26 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



lized appeai-s to be determined more by the num- 

 ber of ripe females than by the number of males. 



Taking advantage of this behavior to realize a 

 maximum production of young with a minimum 

 of brood stock, a ratio of 2 5' -.1$ was tried initi- 

 ally, with results that were considered to be suc- 

 cessful. A later experiment at the pilot plant 

 using a ratio of 5 5 : 1 ■? yielded less-successful 

 results, as will be described later. 



Another important consideration was the den- 

 sity of the brood stock or carrying capacity of 

 the brood tanks. Originally, it was decided to try 

 a total of 48 fish with about 1.9 square feet per 

 individual in one tank, and 96 fish or approxi- 

 mately 0.9 square foot per individual in the other 

 two tanks, each with a sex ratio of 2 9:15 

 (table 1). 



Table 1. — Number of adults of each sex and space allow- 

 ance per male and per individual, in brood tanks at the 

 pilot plant 



On October 12, 1956, the three tanks at the 

 pilot plant were stocked with adult fish aver- 

 aging 20 cm. (8 in.) in length and 150 g. 

 (5.3 oz.) in weight. These fish were supplied 

 by Hawaiian Tuna Packers, Ltd., and obtained 

 from Ewa plantation pond No. 6. The initial 

 mortality among the brood stock was very low, 

 amounting to only three males and three females 

 the first 15 days after stocking. All casualties 

 were replaced. The exact number of adult fish 

 in each of the tanks varied during the course 

 of the test, however, owing to unobserved mor- 

 talities, vandalism, and the recruitment of juven- 

 iles that escaped dipnetting and grew to adult 

 size in the brood tanks. 



The stocking of the Kewalo plant followed an 

 experimental design and will be discussed under 

 factors affecting reproduction, page 37. Stock 

 for the Kewalo plant was also obtained from 

 Ewa pond No. 6. The males averaged 142 g. 

 (5 oz.) and the females about 113 g. (4 oz.) in 

 weight. No length measurements were made. 



FEEDING 



Many investigators have reported on the feed- 

 ing habits of the various species of tilapia. Gen- 

 erally, T. mansarnbica is considered to be omniv- 

 orous (Schuster, 1952; Chen, 1953; Atz, 1954; 

 Panikkar and Tampi, 1954; Brock and Takata, 

 1955; Van Pel, 1955). Vaas and Hofstede 

 (1952: p. 35) stated that tilapia is herbivorous, 

 but will feed on planktonic Crustacea "if such 

 kind of food is more plentiful than vegetable 

 food" and will show a preference for vegetable 

 food when a mixture of the two is present. 



Feeding of the brood stock at the pilot plant 

 was started immediately after stocking was com- 

 pleted. Various types of commercial feed, such 

 as rice bran (powdered), millrun (powdered), 

 chicken starter mash (powdered and granules), 

 alfalfa pellets (%2 in. in diameter), pelletized 

 rabbit feed (%2 in. in diameter), and a pellet- 

 ized pond-fish feed (%2 in- in diameter), were 

 tried to determine which were most acceptable to 

 the fish. With a daily feeding rate of about 2 per- 

 cent of the fish weight, we observed that finely 

 divided, unpelletized feed, such as rice bran, mill- 

 run, and chicken mash, was not efficiently utilized 

 because of the small particle size. The leftover 

 feed contributed to fouling of the tanks. The 

 alfalfa pellets and rabbit feed were also found 

 to be undesirable because of their high content of 

 indigestible fiber, which collected on the bottom 

 of the tanks and also caused fouling. 



Early observations on the feeding habits of the 

 newly emerged fiy indicated that they did not 

 respond to supplementary feed until about 2 weeks 

 after they were placed in the fry tank. The 

 fry grade of pondfish feed (granules) was fur- 

 ther ground to a flourlike consistency to accom- 

 modate the very young. The larger fry and 

 juveniles reared at the pilot plant were fed pond- 

 fish feed in the cnunble grade, which was slightly 

 larger than the fry-grade granules. 



Except in feeding experiments, the brood stock 

 at the Kewalo plant was fed almost exclusively 

 on a prepared trout feed (developer grade, about 

 14 in. in diameter), which in a smaller particle 

 size was also fed to the fry (starter grade, pow- 

 dered; fry grade, granules) and juveniles (small 

 fingerling grade, crumbles). Table 2 gives the 

 composition, supplied by the manufacturer, of 

 the various feeds that were used. 



