244 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



Figure 4. — Mean numbers of pyloric caeca. (Lines in- 

 dicate the 20th and 80th interpercentile range. ) 



If we disregard McGregor's samples the intra- 

 specific variation in the mean caecal count is 

 small, ranging from 150.5 to 165.8 for tshawytscha 

 and from 133.5 to 137 for gorbu^cha. This is a 

 small range in relation to that for the five species — 

 from 75.5 for kisutch up to 205.0 for keta. 



The data for the remaining genei-a are far less 

 extensive so they are combined with the summary 

 for Oncorhynchus .in table 9. In figure 4 the 

 means are given as well as the approximate 20th 



Table 9. — Count of pyloric caeca in North American 

 Salmonidae 



' Upper and lower limits of groups unless given by authors. 

 2 References for Oncorhynchus in table 8. 

 s Belding (1940); eastern Canada. 

 ' Standard deviation, 4.03. 

 » Milne (1948); Sliccna River. 

 Townsend (1944); Oregon. 

 ' DeWitt (1954); nortliern California. 

 ' Vladykov (1964). 



•Morton and Miller (1954); presumably these data include counts for 

 malma and alpinua by DeLacy and Morton (1943), Karluk, Alaska. 

 " Only 1 specimen beyond category of 70-79; distribution extremely skewed. 



and 80th percentiles. Obviously, Oncorhynohus 

 and Cristivomer differ markedly from Salmo and 

 Salvelinus in number of caeca. 



In number of pyloric caeca, as in number of 

 branchiostegal rays, C. namaycush differs 

 markedly from Salvelinus and is close to Oncor- 

 hynchus. 



Fin Rays 



The comparison of fin-ray counts is rendered 

 difficult by differences in counting methods used 

 by different investigators. For instance, for the 

 anal fin counts of 0. nerka in table 10, Foerster 

 andPritchard (1935a, p. 91) write — 



In counting fin rays only developed rays, those which 

 had attained a length of one-half the length of the longest 

 ray, were included. The remainder were considered as 

 undeveloped. Care was taken to ensure that branched 

 rays did not lead to error in the count, 



Milne (1948) apparently used the same method 

 since he comments (p. 73) concerning his differ- 

 ence in average count between 1946 and 1947 — 



. . , it is possible although not probable, that during the 

 first year (1946) less attention was focussed on omitting 

 rays less than one-half the length of the fin or in count- 

 ing branched rays as two with the result that a higher 

 count might have been recorded in error for 1946. 



Chamberlain (1907, p. 89) writes— 



In the fin-ray counts the totals of rudimentary and 

 branched rays are used, but the terminal half ray, which 

 varies greatly in development, is in all cases omitted. 



It will be noted that the counts for 0. nerka 

 given by Chamberlain are about 3 rays higher than 

 the others, owing doubtless to his inclusion of the 

 rudimentary rays. A good summary of this diffi- 

 culty is given by Vladykov (1954, p. 911), who 

 writes — 



, . . there are technical difliculties in counting small 

 simple rays in front of the dorsal and anal fins. The best 

 way is to remove the skin and stain the rays with alizarin. 

 In larger .specimens the stained fins should be dissected 

 and made transparent by placing in glycerine. To avoid 

 error in counting these small rays in unstained speci- 

 mens, some authors, as Kendall (1914, p. 24), counted 

 only "fully-developed" rays in the dorsal and anal fins. 

 Unfortunately there is no definition of the term "fully- 

 developed," Some other authors count only branched 

 rays, which are plainly seen even without staining with 

 alizarin. Unfortunately the number of branched rays in 

 younger fish (parr) is smaller than in older individuals of 

 the same species .... 



