Segregation analyses were performed using a Monte Carlo Markov Chain (MCMC) method implemented using Gibbs Sampling. A total of 8387 test-day milk records obtained from the Sicilo-Sarde dairy flock of the Tunisian National Institute of Agricultural Research (INRAT) were analysed. The major gene was assumed to be purely additive autosomal biallelic (A and B) locus with Mendelian transmission probabilities and priors used for the variance components were uniform. Based on 50 000 Gibbs samples from a chain of 520 000 cycles, the estimated marginal posterior means ± posterior standard deviations of variance components of test-day milk yield were 94.82 ± 136.60, 1892.65 ± 1275.21, 5965.72 ± 1166.13 and 15979.40 ± 251.86 for major gene variance (σ²_G), polygenic variance (σ²_u), permanent environmental variance (σ²_pe) and error variance (σ²_e), respectively. Results showed that the postulated major gene was not significant, the 95% highest posterior density regions (HPDRs) of most major gene parameters included 0, and particularly for the major gene variance. The 95% HPDR for Mendelian transmission probabilities showed that probabilities for 3 genotypes were overlapping. Results indicated that segregation of a major gene was unlikely and that the genetic determinism of test-day milk in Tunisian Sicilo-Sarde dairy sheep is purely polygenic. The estimated polygenic heritability and repeatability were h² = 0.07 and r = 0.32.