FISHERY BULLETIN: VOL. 80, NO. 3 



and Denton 1976). One would predict that her- 

 ring larvae up to 18 mm would have little or no 

 pressure sensitivity. As the bulla becomes filled 

 with gas but before the swim bladder develops, 

 we would expect very high sensitivity to absolute 

 pressure (no adaptation being possible) and her- 

 ring larvae from 26 to 35 mm would be in this 

 category. Larger larvae would retain sensitivity 

 to pressure change, but the development of the 

 adaptation mechanism would prevent its being 

 an absolute sense. One also would predict that 

 herring larvae with gas-filled bullas but no swim 

 bladders would be especially vulnerable to large 

 pressure changes that could cause the mem- 

 brane to burst. Hoss and Blaxter (1979) have 

 shown that herring larvae do appear to be espe- 

 cially vulnerable to large, rapid pressure 

 changes at about this stage of the life history. 

 Blaxter and Hoss (1979) followed the develop- 

 ment of the adaptation mechanism, measured its 

 time constant, and have shown that adaptation 

 usually does not develop until a length of >30 

 mm. 



This paper describes a detailed analysis of 

 pressure sensitivity in herring larvae, using the 

 hypothesis that a larva will swim up to compen- 

 sate for increasing pressure and down to com- 

 pensate for decreasing pressure and that this is 

 due in part to the development of the bulla sys- 

 tem. In the two experiments to be described, par- 

 ticular attention was paid to measuring changes 

 in sensitivity during the development of the 

 bulla-swim bladder system. In addition, the 

 effect of a large, rapid pressure change on subse- 

 quent pressure sensitivity also was investigated 

 in one experiment. 



MATERIALS AND 

 GENERAL METHODS 



Herring were reared from fertilized eggs, 

 using the techniques of Blaxter (1968). The tem- 

 perature during development increased from 

 about 7°C near hatching to 12°C, 4 or 5 mo later. 

 The pressure sensitivity experiments were con- 

 ducted in a constant temperature room at 9°- 

 10°C, using the apparatus of Blaxter and Denton 

 (1976). This apparatus consisted of a Plexiglas 4 

 cylinder 80 cm high and 7 cm in diameter, the 

 transparent wall being marked on the outside to 

 give 16 equal sections numbered 1-16. The sur- 



4 Reference to trade names does not imply endorsement by 

 the National Marine Fisheries Service, NOAA. 



face was designated 0, the bottom as 17. This 

 allowed an observer to record the position of a 

 larva in the cylinder at any given instant with a 

 number from to 17. The pressure in the cylinder 

 could be changed by a preset amount by opening 

 a two-way tap at the top, which exposed the 

 water surface to atmospheric pressure or to posi- 

 tive or negative pressures in a gas reservoir. 



Each larval herring was anesthetized after it 

 was tested and the developmental stage of its 

 bulla (stage I, no bulla; stage II, bulla liquid- 

 filled; stage III, bulla gas-filled) was ascertained. 

 A complication arose that the bulla does not be- 

 come instantaneously filled with gas and may 

 contain only a few or many bubbles. Pressure 

 sensitivity is more likely to be high if the bulla is 

 full of gas. At least 10 larvae of each developmen- 

 tal stage were used. 



EXPERIMENT I 



Design 



Pairs of larvae of approximately equal length 

 and stage of development were tested simulta- 

 neously. After a 2-3 min acclimation period at 

 atmospheric pressure, 10 observations on the 

 position of each fish were made at 15-s intervals. 

 The pressure was then changed and the observa- 

 tions were repeated at the new pressure. 



The pressure sequence selected was based 

 upon prior research (Blaxter and Denton 1976) 

 and involved changing the pressure from atmos- 

 pheric to each of the following pressures four 

 times: ±13, ±39, and ±66 cm H 2 (1 cm H 2 = 

 0.001 atm), for a total of 480 observations and 47 

 changes of pressure (Fig. 1). This fixed sequence 

 of increasing pressure differentials was chosen 

 to avoid the potential danger of larvae becoming 

 overstimulated initially at the higher pressures. 

 Earlier evidence (Blaxter and Denton 1976) indi- 

 cated that larvae moved upwards to compensate 

 for increased pressure and downwards to com- 

 pensate for decreased pressure, and the extent of 

 vertical movement was correlated with the ex- 

 tent of pressure change (AP) applied. Large 

 pressure changes early in the sequence might not 

 only block responses to smaller subsequent pres- 

 sures but might also cause earlier fatigue. There- 

 fore, pressure changes were not randomized and 

 an experiment commenced regardless of larval 

 distribution in the water column. Approximately 

 half the larvae used in Experiment I were pre- 

 exposed for 1 min to an abrupt pressure increase 



568 



