Seed packets carry codes such as F1, F2, S1, BX3 and “stabilised”. Each code names one breeding step. Each step tells you how similar your plants will be to each other and to the photo on the listing. Most sellers print these codes loosely or get them wrong, so you need to know what they should mean.
Part 1 covers the basics. It explains genotype and phenotype and how genes pass from parents to seed. It also covers cannabinoid type, plant sex, what F1 and F2 really mean, and hybrid vigour. Part 2 covers later generations (F3 to F6), inbred lines, selfed seed (S1), backcrosses (BX) and polyhybrids. It also shows how many plants a project needs and how to read a seed listing line by line.
1. Genotype, Phenotype and What a Seed Label Promises
A seed listing makes a promise about variation. It tells you how similar your ten plants will be to each other and to the photograph. Only the generation code carries real information about this. It says how many rounds of selection separate your seed from the breeder’s original cross.
Genotype and phenotype
Genotype is the set of gene versions a plant carries. Phenotype is what you can see and measure, such as height, aroma, resin and flowering time. Phenotype comes from the genotype plus the growing conditions. Two plants with the same genes will look different if one grows in a hotter corner or under weaker light.
So phenotype hunting only works when every plant gets the same treatment. Phenotype hunting means growing many seeds and keeping the best plant. If conditions differ, you may keep a plant that simply had the best spot in the room.
Alleles, loci and the two copies of each gene
Cannabis is diploid. Each plant carries two copies of every gene, one from each parent. The place on a chromosome where a gene sits is called a locus (plural: loci). The different versions of one gene are called alleles.
A plant with the same allele in both copies at a locus is homozygous there. A plant with two different alleles is heterozygous there.
Why seeds from one pack differ
Each pollen grain or egg cell gets only one of the plant’s two copies at each locus. Which copy goes in is random. A parent that is homozygous at a locus always passes on the same allele. A heterozygous parent passes on one allele or the other, each about half the time.
Seeds from heterozygous parents therefore receive different combinations. This splitting is called segregation. A line is stable when its parents are homozygous at most of the loci that matter. Then most seeds get the same combination and the plants come out alike.
Translator’s note: In Thai grow rooms, conditions are rarely the same everywhere. Plants by the air-conditioner outlet, under the centre of the lamp, or beside a sun-heated wall grow in different temperatures. In the rainy season, humidity also varies from corner to corner. These differences change the phenotype, so they can hide real genetic differences or make false ones appear.
What you can do: In your next seed run, number every plant and write down where it stands in the room. Give all plants the same pot size, growing medium and feed. Then, if one plant stands out, you can tell whether its genes or its position caused it.
2. Dominant, Recessive and Codominant Alleles
In a heterozygous plant, the two alleles can work together in three ways. The table shows how each one looks and what it means for you.
| How the two alleles work together | A heterozygous plant looks like | What it means in practice |
|---|---|---|
| Dominant / recessive: one allele hides the other | The parent with the dominant allele | The trait disappears for one generation, then comes back |
| Codominant: both alleles show | Both alleles at once | Nothing is hidden; what you see tells you the genotype |
| Additive or polygenic: many genes each add a little | Something in between, on a smooth range | Many loci, no clean ratios, slow to fix in a line |
Recessive traits surprise growers: the autoflower example
A recessive trait only shows when a plant carries two copies of the allele. A plant with one copy looks normal but can still pass the allele on. That is why recessive traits often reappear when no one expects them.
Autoflowering is a clear example. An autoflowering plant starts to flower because of its age, not because of day length. In 2022, Toth and colleagues published a study in the journal Frontiers in Plant Science. They traced autoflowering in Cannabis sativa to one locus, which they named Autoflower1, and found that a recessive allele at that locus causes it.
Their segregating populations, groups in which the trait was still splitting, had 88 plants. Under long days, 31.8% and 27.3% of the plants flowered anyway. Both figures are close to the 25% expected when one recessive locus controls a trait.
This has a practical result. Suppose you cross an autoflower with a photoperiod plant, meaning a plant that only flowers when nights get long. The first generation contains no autoflowers, because every seed has one autoflower allele and one dominant normal allele. Cross those offspring with each other, and about one plant in four in the second generation will autoflower again.
What you can do: If you cross an autoflower yourself, keep the first generation even though none of its plants autoflower. Grow out the second generation and count the plants that flower under long days. If the count is close to one in four, you know the trait is coming back as expected.
3. Cannabinoid Type: The Best-Documented Single-Gene System in Cannabis
Most cannabis traits depend on many genes and are hard to predict. One trait is simpler. It is the balance between the two main cannabinoids, tetrahydrocannabinol (THC) and cannabidiol (CBD). This balance is called the plant’s chemotype, meaning its chemical type.
In 2003, de Meijer and colleagues published a study in the journal Genetics. They proposed that a single locus, which they called B, sets this ratio. The B locus has two codominant alleles. Each allele codes for a synthase, an enzyme that turns one compound into another.
Both enzymes work on the same starting compound, cannabigerolic acid. One converts it toward cannabidiolic acid, the acid form of CBD. The other converts it toward tetrahydrocannabinolic acid, the acid form of THC. Breeders write the two alleles as BD and BT.
| Chemotype | Genotype | Chemical profile |
|---|---|---|
| Type I | BT / BT | THC-dominant |
| Type II | BT / BD | THC and CBD, roughly balanced |
| Type III | BD / BD | CBD-dominant |
Working out the offspring with a Punnett square
A Punnett square is a simple grid for predicting offspring. Write the alleles one parent can pass on across the top. Write the other parent’s alleles down the side. Each box then shows one possible seed.
The square below crosses two balanced Type II plants (BT/BD). This is the usual way a balanced line goes wrong.
| BT | BD | |
|---|---|---|
| BT | BT/BT | BT/BD |
| BD | BT/BD | BD/BD |
De Meijer’s group found exactly this 1:2:1 ratio in their own second-generation populations. A quarter of the plants became THC-dominant and a quarter became CBD-dominant. Only half stayed balanced.
So a balanced plant can never breed true. It is balanced because it carries two different alleles, and those alleles split again in its seed. Seed sold as “stabilised balanced” does one of two things. Either it splits like the square above, or it comes from a BT/BT parent crossed with a BD/BD parent. In the second case, only that first generation is balanced. Its own seeds will split again.
The exceptions are real
The single-gene model predicts the type, not the exact number. De Meijer’s own paper noted that the exact ratio within one chemotype was progeny specific. In other words, it varied from one seed family to another. The type is inherited cleanly, but the precise THC-to-CBD figure is not.
In 2019, Laverty and colleagues published a study in Genome Research. They mapped where the synthase genes physically sit on the chromosomes. The genes lie in regions full of repeated DNA that are over 250 kilobases long. A kilobase is 1,000 DNA letters.
These regions are highly non-homologous between drug-type and hemp-type chromosomes, which means their layout differs a lot. Large pieces of them have also been moved and rearranged over time. So chemotype is strongly inherited and roughly predictable, but not guaranteed.
Translator’s note: In Thailand, a plant’s THC level decides which legal rules apply to it. See Hemp vs High-THC Cannabis: Where Thai Law Draws the Line. Second-generation seed from balanced parents can give about a quarter THC-dominant plants. You cannot spot these plants by eye. Only a lab test shows them.
What you can do: Before you choose a mother from a CBD or balanced seed batch, have each candidate plant tested for cannabinoids. Read the result with the help of How to Read a Cannabis Certificate of Analysis. The test shows each plant’s real chemotype, so you do not keep a THC-dominant plant by mistake.
4. Plant Sex: The X and Y System
Cannabis is normally dioecious, which means each plant is either male or female. Chromosomes set the sex, as they do in humans. A 2014 study by Divashuk and colleagues in the journal PLOS One examined the sex chromosomes directly under the microscope, using molecular methods (molecular cytogenetics).
The study confirmed that the two cannabis sex chromosomes look different from each other. Females carry two X chromosomes (XX), so every egg cell carries an X. Males carry one X and a distinct, larger Y chromosome (XY).
| Cross | The pollen carries | Expected seeds |
|---|---|---|
| Female (XX) × male (XY) | X or Y | About half female, half male |
| Female (XX) × female made to produce pollen (XX) | X only | All female |
The second row is the whole basis of feminised seed. The breeder treats a female plant so that it makes pollen. That pollen can only carry an X, so every seed is female.
However, a plant’s chromosomes are one thing and the flowers it makes are another. Stress, hormone treatments and disrupted light can all push a genetically female plant to make pollen. For more on this, see Cannabis Sex Determination.
Translator’s note: Power cuts happen in many parts of Thailand and Southeast Asia, especially during storms. A cut during the light period breaks the plant’s light schedule. If the air-conditioner stops too, the room can climb to 30–35 °C. Both stresses can make female plants grow pollen sacs.
What you can do: Record every power cut and every heat spike in your grow diary. Over the next days, check your flowering plants for pollen sacs. This lets you connect any male flowers to a cause and remove them before they pollinate the room.
5. What a First Filial Generation (F1) Actually Is
F1 means “first filial generation”, the first set of offspring from a cross. Formal plant breeding uses the term more strictly. Both parents must be inbred lines, and the two lines must differ from each other.
An inbred line is a family bred within itself for many generations, until it is homozygous at most loci. Crossing two such lines gives every seed exactly the same alleles. All the seeds share one genotype, so the crop comes out uniform.
| Parents | Correct name | Are the seedlings uniform? |
|---|---|---|
| Two distinct inbred lines | True F1 | Yes, genetically identical |
| Two unrelated plants that are not inbred | First cross, loosely called “F1” | No |
| Two siblings from one seed batch | Second filial generation (F2) | No |
| One plant pollinated with its own pollen | S1 (first selfed generation) | Partly |
What you can do: When a listing says F1, ask the seller what the parents were. If they were not inbred lines, treat the seed as a first cross. Then plan for plants that differ in height, timing and aroma.
6. Hybrid Vigour: What the Cannabis Evidence Shows
Heterosis, also called hybrid vigour, means a cross between two inbred lines often grows better than either parent. Most of the classic evidence comes from maize (corn), which has about a century of hybrid breeding behind it. Evidence for cannabis is much thinner. Much of what grower forums repeat is maize genetics applied to cannabis without saying so.
Some cannabis evidence now exists. In 2024, McDonald and Lubell-Brand published a study in HortScience. They compared F1 hybrid lines with selfed inbred lines, meaning lines made by pollinating plants with their own pollen. The hybrids were more uniform and had more total plant dry weight and more flower dry weight.
In 2026, Garcia-de Heer and colleagues published a proof-of-concept study in Horticulture Research. They used single-seed descent: in each generation they took one seed per plant and selfed again. They repeated this for up to six rounds across sixteen lines. Then they crossed the resulting inbred lines with each other. The results were:
- The F1 hybrids were statistically more uniform than the inbred lines or the original lines.
- Seed yield was between 3.9% and 155% higher than the mid-parent value, which is the average of the two parents.
- All F1 plants started flowering within 4 days of each other. In the parent material, the start of flowering was spread over 21 to 23 days.
The most reliable benefit of a true F1 is synchrony, meaning the plants flower and finish together. The yield gain was real, but it varied enormously from one cross to another.
Translator’s note: Synchrony helps a Thai grower because the whole room ripens at once and you can harvest it on one date. This matters most in the rainy season. Ripe buds left on the plant while you wait for late plants run a higher risk of bud rot (Botrytis). True F1 seed from inbred parents is rare in every market, as section 7 explains, so you will seldom find it for sale.
What you can do: In your next seed run, write down the date each plant shows its first flowers. If all plants start within a few days, the batch is uniform. A spread of about three weeks, like the parent material above, suggests the seed was not a true F1.
7. Why Most Seeds Sold as F1 Are Not True First Crosses
To make a true F1, a breeder first has to create two homozygous parent lines. Both lines then have to be kept going for as long as the seed is sold, because each new batch needs both parents again.
That is a lot of work. The single-seed descent study above needed 18 months or more to reach the sixth selfed generation. Only five of its sixteen lines got that far. Almost no one in the seed trade does this work. Seed sold as “F1” is usually a cross between two heterozygous plants, which are often complex hybrids themselves.
“F1 of two elite clones” is the clearest warning sign. A clone, or cutting, is a copy of one plant, and every clone is heterozygous. Seed from two clones therefore cannot be uniform.
One question separates a real F1 from a decorative label: were both parents inbred, and for how many generations? A breeder doing real work can answer at once.
Translator’s note: Seed on the Thai market comes from many sources, and strain names are sometimes copied. Check the seller first with How to Vet a Thai Seed Vendor. For the legal side of buying seed, read Buying Cannabis Seeds in Thailand in 2026.
What you can do: Send the seller the question above before you buy “F1” seed, and save the reply. If there is no clear answer, plan the run as a phenotype hunt rather than a uniform production crop. That way you budget space for selection instead of expecting identical plants.
8. The Second Generation (F2): Where Variation Opens Up
F2 is the offspring of two F1 siblings crossed together. In this generation, variation that was hidden in the F1 comes out in new combinations.
The amount of variation depends on how many loci are involved:
- One locus gives three genotypes in a 1:2:1 ratio.
- Two independent loci give four visible types in a 9:3:3:1 ratio, when one allele simply dominates at each locus.
- Dozens of loci give a smooth range of types.
An F2 is a population to select from, not a finished product. The photo on the listing usually shows one plant that the breeder found in their own F2. The table shows your odds of getting that plant back. Here, “homozygous for all target alleles” means the seedling has two copies of the wanted allele at every locus, so it will pass the trait on reliably.
| Loci involved | Possible genotypes | Chance a seedling is homozygous for all target alleles |
|---|---|---|
| 1 | 3 | 1 in 4 |
| 2 | 9 | 1 in 16 |
| 3 | 27 | 1 in 64 |
| 4 | 81 | 1 in 256 |
Translator’s note: Indoor space in Thai homes and small farms is usually limited, and every extra plant adds to the cooling load. Check the table before you buy seed. If the trait you want depends on three loci, you need far more seedlings than one pack holds.
What you can do: Treat an F2 pack as a hunt. Grow as many seedlings as your space allows, record each plant’s traits, and take cuttings only from the best one or two. Part 2 shows how many seedlings you need for a good chance of success.
Questions about this part? Post them below this topic.
References
- Toth and colleagues (2022). Frontiers in Plant Science. Mapping of the autoflowering locus Autoflower1 in Cannabis sativa.
- de Meijer and colleagues (2003). Genetics. The B locus and the inheritance of chemotype.
- Laverty and colleagues (2019). Genome Research. Physical mapping of the cannabinoid synthase genes.
- Divashuk and colleagues (2014). PLOS One. Molecular cytogenetic characterization of the dioecious Cannabis sativa with an XY chromosome sex determination system.
- McDonald and Lubell-Brand (2024). HortScience. F1 hybrid lines compared with selfed inbred lines.
- Garcia-de Heer and colleagues (2026). Horticulture Research. Single-seed descent and F1 hybrids, proof of concept.
This article is part of the Asiannabis Community educational series on cannabis genetics and breeding. The content is for education only, in places where cannabis cultivation is legally permitted.
Continued in Part 2: Cannabis Genetics Decoded: F3 to F6, S1, Backcrosses and Polyhybrids, and How to Read a Seed Listing (Part 2 of 2)