166 



DEPARTME\' OF THE NATAL SERVICE 



either side of the mid-line with a series of parallel diagonal lines, making an angle of 

 4(r"degrees with the mid-line. By aligning the margins of the tail with these it was 

 possible to spread the tail to an arbitrary, constant angle of 80 degrees. 



Owing to the mixture of herring of different age groups in the samples, it was not 

 feasible to take the average size in treating the material. The smallness of the numbers 

 representing certain age groups in many of the samples rendered the results unsuitable 

 for extensive statistical treatment. The only feasible method was a compromise and 

 therefore somewhat open to objection. 



The relative frequency of the various length groups in a sample indicated whether 

 the sample consisted of more than one age group and also showed the mean size in each 

 group. The various groups could in that way be traced through successive samples and 

 their rates of growth determined. 



The following table gives the results of the measurements : — 



Two of the samples may be said to be homogeneous, consisting of only one age 

 group. They are those of June 28 and October 29. In both the actual range in size is 

 6 cm. (9-14 and 11-16) and the practical range is only 4 cm. (9-12 and 11-14) or perhaps 

 3 cm. (9-11 and 11-13). There can be no doubt that in these cases we have to do with 

 only one age group. The curve for the sample of June 28, obtained by plotting the 

 lengths against the numbers of individuals is given by the continuous line in fig. 1. 

 Evidently too few length groups have been taken to give the most satisfactory curve, 

 but we will not be far astray in taking 10 cm. as representing the mean length of the 

 herring in the sample of June 28 and 12 cm. for those of October 29. 



