FISHERY BULLETIN: VOL. 73, NO. 3 



and unselective fishing. To achieve this com- 

 parison, records of weight and size composition 

 were kept for each catch made during the 

 experiments. 



MATERIALS AND METHODS 



For the experimental animal, Tilapia mossam- 

 bica was chosen. This species is hardy and will 

 grow readily in experimental tanks. It also is used 

 widely in tropical pond culture and thus has some 

 economic importance. Since it is a mouth breeder, 

 handling and exploitation were done only at 

 approximately 2-mo intervals. 



Tanks, feeding, etc. were as reported in Silliman 

 (1970) and represented a modification of the 

 methods of Uchida and King (1962). Briefly sum- 

 marized, the procedures were to raise the two 

 populations in hatchery-type troughs of 850-liter 

 (225-gallon) capacity. Water condition was main- 

 tained either by changing it biweekly or by a con- 

 tinuous dribbling flow into the head of each trough 

 plus bimonthly partial changing. Temperature 

 was maintained at 80° ± 5°F or 26.7°C (weekly 

 means). Illumination was by fluorescent light 12 h 

 per day. Feeding schedules and water condition 

 are detailed in Tables 1 and 2. 



Rectangular enclosures at the standpipe ends 

 were separated from the rest of the troughs by 

 plates with 3-mm holes through which the newly 

 expectorated "fry" could escape, thus furnishing 

 them refuges from cannibalism by the adults. 



After each counting, all fish in the refuges were 

 placed in the main part of the tanks. 



Fishing was done at approximately 2-mo inter- 

 vals by removing each nth fish for fishing rate 1/n 

 {n was always an integer). For the selectively 

 fished population, all fish were placed on one side 

 of a grid consisting of 25-mm diameter vertical 

 glass rods spaced 25 mm (22 mm in latter part of 

 experiment) apart. All fish were provided an op- 

 portunity to swim through the spaces between the 

 rods; only those which could not do so were fished. 

 In the unselectively fished population all sizes 

 were fished except the fry (under 4-mm thick- 

 ness). 



Counting was done simply by netting fish from 

 one container to another. For weighing, fish were 

 drained in a net and then placed in a previously 

 weighed container of water. Fish weight was ob- 

 tained by subtracting the tare from the total. 



All caught fish and the preexploitation stocks 

 were measured for thickness and length. They 

 were categorized as immature (where sex could 

 not be determined by external inspection), male, 

 and female. Sex determination was based on the 

 characteristics set forth by St. Amant (1966). 

 Length (total length to outermost tip of caudal fin) 

 was measured on a board with millimeter scale and 

 head block. Thickness was measured on the same 

 device, plus a sliding block; the fish were held 

 upright between the sliding block and the head 

 block with firm pressure for the thickness 

 measurement. Fish for the pretest measurements 



Table 1.— Food placed in tanks, grams. 



'Diet was varied Initially to achieve optimal reproduction and growth; it was stabilized at the 

 listed amounts on 18 June 1967, month 10.6. 



^Commercal makes of dry food. 



'This feeding was combined with the Friday A.M. feeding in 37 out of 183 wk. and with the Sun- 

 day feeding once. 



496 



