The stocking practices used for brown trout (Salmo trutta) in the South of France result in secondary contact and introgression between populations from two genetically differentiated forms of the same species (KRIEG and GUYOMARD, 1985; GUYOMARD, 1989) : domesticated stocks that originate from the Atlantic form and the wild Mediterranean populations. This paper reviews the protein data for 44 Mediterranean samples examined in the laboratory, with an appraisal of introgression by the domesticated form and a description of the genetic consequences of stocking on the existing populations. The samples were collected from several French departments : Pyrenees Orientates (12 stations), Herault (5 stations), Vaucluse (2 stations), and Corsica (25 stations) (see Table I). Electrophoresis on horizontal starch gel revealed 28 loci. Introgression was quantified using three diagnostic protein loci : lactate dehydrogenase of the eye (LDH-5*), serum transferrin (TF*) and fructose biphosphatase of liver (FBP-1*). The individuals identified as domesticated fish on the basis of their genotype were discarded. Frequencies at diagnostic loci, expected heterozygosity, the deviation from panmixia in each sample, and domestic introgression are given in Table II. Heterozygote deficit was estimated using parameter f of WEIR and COCKHERAM (1984). Likewise, parameter 8 was used to assess whether the structure of the different samples was statistically significant. The values of these two parameters were tested by permutation using the GENETIX 2.0 programme (BELKHIR et al., submitted). The populations analysed can be classified according to their degree of introgression, which ranged from 0 to 77 %. Most introgressed populations displayed substantial and significant heterozygote deficit. Stocking is the most likely cause of these imbalances. However, this interpretation cannot be generalised to all the samples because the deficit did not increase with the severity of introgression. The results as a whole tend to show that stocking has a genetic effect on natural populations (with changes in their genotype, and imbalances in panmixia and linkages). The data also illustrate the weak effect of stocking on population densities. This has already been reported by several authors. The various causes for this lack of success in stocking (stocking techniques, learning problems in the wild environment, domestication) are discussed. Introgression-free Mediterranean populations generally occur at high altitudes and they are practically monomorphic. The advantages of conserving genetically pure populations with no polymorphism are discussed. Two solutions that reconcile objectives as different as supporting trout population densities and protecting biodiversity are examined :the creation of local strains, which requires rigorous genetic management (CHEVASSUS, 1989) and the introduction of triploid individuals, which would result in solely ecological interactions. Finally, the question of the complete interruption of stocking in the whole area or in ''genetic sanctuaries'' that have been identified as genetically original is discussed. Other diagnostic markers distinguishing trout of the two origins have been described recently. They provide information on introgression at levels other than that of proteins. The study of the mitochondrial genome gives data on the maternal origin of each individual and hence on introgression through the mother. Finally, the study of highly polymorphic microsatellite markers will make it possible to distinguish between the natural and domesticated components of the populations of the Atlantic basin, a zone in which there are no diagnostic allozymic loci for these two forms.
