256 



NATURE 



[November 5, 1914 



The diagram shows, for each year of observation, 

 the percentage frequency of the individuals according 

 to the number of rings on their scales, the figures 

 along the abscissae denoting the number of rings, 

 while the corresponding ordinates denote the per- 

 centages of all the individuals falling in each ring 

 group. 



The most prominent feature of this diagram is the 

 regular movement of the primary mode (a) from 

 1908-14, as well as of the secondary mode (b) from 

 1907-09. 



While in 1908 the individuals having 4 rings were 

 the most numerous, in 1909 those having 5 rings, in 

 1910 those having 6 rings, in 191 1 those having 

 7 rings, in 1912 those having 8 rings, in 1913 those 

 having 9 rings, and in 1914 those having 10 rings 

 predominate. 



' In the same manner the secondary mode (5), ap- 

 pearing in 1907 by the relative numerical strength of 

 group 8, moves one class to the right for each of the 

 years 1908 and 1909. 



3 4 5 6 7 



no. of. rings. 



« 11 12 13 14 15 



Fig. 2. — Comparison between the empirical ring curve for 1913 (in Fig. i) 

 and its corresponding normal variation curve. 



It seems difficult to explain this regular movement 

 of the modes, unless we assume that the two pro- 

 minent modes represent two relatively abundant year- 

 classes of herrings. 



If this be true, then the material investigated leads 

 to the conclusion that great fluctuations occur in the 

 stock of Norwegian herrings with regard to the 

 relative abundance of the different year-classes. That 

 the magnitude of these fluctuations is astonishing 

 and "very hard to imagine," is by no means an argu- 

 ment against the observed facts. 



Prof. D'Arcy Thompson emphasises the regularity 

 with which the so-called year-groups arrange them- 

 selves in a unimodal skew-curve, "just as the same 

 fish group themselves also, according to size, in a 

 unimodal but more normal curve." It seems to him 

 " statistically improbable that a dozen separate genera- 

 tions of herrings, spawned in as many years, should 

 have entered into the formation of the composite 

 shoal in these curiously and regularly graded propor- 

 tions." He finds it much easier to explain this fact 



NO. 2349, VOL. 94] 



by assuming that all the herrings (in a sample) were 

 0/ ihe same age and origin, and that consequently 

 the individual herrings vary about a certain modal 

 number of rings, just as they also vary in a normal 

 fashion about a certain modal size. He also em- 

 phasises the probability of an irregular age-curve for 

 a population where large and irregular fluctuations 

 occur in the annual birth-rate. 



It might have been expected, then, that Prof. 

 D'Arcy Thompson would have subjected the material 

 published to a statistical analysis in order to convince 

 himself of the correctness of his opinion regarding 

 the similarity between the age-curves and the norma! 

 curves so usual in biology. He would then have 

 found that no age-curve in the whole material of 

 Norwegian herrings follows the usual law for bio- 

 logical variation. In some cases the age-curves pre- 

 sent themselves by mere inspection as irregular (see 

 Fig. I, 1907, 1908, and 1909), while in other cases 

 (when the curve is unimodal, as in Fig. i, 19 10-14) 

 they are so entirely difterent from the normal fre- 

 quency curve that it is totally impossible to replace the 

 empirical curve by the corresponding theoretical one. 

 This will be apparent from Fig. 2, giving the empirical 

 age-curve for Norwegian herrings in 1913 (see Fig. i) 

 together with the normal curve for the same number 

 of individuals, the same average and standard devia- 

 tion, and the same arrangement of classes. In the 

 following table the same data are given numerically : 



The dissimilarity of the two curves is, in fact, so 

 great as to exclude any idea of the age-curve following 

 the usual law for biological variation. The feature 

 that the curve is apparently unimodal. Is due to the 

 fact that one single group is so overwhelmingly 

 abundant as to depress to a certain degree the irregu- 

 larities actually existing in the frequencies of the 

 other groups. By inspection of the age-curve for 1914 

 (see Fig. i) it will be observed that the frequency of 

 group II, lying close to the modal group, is less than 

 for the more extreme groups 6, 7, and 8. 



For these reasons, as well as because of the regular 

 movement of the modes, it seems to us impossible to 

 explain the observed facts as a result of common varia- 

 tion, even if the help of a mathematical statistician 

 were enlisted. 



JOHAN HjORT. 



EiNAR Lea. 



Scientific Work and the War. 



Dr. Calman's plea (Nature, October 22, p. 198) for 

 the continuance as usual of the work of our learned 

 societies, or even a wider plea for the calm prose- 

 cution of all our scientific studies, may be supported on 

 many grounds, but the only one that I wish to 

 emphasise is the moral effect thus produced upon those 

 neutral nations whom our opponents seek to delude 

 into the belief that we are panic-stricken. To the 



