lo8
channels, to 2,682 kg/ha in vegetated waters. The areas of different types
of habitat, at low and high water, were calculated from planimeter measure-
ments on 1:50,000 maps. Meanwhile tilapia growth rates were determined
by scale ring analyses (Chapman et al. 1971; Kapetsky 1974). S. macrochir
and T. rendalli were considered harvestable at the end of the second year,
S. andersonii end of third year.
The resulting computed biomass, production (total and available), yields
and turnover rates of the tilapia populations on the eastern floodplain
(Table 2) are shown in Figure 5. This brings out clearly that the total pro-
duction is considerably greater than the biomass present in T. rendalli
(turnover rate A/B, 1.59), equal to the biomass in S. macrochir (A/B, 1.0),
and less than the biomass in S. andersonii (A/B, 0.75).
Kapetsky then computed a yield for the whole Kafue floodplain (121,000
ha), weighting for areas of lagoons and channels. This gave an estimated'
total tilapia production of 70,000 t/yr from the whole floodplain, total
yield 13,000 t/yr and available yield 7,000 t/yr. Present catch is 5,000 t/yr
of all species (Dudley and Scully 1980), so if computations are realistic
it looks as though only about half the possible tilapia crop is being caught
(my interpretation of the findings).
The Kariba dam across the Zambezi was closed in December 1958. Here
chemofishing in lake coves was used to determine fish biomasses (Balon
1974). The calculations of production etc., for each species and the whole
lake had, however, to be revised (Mahon and Balon 1977, whose revised
estimates for total production, available production, total yield and available
yield, all species, were 38%,105%, 67% and 107% respectively of the original
estimates). Tilapias were a much less important part of the catch (20 species
considered) than on the Kafue floodplain. However, S. mortimeri (then
called S. mossambicus), though present at a 'low population density' (2,122
fish per inhabited ha), had a mean biomass of 215.7 kg/ha and was estimated
to give the third highest total production (178.5 kg/ha/yr) (after a mormyrid
and Alestes sp.), and the highest total yield (139 kg/ha/yr) of the nine
economically preferred species. T. rendalli was much less abundant, but had
a higher turnover rate than S. mortimeri (1.15 compared with 0.83).
Lake George (Ugapda). In an attempt to determine how this shallow
equatorial lake, mostly less than 3 m deep (250 km2) continues to produce
such good catches of tilapia (S. niloticus, which forms 80% of the catch,
and S. leucostictus) an International Biological Programme team spent six
years studying production at different trophic levels (Dunn 1972; Burgis
et al. 1973; Burgis and Dunn 1978; Burgis 1978; Gwahaba 1975, 1978).
Fish biomasses (of 10 species) were determined using openwater seines; they
were computed to range from 60 kg/ha in openwater to 900 kg/ha inshore
(mean 220 kglha). The determination of fish growth rates proved very
difficult in this equatorial lake, but for S. niloticus was indicated by length
frequency analysis, of a particular cohort of young fish, to be about 1.2
cmlmonth (Gwahaba 1978). The fish yield based on catch statistics over
twenty years fluctuated around a mean of 3,461 t/yr for all (10) species,
2,790 tlyr for S. niloticus, equivalent to 137 kg/ha/yr for all species, 111
kg/ha/yr for S. niloticus. The weight of fish landed had been maintained
but, as already discussed above, the maturation size of S. niloticus had
sean pound
(Sean Pound)
#1