However, in Bangladesh, hatchery-produced stock (mainly carps) have shown adverse effects such as reduced growth and reproductive performance, increased morphological deformities, and disease and mortalities. These effects are probably due to genetic deterioration in the hatchery stocks resulting from poor fish brood stock management, inbreeding depression, and poor hatchery operation (Hussain and Mazid, 2004).

Aquaculture has had positive and negative effects on the environment.

There have been negative and positive impacts of aquaculture on the environment, depending on the intensification of the production systems. An incremental farmer participatory approach to the development of sustainable aquaculture in integrated farming systems in Malawi (Brummett, 1999) found that integrated farming systems are more efficient at converting feed into fish and produce fewer negative environmental impacts. The widespread adoption of integrated aquaculture could potentially improve local environments by reducing soil erosion and increasing tree cover (Lightfoot and Noble, 1993; Lightfoot and Pullin, 1995; Brummett, 1999). Negative environmental effects resulting from the aquaculture industry include threats to wild fish stocks (Naylor et al., 2000); destruction of mangrove forests and coastal wetlands for construction of aquaculture facilities; use of wild-caught rather than hatchery-reared finfish or shellfish fry to stock captive operations (often leading to high numbers of discarded by-catch of other species); heavy fishing pressure on small ocean fish for use as fish meal (depleting food for wild fish); transport of fish diseases into new waters; and non-native fish that may hybridize or compete with native wild fish. Improvements in management can help to reduce the environmental damage (Lebel et al., 2002), but only to a minor extent. However, economic impacts are site-specific. Intensive aquaculture has also had important effects on the landscape, e.g., in Thailand 50- 65% of the mangroves have been replaced by shrimp ponds (Barbier and Cox, 2002).

3.2.1.2.2 Breeding for abiotic and biotic stress tolerance Crops and plants, especially in marginal environments, are subjected to a wide and complex range of biotic (pests, weeds) and abiotic (extremes of both soil moisture and air/ soil temperature, poor soils) stresses. Abiotic stresses, especially drought stress (water and heat) have proved more intractable.

Progress in breeding for marginal environments has been slow.

Progress in breeding for environments prone to abiotic stresses has been slow, often because the growing environment was not characterized or understood (Reynolds and Borlaug, 2006), too many putative stress tolerant traits proved worthless (Richards, 2006), and because the complex

nature of environment-by-gene interactions was not recognized and yield under stress has a low heritability (Baenziger et al., 2006). Drought, for example, is not easily quantifiable (or repeatable) in physical terms and is the result of a complex interaction between plant roots and shoots, and soil and aerial environments (Passioura, 1986). Furthermore, much effort was expended on traits that contributed to survival rather than productivity.

Although yield and drought tolerance are complex traits with low heritability, it has been possible to make progress through conventional breeding and testing methods.

Breeding for marginal and stressed environments has not been easy, especially where wide-adaptation was also important. However, breeding programs that make full use of locally-adapted germplasm and TVs (Ceccarelli et al., 1987), and select in the target environments (Ceccarelli and Grando, 1991; Banziger et al., 2006) have been successful. For example, in Zimbabwe, where soil fertility is low and drought stress common, the careful selection of test environments (phenotyping) and selection indices can increase maize yields across the country and regionally (Banziger et al., 2006). Equal weight to three selection environments (irrigated, drought stress, N-stress), the use of moderately severe stress environments, and the use of secondary traits with higher heritabilities improved selection under stress. In multilocation trials, lines selected using this method outyielded other varieties at all yield levels, but more so in more marginal environments. This would seem to be a successful blue print for conventional breeding for stress environments.

Although drought tolerance is a complex trait, progress has been made with other aspects of abiotic stress tolerance..

Yield is the integration of many processes over the life of a crop, and as such it is unsurprising that heritabilities are low and progress slow. In contrast, the effects of some abiotic stresses are associated with very specific stages of the life cycle (particularly flowering and seed-set) or are associated with very specific mechanisms, and these appear to be more amenable to selection. Progress has been made in breeding for tolerance to a number of stresses, including extremes of temperature (hot and cold), salt and flooding/submergence, and nutrient deficiency. For example, tolerance to extremes of temperature, which are important constraints in many crop species at and during reproductive development (i.e., in the flowering period), have been identified (Hall, 1992; Craufurd et al., 2003; Prasad et al., 2006) and in some cases genes identified and heat tolerant varieties bred (Hall, 1992). These particular responses will be increasingly valuable as climate changes.