FISHERY BULLETIN: VOL. 71, NO. 2 



Q. v> 



"I 

 t 

 01 o 



(A 



O 





 OJ 



c 





X ^"^ 



0) 



c 

 > 



-o 





5 § -5) 



- E 



CNCM-— •OU^lOCnCO'— 

 CSCNCN — •—•—■—■— — 





oooq-ooqcN-q-^O'O 



^ ^' ^' —  ^ ^ ^ ^ o 



CM 



d 



lOhviO'Or^r^'^coco 



00 

 <N 



oooQiDcon^ocoO'O 

 —  ^ ^ _■ ^ _; _■ „■ o 



looioorvrv'tcon 



ooooooo-*-^ o 

 o-^incoooooo "^ 



— CN — CM CN n" — 



■— CNC0CN"^"<J-^lO'— "O 



— ^ O O- — 1^ 



CO lO n CN CO r^ — CO 00 <> 



nioiDLD'— oO'Ouoir) -— 



T — — ID 



ooooo — cNcoo-^r-^ — 

 CNNO'^^CNCo'Tror^ r^ 



oqcN-ocNh^CNoqin'^ 

 hs — CO — CN — r^cJcS 

 inNOio<5to«OiO"^cN 



00 



OOOCNOOOCO-^O o 



oocoocNooioirjo o 



■^OCOOCO— COOOCN -^ 



— 0*>0G0fv00tOOlO 



nncocoiOTrcocom 



^ tS *- •- 



~ 3 C C = 



0.55 a 2 a. ii ? ? S 



~ > i 2 .c .0 -o 



3 s ~ p -c 



2"^ 





01 



^^_ 



'L.2I 



•^ 

 u 



OD O 



0-0 



<u 



OJ 





3 4) 



- ~: > 

 I 2.01 



d 



CN 



X 



CO 



01 



c 



^° X 



= °.?cn5 



T3 (DCS 



c ' .— ^ 

 •=X HI 



oB.|| S- 



Q. C 

 TJT55 E 



E E £^ 

 - - o »i 



LUUJ U-CO 



T in •£ t^ 



Lift coefficients vary with sweepback, angle 

 of attack, aspect ratio, the ratio of fin span 

 to body width, and taper ratio (Lawrence and 

 Flax, 1954). The differences in these param- 

 eters among species or within species were 

 not evaluated, but comparisons made below 

 between observed and calculated speeds sug- 

 gest that the assumption about a constant 

 lift coefficient based on total lifting area was 

 not too bad. The coefficients used here should 

 be more accurate for scombroids with short 

 pectorals because E. afftuis has short pec- 

 torals. The span of the pectorals, tip to tip, 

 for E. affinis is about 2.4 times the body width. 



Correlation coefficients (Table 5) were com- 

 puted to discover how closely the observed 

 swimming speeds were associated with fork 

 lengths of the fish, the mass of the fish, the 

 required lift to keep the fish from sinking, 

 the lifting area of pectoral fins, the surface 

 loading of pectoral fins, and the predicted 

 swimming speed. Relations between observed 

 speed and fork length, mass of fish, required 

 lift, and lifting area of the pectoral fins were 

 all nonsignificant statistically. But when re- 

 quired lift and the lifting area were con- 

 sidered jointly as surface loading of the pec- 

 torals, the correlation was significant (P < 

 0.005). Surface loading is the weight carried 

 per unit area of lifting surface. Predicted 

 speeds were also highly correlated with ob- 

 served speeds (Table 5, Figure 7). The highest 

 correlation, +0.89, between observed and pre- 

 dicted speeds in body lengths per second 

 suggested that the hydrostatic equilibrium 

 requirement of continuous swimming accounts 

 for approximately 80% of the variation in 

 observed swimming speeds. The model pre- 

 dicts typical swimming speeds of scombroids 

 well and suggests that the variations in mass, 

 density, and the lifting area of the pectoral 

 fins can account for almost all variations in 

 observed swimming speeds among scombroids. 



Observed speeds averaged 9.8 cm/sec or 

 0.2 //sec faster than computed speeds. The 

 computation of Fioo assumed that the pectorals 

 were continuously extended but this was not 

 always the case. Thus, the fact that observed 

 speeds were greater than Vioo is not sur- 

 prising. E. afftuis, for example, had the pec- 



344 



