Statistical modelling/Sample size estimation

Note that the residual mean square for slides/buds(bud.slide.stratum Residual) is much greater than the mean square for fields/slides (bud.slide.*Units* stratum). We can equate the mean squares in the analysis of variance to expected values containing variance components s₁², s₂², s₃² at each level of variation. The variation among the 6 sets of 120 observations will not only be associated with the buds themselves but also to variation in mean slide results and to variation among fields of view. In contrast, there is only one source of variation for fields within slides. The coefficients for the expected mean square (E.M.S.) values reflect the number of times each item is replicated. Thus each bud has 120 (8x15) observations and each slide 15 observations. The s.e. between means has remained at 5.653. Again why?