WEBB and COROLLA: BURST SWIMMING OF NORTHERN ANCHOVY 



61 



-? 4 



> 



< 

 t- 



a 



1 - 



. 240 ms 



100 ms 



:i*' — '-*t-i-i 



0.0 Q2 04 0.6 0.8 10 1.2 1.4 



TOTAL LENGTH (cm) 



Figure 2. — Exemplary relationships between the distance 

 traveled and total length of northern anchovy larvae after 20, 

 100, and 240 ms of burst swimming initiated by am electric shock 

 stimulus. Vertical bars show ± 2 SE. Open squares show indi- 

 vidual data points for individuals of groups of larvae when few 

 swam for 240 ms. 



Table 2. — Summary of relationships between distance traveled 

 (S in centimeters) in various elapsed times as functions of total 

 length (TL in centimeters) of northern anchovy larvae stimu- 

 lated to maximum performance by an electric shock. 95% con- 

 fidence intervals about the slope are shown. 



Elapsed time from 



start of swimming 



(ms) 



Best-fit equation relating 

 distance traveled 

 with total lengtfi 



20 



40 



60 



80 



100 



140 



180 



220 



260 



S = 0.09 + 0.10±0 04TL 

 S = 0.15 + 0.37±0.08TL 

 S = 0.17 -I- 0.80 + 0.15 TL 

 S = 0.21 -I- 1.17±0.22TL 

 S = 0.28 + 1.46±0.27TL 

 S = 0.54 + 1.75±0.37TL 

 S = 0.82 + 2.06±0.44 TL 

 S = 1.07 -I- 2.25 ±0.60 TL 

 S = 1.26 + 2.52±0.53TL 



These distances were traveled by continuous 

 swimming because gliding is limited for or- 

 ganisms the size of the larvae moving at low 

 speeds (see Weihs 1980). 



Although the relationships between distances 

 traveled and total length were best described by 

 linear regression equations, the coefficients of de- 

 termination indicated that the relationship usu- 

 ally described only 65 to 75% of the variability. A 

 major contributor to the variability was the 

 apparent low performance of larvae with mean 

 total lengths of 0.65 and 0.72 cm. For example, r^ 

 for the relationship between distance traveled in 

 100 ms and total length increased from 0.71 to 0.83 



on deletion of the observations for larvae in these 

 two groups. However, the larvae appeared healthy 

 and there were no apparent reasons to assume 

 these data were anomalous. The reason for their 

 lower performance is unclear. Another factor con- 

 tributing to variability in the data may have been 

 the use of larvae from several spawnings at dif- 

 ferent times from different small sets of only 25-50 

 parents from laboratory stock. The very low r^ for 

 the distance-total length relationships at small 

 elapsed times can also be attributed in part to 

 greater measurement error. Larvae traveled small 

 distances, of the order of 1 mm in 20 ms for small 

 fish, and even with magnification of 2 x these 

 small distances were obviously subject to greater 

 measurement error than larger distances. 



Swimming speeds were not constant. Speed in- 

 creased with time to reach the maximum burst 

 speed after 80-100 ms and then speed declined for 

 the remainder of the burst (Figure 3). The time to 

 reach maximum burst speed was not affected by 

 total length, but the decrease in speed from the 

 maximum was greater for larger larvae, and ex- 

 tended over a longer period of time. The maximum 

 burst speed increased linearly with total length 

 (Figure 4), and since time to maximum speed was 

 independent of length, mean acceleration rates 

 will also be proportional to total length. Mean 

 speeds during a burst of swimming (U) also in- 

 creased linearly with length (Figure 5), but at a 

 lower rate than maximum burst speeds [Umax) 

 where 



U = 4.89 + 8.18±1.1L (r^ = 0.861; n = 85) 

 Umax = 1.95 + 20.8±2.5L (r^ = 0.891; n = 85). 



The difference between maximum and mean 

 speeds increased with total length because varia- 

 tions in speed in a burst increased with total 

 length (Figure 3). Maximum and mean speeds 

 during a burst were consequently similar for the 

 smallest larvae. 



Figure 5 also shows other relationships between 

 swimming speed and length reported in the litera- 

 ture for comparison. The mean speeds measured 

 for anchovy larvae were greater than those mea- 

 sured for other larvae, exceeding the ". . . theoreti- 

 cal 10 body lengths/sec ..." (Blaxter 1969) 

 maximum. Intermittent swimming speeds 

 (Hunter 1972), the normal voluntary swimming 

 pattern of larvae, and voluntary bursts were lower 

 than the burst speeds measured here, presumably 

 because of the absence of threatening stimuli. 



145 



