MAGNUSON and HEITZ : GILL RAKER APPARATUS 



ing area increased as the 2.2 i^ower of fork length 

 for A', pelamis and the 1.9 power for T. alba- 

 cares. Forcing the regressions to pass thi'ough 

 the origin may have decreased the slope. 



To facihtate comparison of different species, 

 the mean gap and filtering area were cominited 

 from the regression in Table 1 for fish with a 

 fork length of 35 cm. These are listed in Table 2 

 in order of decreasing number of gill rakers, 

 increasing gap, and decreasing filtering area. 



As expected, the number of rakers and gill 

 raker gap were closely related (Table 2). Lack 

 of comjilete correspondence may have resulted 

 from differences in the thickness of gill rakers, 

 differences in the length of the gill arch, or both. 



Among scombrids no relation was evident be- 

 tween filtering area and number of rakers or 

 between filtering area and mean gill raker gap 

 (Table 2) . Apparently, the length of raker was 

 an important variable determining differences 



Table 2. — Scombrid and coryphaenid species (35-cm fork length) listed in order of increasing numbers of gill rakers, 



and decreasing mean gill raker gap and filtering area. 

 (Data on S. orientalis from one fish, S. japovicus from two fish.) 



Rank 



Mean 



number of 



rokers 



(») 



Species 



Mean gill 

 raker 

 gap 

 (mm) 



Species 



Filtering 

 area 

 (mm-) 



Species 



100 

 FORK LENGTH 



Figure 3. — Comparison of the mean gill raker gap and 

 fork length relationship for various scombrid and cory- 

 phaenid fishes. Lengths shown approximate ranges 

 known for each species. 



in filtering area among species. Coryphaenids 

 had a larger gill raker gap and smaller filtering 

 area than any scombrid except striped bonito, 

 Sarda orientalis (Temminck and Schlegel). 



Among scombrids 35 cm long, Sarda had the 

 largest gaps (1.8-3.3 mm) and Auxis and Katsu- 

 ■ivoniis the smallest (0.51-0.74 mm). Thunmis, 

 Euthynnus, and Scomber had intermediate gap 

 widths (1.2-1.4 mm). Among Sarda, Auxis, and 

 Thunnus represented in our samples, species 

 within genera had more similar gill raker gaps 

 than those in different genera. On this basis 

 alone food habits for fish of the same length 

 would be expected to be more similar within 

 genera than among genera. 



Mean gill raker gap differed markedly with 

 species and length of fish (Figure 3). For ex- 

 ample, a 50-cm K. pelamis, a 30-cm T. albacares, 

 and a 10-cm S. orientalis all had a mean gill raker 

 gap of approximately 1 mm. Conversely, gill 

 raker gaps of these three species differed mark- 

 edly at the same fork length. Gaps of 50-cm K. 

 pelamis, T. albacares, and S. orientalis were ca. 



365 



