Heads up, this is going to be a little long....
I'm always trying to expand my knowledge on breeding and in particular the basic/foundational stuff (easy to get lost in all the new high tech stuff) and something I've been thinking about several times is variety vs stability and open pollination (OP) vs a more intense selection. After a long discussion with another small breeder and some more reading I ran into an article that sums it up quite nicely. Figured I share them here:
http://howtosaveseeds.com/breeding.php - About selecting (individuals) and the effect on the genetic diversity.
http://www.garden.org/subchannels/care/seeds?q=show&id=293 - Hybrid or Open Pollinated
An OP strain is more like a population than a variety (what in cannabis world is a strain), the latter requires more intense selection and is more uniform. The OP is however more genetically diverse which apart from possibly being less uniform also has its advantages, like more possible new combinations of genes.
The reason for the interest is to deal with the risk of inbreeding depression. Assuming I find some great plants amongst the F3, I want to continue stabilizing it (basically the non-whorling line) till it's suitable for creating an F1 hybrid and function as the recurrent parent in backcrossing, ie. till it's homozygous for its key traits. So I want it to be stable for growing
and breeding. Given limited space and plant count the risk of selecting out good traits increases with every generation. That can be somewhat prevented by doing an extra round of the same generation if the current batch doesn't produce enough plants that have the desired traits. And also be using multiple females/males. Creating two or more lines and crossing those together later in the line would be another option. What inherently cannot be avoided when locking in genes is locking out some others by locking in inferior genes. For example for something less visible like pest or disease resistance.
It always starts with an IBL (ideally, which I'm aiming for). In case of backcrossing one can, and in many cases should, cross multiple plants back to the recurrent parent. That way you can select intensely yet retain genetic diversity, i.e. without narrowing the gene pool. I can see why backcrossing is so attractive :) In case of creating the IBL itself, the pool becomes more narrow by default... hence, hybrid vigor once you introduce another.
Obviously it's already limited to the genetic pool based on the original parents (CH and chunk). I want to narrow it as little as possible, while locking in the important traits (like same structure, flowering time, frostiness, bud structure, yield etc.). When it comes to inbreed generations, sometimes less is more. Ideally one picks two homozygous (for the relevant traits) plants in F2 already (since the traits are segregated they are rare but do exist), and then in case of a small plant count going from F4 isn't necessarily an improvement unless the goal is to lock in more traits.
In practice the amount it matters depends obviously on what I cross the P with once homozygous. If that one has a narrow genep ool (like a cross from a cross from a cross)... So if I get to the point of creating a F1 hybrid, I need to get and stabilize something very old and or diverse (like a double cross, crossing two different F1s, or land race even).
However, from
http://en.wikipedia.org/wiki/Cannabis : "Dutch
Cannabis researcher E.P.M. de Meijer and coworkers described some of their RAPD studies as showing an "extremely high" degree of genetic polymorphism between and within populations,
suggesting a high degree of potential variation for selection, even in heavily selected hemp cultivars."