44 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



water control tanks compared favorably with the 

 o^rowth rates of the 1,000-fish lot in the space- 

 growth experiment. We expected that the growth 

 rates in the control tanks would be somewhat simi- 

 lar to those experienced in the 4,000-fish lot of the 

 space-growth experiment. The growth rates in 

 the two control tanks were 3.2 and 2.8 mm. per 

 week over an 8-week period, while the growth rate 

 of tlie 4,000-fish lot was 2.1 nnn. 



The fish were raisetl under seemingly identical 

 conditions, except for the time of year that the 

 experiment was conducted. The 4,000-fish lot was 

 reared from mid-May to mid- August, while the 

 control-tank lots were reared in August and Sep- 

 tember. Although there are no temperature rec- 

 ords for these gi'owth experiments, the tempei-a- 

 ture record of the heating experiment (table 8) 

 discloses a difference of about 4° F between May 

 and August in control tank 13. Presumably, this 

 difference also applied to the assault-boat tanks. 

 Therefore, it is quite reasonable to conclude tliat 

 water temperature is important in obtaining fast 

 growth; however, the optimum temperature has 

 not been determined. 



Some other factor, possibly environmental, 

 chemical, or genetic, acting singly or in combina- 

 tion with temperature may also have contributed 

 to this discrepancy in growth rates. The end re- 

 sults of the salinity-growth experiment paralleled 

 the results of the salinity-spawning experiment, 

 however, indicating that a commercial rearing 

 plant can, and sliould be operated on a brackish- 

 water system, thereby reducing or eliminating the 

 high cost of using fresh water. 



CANNIBALISM AND PREDATION 



Cannibalism and predation are important fac- 

 tors in the successful rearing of young tilapia. 

 Our initial plan to rear tilapia fry to bait-fish size 

 in a single large iry tank met with a major setback 

 when about 2 months after production started we 

 discovered juveniles, ranging in length from 25 

 to 38 mm. (1 to 1.5 in.), chasing and consuming 

 newly emerged fry that were being released into 

 the tank. This situation was remedied by install- 

 ing Monel-screen partitions to separate tlie tank 

 into three compartments and segregate the young 

 according to size. Cannibalism among tilapia was 



also reported by Chen (1953). The results of his 

 observations are summarized in table 18. 



Table 18. — Size relation between intraspecific predator and 

 prey in T. mossambica 



[Data from Chen (1953)] 



Size of predator 



0.75 Inch 



3 to fi inches.. 

 7 Inches 



Size of prey 



Up to 0.38 inch. 

 Up to 0.5 inch. 

 Up to 1.25 inches. 



In order to extend the findings of Chen and 

 further define this intraspecific, predator-prey 

 size relation, the following experiments were con- 

 ducted in 35-gallon aquariums. Eight juvenile 

 groups (predator) of different average lengths 

 were selected. The average length of the groups 

 ranged from 20.4 to 64.4 mm. (about 0.75 to 2.5 

 in.). A predator group consisting of 10 juveniles 

 was first measured and placed in an aquarium. A 

 second group of 20 to 30 young (prey), all fry 

 or fry and juveniles, was also measured and 

 placed in the same aquarium. Each of the eight 

 experiments was carried out for a period of 72 

 liours, after which time the remaining prey were 

 removed and measured. The fish did not receive 

 any supplementary feeding during the period of 

 the experiment. By comparing the lengths of the 

 remaining young with the lengths of the young 

 that were put into the aquarium,' we were able to 

 determine the maximum size of the young that 

 were killed or consumed by each size group of 

 predators. The results are given in table 19 and 

 figure 17. 



20 30 40 so 60 



AVERAGE LENGTH OF JUVENILE GROUP IHM) 



Figure 17. — Relation between size of jnvenile groups 

 (predators) and maximum size of young killed or 

 consumed. 



