FISHERY BULLETIN: VOL. 73, NO. 3 



Table 3.— Population and catch, Tilapia mossambica. 



10 = 1 August 1966. 



^Because of selection problems, actual rates varied considerably from these. In analyses, effective rates 

 in terms of weight were used. 



3Populations re-equalized after accidental mortality from temporary relocation of fish. 

 "New fish added for genetic variability. 



sufficient genetic variability might exist in the 

 populations to permit a genetic effect of selection. 

 Since one of the major objectives of the 

 experiment was to detect such an effect, I decided 

 to add outside stock. During the 2-mo period 

 preceding month 47.3 (Table 3), I added 45-47 (2 

 fish uncertainty due to counting difficulties) im- 

 mature Tilapia mossambica from Arizona to each 

 population. These fish were descendants of ones 

 from Malacca. 



Exploitation was started at month 39.2 (Table 3, 

 Figures 1, 2) at a conservative 10% per 2 mo. This 

 exploitation period included 1.0 to 2.6 of the brood 

 intervals reported by various authors (Kelly 1957, 

 30-40 days; Swingle 1960, 30-40 days; Uchida and 

 King 1962, 23-61 days). Because of irregularities in 

 recruitment, population numbers (Figure 1) 

 reflected population responses less well than 



population biomasses (Figure 2). The latter, 

 however, generally reflected the expected popula- 

 tion decrease from imposition of the 10% exploi- 

 tation rate. When the rate was increased to 20% at 

 month 63.2 (Table 3, Figure 2), further declines in 

 population biomasses occurred. 



BASIC RELATIONS 



Recruitment was estimated from changes (/?int) 

 in stock number and data of mortality and catch 

 (Table 4), using the approach of Silliman (1972). 

 Because of variations in length of period between 

 counts, values of R^^jwere adjusted to a standard 

 2-mo interval (Table 4). Observation of large 

 numbers of fry in the refuges, followed by the 

 later appearance of peaks of recruitment (Table 4), 

 indicated that the "reproductive lag" was about 2 



498 



