MAGNUSON and HEITZ: GILL RAKER APPARATUS 



our samples had 53 to 64. Longest rakers were 

 near the joint between the upper and lower 

 branches of the arch. They became progres- 

 sively shorter toward the ends of the arch. For 

 example, a K. pelamis 50 cm long had gill rakers 

 21 mm long at the joint but only 2 and 8 mm 

 at the ends of the upper and lower branches, 

 respectively. The largest gap (1.8 mm) was 

 near the center of the lower branch. Gaps were 

 smaller on the upper branch than lower branch 

 and were most narrow at the ends of the arch 

 (0.2 mm and 0.9 mm for the upper and lower 

 branches). Often the gap between the first 

 raker of the upper and the first raker of lower 



arch was as great as the widest gap on the lower 

 arch. 



Most of the filtering area of scombrids was 

 confined to the lower branch of the gill arch. 

 The lower branch comprised 73 to 80 '^f of the 

 total. The filtering area of coryphaenids was 

 essentially restricted to the lower arch. Dol- 

 phin, Coryphaena hippurus Linnaeus, had no 

 rakers on the upper arch, pompano dolphin, 

 Coryphaena eqnlsetis Linnaeus, had only one. 



Gill rakers of the two coryphaenids were 

 shorter and more uniform in length than those 

 of scombrids. The longest gill raker from a 

 55-cm C. equisetis was only 9 mm contrasted 



E 

 E 



a. 



< 



a: 

 lij 



< 



z 

 < 



UJ 



Katsuwonus pelamis 



y=0.034x 



"T 1 1 1 1 1 1 1 1 1 1 1 1 1 1 r 



Euthynnus offlnls 



-y = 0.038x 



Thunnus alalunga 



1 1 1 1 r 1 1 r r- 



50 100 



FORK LENGTH (cm) 



1 1 — I — I — I — I — I — I — I — I — 1 — I — I — 



50 100 150 



FORK LENGTH (cm) 



Figure 2. — Relation between mean gill raker gap and fork length showing the advantage of using 

 regression through the origin for predicting mean gill raker gap especially when sample sizes are 

 small and restricted in length range. 



363 



