FISHERY BULLETIN: VOL. 78, NO. 1 



tween 60 and 407f oxygen concentration ( Figure 3 ) 

 can now be understood in terms of the theoretical 

 results above. When the oxygen concentration in 

 this water is 60% or higher, diffusion alone can 

 satisfy the respiratory requirements of both day 

 and day 1 motionless larvae. Thus, the swimming 

 activity at the higher concentrations is due to 

 other factors, such as depth control. The measured 

 activity level (Figure 3) does not change between 

 60 and lOO'/f oxygen concentration, as expected 

 from the theoretical model's predictions. 



The increased swimming activity observed 

 when the concentration drops below the 40-60% 

 level must therefore be a respiratory reaction. Ac- 

 tive swimming causes convective diffusion (which, 

 as shown in the Analytical Model section, leads to 

 much higher oxygen transport rates) and moves 

 the larva to a new, nondepleted position. As ex- 

 pected from this mechanism, activity increases 

 with decreasing ambient oxygen concentration, as 

 oxygen transport rates drop below the required 

 level faster at low ambient concentration, initiat- 

 ing motion more often. The dashed lines in Figure 

 3 verify this theoretical reasoning and the ob- 

 tained values of 59% concentration (day 1) and 

 55% concentration (day 0) for the beginning of 

 respiration-driven swimming are in very good 

 agreement with Figure 6, especially considering 

 the experimental errors involved in the various 

 data sources. 



Next, I consider the significance of swimming 

 activity at higher oxygen concentration. The most 

 plausible reason for the swimming behavior is to 

 keep the larvae, which are negatively buoyant, 

 from sinking out of the preferred depth zone in the 

 sea. Day 1 larvae swim at an average angle of 39° 

 upwards from the horizontal with no significant 

 variation with oxygen concentration (Table 1). 

 The large standard error is an indication of the 

 wide spread of observed directions. The average 

 swimming speed at this stage is 5.2 ±4.1 cm/s 

 (Hunter 1972) and the average duration of a 

 swimming bout (at oxygen concentration of 60- 

 100%) is about 2.1s (Figure 4). The average verti- 

 cal component of the distance moved during a 

 single bout is therefore 



/i„p = Vt sin a = 5.2  2.1 sin 39° 



6.9 cm. (16) 



The uncertainty in this value is large due to the 

 standard errors in both the swimming angle and 



116 



the average swimming speed, but it is probably 

 accurate at least to an order of magnitude. Be- 

 tween swimming periods, the larvae sink at a 

 speed of 0.12 ±0.03 cm/s (Hunter and Sanchez 

 1976). The average number of swimming bouts per 

 5-min period was found to be about 7 (Figure 4), 

 i.e., giving an average sinking time of 43 s. This 

 leads to a vertical distance of 5.2 cm, which is close 

 enough to the value of 6.9 cm of Equation (16) to 

 show that the swimming of day 1 larvae at high 

 oxygen levels most probably is a depth-control 

 mechanism. 



The newly hatched (day 0) larvae present a dif- 

 ferent situation. Pelagic eggs are slightly posi- 

 tively buoyant (Blaxter 1969) while the chorion, 

 which is shed during hatching, is somewhat nega- 

 tively buoyant. Thus, while no measurements in- 

 dependent of the present ones exist, it is reason- 

 able to assume that these newly hatched larva are 

 approximately neutrally buoyant due to their 

 large yolk sac. As the yolk is consumed, the 

 specific gravity increases and the sinking rates for 

 day 1 are obtained. The larvae are approximately 

 neutrally buoyant during the first hours after 

 hatching so that no net sinking or upward swim- 

 ming is expected. Table 1 shows that this is actu- 

 ally the case at day 0, where the average direction 

 is very close to horizontal and the large error indi- 

 cates almost random swimming direction. Some 

 upward swimming may be discerned at the very 

 low O2 concentration experiments (20% O2). This 

 may be a result of an inadvertent oxygen gradient 

 in the tank or a phototactic response induced by 

 the low oxygen concentration. Phototaxis is prob- 

 ably the means by which the older larvae choose 

 swimming direction, and is directed upwards as 

 light in the present experiments comes from the 

 surface. 



ACKNOWLEDGMENTS 



This paper was written while I was a NRC- 

 NOAA Senior Research Associate, on leave from 

 the Department of Aeronautical Engineering, 

 Technion, Haifa, Israel. I would like to thank John 

 R. Hunter and Reuben Lasker for reading the 

 manuscript and various discussions, E. H. 

 Ahlstrom and Gail Theilacker for generously al- 

 lowing me to use their unpublished data, and Bob 

 Millman and Steve Lucas for help with the exper- 

 iments. 



