FISHERY BULLETIN: VOL. 84, NO. 4 



The t m can then be computed from L m with the von 

 Bertalanffy growth equation. The t m for the seven 

 species, which is assumed to be the same for both 

 sexes of a species, is given in Table 1, and the ratio 

 of spawning stock biomass under exploitation to the 

 unexploited spawning stock biomass is presented for 

 three levels of F (Table 10). As expected, the ratio 

 decreases as F increases. However, without the 

 spawner-recruit relationship, it is difficult to deter- 

 mine the extent that the spawning stock biomass 

 can be reduced before recruitment is substantially 

 affected. It has been suggested that as a lower 

 bound, the spawning stock biomass should not be 

 reduced below 20% of its unexploited level before 

 there is a deleterious reduction in recruitment 

 (Beddington and Cooke 1983). The level of F = 1.0 

 is the largest level of F which insures that the 

 relative spawning stock biomass for all the species 

 does not fall below 20% and hence the spawning 

 stock approach also estimates the MSY for the bot- 

 tom fish in the Marianas at 109 t/year. 



Table 10. — The ratio of spawning stock biomass to unexploited 

 spawning stock biomass for three levels of fishing mortality (F) at 

 the age of entry which maximizes the yield per recruit. 



DISCUSSION 



The assessment proposed here is a multispecies 

 approach which is most suitable for resources where 

 prey-predator interactions are negligible. Two 

 assumptions initially required to implement this 

 program, i.e., constant recruitment and that the 

 resource be essentially unexploited, can in some 

 instances be relaxed. Simulation results suggest that 

 if recruitment is seasonal and a pooled length fre- 

 quency is constructed from individual length- 

 frequency samples collected over the year, the 

 length-frequency based method used here to esti- 

 mate mortality produces an essentially unbiased 

 estimate (Ralston 4 ). Furthermore, the assumption 



that stocks be unexploited can be relaxed if an 

 estimate of the average of F for the archipelago can 

 be obtained. Then M can be estimated by the dif- 

 ference between F and total mortality, and instead 

 of estimating unexploited recruited biomass from 

 the CPUE survey, the biomass under F will be 

 estimated, and yields calculated as the product of 

 exploited biomass with the ratio of yield/biomass 

 resulting from F computed from the Beverton and 

 Holt yield equation. 



The estimate of maximum equilibrium yield from 

 the Beverton and Holt (1957) equation for the deep 

 slope snappers and groupers from 22 banks in the 

 Mariana Archipelago is 109 1 annually with a fishing 

 mortality of 1.0. About 70% of this yield would be 

 expected to come from the southern islands of the 

 chain, including Guam and Saipan. Another 27% 

 would come from the northern islands and only 3% 

 from the seamounts (Table 11). 



The mean of the annual sustainable yield levels 

 per nautical mile of 200 m contour for the northern 

 banks, southern banks, and western seamounts are 

 212.9, 228.5, and 264.4 kg, respectively, with a ratio 

 of total yield for the archipelago to the total length 

 of the 200 m contour of 222.4 kg/nmi (95%) C.I. of 

 165.3-279.6) (Table 11). Detailed bathymetry data 

 to establish a correspondence between contour 

 length and area are available from Guguan Island 

 in the northern Marianas, and it is estimated that 

 1 nmi of 200 m isobath corresponds to 0.23 nmi 2 of 

 habitat in the 125-275 m depth range (Polovina and 

 Roush 5 ). Based on this correspondence the unit MSY 

 of 222.4 kg/nmi of 200 m contour for the Marianas 

 is equivalent to about 1.0 t/nmi 2 or 0.3 t/km 2 . 



These values suggest that the Marianas may be 

 slightly less productive for bottom fishes than the 

 Hawaiian Archipelago where a lower bound esti- 

 mate for MSY of 272 kg/nmi of 200 m contour was 

 obtained from a stock production model applied to 

 commercial catch and effort data that did not include 

 the recreational fishing component of snappers and 

 groupers. Also, an estimate of 286 kg/nmi of 200 

 m contour was derived from an ecosystem model ap- 

 plied to an island system in the Northwestern 

 Hawaiian Islands (Ralston and Polovina 1982; 

 Polovina 1984). 



The species composition of the catch should 

 depend to some extent on levels of F and t c . As F 

 increases and t c decreases, the contribution of 



4 Ralston, S. The effect of pooling length-frequency distributions 

 on mortality estimation in seasonally breeding fish populations: 

 A Monte Carlo simulation. Manuscr. in prep. Southwest Fish- 

 eries Center Honolulu Laboratory, National Marine Fisheries Ser- 

 vice, NOAA, Honolulu, HI 96822-2396. 



6 Polovina, J. J., and R. C. Roush. 1982. Chartlets of selected 

 fishing banks and pinnacles in the Mariana Archipelago. South- 

 west Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, 

 Admin. Rep. H-82-19, 7 p. 



768 



