214 



Part III. — Eighteenth Annual Report 



important part in causing differences in structure. It is possible, but at 

 the earlier portion of the spawning period, when there is little nutrition to 

 be had, there are also relatively few young fish, and at the later portions 

 of the spawning period, when food is abundant, so also are the young fish. 

 For such parts as the vertebras and tin-rays, therefore, it is probable that 

 the range of variation which is observed is mostly, if not eutirely, due to 

 the physical and chemical conditions. 



In the case of other characters, length of fish, length of skull, and so 

 on, the range of variations observed is probably due not so much to the 

 physical and chemical conditions, nor to the differences in nutrition, but 

 to the known difference of age. If the spawning period extends over 

 three months and more, all dimensional parts will show in the range of 

 their variations the effect of this range of variation in the time at which 

 the individuals were born. If two individuals start out on life from the 

 same place and about the same time, then the chances are that they will 

 differ much less from one another than two other individuals which start 

 out on life from the same place but at different times. Hence, the range 

 of variation of any organ at any particular period of life may be considered 

 as corresponding to the initial difference of age. Each period of life, 

 however, has a certain average-condition for each character, and in order 

 to obtain a measure of the growth-variability attention has to be directed 

 to the change in this average-condition and not to the ranges of variation. 



The converse problem of variations due to differences in place may be 

 stated here for the sake of contrast. If two individuals start out on life 

 about the same time but at different places, then the chances are that they 

 will differ more from one another than two individuals which begin life 

 at the same place, and the differences between them will be the greater 

 the greater the distances between the two regions. Similarly for time, 

 and if the two individuals in either case which have begun life at some 

 distance from one another are the representatives of many others— of 

 groups, in short — they may possibly represent two distinct races. The 

 further condition necessary is that the offspring of the individuals should 

 arise constantly at similar distances apart from one another as the two 

 individuals mentioned. The external physical and chemical conditions 

 will thus be constantly different, and the differences in the characters which 

 depend upon these conditions will likewise be constant. 



It follows from this that racial differences only differ in degree from 

 the differences due, on the one hand, to age or time, and, on the other 

 hand, to place, if it be understood that differences in time and place mean 

 differences in the external conditions. Races may be recognised by other 

 signs, such as differences in the average size at maturity, distinct differences 

 in the times of spawning, distinct regions within which the migrations of 

 the different groups are restricted; but from the examination of characters 

 alone it is evident that certain limits must be ascribed to the ranges of 

 variations due to age and place. The simplest method of picturing how 

 these limits are obtained is that of arranging the different ranges of 

 variations in the form of curves along the same axis. These curves will 

 almost certainly overlap, but the centres of the curves -will be at some 

 distance from one another. The question then becomes — What must be 

 the least difference between the centres so that we may say that these 

 centres respectively represent distinct races 1 According to the method of 

 treating variations already explained, these curves may be taken to 

 represent distinct races if the probable fluctuations of the averages of 

 each group do not overlap. 



Before going further it has to be noted that the difference of "age" 

 here spoken of is distinct from the difference of " age " referred to in the 

 first few paragraphs of this section. A reference to the following Tables 



