marco, the answers to your questions are no, yes, no.
but to expand, flower color, like most other aspects of phenotype, is not monogenic, nor is it inherited in a purely Mendelian fashion. this leads to occasional frustrations for breeders, as efforts to make an aesthetic, desirable cross from 2 strong parents will sometimes meet with failure (depending on who you ask, the likelihood of this occuring varies
). to use an example along my previous lines, this would be paph roth 'awesome-o' x paph roth 'even better' = paph roth 'uggo.'
as for the reasons why this occurs, part of it is not fully understood, and the rest has to do with the fact that genetics is far, far more complicated than Mendel realized. for an example of the above, watch that show on the Learning channel about the family with dwarfism--while both parents are achondroplasts, and some of the children are, not all are, despite the fact that this is a supposedly recessive trait.
as for the effect of line breeding in inducing heterogeneity, this is definitely true, for a simple reason. the first time you make a cross, they have distinct genetic contributions from the father and mother. for instance, paph roth 'awesome-o' x paph sanderianum 'rockin' = paph prince edward of york 'awesome rocker.' clearly evident will be a certain percentage of roth and a certain percentage of sanderianum. now if you self or sib 'awesome rocker,' you're going to get highly variable results, because each parent plant will contribute a varying amount of its respective parents' genes to the offspring. so one plant may be vary roth dominated, one may be very sanderianum dominated, and the majority will be somewhere in the middle, but won't necessarily look like prince edward of york. in the next generation, there will be less variability. over time, if you continue to self the offspring (especially if you selectively self those that fit the phenotype you desire), you will have completely homogeneous offspring. the analogy that I find most helpful for understanding this process is that of mixing a chocolate cake. the batter is first white, and then chocolate is poured in. initially, the contributions of the batter and the chocolate are readily apparent, and if you were to bake it without stirring, you'd get a half white, half chocolate cake. but if you were to gently stir that, you'd get some parts of the cake white, others chocolate, and the majority kinda muddy. over time, and with more stirring, the cake becomes smooth and even.
anyway, hopefully this is clear. If not, let me know and I'll try to clarify.
paphiness, I personally think that if a vendor can guarantee a flask is all alba, then they are at least one of the following: very confident, very deceptive, very ignorant, or a masterful scientist who has managed to mericlone a paph.
