Short answer. You reverse one branch (or a clone) of a female so it makes pollen, then use that pollen on the same genotype. The seed is an S1. S1s are female (both gametes carry X) [1][2] and they are related to the parent like F2s are to an F1: they reshuffle the parent's own alleles. Expect variation, some plants close to the parent, and some recessive surprises. Used carelessly, selfing is also one of the main ways herm-prone lines get made.
Breeder's standard. Only self a female that has stayed fully female for at least two runs, reversed under control (never pollen from a stress herm). Grow out the S1s, cull every plant that shows intersex flowers, and don't self again from a line that throws hermies. Never stack S1 on S1 without that culling, and disclose the herm rate with any S1 seed you release. Cannabis is a dioecious species; keeping it that way is the breeder's job.
House position: selfing is mostly frowned upon. Real breeding needs real males, selected as carefully as the females. Males are encouraged and rewarded on GenetixBay, and a well-documented male line is a prized contribution. Selfing is mostly frowned upon: it was historically an emergency measure to save a line, not a shortcut. GenetixBay holds the best breeding standards: snobby, and proud of it.
What's happening
At every gene where the parent is heterozygous (Aa), S1 offspring segregate 1 AA : 2 Aa : 1 aa. On average half the heterozygosity is lost per generation of selfing [4]. That does two things:
- Reveals recessives. Useful hidden traits (a colour, a terpene type, autoflowering) appear in ~25% of plants. So do harmful ones (weak growth, poor fertility, deformities): this is inbreeding depression [3].
- Fixes faster. A selfed line becomes uniform much faster than a sib-mated one, if you can live with the inbreeding.
Punja and Holmes (2020) found the genetic variation among progeny after one generation of selfing was similar to that among cross-fertilised progeny in their markers [1]. That is a reminder: S1 seed is not a clone in a seed.
Numbers that matter
| Generation | Expected heterozygosity left (from a fully heterozygous start) | Inbreeding coefficient F |
|---|---|---|
| S0 (the parent) | 100% | 0 |
| S1 | 50% | 0.5 |
| S2 | 25% | 0.75 |
| S3 | 12.5% | 0.875 |
| S4 | 6.25% | 0.94 |
Single-locus expectations [4]. Real clones are not 100% heterozygous, so start values are lower.
When S1s make sense
| Goal | S1 useful? | Why |
|---|---|---|
| Back up a clone in seed form | Partly | You preserve its alleles, not its exact combination |
| Find what recessives a clone carries | Yes | ~25% of S1s show each recessive |
| Make a feminised production line fast | Sometimes | Uniformity rises quickly; vigour may fall |
| Create a breeding parent with fewer hidden recessives | Yes, with selection over S1-S3 | Purging by selection |
| Sell "a seed version of clone X" | Only with honest labelling | Must say S1 and show grow-out variation |
Why careless selfing makes hermies
Historically, selfing was an emergency measure: a way to save a line when no male was left, not a trait to want or a shortcut to sell. Treat it the same way today.
Cannabis has been dioecious and stable for millennia. The GenetixBay breeder position is that herm-prone lines are mainly made, not born, and repeated selfing is a prime route:
- Selfing a plant that has already hermed puts its readiness to make male flowers into every seed, and with no male parent there is nothing to dilute it.
- S1 on S1 (S2, S3...) without strict culling fixes whatever the line carries, the tendency included, at the same speed it fixes the traits you wanted.
- Hermaphrodite pollen gives female seed [1], so the defect hides inside an "all-female" pack until the buyer's light leak exposes it.
Punja and Holmes found one generation of selfing had no measurable effect on genetic diversity in their markers and said more sequential selfing would need study [1]. That supports the point that one S1 from a proven-stable mother is a tool; it is the unselected stacking and selfing of unstable plants that does the damage.
Step by step
- Choose a female with no intersex tendency across at least two runs. Selfing a plant that hermaphrodites readily concentrates the trait. Don't self it; cull it.
- Reverse one clone or branch with silver thiosulfate (see feminised seed and sex reversal for method and safety).
- Collect the pollen in a separate space.
- Pollinate tagged branches on an untreated clone of the same plant.
- Harvest seed 5-7 weeks later; record as an S1 batch with the parent's GB-IND/GB-CUT ID.
- Grow out at least 20-30 S1s and record vigour, uniformity and any off-types or intersex plants. Cull intersex plants and never use them as parents. Publish those numbers, including the herm rate, with the seed.
Common mistakes
- Selling S1s as "the clone in seed form".
- Selfing for several generations without selecting for vigour and fertility.
- Using pollen from a stress-induced hermaphrodite rather than controlled reversal. You can't separate "the trait" from "the tendency".
- Stacking S1 on S1 and releasing the result untested. That is how pollen chuckers spread herm-prone genetics.
Evidence
- S1 seed is female and genetically variable: Established [1][2].
- Loss of heterozygosity and inbreeding depression under selfing: Established [3][4].
- Careless and repeated selfing as a main source of herm-prone lines: GenetixBay house position (breeder experience). Risk depends mainly on the parent's own stability and on culling; one generation of selfing did not reduce marker diversity [1]. Controlled cannabis comparisons of intersex rates in S1 vs outcrossed lines are lackingbeing reviewed.
