The aim of this paper is to combine different but complementary approaches to check the neutrality of a given locus in a selected population. Analysis was undertaken through the polymorphism's evolution compared with that predicted under the effect of drift and through the analysis of the variance components of the measured traits, considering the effect of the locus as either a fixed or a random effect. This study deals with the case of the MHC locus, using both data from experimental lines of chicken selected for three different criteria of immune response, and frequencies of the genotyped haplotypes over time. Both the evolution of the polymorphism and the variance components approach have led to the conclusion that the MHC locus has an effect on the trait affecting antibody production against the Newcastle disease virus. Results have also highlighted the interest in using various methods in the case of low allelic frequencies. However, none of the common hypotheses, overdominance or frequency-dependent selection, was sufficient to explain the observed variation of the MHC polymorphism, which was displayed by the temporal variation of the allelic frequencies.