FISHERY BULLETIN: VOL. 87, NO. 3, 1989 



think that the downward bias is due to extrusion 

 through the mesh. The mesh retention rate of 

 larvae in the first two size classes could be esti- 

 mated from the ratio of the observed catch/tow/ 

 day and the predicted larval production com- 

 puted from the overall average mortality curve 

 of 15 years. The estimate of mesh retention rate 

 for larvae of 7.5 mm (3.2 days old) is, 



Observed: 0.158 (larvae/tow/day) 

 Predicted: 0.978 (larvae/tow/day) = 



exp(-0.078 X 3.2) 

 Retention rate: 0.16 = 0.158/0.978 



1.255 



and for larvae of 12.5 mm (10.5 days old) is, 



Observed: 0.345 (larvae/tow/day) 

 Predicted: 0.553 (larvae/tow/day) = 1.255 



exp(-0.078 X 10.5) 

 Retention rate: 0.62 = 0.345/0.553. 



DISCUSSION 



Production and mortality rates of fish larvae 

 and juveniles cannot be obtained without good 

 growth models, which are indispensable in calcu- 

 lating stage durations. A new growth model 

 based on growth increments of otoliths has re- 

 cently been estabhshed (Watanabe et al. 1988) 

 enabling calculation of mortahty rates. The cur- 

 rent results are virtually the first attempt to 

 calculate larval production at hatching. These 

 should provide the best index of reproductive 

 level ever obtained. 



Sabhn (1978) first calculated mortality rate of 

 juvenile saury for 8 individual years (1968-75), 

 using catch/tow values of two size groups, 26-30 

 mm and 46-50 mm, and 20 mm/mo as a growth 

 rate of this size range. The average monthly 

 IMR estimated by Sablin was 1.15 with a range 

 from 0.76 to 1.62. He obtained the IMR from 2 

 size groups for individual years, whereas we 

 used catch data from 9 size classes in computing 

 the IMR. We believe the estimates of IMR re- 

 ported here are an improvement over those of 

 Sabhn. 



Monthly mortality rates using the Sablin 

 IMRs (1978) ranged from 53.2% to 80.2% with an 

 average of 68.3%. These values were much lower 

 than the monthly rates calculated from the daily 

 IMRs in this paper, 70.8-96.8% with an average 

 90.4%. The discrepancy between the two seems 

 to derive from the difference in growth rates 

 used for the computations (Fig. 7). Sabhn used a 

 rate of 20 mm/mo (0.67 mm/d) for the size range 



from 26-30 mm to 46-50 mm. We used the 

 growth model by Watanabe et al. (1988), which 

 shows that sauries gi'ow from 27.5 mm to 47.5 

 mm in 17.2 days (1.16 mm/d). This was 1.7 times 

 faster than growth rate of Sablin. We recalcu- 

 lated monthly mortality rates using the Sabhn 

 IMRs and the Watanabe et al. growth rate of 

 1.16 mm/d producing estimates close to ours, 

 73.4-94.1%. with an average of 86.5%. 



The absolute value of total annual production 

 of newly hatched larvae of the Pacific saury can 

 be calculated from our results using the assump- 

 tions that 1) all fish in a vertical water column 

 are in the upper one meter, and 2) those in a 

 volume of water strained are captured by the net 

 during the NIT period. The values of larval pro- 

 duction at age are on daily basis over the area 

 covered by a net tow. Thus, the annual total 

 production of the hatched larvae can be com- 

 puted as below: 



Annual Total Production of Hatched Larvae = 

 Po- 365 •A/401.3 



where Pq is larval production at hatching, A is 

 total spawning area in m^, and 401.3 is surface 

 area in m^ covered by one net tow. However, the 

 distribution of saury larvae and juveniles in the 

 northwestern Pacific extends far to the east of 

 Japan, and we cannot delimit the total spawning 

 area. Thus, calculation of the absolute larval pro- 

 duction of saury at hatching for the entire area is 

 not practical. However, the calculation of abso- 



60 



'50 



il40 



;3o 



20 



10 







'(11 



10 20 30 ao 50 60 



AGE IN DAYS 



Figure 7.— Growth curves used by Sablin (1973) (1) and 

 by us (2) for Pacific saury. 



610 



