160 



DE PART M EXT OF THE XATAL SERVICE 



The features noted were as follows: Total number of vertebra; (Vert. T.); 

 numerical order of first vertebra with closed h^mal arch (Vert. H.) ; number of fin- 

 rays in dorsal (Dors.); and in anal fin (An.); and nxnnber of keeled scales between 

 the ventral fin and the anus (K.). 



I hiave worked out the standard error of the averages for each of these characters 

 in the several samples, and found that the resulting error of the differences in all 

 cases may, without risk of serious inaccuracy, be taken as about 0.17, the actual 

 figures being sometimes slightly higher, sometimes a little lower than this value. In 

 making comparison between two averages, then, a value of at least 0-5, or thereabouts, 

 will be required before the differences can with any great degree of probability be 

 regarded as statistically significant, i.e., due to other causes than accidental fluctuation. 

 Where the differences of several averages tend in the same direction, smaller values 

 may naturally be taken as significant. 



In the autumn of 1915, however, I made an observation which seems to suggest 

 that it may, in certain cases, be necessary to demand a higher order of magnitude 

 in the differences between two averages before deducing therefrom the existence of 

 variety in race. And as this particular feature has not, as far as I am aware, been 

 touched upon before, I have thought it advisable to mention it here. 



While investigating the number of vertebrae in a sample of young herring from 

 the north of Norway, in the autumn of 1915, I found that the fish of about 1-| years 

 old (numbering forty-nine in all) averaged 57-37 vertebra% while those of about 2| 

 years (241 specimens) showed an average of 57-72. The difference between the two 

 age groups in this respect was 0-36 z±z 0-11 giving 0-9995 for the probability that it 

 was not due to accidental fluctuations of sampling. It would thus seem that variation 

 may occur in the number of vertebrae of the different year-classes (year-class variation 

 as opposed to age variation). In samples where a large majority is furnished by a 

 single year-class, therefore, such year-class variation may possibly affect the results 

 to a certain degree, while on the other hand, this is hardly likely to be the case in 

 samples consisting of several year-classes, all more or less equally represented. As 

 samples of young herring are particularly liable to exhibit such majority of a single 

 year-class, it will be prudent to regard them as less truly representative of the race to 

 which they belong. 



Table 47 shows the average values for the seven samples examined as to number 

 of vertebrae, etc. The values are here carried to the first decimal point, a value of 

 about 0-5 being, as we have seen, required to make the difference statistically signifi- 

 cant. There are one or two slight discrepancies between this table and the correspond- 

 ing one in Dr. Hjort's preliminary report. One of these (West Ardoise, number of 

 vertebrae) is due to the omission there of an extreme variate which has been included 

 here ; another owing to a slight error in the arrangement of the variates ; the remain- 

 der to the loss of some of the sheets on which the figures had been noted; none arp. 

 however, of any grave importance. 



Table 47. — Eacial characters. Averages for samples from different localities. 



Locality and Date. 



Characteristics. 



Dors. 



An. 



K. 



Vert. 

 H. 



Vert. 

 T. 



Newfoundland, west coast. Autumn '14 



Magdalen Islands, May 1914 



Northumberland S-trait, May 21, 1914 



W. Ardoise. Aug. 10, 1914 



Lockeport, Nov. 1914 



BayofFuiiHy, Nov.. 1914 



(jloucester, Mass. Nov. 1914 



Old, mature, developing sex. org. 



Old, mature, ripe 



Old, mature, ripe 



Younger, mature, ripening 



Younger, mature, ripe? 



Young, immature 



Young, immature . 



200 

 18-3 

 1.S 3 

 18 5 

 20-3 

 18-7 

 19-9 



13 2 



25 

 2.0-4 

 2.5-2 

 2.5 4 

 25-2 

 25-0 

 24 9 



56-8 



