of Georgia (east coast of Vancouver Island and 

 along adjacent mainland), six along the west 

 coast of Vancouver Island and six from the outer 

 coast above Vancouver Island to the Alaskan 

 border, including one locality in the Queen Char- 

 lotte Islands . The material was collected over 

 a 4 -year period. 



Morphometric characters studied were the 

 following: standard length, head length, and 

 distance from snout to insertion of dorsal fin. 

 Meristic characters studied included vertebral 

 counts, keeled scale counts and pectoral fin-ray 

 counts . In addition, age determinations, based 

 on scales, were made. As in other studies on 

 herring populations, principal emphasis was 

 placed on variation in number of vertebrae in 

 different sampling areas . 



Tester found a variation in mean vertebral 

 count with latitude which conformed with the gen- 

 eral gradation found along the west coast of 

 North America by Rounsefell . During the period 

 investigated, herring at Barkley Sound (49°N.lat.) 

 averaged approximately 52.8 vertebrae; those at 

 Jap Inlet (54°N.lat.) averaged approximately 

 53.2 vertebrae, or 0.4 vertebra higher. 



The variation in mean vertebral number 

 with year class is summarized in table 4. For 

 most localities, data are available for 4 to 7 

 year-classes. The extent of the variation found 

 between year -classes in a locality (4 or more 

 year-classes considered) was 0.42 vertebra as 

 the extreme, 0-10 vertebra as the lowest. Tester 

 marshalls data to show that variation in mean 

 vertebral count of successive year -classes in 

 the same locality is related in some way to varia- 

 tion in water temperatures during the spawning 

 and early developmental periods . 



Variance of samples grouped by year- 

 classes for each locality gave proof of the heter- 

 ogeneity of the material and hence of the presence 

 of more than one "population" in British Columbia 

 waters. 



When vertebral counts were divided into 

 abdominal and caudal counts , some differences 

 were found that were not apparent from total 

 counts. Thus, a significant and constant differ - 

 ence was found between Saltspring Island and 

 Barkley Sound herring which was not apparent 



from a consideration of the total count . Similar 

 results were obtained on comparing the count of 

 the Saltspring Island material with that of other 

 localities on the west coast of Vancouver Island. 



Mean keeled scale counts, a character 

 widely used in investigation of the Atlantic her- 

 ring, were made by Tester on many of the samples 

 studied. He found that in all seasons the counts in 

 northern latitudes averaged lower than those in 

 southern localities. The range in mean keeled 

 scale counts was from 1 1 . 75 to 12.13. The mean 

 counts from Barkley Sound were about 0.25 scale 

 higher than counts from northern British Columbia 

 localities . 



Pectoral fin ray counts were made on her - 

 ring from all localities sampled in 1932-33. 

 Individual counts varied from 14 to 20 fin rays, 

 mostly 17 or 18. Mean counts ranged from 17.55 

 to 17.68. No uniform variation was found with 

 latitude. Tester concluded that this character 

 does not appear to vary significantly among her- 

 ring from different localities in British Columbia. 



Differences in rate of growth were found 

 between herring in the vicinity of Vancouver and 

 those in northern localities, with the higher rate 

 found in southern British Columbia. 



Tester used the same methods for treating 

 data on head length as Rounsefell (1930) . He also 

 found a gradual decrease in the mean percentage 

 head length with increase in body length . He found 

 no consistent difference in this character between 

 the sexes . Differences were found with age when 

 comparing three to six year old fish collected 

 during several seasons. Regressions computed 

 from his data in table XV, of head length on 

 standard length have the following statistics : 



52 



