Field Guides

What are terpenes, and what can they tell you?

By SŌMA 10 min read Updated

Terpenes are a large family of aromatic compounds made by many plants, cannabis among them. New York's Office of Cannabis Management (OCM) describes them as the compounds that provide the aroma of cannabis and of a variety of other plants, and gives four examples: limonene smells like citrus, pinene like pine, linalool like lavender and caryophyllene like black pepper.

They are not the whole story of a flower's smell, and what they do beyond smell is contested. This guide covers what terpenes are, what a terpene panel on a lab certificate can and cannot tell you, and what research does and does not show about their effects. Its sources are listed at the end.

What terpenes are

All terpenes are built from the same five-carbon building blocks. In cannabis the word usually covers the monoterpenes and sesquiterpenes and their oxygen-containing relatives, the terpenoids, and together they are the plant's major volatile compounds. Cannabis makes them in a resin, inside glandular trichomes: tiny glands that are abundant on the surface of the female flower and that produce the cannabinoids too. Terpenes are not unique to cannabis. Limonene, one of those most often found in it in the greatest amounts, is ubiquitous in citrus fruit, and the terpenes of hops, a close relative of cannabis, are an important flavouring in brewing.

By weight they are a small part of the flower. In a study of lab results for flower from six US states, among nearly 43,000 samples with terpene data, total terpenes averaged 2% of the weight; myrcene, β-caryophyllene and limonene were present at the highest levels on average, and a single terpene was rarely above 0.5%.

Terpenes and smell

Terpenes account for much of the scent of cannabis flowers, and OCM's testing rules call them primarily responsible for its aroma. But a list of terpenes is not a description of a smell. Some of the most distinctive notes have been traced to other compounds, present in far smaller amounts:

  • The skunk-like smell has been traced to a family of sulfur compounds. One of them, the compound known for the smell of “skunked” beer, tracked the aroma of 13 cultivars most closely.
  • In 31 concentrates with very different aromas, the terpene profiles were often very similar; the differences came from minor, non-terpene compounds, often less than 0.05% of a sample's weight. Sulfur compounds of another kind went with strong citrus and tropical notes in some samples, and a compound called skatole with a savoury, “chemical” note in others.
  • In concentrates made from five plants grown from the same cross, aroma differences came from minor compounds rather than from the dominant terpenes, and two fatty acids were implicated in cheese-like notes.

These three studies come from researchers at one company, which funded the two later ones; members of the group have filed patents related to the findings, and two of the studies analysed concentrates rather than flower. An independent 2025 study of dried flower from six CBD-rich hemp cultivars found both kinds of compound among the most potent odorants: terpenes such as α-pinene, myrcene and linalool, and sulfur compounds, the skunk-linked one among them. An earlier study of how consumers describe the smell of flower had already noted that it was not obvious how terpenes could produce the skunk, diesel and earthy notes people reported.

Another 2025 study tested directly how well chemistry predicts the smell of intact flower. A trained panel described 91 samples with 25 defined aroma words: 59 of CBD-rich hemp and 32 of THC-rich flower from a licensed shop. Terpene profiles alone poorly predicted how the panel described the samples, and in the hemp samples, the only ones in which sulfur compounds were measured, neither the terpenes nor the sulfur compounds strongly predicted it. One group of samples especially high in limonene lacked a strong citrus character. The authors say their sample may have been too small, and that key aroma compounds may have gone unmeasured; the study was funded by a pharmaceutical company which, they say, had no role in it.

What this means in practice: a terpene panel lists the terpenes a lab measured in a sample, which are some of the compounds behind its smell. The sulfur compounds are not terpenes, and the studies that found them used a detector built for sulfur. A high limonene figure does not guarantee a citrus nose, and a terpene panel cannot show you a skunk or fuel note.

Terpenes on a New York certificate

OCM's testing rules require every final cannabis product, flower included, to be tested for terpenes, and set a minimum list of 20 terpenes to test; its laboratory guidance also lists terpenes among the tests required for every product type. So the lab certificate, reached through the label's QR code or link, should carry a terpene panel. The label itself must show a terpene profile only if terpenes are marketed. Our QR code guide shows where terpenes appear on the Metrc Retail ID record, which may list none. Either way, a panel describes the sample the lab tested, not your package today.

Do terpenes fade first?

This guide used to say that terpenes, being volatile, are the first thing to fade as flower ages. The evidence is thinner than that. A one-year laboratory study of flower kept in the dark at four temperatures measured terpenes in a single CBD-rich variety: there they fell under every condition and were down by more than half on average after four months, while the cannabinoid acids, the forms in which the plant makes its cannabinoids, fell by 26% on average over the year, mostly by turning into THC and CBD.

Two other findings point the same way, each with limits. A 1996 study of indoor-grown flower compared the oil distilled from fresh and from air-dried buds; we could read only its abstract, which says drying did not change which compounds were present, or stop people familiar with the smell from recognising it. A later paper, by authors who hold shares in a cannabis company, reports that the same study measured terpene losses of 31% after a week, 45% after a month and 55% after three months. And in the 2021 sulfur study's own growing trial, in four plants of one variety, most of the skunk-like sulfur compounds peaked during curing and had dropped substantially ten days after curing ended.

So aroma compounds can fall quickly in storage, but each of these studies looked at one variety or one set of samples, and none shows that terpenes in general fade before anything else, or how fast the aroma of a given package fades. Why freshness matters and How to store cannabis set out what is known.

The aroma families

You do not need terpene names to describe a nose. In a study in which 61 people chose from a list of 48 aroma words to describe dried flower, the words chosen most came from everyday categories, such as woods, spices, florals and citrus fruits, along with notes such as diesel, skunk, pungent and cheese; the single word chosen most was “earthy”. Families like these are a working vocabulary, not a chemical map:

  • Earthy, woody and herbal: earth, wood, sage, tea.
  • Pine: pine, fresh forest, resin.
  • Citrus: lemon, lime, orange, grapefruit.
  • Floral: lavender, rose, violet.
  • Pepper and spice: black pepper, a warm sharpness.
  • Fuel, skunk and funk: diesel, skunk, cheese. The studies above tie skunk and fuel notes to sulfur compounds, and cheese-like notes to fatty acids, rather than to terpenes.

Our sensory vocabulary and How to taste cannabis put these families to use.

Terpenes and effects: the “entourage” idea

The entourage effect is the idea that a plant's compounds act together, so that the whole differs from the sum of its parts. The term was coined in 1998, for a laboratory finding about the body's own cannabinoid-like compounds; its authors suggested, as an afterthought, that it might apply to cannabis products too, and terpenes were later proposed among the compounds involved. It is widely repeated, and unproven.

What the evidence shows so far:

  • Cell studies disagree. Two found that terpenes common in cannabis neither acted directly at the cannabinoid receptors CB1 and CB2 nor changed how THC acted there, beyond a possible weak action of β-caryophyllene at CB2; a third found that five terpenes, at high concentrations, activated CB1.
  • In mice, that third study found that some terpenes produced cannabinoid-like effects and added to the effects of a synthetic cannabinoid.
  • In people, controlled studies are few. In one trial, 20 healthy adults inhaled vaporised THC alone and with limonene. Only the highest limonene dose, given to 12 of them in a session that always came last, led to significantly lower ratings of feeling “anxious/nervous” and “paranoid” than the same THC alone; THC's other effects were unchanged, and limonene alone did not differ from placebo. The researchers used pure compounds, at amounts and a ratio they say are unlikely in unadulterated flower; they note that the fixed order of that last session may have affected the result; and a patent application has been filed on their behalf.
  • A 2023 scoping review of the research for and against the idea found the literature “in its infancy”, with simplistic methods and contradictory findings.

So treat the entourage effect as a hypothesis. The research so far does not support reading a feeling from a terpene figure on a certificate.

Terpenes and the THC number

THC is the cannabinoid known for the high, in OCM's words, and it is not what you smell. THC itself is non-volatile and odourless, as the authors of the consumer study above note, and in that study people's smell-based guesses of potency did not track the THC the lab measured. Nor does the THC figure tell you the terpene mix: in the six-state study above, the three main terpene profiles of THC-dominant flower had similar THC distributions, so packages with the same THC figure can carry quite different terpene profiles. To choose by aroma, use your nose, and the terpene panel for what it measures, not the THC number.

How uses terpenes

SŌMA describes strains in a fixed vocabulary of aromas, flavours and effects. Its full records, researched by hand, are written from growers', breeders' and other published descriptions, and a lab panel does not change them: a panel measures amounts of terpenes, while a record describes how a strain is perceived. For a name with only a provisional record, or none, SŌMA reads aroma and flavour from New York lab terpene panels where it holds them, averaged across the lots of that name from the grower the menu names, and never from another grower's lots; a name several growers sell in New York is not read from a panel until the menu says whose it is.

A panel cannot produce a gas or skunk note, and, as the 2025 study found, terpene profiles predict how flower smells only loosely, so a read built from a panel is an approximation of the smell. Where SŌMA has neither a record nor a panel, it estimates from the plant's parents, a model's reading or the name itself, gives that read lower confidence, and labels it an estimate. Each scored read says what it rests on.

Sources

Frequently asked questions

What are terpenes in simple terms?

A large family of aromatic compounds made by many plants. In cannabis they are made in the resin glands on the flower, alongside the cannabinoids, and they are its major volatile compounds. They account for much of its smell, though some of the most distinctive notes, such as skunk, have been traced to other compounds, and in one study of intact flower, terpene profiles poorly predicted how a panel described the smell.

Do terpenes get you high?

THC is the cannabinoid known for the high. Whether terpenes change the character of the experience is contested: cell studies disagree on whether common terpenes act at the cannabinoid receptors, and studies in people are few and small. In one small trial using pure compounds, inhaled limonene taken with THC led to lower ratings of feeling anxious or paranoid than THC alone, but only at the highest dose, given to 12 people in a session that always came last, and at amounts its authors say are unlikely in flower; a patent application was filed on the result.

Is the entourage effect real?

It is unproven. The idea that terpenes and cannabinoids act together is widely repeated, but a 2023 scoping review found the research in its infancy, with simplistic methods and contradictory findings. Treat it as a hypothesis, not as a reason to expect a particular feeling from a particular terpene.

Does the THC percentage tell me how flower will smell?

Not which aromas. THC itself has no smell; in one study people's smell-based guesses of potency did not track measured THC, and in a six-state study of lab results the main terpene profiles of THC-dominant flower had similar THC levels (total terpene content did rise modestly with total cannabinoids there). Your nose, and the terpene panel for what it measures, say more about aroma.

Where can I see a flower's terpenes in New York?

On the lab certificate the label's QR code or link leads to: OCM's testing rules require every final product, flower included, to be tested for terpenes. The Metrc Retail ID record may show them too; our QR code guide shows where. A panel describes the sample the lab tested, not your package today.

Corrections

  • Our title said terpenes shape your experience, and the description called them the compounds behind a strain's smell and much of its character. Whether terpenes change how cannabis feels is contested, and some of the most distinctive notes in its smell have been traced to other compounds; the title and description now say what terpenes can tell you.
  • We said that what makes one strain smell like lemon and another like diesel is mostly terpenes, and, in the FAQ, that terpenes are responsible for a strain's smell and much of its flavour. Studies have traced skunk- and fuel-like notes to sulfur compounds and other minor compounds that are not terpenes, and in a 2025 study of intact flower terpene profiles poorly predicted how a panel described the smell. The guide now says so, with the limits of those studies and an independent one that found terpenes and sulfur compounds both among the most potent odorants. The citrus, pine, lavender and pepper examples now follow OCM's guide to lab certificates.
  • We said terpenes are why two batches with the same THC number can feel like different plants. We found no study showing that and removed it; the guide now says that packages with the same THC figure can carry quite different terpene profiles, as a six-state study of lab results found.
  • We said that, because they are volatile, terpenes are the first thing to fade as flower ages or is stored badly, so freshness shows up first in the nose. The storage studies we found each looked at one variety or one set of samples, and none shows that terpenes fade before anything else; the guide now gives their findings rather than a rule.
  • In the aroma families we described earthy notes as “a settled heaviness” and citrus as “bright and lifting”, which suggest effects. We found no source linking those families to those feelings and removed both phrases; the families now draw on the words chosen most in a published study in which people described the smell of flower from a list of 48.
  • We called the entourage effect a plausible hypothesis, and said terpenes are thought by some to influence the character of the experience. The guide now sets out what cell, animal and human studies do and do not show, that the cell studies disagree, and that a 2023 review found the research in its infancy, with contradictory findings.
  • We said a single THC figure tells you almost nothing about how flower will smell, taste or feel, and that the terpene picture describes the experience you are choosing and matters more than the THC percentage (in the body and the FAQ). For smell, the guide now gives the evidence; the claims about feel are removed, since THC is the cannabinoid known for the high and we found no source for them.
  • We said SŌMA describes strains in this aroma language and matches them to the families you prefer. SŌMA weighs effects as well as aroma and flavour; its researched records are written from published descriptions, and for other names it reads aroma from New York lab terpene panels where it holds them, or estimates from the plant's parents, a model's reading or the name. The guide now says so.
  • We said a working sense of the main aroma families is enough to describe what you like and find more of it, and that noticing which families you reach for again and again is the start of a profile you can use. Both were advice rather than findings: the first is now “You do not need terpene names to describe a nose”, and the second stays in our sensory vocabulary.