[Question] adonis permanova with multiple factors and semi-nested design

Hi everyone!! I'm trying to come up with the proper model for my adonis permanova but getting confused on whether the strata function is appropriate or not. I have 14 sites representing 4 different forest types (3 or 4 sites per forest type). In each site, there are 4 plots representing different densities of the same soil inoculation type. I am interested in whether the inoculation densities have an effect on the soil community composition and whether this is dependent on forest type (interactions). In this case, each inoculation density is represented in each site thus using site as a blocking (strata) category might not be appropriate as it would constrain the permutation to be within each site and not between (is this right?). However, I am sure that the site will explain part of the variation and should be included in the model. I am running this code but not sure if it is the correct one. adonis(spp.matrix ~ inoc.density*forest.type + site, distance="bray") The other option would be to use strata: adonis(spp.matrix ~ inoc.density*forest.type, strata = site, distance="bray") I've also seen some codes where they use something like inoc.density/site but not sure what is that accounting for. Thaaanks in advance :)