252 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



Table 20. — Count of gill rakers on first gill arch, left side, in Salmo, Salvelinus, awl Cristivomer 

 Ix indicates gill rakers present in frequency column, but numbers not given] 



1 Kendall (1936); Penobscot River. 



' McCrimmon (1949). 



» Milne (1648); Skeena River, British Columbia. 



' Mottley (1936); Kootenay Lake, British Columbia. 



» Shapovalov (1947). 



• Vladykov (1954). 



The gill-raker counts of tables 19 and 20 are 

 summarized in table 21, in which I have endeav- 

 ored to give some indication of dispersion. Many 

 of the samples were so small, with the distribution 

 either discontinuous or skewed, that the standard 

 deviation was discarded and instead I have shown 

 the range and the interpercentile range from the 

 80th to the 20th percentile (see fig. 6). 



It is interesting to note that ti'utta shows the 

 lowest average for gill rakers (fig. 6), as it also 

 does for branch iostegal rays and vertebrae (fig. 

 3 and 5) . Fonfinalis, which is next to the bottom 



TSHAWrrSCHA  



•  «T4 



. « KISUTCH 



O ALPWUS 



-)  UARSTONI 



— ' WMAYCUSH 



SALAR 



— GAfRONERI 

  6 KAULOOPS 

  AURCOll/S 



. ., FONTINALIS 



X TRUTTA 



DOX# Mean «0« SPECIES 



I  BO-ZO PEBCENTILE RANGE 



TaBLE 21 

 ., OF INOIWII 



AAKERS ON riRST GILL ARCH 



FiGUBE 6. — Gill rakers on first gill arch. 



' DeLacy and Morton (1943); Karluk River, Alaska. 



* Wilder (1952); anadromous stock, Moser River, Nova Scotia. 



• Wilder (1952); resident stock, Moser River, Nova Scotia. 



10 Wilder (1952); from 3 brooks in Nova Scotia. Sample from Bocabec 

 Brook in New Brunswick omitted because of small size of the flsh. 



in gill-raker count, occupies the same position for 

 number of pyloric caeca and is quite low in num- 

 ber of branchiostegal rays and vertebrae. 



The question of gill rakers on other than the 

 first gill arch will be discussed later. 



Scales 



Although scale counts are widely used in taxo- 

 nomic work they must be used cautiously because 

 of the variation in counting practice among differ- 

 ent investigators. Neave (1943) gives an excellent 

 discussion of the various counting methods in 

 vogue. One difficulty arises from the failure of 

 many authors to recognize that the number of 

 scales in the lateral line does not usually corre- 

 spond either to the mmiber of diagonal (oblique) 

 rows just above the lateral line or to the number 

 of diagonal rows counted along any horizontal row 

 several rows above the lateral line. As a result 

 many published data on the count of lateral-line 

 scales, or "scales along the lateral line," actually 

 refer to a count of diagonal rows made either just 

 above the lateral line (usually a somewhat higher 

 count) or of diagonal rows counted several longi- 

 tudinal rows above the lateral line (usually a still 

 higher count). 



Some investigators have varied these practices 

 by counting the lateral-line tubes or sensory pores 

 and considering them equal in number to lateral- 

 line scales. A fifth method has been to count the 

 rows of diagonal scales 10 or 15 rows above the 



