388 



Fishery Bulletin 94(3). 1996 



tected (Methot, 1983; Karakiri et al., 1989; Maillet 

 and Checkley, 1991). Estimated standard lengths at 

 hatching and metamorphosis were calculated and 

 compared. 



We also examined how developmental changes in 

 the acoustico-lateralis system might relate to sen- 

 sory thresholds that are critical in predator avoid- 

 ance. Larval herring begin to demonstrate their char- 

 acteristic C-shape startle response to environmen- 

 tal stimuli at a length of approximately 27-30 mm 

 SL (Batty, 1989). This behavior does not become fully 

 functional until the otic bullae are filled with air (22- 

 30 mm), the air bladder and its connection to the 

 acoustico-lateralis system are complete, and the lar- 

 vae acquire the ability to swallow air at the surface 

 (30-40 mm). Blaxter and Fuiman (1990) state that 

 "filling the bullae with gas is probably the most sig- 

 nificant event in the development of the anatomy of 

 the sensory systems so far as predator evasion is con- 

 cerned." They observed that successful attacks by 

 predators were reduced from 77% to 32% following 

 the filling of the bullae. Without a mechanism to 

 adjust the volume of air with pressure changes, the 

 fish are restricted in their ability to migrate verti- 

 cally, limiting their ability to search for prey, to avoid 

 predators, and to control transport (Hossetal., 1989). 

 Because it is virtually impossible to determine the 

 condition of the otic bullae in field-collected larvae, 

 the otolith to fish-length relationship was analyzed 

 to determine its usefulness in estimating the timing 

 of this critical developmental landmark. 



Methods 



Atlantic herring larvae were collected on 24 cruises 

 at selected stations within a standard grid of sam- 

 pling stations covering the western Gulf of Maine, 

 Georges Bank, and Nantucket Shoals during the 

 autumns and winters of 1976-77, 1988-89, 1989- 

 90, 1990-91, 1991-92, 1992-93, and 1993-94 (Fig. 

 1; Table 1). The samples were collected with a con- 

 tinuous double-oblique haul by using a 61-cm bongo 

 net sampler (0.505- and 0.333-mm mesh) deployed 

 to a maximum depth of 200 m ( 100 m in 1976-77) or 

 to within 5 m of the bottom in shallower areas. Fur- 

 ther details of the sampling gear and protocols can 

 be found in Posgay and Marak (1980) and Sibunka 

 and Silverman (1984). Atlantic herring larvae were 

 removed for otolith analysis immediately following 

 the haul and preserved in 95% ethanol. 



In the laboratory, larvae representative of the size- 

 range collected were selected for analysis. Standard 

 length was measured to the nearest 0.1 mm prior to 

 removal of the otoliths. Differential shrinkage with 

 respect to standard length was corrected by using 

 Theilacker's (1980) algorithm, which was specified 

 and discussed in Bolz and Lough (1983). The 2 

 sagittae and 2 lapilli were dissected from each fish 

 and mounted whole on microscope slides with Per- 

 mount. The growth increments on most of the otoliths 

 were discernible without any further preparation. 



The sagittae were viewed under a Zeiss compound 

 microscope with transmitted light. The number of 



40 



72 



71 



70 



I » 



67° 



t,H 



... 



Figure 1 



Station locations in the Gulf of Maine-Georges Bank region where larval herring were 

 collected for otolith ageing. 



