SILLIMAN: EFFECT OF CROWDING 



i 3 



2- 



» I 



(9 

 O 



oL 



.8 12 16 2,0 



EFFECTIVE EFFORT 



2 4 



28 



32 



200 



400 600 800 1,000 1,200 1,400 1,600 

 BIOMASS (groms) 



Figure 3. — Effort-CPUE line (least squares fit) and 

 biomass-yield curve, fitting of Fox (1970) model. CPUE 

 assumed proportional to population size. L indicates 

 population in 155.2-liter tank and S in 77.6-liter tank. 



smaller population, L has a positive deviation 

 and S a negative. 



An explanation for this seemingly paradoxical 

 finding may be found in the source of the yield 

 at each exploitation rate. At the lower rate, 

 during months 0-14, the populations were large 

 and no recruitment occurred (Figure 2, Table 3) . 

 The yield that did occur, then, resulted entirely 

 from growth. Under the higher exploitation 

 rate prevailing during months 16-24.3, the pop- 

 ulation was smaller and recruitment did occur 

 (Figure 2, Table 3) , furnishing part of the yield. 

 Growth could have been retarded in the larger 

 tank by greater expenditure of energy in swim- 

 ming. Recruitment could have been favored in 

 that tank by psychological factors controlling 

 spawning, by greater opportunity for newly re- 



leased young to escape cannibalism, or by a com- 

 bination of these factors. 



The finding of greater growth in the smaller 

 tank is in contradiction to the conclusions of 

 Chen and Prowse (1964) for T. mossambica. 

 Examination of their data, however, shows that 

 these conclusions were based on results from 

 their largest ponds (0.10-0.40 ha) after a year 

 of growth. Data comparing their smaller ponds 

 were available only to 6-months' growth and are 

 less clear cut. For instance, at 6 months their 

 0.04-ha pond showed greater growth than their 

 0.20-ha pond. Even their smallest pond (0.004 

 ha and 76 cm depth) had a volume of 31,000 li- 

 ters, or 200 times that of the larger tank in the 

 T. macroce])hala experiments. With a diflferent 

 species (even though in the same genus) and 

 such a vast difference in size of habitat it is per- 

 haps not surprising that their findings are at 

 variance with those reported here. 



In both the present experiment and that of 

 Chen and Prowse, the term "crowding" can be 

 used only in a relative sense and not in the sense 

 of lack of space to move about. Assuming the 

 fish have the same density as water (nearly 

 true) , the maximum population in the smaller 

 tank of the present experiment occupied only 

 1.72 9f of the water. The maximum concentra- 

 tion in Chen and Prowse's experiments was 

 0.02%. Thus, whatever effects occur must re- 

 sult from such factors as relative distance swum 

 or sociological phenomena, like aggression. Chen 

 and Prowse chose the latter, although it is a bit 

 difficult to imagine when there are 2 or less kg 

 of fish per 10,000 liters of water. 



Finally, yields may be compared as measures 

 of conversion' of food to fish flesh. The maxi- 

 mum bimonthly sustainable yield from the gen- 

 eral population curve (Figure 3) is about 140 g 

 (calculated value, 137.8 g) . Food provided dur- 

 ing each 2-month period weighed 749 g {S% 

 weeks X 86.4 g weekly total fed, as shown in 

 Table 1). Apparent conversion was thus 18%. 



The above results may be compared with those 

 from the growth experiments of Swingle (1960) 



= Used here in the sense of net growth in weight 

 (^ sustainable yield) expressed as a percentage of 

 weight of food made available to the population. 



697 



