FISHERY BULLETIN: VOL. 74, NO. 4 



5,000 



- 4,000 



Table 2. -Mean swim bladder volume and mean length of 

 northern anchovy larvae in sealed and unsealed containers. 



1,000 



3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 

 LENGTH (mm) 



FiGfRE 6. -Change in sinking speed from hatching through 

 threshold of the initial inflation of swim bladder (arrow) of 

 northern anchovy larvae. For larvae <10 mm, mean sinking speed 

 ± 2 SE were plotted against the midpoint of 1-mm length 

 classes, except for the first point which was for yolk-sac larvae 

 and is plotted at average length for the class. For larvae >10 mm, 

 sinking rates are estimated from the mean swim bladder volume 

 given in Figure 3 using the general equation for sinking speed 

 given in Figure 5. Open circles are estimates for the day and solid 

 triangles for the night. Line is the cube of the length (U) plotted 

 against length. 



in the volume of the larva and its sinking speed. 

 Thus, the timing of the swim bladder inflation 

 may be related to these events. 



The effect of night inflation of the swim bladder 

 is also illustrated in Figure 6. For larvae 12 mm 

 and larger, the average sinking speed at night 

 appears to be relatively constant at about 0.6 £ 0.1 

 cm/s (22 m/h) while the sinking speed in the day 

 increased exponentially with length. A larva 16.5 

 mm long had a sinking speed at night nearly half 

 that of the day. Swimming speeds of larval an- 

 chovy while searching for food in the day, range 

 from about 0.6 to 1.0 body length/s (Hunter in 

 press). If a larva did not inflate the swim bladder 

 at night, the swimming required just to maintain 

 a position in the water would be equivalent to that 

 used in the search for food in the day. Since larvae 

 do not feed at night, filling the swim bladder would 

 clearly be advantageous as an energy conserving 

 mechanism. 



Mechanism of Swim Bladder Inflation 



It was not possible to determine if larvae in the 

 tank sealed with mineral oil swam just below the 

 layer of oil or into it because our view was from 

 above rather than from the side. However, the 

 mean swim bladder volume for larvae sampled at 

 night in the sealed tank was less than that for 

 larvae sampled on the previous night when the 



tank was not sealed {\ test, P = 0.001, Table 2). 

 The mean volumes of the swim bladders for larvae 

 in the day and at night in the sealed tank were not 

 different. This experiment suggests that anchovy 

 larvae in the laboratory fill their swim bladders by 

 swallowing air at the surface. 



An analysis of the oxygen content of the swim 

 bladder could suggest whether or not the gas in the 

 swim bladder was secreted or taken from the air. 

 Newly secreted gas would be expected to be 

 oxygen (Wittenberg 1958), but if larvae were 

 swallowing air the concentration should be about 

 lY7c oxygen. Our analysis did not agree with either 

 pattern even though some measurements were 

 made 30 min after the onset of dark. Samples 

 averaged about 11% oxygen, consistently less than 

 the atmospheric concentration (Table 3). Carbon 

 dioxide levels (0.9 to Ifl'Jc) were higher than 

 atmospheric levels but little can be concluded 

 because our experimental reading error was 1 to 

 2% owing to the small volumes used. It is probable 

 that oxygen was lower than atmospheric concen- 

 tration because it was absorbed from the bladder 

 by the larva. Except for the first two observations 

 in Table 3, oxygen concentration tended to 

 decrease with time from the onset of dark. It 

 should be noted that preferential removal of 

 oxygen from swim bladder gases is not unique to 

 anchovy larvae but is found in most fishes which 

 have been studied (Wittenberg 1958). 



The rate at which the swim bladder was filled 

 also suggests that the filling is accomplished by 

 gulping air. Larvae with filled swim bladders were 

 captured 20 to 30 min after the onset of dark and 

 the means were at a maximum by 2 h after dark. 

 Uotani (1973) reported ioxE. japonic us that filling 

 was completed by 1 h in the sea. Fishes that fill the 

 swim bladder by secretion require much more time 

 to fill the bladder, for example, Stenotomus ver- 

 sicolor (Mitchill) requires 10 to 12 h; Anguilla 

 rostrata (LeSueur), 12 to 24 h; Opsanus tan 

 (Linnaeus), Prionotus carolinus (Linnaeus), and 

 P. evolans (Linnaeus), 24 h; and Tautoga onitis 



852 



