Verreth, Johan, Promotor van Arendonk, Johan, Promotor
Komen, Hans, Co-promotor

Among the tilapias, Nile tilapia is the most important species farmed in the world and is the mainstay of many resource-poor fish farmers. The majority of its culturing is carried out in the tropics in semi-intensive environments with a wide array of pond inputs from the farm. To increase production, the overall efficiency of fish reared in these systems needs to be improved. Studies have shown that whereas about 15-30% of the nutrient input in pellet fed-pond systems is converted into harvestable products, only 5-15% of the nutrient input in fertilizer-only pond systems is converted to harvestable products. This thesis was a part of the INREF-POND project (Chapter I), which consisted of sub-projects at the fish, pond and farm level with experimental and modelling studies. The aim of the project was to find ways of increasing the efficiency of nutrient use of integrated systems. Efficient breeding programs are needed to improve the overall nutrient use efficiency of fish in fertilized ponds. This thesis aims at (i) studying the feasibility of selective breeding for the improvement of growth of Nile tilapia in low-input extensive farming conditions, (ii) determining the effects of selection on other performance traits and (iii) investigate genotype bye environment interaction in Nile tilapia. A common practice in tilapia farming is to nurse fry in hapa-in-pond systems with high protein pellets. To carry out a selective breeding in low-input condition, knowledge of factors determining growth in these systems is important. Chapter II presents results of an experiment that was carried out to determine the optimal conditions for early rearing of tilapia fry in hapa-in-pond systems. The aim of this study was to quantify the environmental and genetic effects on early growth of Nile tilapia, in hapa-in-earthen pond systems. In a pilot study, we grew swim-up fry with or without supplementary feed in hapas suspended in fertilised ponds at 5, 10, 15, and 20 fry/m2 densities. In the main experiment, we reared swim-up fry from 25 full-sib families separately for 42 days at 15 fry/m2 density in hapas suspended in two earthen ponds. Hapas were arranged in two column arrays along the sides of the ponds. Ponds were fertilized daily with chicken manure. In addition, fry in one column in each pond were fed twice daily on 40% protein pelleted feed. Results from the pilot study indicated significant effects of stocking density and treatment on fry growth. In the main experiment, the dietary treatment effect was not significant but there were large differences in growth between ponds: mean body weight at 42 days was 1.7 g in pond A and 0.4 g in pond B. Heritability (h2 ) of 42-day fry body weight estimated from the whole data set using a univariate model was 0.01±0.06. The bivariate heritability estimates were 0.59±0.19 in pond A and 0.05±0.11 in pond B. The common environmental / hapa (c2) effects were 0.14±0.06 and 0.29±0.07 in respective ponds. We found significant positive spatial autocorrelation (P = 0.02) indicating resemblance in growth of fry in neighboring hapas. Analysis of environmental variables showed that the two ponds differed significantly in dissolved oxygen. The low genetic correlation (rg = -0.27±0.69) between body weights of fry in both ponds therefore might suggest genotype by environment interactions for tolerance to low dissolved oxygen in Nile tilapia juveniles. The inability of tilapia to tolerate low temperatures is of major economic concern as it reduces their growing season and leads to over winter mortality. In Chapter III, cold tolerance of juvenile Nile tilapia was investigated and heritability estimates obtained. Eighty full-sib families were produced by mating each sire with two dams. Fry were grown in hapas suspended in earthen ponds fertilized with chicken manure, and were 41-91 days post-hatch at the start of the experiment (mean standard length 50.6 mm; mean body weight 5.1 g). Fry were tagged and exposed to low temperature in an indoor facility. Temperature was lowered from 16 °C to 11°C in 48 hours and from 11 °C to 8 °C at the rate of 1°C/day. Cold tolerance was expressed as Temperature at Death (TAD) and Cooling Degree hours (CDH). Fish mortality started at 13.6 ºC and total mortality occurred at 8.6 ºC. Mean TAD and CDH were 10.1 ºC and 298.07 respectively. Fish body weight (BW) had a highly significant effect on cold tolerance (P <0.0001). Smaller fish (<5g) were more susceptible to lower temperature than larger fish. The heritability of cold tolerance was 0.08±0.17 for CDH and 0.09±0.19 for TAD, estimated with an animal model. There was a considerable common environmental/full-sib effect for this trait (0.33±0.10 for CDH and 0.27±0.09 for TAD). These values indicate that estimation of genetic parameters for cold tolerance in tilapia should include both direct additive and common environmental effects. Based on the results of this study we conclude that the most appropriate way of enhancing cold tolerance of tilapia juveniles is by husbandry practices that increase pre-winter body weights. The effects of genotype, age, size, condition factor and diet (natural phytoplankton versus formulated protein pellets) on low-temperature tolerance of juveniles were investigated and reported in Chapter IV. This chapter compared the results of the experiments reported in Chapter III with a second experiment carried out to determine the effect of the environment and diet on cold tolerance. In the first experiment, 775 juveniles from 43 sires and 80 dams were reared under mid-summer conditions for 41-91 days. In the second experiment, 393 juveniles were produced by single-pair mating of 20 dams and 20 sires from the same brooders as in the first experiment. These fish were reared for 42 days under autumn conditions with either high protein (40%) pellets or natural tilapia diet. At the end of the growth period fish from each experiment were tagged and exposed to gradually lowered temperatures. Cold tolerance was significantly affected by genotype, size, aquarium, condition factor (P= 0.0001) and diet (P=0.0547). In both experiments, smaller fish were more vulnerable to cold stress. Age did not significantly affect cold tolerance. Fish reared under mid-summer conditions died between 13.6 ºC and 8.6 ºC while those reared under autumn conditions died between 11.7 ºC and 7.5 ºC. This suggests that acclimatization to lower temperatures before cold stress can improve the cold tolerance ability of O. niloticus. Knowledge of the heritability for growth and the correlated changes that occur in other traits due to selection for growth is important for the design of an efficient genetic improvement program. In Chapter V and Chapter VI estimates of phenotypic and genetic parameters for growth, body measurements, reproductive traits and gut length from two generations of selection for growth in a low input environment are presented. The selection environment consisted of earthen ponds which were daily fertilized with 50 kg dry matter (dm) /ha chicken manure. No supplementary feeds were provided. In total, 6429 fully pedigreed experimental fish were included in the analysis. Survival till harvest was highly variable ranging from 35% to 77% and was affected by initial weight, pond, and age effects. Body weight at harvest (BW) increased from a mean of 67.4 g in the grandparental (unselected) population (G0) to 129.5 g in G2 was affected by initial weight, pond, sex and age effects. Generations were discrete and therefore genetic parameters were estimated separately for each year. Heritability estimates for BW ranged from 0.38 to 0.60, and the heritability for survival ranged from 0.03 to 0.14. The estimated selection response was 23.4 g (34.7%) between G0 and G1 and 13.0 g (14.9%) between G1 and G2 (Chapter V). Chapter VI presents the estimates of phenotypic and genetic parameters for body size measurements, reproductive traits, and gut length. Heritability estimates for body measurements ranged from 0.4-0.6 for standard length to 0.69-0.79 for head length. Phenotypic correlations between body weight and body measurements ranged from 0.64 to 0.89. Genetic correlations were close to unity. The heritability estimate for maturity at harvest was 0.13. Heritabilities for carcass traits were estimated from G1 only and were 0.16 for gutted weight and 0.06 for dressing percentage. Phenotypic correlation between body weight and gutted weight was 0.84 and the genetic correlation was 0.20. Gut length increased with selection for body weight. Heritability estimate for gut length was 0.22. Moreover, gut length and body weight were genetically highly correlated. These results demonstrate the feasibility of selection for growth of Nile tilapia in low-input environments. The effect of selection environment on the performance of selected strains over a range of potential production environments is a fundamental breeding question. In the presence of genotype by environment interaction, genetic improvement obtained by selection in one environment may not be realized in other environments. In Chapter VII, the results of a study designed to determine the extent of genotype by environment interaction in Nile tilapia are presented. We compared the performance of the low line with a high line selected for growth in ponds where fish received 25% protein. The 2 lines were tested in five test environments: 40% protein pellets feed (P200), 25% protein pellets (P100), 16% protein pellets (P50), 50 kg/ha chicken manure (M100) and 25 kg/ha chicken manure (M50). Nitrogen input was similar in P50 and M50, and in P100 and M100 treatments respectively. Survival from stocking to harvest ranged from 70-75% in the high line and from 62 to 76% in the low line. Analyses revealed significant differences in growth performances of the two lines across test environments. The phenotypic mean body weight at harvest was highest for test environment P200 (123.4 g in the low line, and 131.7 g in the high line). Lowest phenotypic body weight means were 92.1g (test environment M50) in the low line and 82.4g (test environment P100) in the high line. Although the high line performed better in more test environments, there was a significant line by test environment interaction indicating that both lines were sensitive to the environment. Family by test environment interaction was significant only in the low line. This genotype by environment interaction was related to the interaction of the families to nitrogen and dissolved oxygen in the ponds. These results described in this thesis show that there are good prospects for setting up sustainable breeding programs for resource poor tilapia farming conditions without requiring expensive supplementary protein pellets. The prospects and requirements for establishing such a practical breeding scheme for the small-scale farmer in resource poor regions of the world are discussed in Chapter VIII. Because poverty alleviation and food security are primary goals in the developing world, it is important that cheaper breeding programs are initiated and implemented and the genetically improved material made accessible to the rural fish farmer. High genetic variability is crucial in the breeding program because it is critical for both the short-term and long term limits of response. Therefore, strategies to maintain the variability through generations of selection should be part of the breeding program. Appropriate breeding goals for these breeding schemes need to be set-up with the involvement of the local farmers, preferably for each agro-ecological zone to ensure that the breed meets the requirement of local farmers. The formation of multiplication centres where farmers can collect fry or a decentralised breeding program should ensure that the improved fish seed gets to the farm gates when required. Full benefit of the breeding program will be realised by promoting husbandry practices that use locally available crop residues and manure to boost the productivity of the pond and provide ample nutrients for the genetically improved breeds of Nile tilapia.
_________________________________________________