What terpenes are

Five-carbon bricks¶

Every terpene is assembled from the same five-carbon unit. Two activated isoprene equivalents, isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP), are joined head-to-tail by prenyltransferases into linear chains of ten, fifteen, twenty, or more carbons. Those chains are then folded, cyclized, and rearranged by a family of enzymes called terpene synthases into the thousands of ring systems that give the class its diversity. The carbon count fixes the top-level vocabulary:

In strict usage a terpene is a hydrocarbon (only carbon and hydrogen) and a terpenoid is a terpene that has been oxidized, rearranged, or otherwise functionalized. Linalool, an alcohol, is technically a monoterpenoid; limonene, a hydrocarbon, is a true terpene. Industry and most of the literature use “terpenes” for both, and so does this paper except where the distinction changes a count or a regulatory status.

How plants make them¶

Plants run two independent routes to the C5 precursors Christianson, 2017Pichersky & Raguso, 2018. The mevalonate (MVA) pathway in the cytosol supplies precursors mainly for sesquiterpenes (C15) and triterpenes (C30). The methylerythritol phosphate (MEP) pathway in the plastids supplies precursors mainly for monoterpenes (C10), diterpenes (C20), and carotenoids (C40). Prenyltransferases condense IPP and DMAPP into geranyl diphosphate (GPP, C10), farnesyl diphosphate (FPP, C15), and geranylgeranyl diphosphate (GGPP, C20). Terpene synthases then act on those diphosphates.

Terpene synthases are the source of the class’s astonishing diversity, and of much of the confusion around it. A single synthase frequently yields a dozen or more products from one substrate, because the carbocation chemistry it initiates can branch at several points before the reaction is quenched Christianson, 2017. Small changes in the enzyme’s active site change the product spectrum. Plants also carry large, rapidly evolving synthase gene families; Cannabis sativa alone has dozens of terpene synthase genes, and its inflorescence has been reported to accumulate well over a hundred distinct terpenes Booth & Bohlmann, 2019. Downstream cytochrome P450s, dehydrogenases, and transferases decorate the skeletons further. The ecological logic is that a cheap, combinatorial system for making many volatile and semi-volatile molecules lets plants signal to pollinators, deter herbivores, recruit the enemies of herbivores, and protect tissues from oxidative and thermal stress Pichersky & Raguso, 2018.

For a product team the practical lessons are these. First, the terpene profile of a plant is a population of molecules produced by a network of enzymes, so it varies with genotype, tissue, development, environment, and post-harvest handling. Second, minor constituents are not noise; a synthase’s side products are real, reproducible, and often odor-active. Third, “the same terpene” from two botanical sources may differ in enantiomer ratio, and enantiomers can smell and behave differently.

How many terpenes are there?¶

The honest answer is that there is no defensible single number without a counting rule Terpedia, LLC, 2026. Reviews have variously reported more than 20,000, more than 30,000, roughly 55,000, and more than 80,000 known terpenes or terpenoids. These figures count different chemical universes: some include only plant compounds, some include stereoisomers as separate entries, some include glycosides and conjugates, some count database records rather than characterized structures, and some quietly conflate terpenes with terpenoids.

Terpedia’s own census takes a different approach. Rather than asserting a total, it unions the source-declared terpenoid records of two independent open databases, COCONUT Sorokina et al., 2021 and TeroKit Zeng et al., 2020, on exact standard InChI, and reports the result as a dated, scoped tuple: 268,924 candidate terpene/terpenoid identities in the September 2026 snapshot, of which 225,905 have an exact PubChem CID match. An independent audit against the ChEBI ontology confirms 6,585 of those by exact InChIKey, and adding 54 directly characterized terpene-synthase products from the MARTS database brings the ontology-confirmed set to 6,605 Terpedia, LLC, 2026. Those are candidate identities, not independently validated natural terpenes. The gap between the 268,924 candidates and the 6,605 ontology-confirmed identities is itself informative: it is the amount of terpene chemistry that exists in databases but has not yet been organized into a curated classification.

The point for industry is not the number. It is that a supplier, a lab, a regulator, and a marketing team who each quote a different “number of known terpenes” are not disagreeing about facts; they are using different counting rules, and a platform that wants to be trusted has to make the rule explicit.

Where terpenes occur¶

Terpenes are produced by plants, fungi, bacteria, marine organisms, and insects. Commercially, the plant kingdom dominates:

• Conifers supply turpentine and its fractions (α- and β-pinene, camphene, 3-carene), the historical bulk source of monoterpene feedstocks.

• Citrus supplies d-limonene at industrial scale as a by-product of juice production, together with valencene, γ-terpinene, and citral.

• Lamiaceae herbs (mint, lavender, rosemary, thyme, oregano, basil, sage) supply menthol, linalool, linalyl acetate, 1,8-cineole, camphor, thymol, carvacrol, and thujone.

• Myrtaceae (eucalyptus, tea tree, clove, allspice) supply 1,8-cineole, terpinen-4-ol, and β-caryophyllene.

• Hops and cannabis share a characteristic profile dominated by myrcene, β-caryophyllene, α-humulene, and limonene, with linalool, α-pinene, β-pinene, ocimene, terpinolene, and nerolidol as frequent secondary constituents Booth & Bohlmann, 2019.

• Spices such as black pepper, cinnamon, cardamom, ginger, and nutmeg supply β-caryophyllene, sabinene, zingiberene, and myristicin-adjacent aromatics.

Occurrence is a report, not a concentration. The LOTUS initiative, an open compound-organism occurrence resource that Terpedia ingests, records which molecules have been reported in which organisms and cites the paper that reported each pair Rutz et al., 2022. It is invaluable for asking “where else has this molecule been found?” but it does not tell you that an organism is a meaningful commercial source. Terpedia keeps that distinction explicit in its data model.

Physical chemistry that matters for products¶

Four properties govern how terpenes behave in a product and how reliably they can be measured.

Volatility. Monoterpenes have boiling points around 150–190 °C and appreciable vapor pressure at room temperature; sesquiterpenes are heavier and less volatile. This is why terpene profiles change during drying, curing, storage, and heating, why headspace and vapor composition differ from the bulk, and why a certificate of analysis is a snapshot of a moment rather than a property of a product.

Lipophilicity. Terpenes are hydrophobic. Limonene’s measured logP is about 4.4 and it is practically insoluble in water. This drives their partitioning into oils, membranes, and packaging, their need for emulsification in aqueous products, and their tendency to be lost to plastics.

Oxidation. Unsaturated terpenes autoxidize on exposure to air and light. For limonene and linalool the primary oxidation products are allylic hydroperoxides, which are potent skin sensitizers even though the parent molecules are weak or non-sensitizing Dittmar & Schuttelaar, 2019. Oxidative aging also changes aroma. Stability, antioxidant systems, and storage conditions are therefore quality and safety questions, not just shelf-life questions.

Stereochemistry. Many terpenes are chiral, and enantiomers can differ in odor and biology. (R)-(+)-limonene is the orange note; (S)-(−)-limonene reads as pine or turpentine. Analytical methods that do not resolve enantiomers, and databases that do not distinguish them, lose information that a perfumer or a flavorist relies on.

Each of these properties creates a specific data-management requirement: stereo-aware identity, time-stamped measurements, oxidation-product tracking, and matrix-aware concentration units. Those requirements are the reason Terpedia’s data model is built around exact chemical identity and provenance rather than around names.