Selecting for Terpenes

Growing, buying or possessing cannabis and seeds is legal only in some places. Check the law where you live before acting on this information.

What it is

Selecting for terpenes means choosing which plants to keep, clone, or breed forward based primarily on their aromatic chemistry—the volatile compounds that give cannabis its smell and contribute to flavor. Instead of picking the biggest yielder or the highest-THC phenotype, you're picking the one that smells like gas, or citrus, or pine, or whatever target chemotype you're hunting.

Terpenes are a large family of volatile hydrocarbons produced in the trichome heads alongside cannabinoids Strong evidence^[1]. Common ones in cannabis include myrcene, limonene, caryophyllene, pinene, linalool, terpinolene, and ocimene, though dozens more appear in smaller amounts Strong evidence^[2]. A plant's terpene fingerprint is largely heritable, which is why selection works at all Strong evidence^[3].

Note: the claim that a specific myrcene percentage predicts sedation or strain classification is folklore without peer-reviewed support. See Why Terpenes Matter (Or Don't) for a full explainer.

Why growers use it

Three main reasons:

1. Market differentiation. In legal markets, flower with a distinctive, clean aroma—identifiable as a specific chemotype—tends to command higher retail prices than generic bulk flower. This is particularly true for regulated markets where consistent terpene profiles help with shelf recognition and repeat purchase Weak / limited.

2. Flavor in the finished product. Terpenes survive curing and are responsible for most of what you taste when you smoke or vape cannabis. Selecting for terpene intensity and complexity directly improves the sensory experience Strong evidence^[6].

3. Proposed synergy with cannabinoids. Some research suggests terpenes may modulate the subjective effects of THC-dominant cannabis through receptor interactions Disputed^[4][5]. However, controlled human trials are minimal. A 2020 study found that single terpenes added to THC showed no detectable behavioral effect in mice Weak / limited^[5], while other work proposes terpene-cannabinoid binding at specific receptors—a mechanism not yet confirmed in human subjects.

What selecting for terpenes will not do: increase yield. It often reduces yield, because the most aromatic phenotypes are frequently smaller, slower, or more finicky than commercial workhorses.

When to start

You can begin rough screening as early as late vegetative growth—rub a stem or a fan leaf petiole and smell your fingers. This catches gross differences (fuel vs. floral vs. neutral) but is unreliable for fine distinctions or intensity.

Serious evaluation happens mid-to-late flower, roughly week 5 onward, when trichomes have built up and the plant is producing its mature aromatic profile. Even more informative is evaluation after harvest, drying (10–14 days), and at least a 2-week cure, because curing dramatically changes the terpene profile—some compounds volatilize off, others become more prominent as chlorophyll and 'green' notes fade Strong evidence^[6]. The cured product is what you and your customers will actually consume.

If you only smell plants while they're standing in the tent, you are selecting for the living-plant aroma profile, which does not reliably predict the cured flower experience.

How to do it (step-by-step)

Step 1: Define your target. Write down what you're hunting for before you start smelling—for example, 'gas with a citrus top note,' 'pure pink grapefruit,' 'cake and vanilla,' etc. Vague targets produce vague results.

Step 2: Run enough plants. Phenotype variation within a seed line is real. To find rare terpene expressions you typically need at least 10–30 seeds of a cross; polyhybrids may require more. Clones of the same plant are not phenotype hunting—they contain no variation.

Step 3: Tag and track. Label every plant with a unique ID. Keep a notebook or spreadsheet. Record vigor, structure, flower time, and aroma notes at multiple checkpoints.

Step 4: Smell at multiple stages.

• Late veg: stem rub, leaf petiole rub.
• Week 3–4 flower: stem rub, light squeeze of a lower bud.
• Week 5–7 flower: fresh bud squeeze, sample-jar test (snip a small bud, place it in a glass jar, close the lid for 30 minutes, open and smell).
• Post-dry (10–14 days): jar test on a representative bud.
• Post-cure (2–4 weeks in glass with burping): final evaluation.

Step 5: Reset your nose between samples. Fresh air and 1–2 minutes of waiting is most reliable. Evaluate no more than 5–6 plants per session before your discrimination collapses due to olfactory fatigue.

Step 6: Use blind ranking. Have a friend relabel the jars and rank them without knowing your original top picks. If your rankings change between sessions, your nose is less reliable than you think.

Step 7: Send finalists to a lab. A terpene panel by GC-MS (gas chromatography–mass spectrometry) or HPLC (high-performance liquid chromatography) from a licensed cannabis testing lab will give you actual percentages of individual terpenes Strong evidence^[7]. This is the only way to confirm that a 'gassy' phenotype is high in caryophyllene rather than low in chlorophyll smell, or that a 'lemon' phenotype actually has elevated limonene.

Step 8: Keep clones of finalists, not just memories. Take cuts of anything you're seriously considering before you chop. You cannot un-harvest a plant.

Common mistakes

• Selecting on living-plant smell only. The aroma of a standing flowering plant correlates only loosely with the cured product. Stem rubs release crushed-leaf volatiles, not the terpene profile you'll encounter after dry and cure.
• Confusing intensity with quality. A plant that screams aroma in the tent is not necessarily the most flavorful when smoked. Some of the most pleasant-tasting cultivars are quiet on the plant.
• Ignoring total terpene content vs. profile. Two plants with the same 'lemon' character may have 1.5% total terpenes and 0.6%, respectively. A lab panel reveals this; your nose does not.
• Assuming strain name or indica/sativa predicts terpenes. The indica/sativa split has been repeatedly shown to be an unreliable predictor of chemical profile; genetic and chemical analyses group cultivars in ways that ignore those traditional labels Strong evidence^[8].
• Skipping proper cure before making final decisions. Wet-trimmed, fast-dried flower smells different—often simpler and less refined—than the same flower cured for 2–4 weeks in glass.
• Selecting one plant from one seed run and calling it stable. A single F1 standout is a starting point, not a finished line. Stabilizing a terpene profile across a population typically takes 3–6 generations of selection and controlled breeding.

Related techniques

Terpene selection rarely happens in isolation. It usually pairs with:

• Selecting Mother Plants — choosing which individuals to propagate and maintain.
• Backcrossing — the breeding method most often used to lock in a desirable terpene phenotype across a population.
• Curing flower — the post-harvest process that determines how much of your selected terpene profile actually survives to the jar (note: while Boveda and similar humidity packs are convenient, they can accelerate terpene loss over long storage).
• Lab cannabinoid and minor-cannabinoid testing — pairing terpene panels with THC/CBD data for a complete chemotype picture.
• Trichome visual inspection — harvest-readiness assessment, since over-mature trichomes have already lost significant volatile content.

Sources

How this page was made

Related

• Backcrossing — A breeding technique used to lock in a desired trait by repeatedly crossing offspring back...
• Boveda Packs Ruin Terpenes — A popular claim with no evidence. What really degrades terpenes—and why humidity packs hel...
• Pheno Hunting — Deliberate selection of the best plant phenotype from a seed population to clone and prese...
• Selecting Mother Plants — Identifying and preserving the best phenotype from seed for reliable, consistent cloning.
• Why Terpenes Matter (Or Don't) — Terpenes drive scent and flavor; effects claims are mostly preclinical or anecdotal.