity of pelamyds in one year to 

 be followed by a failure of the tunny-fishery in the nextf. 



Shells lend themselves to this method, as Louis Agassiz found when 

 he gathered periwinkles on the New England shore. Winckworth 

 found the Paphiae in Madras harbour "of two sizes, one group just 

 under 15 mm. in length, the other nearly all over 30 mm. A small 

 sample, dredged ^ve months earlier hora the same ground, was inter- 

 mediate between the other two." When the mean sizes of the two 

 groups were plotted against time, the lesser group being shifted 



* Cf. C. A. Wingfield, Effect of environmental factors on the growth of brown 

 trout, Journ. Exp. Biol, xvii, pp. 435-448, 1939. 

 I Aristotle, Hist. Anim. vi, 571 a. 



