This paper has is view the joint analysis of augmented trials in randomized blocks. Each experiment had t = c + z, in r blocks where we have c commom treatments, that is, treatments present in each blocks, and z regular treatments, which appear in only one of the r blocks. Each block has k = c + pj plots, where pj (j = l, ..., r) is the number of regular treatments in it. The analysis was carried out, taking the whole set of trials as one experiment with incomplete blocks, assuming that the trials had similar variances. In the example presented, treatments belonged to six classes of association, with least significant difference (by Tukey's method) between two of them (Δ) given below: 1. Two common treatments: Δ 1% = 15.955 t/ha; Δ 5% = 13.755 t/ha. 2. A common treatment and a regular one: Δ 5% from 23.047 to 23.216 t/ha; Δ 1% from 26.733 to 26.930 t/ha. 3. Two regular treatments, with λ = 3: Δ 5% from 28.515 to 28.659 t/ha; Δ 1% from 33.075 to 33.243 t/ha. 4. Two regular treatments, with λ = 2: Δ 5% from 29.064 to 29.950 t/ha; Δ 1% from 33.711 to 34.739 t/ha. 5. Two regular treatments, with λ = 1: Δ 5% from 29.160 to 29.994 t/ha; Δ 1% from 33.823 to 34.791 t/ha. 6. Two regular treatments, with λ = 0: Δ 5% from 30.281 to 30.628 t/ha; Δ 1% from 35.124 to 35.526 t/ha. We realize, therefore, that these leat significant differences are rather smaller in the case of two common treatments, but are not too different in the other cases.