[Illuminate]

If i took two hybrids of different P1 landrace(cough), crossed them together using regular m/f pistillate/staminate donors. Then from this 4 way picked two very different but similar expressions from the girls, and reversed them each to each other and each to itself to make 4 lines. aa ab ba bb

Then selfed the ones from each line that fit different criteria, until all the four lines true breed for a trait then re combine to add vigour and present the traits picked in higher %.

is this sound logic?

[Only Ornamental]

Could you draw a picture...
Seriously, I don't really get what you're trying to do; first you start with 4 different lines (the parentage is of minor concern as long as it's healthy and not too closely related), then you mix them mumbo-jumbo, you pick some plants and go through the trouble of isolating 4 new lines each stably expressing a specific trait, and then you cross the 4 lines with each other to reintroduce the vigour you messed up during the former selection only to lose the 4 traits you've been stabilising? WTF... or did I get something wrong?
You start with landraces for a reason . Do open pollination and spend some time in stabilising the wanted traits (given they are found in them) or start with no matter which plant (inbred or not) and do a forced inbreeding (selfing to fast success), lose any vigour and obviously have to do an outcrossing preferably to a line also stably expressing what you seek.

[Illuminate]

I'm just trying to get my head round what tom etc are all pipes about. If u self this plant c x a x p x c, you get s1 that segregation affects, from those s1 some will hopefully be true breeding for (a) trait no?

[Only Ornamental]

Right, this is correct. And yes, Tom et al. propose using parallel lines so you A) have a backup plan if one line goes down the drain for whichever reason and B) can cross these if they get too inbred.

[sourpuss]

I think u need to look deeping into traits. U cant say a whole plant is aa recessive dom or Aa.

[Illuminate]

the Aa bit was just labels not mendelian squares. im thinking that if i x the lanky one and the more dense flower one, its gonna give out a hella lotta variation like throw back to the grand parents and the inbetweeners bieng where its at.

[KiefSweat]

the last plant i selfed i ended up with about a 9:3:3:1 distribution of phenos.

i would say its more segregated into different families with them sharing about 80% of the overall genes for what we where looking for. and where stable for overall flavor and chemical profile.

[Illuminate]

awesome news.. so the segregation is obviously still limited to the genes to re shuffle.