How many seeds should be grown to get a good representative sample of a cross?

[de145]

If one were to cross an unrelated male and female is there a good rule of thumb for how many seeds you really need to grow to get a good representation of all that cross will result in?

I.E. if I cross two unrelated parents and get 100 seeds how many of those must I grow to get a representative of all the common phenotypes?

Or is this one of those "it depends" questions?

 

[HidingInTheHaze]

I'd say it depends on the parentals used. Like using polyhybrids vs true inbred lines. The crosses made from 2 polyhybrids might produce a wider range of diversity then using 2 true inbred lines.

I don't think anyone could tell you exactly how many seeds you would need to see all of the pheno's that crossing could offer.

 

[Dropped Cat]

Results.

I have grown less than twenty from a cross of two Afghans.
I liked the results and made back crosses and F2's. So far
I have enough of the traits to attempt stability.

I'm micro, non commercial.

What are your initial objectives?

 

[Snype]

If one were to cross an unrelated male and female is there a good rule of thumb for how many seeds you really need to grow to get a good representation of all that cross will result in?

I.E. if I cross two unrelated parents and get 100 seeds how many of those must I grow to get a representative of all the common phenotypes?

Or is this one of those "it depends" questions?

It depends on what the parents are. If 1 of the parents is a true IBL or landrace then you would need less seed to look through but as HITH said about polyhybrids, you would probably have to look through a lot of seed. In one of the pheno's that I'm looking for in one of my seed crosses, I'll be lucky to find the pheno that I want in every 20 females. That's if I'm lucky.

 

[de145]

Interesting; in this case the male parent is a landrace Kerala and the other is a hybrid Northern Skunk x Northern lights. So in theory less variability then?
I'm not doing this with any particular goal other than finding out what happens at this point, I just want to be sure I've got an idea of what happened so I was wondering how many to grow out. So far I have three seedlings and the only variation I can tell at this point is the degree of smell when you rub the stalk.

 

[de145]

According to DJ Short:
http://www.cannabisculture.com/articles/1511.html
"I cannot overstress the importance of the two P1 parents being as genetically different as is possible. It is this initial genetic diversity that leads to the most possibilities in succeeding lines.
If the P1's are sufficiently diverse, then the f1 will be a true hybrid, expressing a near total uniformity and great vigor. It is in the crosses beyond the initial f1 (especially the f1xf1=f2 cross) that specific traits are sought. There will be a tremendous amount of variance in the f2 crosses of f1's obtained from a female pure sativa and a male pure indica."

Just like tomatoes or flowers I guess, so if I have two very diverse close to landrace strains then I would expect nearly zero variability and could grow out very few to get a feel for them. And more accordingly if they are hybrids.

On the other hand the f2 generation would have wild diversity and I would likely need to grow a great many but that's apparently where the most interesting things could be found.

 

[BredForMeds]

we try to not make less then a few hundred seeds per cross. and pop no less then 50 . and then breed best male with best female.. before I even attempt to give any seeds out..

 

[Pantagruelion]

The more the merrier. At least ten. The more you do,the more likely that you will find what you are after.

 

[Kalbhairav]

According to DJ Short:
http://www.cannabisculture.com/articles/1511.html
"It is in the crosses beyond the initial f1 (especially the f1xf1=f2 cross) that specific traits are sought. There will be a tremendous amount of variance in the f2 crosses of f1's obtained from a female pure sativa and a male pure indica."

The above is good advice.

I once asked Chimera a very similar question and he said that 200 of the F1/F2 offspring was sufficient in exploring all genetic possibilities within a line. I do think he was referring to a real F1 cross where the P1 parents were stable and line bred though.

Another element to keep in mind is how many males your using in the initial P1 to P1 cross. If the ratio is 1:1 then the diversity of the offspring will be limited to whatever that male has to offer. If you intend to line breed then using more than one male may be better to stop bottle necking further down the line. The problem with multiple males in the initial cross is the diversity of genetic material. You will have to venture further to find F1 females that show promise. The clinch is that you may find more than one very beautiful female with different traits to work with

This topic has been hacked out many times and many here have varying opinions to which is the best method for each individuals space/time etc. Which ever way you choose have fun selecting and exploring..

 

[Only Ornamental]

My
An inbred line is nothing like a landrace. Inbreeding results in less overall variability whereas a landrace shares certain (race determining) traits and expresses a huge diversity in others.

If you'd take two 'true inbred lines' (this does likely not exist in cannabis) and crossed them, the F1 would be VERY similar (with perfect inbred parents all F1 siblings would theoretically be identical). That means, in the F1 you select for trait mixes resulting from heterogeneous parental alleles. Say, the male is a cross of purple x green, its F1 offspring may be green or purple depending on the mother plant. You should search for that trait in the F1 if it follows dominant/intermediate inheritance. All traits stable in both parent lines will be rather equal in the F1 and you risk less not to select those right away.
Starting with two inbred parental lines which are completely different one from another gives you, as said, identical F1s but in the F2 this results in mathematically (with 20 chromosomes and neglecting cross-over effects and alike) 3^20 possibilities (homozygous for one or the other and heterozygous when neglecting that the father or mother chromosome could dominate over the other). This is ~3.5 billion possibilities only based on chromosomes! Fortunately, not all chromosomes carry traits which are important to you.
That's why you need a lot; if you can't, search for a few traits eventually even using also non-perfect plants for the cross if the trait doesn't show in the keeper siblings. Cross those and sift through the F3 and even the F4. Using several generations for the selection process reduces the number of needed plants considerably (but you'd also lose in hybrid vigour; it's a double-edged sword). Finally, looking at the F1 to learn the heterozygous traits and how to distinguish them from homozygous (parental) ones helps in selecting F2s (it's up to you if you want likely more vigorous heterozygous plants of stable and predictable homozygous ones).


I have a little game for you to illustrate that:
Imaging having 12 identical balls, one of which is a bit lighter than the others. You only have a pan balance and should use as little weighing steps as possible.
If you weigh every single ball, you need up to 11 steps (cause 1 ball remains on the second pan) or a lot of luck, if you use the 'several generations' approach, you're done in 3 steps (but luck will not help this time) .
Let me know if you need a hint LoL.

 

[mofeta]

I have a little game for you to illustrate that:
Imaging having 12 identical balls, one of which is a bit lighter than the others. You only have a pan balance and should use as little weighing steps as possible.
If you weigh every single ball, you need up to 11 steps (cause 1 ball remains on the second pan) or a lot of luck, if you use the 'several generations' approach, you're done in 3 steps (but luck will not help this time) .
Let me know if you need a hint LoL.

1. Select the lighter of two groups of six
2. Select the lighter of two groups of three made out of the first six
3. Weigh two of the last three

Fun problem, but it only works if you know beforehand that there is for sure one, and only one light ball.

Would you elaborate on what plant breeding principle this illustrates? I can only think of a few limited applications in weed breeding for this.

Thanks
mofeta

 

[Only Ornamental]

Congrats, you just revealed the solution ;( .

It's only an example: It might be better to propose that you may buy additional balances. So either you use six at a time or just one for three times (or a mix thereof).
It was only meant as a parable to illustrate this thing with quantity v.s. intelligence .

 

[Hmong]

Interesting; in this case the male parent is a landrace Kerala and the other is a hybrid Northern Skunk x Northern lights. So in theory less variability then?

I am growing a hybrid thats from two IBLs
As people said before, when all traits in both parents are stable the F1 will be a total new plant. a true mutant combining all dominatnt traits.

Rosa: (Cheese (Exodus Cheese Cut/Skunk#1 IBL x Cranberry Haze IBL)

the F1 i purchased was a perfect example for that. bud and trichome building were typical for a skunk, so very dense and resinous. But the high, smell and colour were from the Cranberry Haze male. I only bought 20 seeds but more werent needed to find what I was looking for. I got 2 good males and one real F1 girl. you know those who are at least 20% better in every aspect than the other individuals. Also I had some difference in colouration and flowering time, more purple seems to be more Chitral male (Indice) influence and therefore a shorter flowering time than the slowest one, which had a lime green coloration of budleaves and calyxes. (Colombian like trait)
Of course her high was a littler more clear too.
Smell/Taste were like predicted. strong odor over all very pungent smell which stayed long on everything (skunk) but with the cranberry haze flavour (red cranberry/very sour).

So in conclusion, you can easily start with a pack of seeds for the F1 and work along with it. In your case i would say, i fyou have 100 Seeds, start with 50 which gives you even a good test for germ capacity and then after 3 weeks keep on with the best 20.
thats my rule of thumb, i would always popp 50 if i use my own ones or lets say have beans of such quantity.

Now in the F2 it is like as iad here before, the first real diversity appears. I have some mostly sativa phenos, about 50% F1 like phenos, and so far one I would say 95% indica pheno. So to go on further with this strain you would now have to choose a "lineage" in the F2. there are now more different traits you have to select on. Lets say in my case i want to keep on with the "Rosa type" so I would also again need 20 to do a propper F1 like selection for my F3. Since its the rezessive genes which appear in the F2 there might also be more difference especially in taste/smell between each individual of a phenotype.
So I would prefer to even double that 20 plants rule in the F2. So to archieve that you would need to popp 150 seeds then. I dont have this kind of capacity so I allways keep on with 50 @ once and make perpetuating runs. depending on my motivation of course.

So for me as a breeding amateur this kind of strain is quite ideal, since it is very predictable.
to do real polyhybrid crosses which of course arent that far predictable, a huge number of plants will be needed instead. Most (dutch) seedbanks are not willing to put that effort into a new line and just throw one strain after another out, since no selection of all is needed to create a new Fem Strain. Always keep that in mind when buying seeds. I would say at least 50% of all dutch "strains" are a fraud, thats why I only purchase landraces or their hybrids any more.

I hope this info helps you with your project. good luck, let it grow.

 

[Hmong]

I also noticed that there are many "So's" in my posts. So I need to work on that I guess hahaha

 

[Homebrewer]

If I can't find something nice in 12-15 seeds then it's not worth going through any more of them, IMO.