From the history of genetic improvement in aquaculture, particularly that for Atlantic salmon in Norway (footnote 14), ICLARM and its partners had anticipated that selective breeding would be the more appropriate strategy for the development of GIFT. However, it was also necessary to explore the possible advantages of crossbreeding among the eight assembled Nile tilapia strains. This required another experiment on a scale never before undertaken in Asian aquaculture. The weights at harvest after 90 days of all 64 possible pure- and crossbreeds among and within the four African and four Asian strains were measured in different test environments, again including ponds and cages as well as lowland and upland locations. This involved the tagging of 23,000 individual fish. Hybrid vigor (harvest weight advantage of crossbreeds over parents) was low (average 4.3%) and significant in only 22 crosses, of which only 7