382 



Fishery Bulletin 93(2), 1995 



eggs were the youngest to appear in significant quan- 

 tities (17.9% of total eggs in the stratum; Table 3). 

 The thermal history model (Table 2) predicted that 

 stage-13 eggs should have occurred between 370 and 

 100 m depth, which is within the sampling ranges of 

 the nets (tow 537, descent) which caught the great 

 majority of them (Fig. 7). Stage-14 eggs were pre- 

 dicted to be from 285 to 45 m (Table 2), and these 

 eggs were found evenly distributed between the 1— 

 200 and 200-400 m strata (Table 3). Stage-15 to 

 stage-26 eggs approached neutral buoyancy (Fig. 3) 

 and accumulated in the mixed layer on the basis of 

 their high proportional occurrence (Table 3; summary 

 column in Fig. 7). 



In the 200 to 400 m MOCNESS stratum, stages 

 12 and 13 were most common (Table 3), and the model 

 (Table 2) predicted that these stages would be cen- 

 tered in this stratum. Stages 11 and 14 were also 

 common (Table 3). Their expected mean depths were 

 50 m deeper and shallower than this stratum, re- 

 spectively (Table 2), and most of the stage-11 and 

 stage-14 eggs in this stratum were caught in the 

 deeper and shallower parts of it (tows 544 and 537, 

 ascent, respectively; Fig. 7). Eggs approaching neu- 

 tral buoyancy were caught in the upper part of this 

 stratum (tow 537, ascent), which could be expected 

 since this net sampled the lower mixed layer. 



In the 400 to 600 m MOCNESS stratum, the domi- 

 nant stages were 10, 11, and 12 (Table 3). The model 

 (Table 2) predicted that all these stages would be 

 most common in the upper half of this stratum, and 

 the nets that caught the great majority of them were 

 in tows 544 and 545, fishing between 467 and 374 

 m; Fig. 7). Of the total eggs in this stratum, 4.3% 

 were in stage 13, which was unexpected from the 

 model, even taking into account the modelled ranges 

 of occurrence (Table 2). It is likely that a low rate of 

 misidentification of stages (between stages 12 and 

 13) caused these unexpected results. This would be 

 expected of any staging system that attempts to place 

 eggs into discrete categories when they are actually 

 undergoing a continuous growth process. 



In the 600 to 800 m MOCNESS stratum, the domi- 

 nant stage was 4 (Table 3). This stage was predicted 

 to occur between 745 and 645 m (Table 2) which was 

 within the sampling range of the nets that caught 

 the great majority of them (tows 545 and 546, fish- 

 ing between 783 and 670 m; Fig. 7). The next most 

 abundant stage was 3 and was predicted to occur 

 between 795 and 695 m. The deep net on tow 545 

 caught most of these eggs and fished between 780 

 and 670 m. Tow 537 (ascent) caught some older eggs 

 (stages 20-24) that evidently were sinking. These 

 eggs were somewhat younger than were predicted to 

 sink (Fig. 3), suggesting that there is some variabil- 

 ity in the stage when sinking begins. 



In summary, over all depth strata the predictions 

 of modelled depth-at-stage were consistent with 

 depth distributions of the dominant and subdomi- 

 nant stages caught in the MOCNESS tows. This con- 

 sistency suggests that the basic parameters of the 

 model are sound and are reasonably well estimated. 

 One factor that might have caused the expected ver- 

 tical distributions of eggs to deviate from the observed 

 distributions may have been significant amounts of 

 spawning far off the bottom. However, there was no 

 evidence of this, in that significant numbers of very 

 young eggs were seen only in the deepest stratum 

 (Fig. 7, Table 3), and no "spawning plumes" of fish 



