It’s exactly the same problematic that with a normal backcross (without selfed intermediary), but in far more radical context (because the selfed intermediary). Because with the S1 you have forced the phenotype to intensify its expression “totally” and in one shot, at this point you’re knowing the inherent equilibrium of the phenotype in term of traits prevalence. But not its potential in term of segregative dynamic, which is locked and catalyzed by the process itself.
I’m supposing in the example that your reference is a given elite clone that you want to produce in a stable form in seeds (just because it’s the most tryed option since ages, it add indirectly an extra of rationality in the discussion). So, the genotype carryed by the reference is wider initially, not allready narrowed by the potential of an unique phenotype.
Always the same mantra. You don’t have to justify your methods to anyone if you’re obtaining decent results; seriously if the growers/friends/yourself testing it are happy and that the weed is good : no one give a fck how it’s made ^^
I think that i’m seeing the pattern of the bias now. The choice to use selfed specimens is specifically to reduce drastically the need to deal with variations. Even with a simple BX process the difference is enormous.
You’re already stabilizing what you’re doing, just because the process itself. The “variations” are the allelic dynamic you have locked in one shot. They are not really variations, but the direct output of your selection in its purest form. It’s why to cumulate it with a backcross is counter-productive if you think that the variation is carried in this case by the S1, and not the P1. How to read the output rightly in this case and to improve your selection of S1 ?
As I’ve said, selfing is an effective tool to study the homozygosity dynamics in optimal conditions. But you have to decypher the trait and their leverage for that.
If every single plant of your S1 have citrus scents, it’s not the garantee that the P1 will pass directly to your S1 the intensive form of expression of this scent.
Not easier. The easier way is to stabilize the traits before locking it, with genotypes already streamlined in term of stability. And it become barely industrial, but it’s another subject.