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



amount of time that larvae or juveniles are vul- 

 nerable to capture. Because the growth rate is 

 not constant over the size range of the young fish 

 sampled in this study, we must correct for this 

 bias. The production of larvae by age is defined 

 as the abundance of larvae by size divided by the 

 duration of gi-owth through the size class, which 

 is expressed as catch/tow/day. This allows us to 

 calculate instantaneous mortality rate (IMR), 

 and thus, construct a mortahty curve. We need 

 to know age and growth rate of saury to calcu- 

 late production at age and mortality rates. 



In northern anchovy, Engraulis mordax, 

 there have been extensive studies of eggs and 

 larvae, and the methods to calculate the produc- 

 tion and mortality rate were established (Zweifel 

 and Smith 1981; Lasker 1985; Lo 1985; 1986). 

 Ichthyoplankton survey data, embryonic incuba- 

 tion times, and larval growth rates are the es- 

 sential parameters for this method. Further, one 

 needs to have information on possible biases of 

 tow data. For the Pacific saury, we have 15 

 years of larval net-tow data, a newly developed 

 growth model based upon otolith growth incre- 

 ments (Watanabe et al. 1988), and information 

 on bias correction of tow data for day and night 

 differences. All these make it possible to com- 

 pute fish production at different ages and mor- 

 talities. Extending the mortality curve to age 0, 

 we can calculate larval production at hatching, 

 which might be the best index of reproductive 

 level so far available. In this paper we selected 

 the exponential decay mortality curve and used 

 it to calculate the larval production at hatching 

 and the daily IMR of Pacific saury in the north- 

 western Pacific Ocean for 1972-86. 



METHODS 



Data Source 



The Pacific saury spawns nearly all year round 

 in the entire northwestern Pacific Ocean. The 

 sampling areas and seasons were somewhat dif- 

 ferent from year-to-year, but our data set of 15 

 years (1972-86) included around-the-clock sam- 

 ples taken all year from a large area of the north- 

 western Pacific, lat. 29-45°N, long. 129-174°E 

 (Fig. 1). 



Saury larvae and juveniles were collected by a 

 surface ring net that was towed for 5 minutes at 

 2 knots. The mouth diameter was 1.3 m, and 

 therefore the surface area covered by one tow 

 was 401.3 m^. A mesh size of the 3.0 m forward 

 part of the net was 2.0 mm, and the 1.5 m rear 



part was 0.33 mm. Samples were washed down 

 after the 5 min tow and preserved in 10% unbuf- 

 fered formalin. Larvae and juveniles of saury 

 were measured to the nearest 0.1 mm knob 

 length (KnL), the distance from the tip of the 

 lower jaw to the posterior end of a muscular 

 knob at the base of the caudal peduncle. They 

 were then grouped into 11 size classes from 7.5 

 mm (including 5.0-9.9 mm) up to 57.5 mm 

 (55.0-59.9 mm). The midpoint and both Umits of 

 each class were then converted into the capture 

 size before preservation, using a shrinkage fac- 

 tor by formalin preservation (0.97) reported by 

 Theilacker (1980) for northern anchovy. The 

 midpoint of each size class was used to represent 

 the class, though it is not a mean age of the class. 

 The capture size was converted to age, and dura- 

 tion of each size class was obtained using the 

 growth model developed by Watanabe et al. 

 (1988). The growth equation from hatching to a 

 100 mm juvenule is 



KnL = 5.90 exp((0. 0865/0. 0293) 



X (1 -exp( -0.0293 0)) 



where KnL is the fish knob length in mm after 

 capture (before preservation) and t is the age in 

 days from hatching. 



Bias Corrections 



An ichthyoplankton survey is essential for 

 most pelagic fish research (Smith and Richard- 

 son 1977), but it is not bias free. The most com- 

 mon biases in catch of fish eggs and larvae are 

 extrusion of eggs and larvae through net mesh 

 and evasion of a net frame by fish. For the Pa- 

 cific saury, the mesh size of the anterior part of 

 the net, 2.0 mm, might be large enough to lose 

 larvae by extrusion. Vulnerability of the fish to a 

 net tow varies throughout a day due to changes 

 in evasion abilities. Availability of the fish to a 

 tow changes as well due to diel vertical move- 

 ment. These factors result in differences in num- 

 bers of fish captured by net tows during the day 

 and night. For the analysis, we defined 10 diel 

 time periods based upon angles between the cen- 

 ter of the sun and the celestial horizon. We used 

 four periods in the morning — DAWN (DWN), 

 MORNING TWILIGHT (MTW), SUNRISE 

 (SRS), and MORNING (MRN)— and another 

 four periods in the evening — AFTERNOON 

 (AFT), SUNSET (SST), EVENING TWI- 

 LIGHT (ETW), and DUSK (DSK). Time dura- 



602 



