FISHERY BULLETIN: VOL. 73, NO. 3 



express genotypic differences which otherwise 

 might cause changed growth patterns. In other 

 words, the degree of selection imposed did not 

 work against females because they were unable to 

 achieve extra large size in any event. This 

 hypothesis cannot be tested with data from the 

 present experiment. 



Fishing in the experiment was similar to that in 

 a commercial fishery, with the vertical slots in the 

 apparatus corresponding to the meshes of com- 

 mercial gear. Results may, therefore, be of some 

 significance in fishery management. They suggest 

 that as wide a range of sizes as possible be included 

 in the catch. An appropriate balance should be 

 struck between the possible higher market value 

 of large fish and the lesser yields that may be 

 achieved under selection. Also, the possibility of a 

 genetic change in growth pattern under selection 

 should not be overlooked. 



SUMMARY 



1. Two populations of Tilapia mossambica were 



grown with as nearly identical space, water 

 condition, and food as possible. 



2. After a period of initial growth each population 



stabilized at a weight of about 10 kg. Numbers 

 were less stable at this time and ranged from 

 173 to 218 fish. 



3. To increase genetic variability, 45-47 immature 



fish of Malacca descent were added to each 

 population at month 47.3. 



4. Exploitation was started at month 39.2, at 10% 



per 2 mo (1.0-2.6 brood intervals) and 

 increased to 20% at month 63.2. Selective 

 fishing was from fish which could not pass 

 through 25-mm (later 22-mm) vertical slots 

 between glass rods. Unselective fishing was 

 from all fish except fry (under 4-mm thick- 

 ness). 



5. Recruitment was estimated from data of stock 



number, mortality, and catch. Reproductive 

 lag was 2 mo. The stock-recruitment relations, 

 roughly fitted with parabolas, suggested 

 greater recruitment in the selectively fished 

 stock than in the unselectively fished one. 



6. Two rectilinear thickness-length relations were 



calculated, one for immature and male fish and 

 another for females. 



7. Catch thickness frequencies for the unselec- 



tively fish population revealed modes corre- 

 sponding to peaks of recruitment. 



8. Catch thickness frequencies for the selectively 



fished population, compared with those for the 

 unselectively fished population, demonstrated 

 that the device for selection at 25 and 22 mm 

 was almost completely effective. 



9. The exploitation-yield relation was assessed by 



fitting Fox exponential surplus-yield models 

 to data from both populations. Fitted models 

 indicated a higher maximum sustainable yield 

 in weight for the unselectively fished popula- 

 tion than for the selectively fished one. 

 Eflficiency of food conversion was 29-36%. 



10. Growth rates from catch thickness frequen- 

 cies, together with the 2-mo reproductive lag, 

 suggested a generation length of 6 mo. 

 Recruitment records indicated three genera- 

 tions under exploitation for the selectively 

 fished population and four during the same 

 period for the unselectively fished one. 



11. To test for genetic effect of selection, a group 

 of 46 fish was selected from each population. 

 These were matched as closely as possible by 

 size and sex composition and grown under 

 previously established standard conditions. 



12. Growth in length over a period of 150 days was 

 significantly greater among males from the 

 unselectively fished population than among 

 males from the selectively fished one. Growth 

 for females in the two groups was practically 

 identical. 



13. GroMd:h in total weight was distinctly greater 

 for the group from the unselectively fished 

 population than in that from the selectively 

 fished one. 



14. It was concluded that these experiments 

 demonstrated diminished total yield and re- 

 tarded male growth in the selectively fished 

 population compared with the unselectively 

 fished one. An hypothesis based on epistasis 

 was advanced to explain lack of growth re- 

 sponse among females. 



15. As applied to commercial fisheries, the 

 experimental results suggest fishing as wide a 

 range of sizes as possible. If economic gains 

 from selection are indicated, they should be 

 balanced against possible costs in reduced 

 total yield and retarded growth rate. 



ACKNOWLEDGMENTS 



I am grateful for the advice and encouragement 

 rendered by Francis M. Fukuhara, Frederick M. 



506 



