NORTH AMERICAN SALMONIDAE 



241 



MERISTIC CHARACTERS 



In using mcristic characters to distinguish be- 

 tween anj' two populations there are certain things 

 to l)ear in mind. Several investigators have estab- 

 lished that in some species some of the meristic 

 characters exhibit phenotypic variation induced 

 by variations in environmental factors during 

 early developmental stages. For a review of these 

 studies see Tuning (1952) and Seymour (1959). 



By incubating and rearing chinook salmon, 0. 

 tshawyturha, at constant temperatures, Seymour 

 (1959) showed that the fish formed the lowest 

 number of vertebrae at intermediate temperatures 

 (4-5°-55° F.), and higher vertebral numbers at 40° 

 and at 60°. He found, however, that this pheno- 

 typic variation was much less than the genotypic 

 variation when lots of eggs from four rivers, the 

 Sacramento, Green, Skagit, and Entiat, were in- 

 cubated and the fish reared at several constant 

 temperatures. The mean number of vertebrae for 

 all temperatures was about 66 for the Sacramento, 

 68 for the Skagit, 69 for the Green, and 72 for the 

 Entiat River. As the spawning season in different 

 localities tends to conform to the optimum local 

 conditions, the temperature-induced variation is 

 probably of even less importance than these con- 

 trolled experiments might suggest. The number 

 of individuals with abnormal vertebrae increased 

 in temperatures above 60° and below 40° F. 

 Seymour also found that low oxygen content of 

 the water during incubation increased the number 

 of vertebrae. 



Branchiostegal Rays 



Most meristic data on Salmonidae have not 

 been collected in such a manner, or are not suffi- 

 ciently extensive, as to yield a reliable measure of 

 the range of variation to be expected between 

 samples taken in different years or in different 

 localities. One of the best series of data is from 

 Chamljerlain (1907) for sockeye salmon from six 

 streams in the southern portion of southeastern 

 Alaska for the years 1903 and 1904. Since none 

 of his samples had less than 100 individuals we 

 have made an analysis of his data, shown in 

 table 6, for the mean branchiostegal i-ay counts 

 on 4,686 specimens. 



The number of rays is usually higlier on the left 

 side as the left membrane normally overlaps the 



Table 6. — Mean count of branchiostegal rays in sockeye 

 salmon, southeastern Alaska, 1903 and ISOJ, 



Note.— Data from Chamberlain (1907); total of 4,686 specimens, samples of 

 100 to 513 individuals each. 



right. Chamberlain states that "In no instance 

 was a clearly defined case of right overlapping 

 seen, though occasionally the right membrane cai*- 

 ries the higher number of rays." Similarly, 

 Vladykov (1954, p. 909) found the number of 

 branchiostegals on the right side in all charrs 

 somewhat smaller than on the left. 

 The analysis of table 6 follows. 



Source of variation 



Total 



Between sides 



Between years 



Between localities. 

 Interaction (error) 



D.F. 



Sum of 

 squares 



2. 423393 

 1.802920 

 0.007921 

 0. 555203 

 0.057349 



Mean 

 square 



0. 105365 

 1. 802920 

 0. 007921 

 0.111041 

 0.003584 



503.047" 

 2. 210 N.S. 

 30.982" 



The significant difference in the mean number of 

 i-ays between the left and right sides was con- 

 firmed, as well as a significant difference between 

 localities, but the difference between years was 

 very small. 



Eepeating this analysis, but employing only the 

 number of rays on the left side, a significant dif- 

 ference is again shown between localities, but not 

 between years. If we ignore the possibility of 

 greater differences occurring between ye<ars, we 

 still find a maximum mean difference for the left 

 side of 0.506 rays between samples (13.986-1:5.480) . 

 This suggests use of great caution in forming con- 

 clusions about interspecific differences in a meristic 

 character on the basis of small samples, especially 

 if the samples are not geographically representa- 

 tive. 



If one compares this mean branchiostegal count 

 for 0. nerka from southeastern Alaska with the 

 average given by Foerster and Pritchanl ( 1935a) 



