Cannabis Genetics: Sativa, Indica and Modern Classification

The division of cannabis into sativa and indica has shaped dispensary menus, seed bank catalogues and the everyday vocabulary of consumers for decades. In dispensaries and coffeeshops it serves as basic orientation: sativa is supposed to make you alert and creative, indica to relax you and help you fall asleep. In the current scientific literature, however, this two-way split has largely been superseded. Plant research, genetics and chemotaxonomy agree in concluding that the botanical categories reliably predict neither the chemical composition nor the pharmacological effect of a plant. At the same time, the language remains firmly anchored in the market, and the categories have a real core in the history of where the landraces came from. This article places both perspectives in a systematic context and shows what a modern classification by chemovar looks like.

1. The classic types at a glance

The traditional division rests on three observable bundles of traits: plant morphology, the geographical origin of the parent lines and – added in the trade – presumed effect profiles.

In commercial usage, sativa refers to tall, slender plants with narrow, long leaflets and a long flowering time of ten to 16 weeks. The parent lines come from equatorial and subtropical regions – South Asia (India, Thailand, Laos, Vietnam), Central and South America (Mexico, Colombia, Jamaica) and parts of Africa (South Africa, Morocco). The type is credited with cerebral, stimulating effects, the so-called “head high”.

Indica refers to compact, bushy plants with broad, dark green leaflets and a short flowering time of six to ten weeks. The parent lines come from the mountainous dry regions of Central Asia, above all Afghanistan, Pakistan, northern India and the adjoining parts of Turkestan. In the trade, indica stands for a physically sedating effect (“body high”, “couch-lock”).

Hybrid is not a botanical category but an umbrella term for crosses between the types named. After decades of commercial breeding, practically all strains available today are hybrids; pure sativa or indica genetics can hardly be found any more outside specialised landrace projects.

2. How the classification came about

The botanical separation goes back to two taxonomists of the eighteenth century. In 1753, Carl Linnaeus described cannabis in Species Plantarum as the single species Cannabis sativa L., with “sativa” derived from the Latin sativum (“cultivated”). In 1785, Jean-Baptiste Lamarck introduced Cannabis indica as a second species and distinguished it from C. sativa by its lower height, woody stems, narrower leaflets and a hairy calyx. The assignment was geographically motivated: “indica” literally meant “from India”.

The three-way division common in the trade today is, however, considerably younger than these botanical sources. In the 1970s, the Harvard ethnobotanist Richard Evans Schultes transferred the term “indica” to Afghan plants that did not correspond morphologically to the Indian plants originally described by Lamarck. The modern vernacular taxonomy was finally cemented by a line drawing that Loran Anderson published in 1980 and that presented indica and sativa as morphologically and geographically contrasting types. From then on, the division was adopted by breeders in California, the Netherlands and later worldwide. John McPartland therefore calls this younger form a “vernacular taxonomy” and explicitly distinguishes it from formal botanical nomenclature.

3. Geographical centres of origin and spread

On the basis of pollen analyses and chloroplast DNA studies, the region where cannabis originated is today placed on the north-eastern Tibetan Plateau. A study by Zhang and colleagues, published in Frontiers in Plant Science in 2018, analysed the chloroplast DNA of a broad global sample and identified three haplogroups that correlate with the latitudes of the regions of origin: a high-altitude group (H) with the shortest growing time and low height, a middle group (M) and a lowland group (L) with the longest maturation time and tallest growth. This climate-driven differentiation is the botanically real basis of the indica/sativa distinction that was labelled later.

Cannabis spread along early trade and migration routes: south into present-day India, west along the Silk Road through Central Asia to Persia and the eastern Mediterranean, east to China and, over thousands of years, to Europe. In each of these regions, landraces developed – genetically relatively stable local populations that had adapted to their particular microclimate over many generations. Among the best-known landraces are Afghan (Hindu Kush), Hindu Kush (Pakistan/India), Kashmir, Nepalese, Thai, Colombian Gold, Acapulco Gold, Panama Red, Malawi and Durban Poison.

In 2020, after analysing 1,100 herbarium specimens, McPartland and Small presented a taxonomic reorganisation in PhytoKeys: according to it, what is colloquially called “sativa” corresponds to Cannabis sativa subsp. indica var. indica, originating in South Asia. What is called “indica” corresponds to C. sativa subsp. indica var. afghanica, originating in Central Asia (Afghanistan, Pakistan, Turkestan). The wild relatives are var. himalayensis (South Asia) and var. asperrima (Central Asia).

An often overlooked finding is important here: the varieties were never sharply separated. In the geographical contact zone – especially in the Kunar and Chitral catchment in northern Pakistan – intermediate forms have occurred for millennia. Landraces are interfertile and cross into each other naturally in overlap zones. McPartland and Small speak of “introgressive hybridisation” and document that all four described varieties regularly occur side by side in herbarium specimens from Pakistan. A landrace field along the Silk Road therefore often shows plants with sativa traits and plants with indica traits in the same stand. This is not the result of modern breeding but natural gene flow between neighbouring varieties – an important nuance that gets lost in the trade narrative (“pure” landraces).

4. The commercial wave of hybridisation since the 1970s

Modern breeding has radically reshaped the landrace diversity that grew up over history. Three developments were decisive:

First, in the late 1960s and 1970s, travellers of the hippie era (“Hippy Trail”) brought seeds from Afghanistan, Nepal, Thailand, Colombia and Mexico to the USA and Western Europe. Crosses of these lines began in California and later in the Netherlands.

Second, McPartland and Small document in their 2020 paper that from the 1970s genetics were also imported in the opposite direction into the countries of origin: “Mexican Gold” to Afghanistan around 1972, Afghan landraces to Nepal, Jamaica and Thailand in the 1970s, and from 1980 onwards to southern Kashmir, South Africa and Morocco. As a result, the “pure” landraces in their centres of diversity are today contaminated in many cases.

Third, as outdoor growing was pushed indoors (triggered by intensified law enforcement in the USA in the 1980s), a selection pressure set in that favoured compact, fast-flowering plants – in other words, Afghan genetics. In this phase the Netherlands became the epicentre of commercial breeding. The most influential breeding results to this day date from the 1980s. In the breeder literature (Seedfinder database, interviews in trade magazines), Northern Lights, Haze and Skunk #1 are regarded as the “three pillars” of modern hybrid breeding. According to these sources, Northern Lights was developed from Afghan landraces in the US Pacific Northwest in the late 1970s and was stabilised and commercialised by Sensi Seeds in the Netherlands from 1985 onwards. Haze was created in California in the 1970s from crosses of Colombian, Mexican, Thai and South Indian sativa landraces. It should be noted that these breeder stories come almost exclusively from industry and seed bank sources, are not documented in the peer-reviewed literature, and are partly classed as legendary even by the seed banks involved (for instance Sensi Seeds in its 40-year retrospective).

The consequence of these three developments is consistent: as early as 30 years ago, unhybridised landraces were hard to obtain in Europe and the USA. Today the Seedfinder database alone lists more than 14,000 hybrid strain names. The hybrid “AK-47” won the prize for best sativa at the Cannabis Cup in 1999 and the prize for best indica in 2003 – an often-cited example of how arbitrary the two categories are in practice.

5. The scientific criticism of the sativa/indica division

Current criticism of the two-way split in plant science rests on three independent lines of evidence: chemical analyses, genetic studies and taxonomic arguments.

5.1 Chemical analyses

If the labels sativa and indica are to allow a prediction of effects, they would have to correlate with systematically different cannabinoid and terpene profiles. That is not the case. A study by Smith and colleagues, published in PLOS ONE in 2022, analysed 89,923 commercial cannabis bud samples from six US states – the largest chemotaxonomic analysis of commercial material to date. The authors did find reproducible chemical phenotypes (chemotypes), but the trade labels “indica”, “sativa” and “hybrid” did not match these chemotypes consistently. Individual label assignments did show statistical tendencies (such as a weak association of “indica” with myrcene-dominant profiles), but no reliable predictive power for the individual sample.

In 2019, the neurologist Ethan Russo summed up this finding in an influential perspective paper in Frontiers in Plant Science: since cannabis plants cannot be characterised by height or leaf width, he argued, the only meaningful classification is one by biochemical and pharmacological traits. For this, Russo coined the term chemovar – “chemical variety” – and argued for replacing the botanically untenable term “strain” (which strictly speaking applies only to bacteria and viruses).

5.2 Genetic evidence

In parallel, genetic studies showed that the labels have hardly any discriminating power at the molecular level either. In 2015, Sawler and colleagues (including Jonathan Page) analysed 81 marijuana and 43 hemp samples with 14,031 SNPs in PLOS ONE. They found only a moderate correlation between the genetic structure of marijuana strains and their reported C. sativa/C. indica ancestry and stated that strain names often do not reflect a meaningful genetic identity.

In 2021, Watts and colleagues published the most precise analysis of this question so far in Nature Plants. They genotyped 137 samples with more than 100,000 SNPs and reanalysed 297 samples for terpenes and cannabinoids. The result: at the genome-wide level, samples labelled sativa and indica could not be told apart. At the same time, the authors found an important nuance: the labels correlated with variation at terpene synthase genes on chromosomes 5 and 6. This means that something chemically real stands behind the distinction – namely a difference in aroma (sweet vs. earthy) – but it is not a reliable categorisation of effects at the genome level.

In 2014, McPartland and Guy had shown by DNA barcoding that the divergence between samples labelled C. sativa and C. indica is only 0.41 % – compared with an average of 3.0 % between five other species pairs in the family Cannabaceae (for example Humulus lupulus and H. japonicus). This is well below the threshold that would justify separation at species level.

5.3 Taxonomic consequence

From these findings, a predominant consensus has emerged in the specialist literature: cannabis is a monotypic genus with a single species, Cannabis sativa L., which is divided into subspecies and varieties. In 2018, McPartland described the indica/sativa categorisation in Cannabis and Cannabinoid Research as an “exercise in futility” – the ubiquitous hybridisation had made the distinction meaningless.

6. The modern approach: chemotypes and chemovars

Classification by chemical phenotype has become established as the scientifically sound alternative. It works on two levels:

Chemotypes divide cannabis roughly by the dominant cannabinoid:

  • Chemotype I: THC-dominant (THC:CBD ratio > 1)
  • Chemotype II: intermediate (roughly balanced ratio)
  • Chemotype III: CBD-dominant (THC:CBD ratio < 1)
  • In addition, chemotype IV (CBG-dominant) and chemotype V (cannabinoid-free, for fibre hemp) are distinguished less often.

Chemovars go further and also take the terpene profile into account. Dominant terpenes include myrcene (earthy, associated with sedation), limonene (citrusy, associated with lifting mood), β-caryophyllene (spicy, with direct binding to CB2 receptors), pinene (pine-like), linalool (floral, as in lavender) and terpinolene (fresh and herbaceous). Arno Hazekamp and colleagues worked out the chemovar approach methodically in two influential studies (“Cannabis: From Cultivar to Chemovar”, 2012, and “Cannabis: From Cultivar to Chemovar II”, 2016). The second study analysed extensive samples for cannabinoids and terpenes and showed that reproducible clusters with clinical relevance can be derived from them – unlike the sativa/indica label.

The analytical gold standard for determining chemovars is high-performance liquid chromatography (HPLC). It allows the quantitative separation of all cannabinoids in their native acid form – that is, above all THCA and CBDA, as they occur in the living plant. Gas chromatography (GC), which was frequently used in the past, breaks these acids down by heat and therefore only provides a combined value for THCA and THC. For regulated medical and commercial applications, HPLC is therefore the method of choice.

7. The entourage effect – why terpenes matter

The chemovar approach is closely linked to a further debate: the so-called entourage effect. Raphael Mechoulam and Shimon Ben-Shabat coined the term in 1998 for the observation that the body’s own endocannabinoid system is modulated in its action by “inactive” metabolites. Applied to the cannabis plant, the hypothesis runs: the pharmacological effect does not arise from THC alone but from the synergistic interplay of major cannabinoids (THC, CBD), minor cannabinoids (CBG, CBN, THCV, CBC) and terpenes.

In a much-cited paper in the British Journal of Pharmacology in 2011 (“Taming THC”), Russo argued that individual terpenes have specific pharmacological effects. β-caryophyllene binds directly to the CB2 receptor of the endocannabinoid system and has an anti-inflammatory effect. In preclinical models, linalool shows anxiolytic (anxiety-relieving) properties, similar to those of lavender oil. Pinene can lessen the memory impairment caused by THC by acting as an acetylcholinesterase inhibitor. Myrcene enhances the muscle-relaxing effect of THC. In animal studies, limonene has a mood-lifting effect.

The evidence for these individual terpene effects is mixed. A review by Christensen and colleagues in Biomedicines (2023) argues that the entourage effect is best explained by established pharmacological concepts such as synergistic interaction and bioenhancement and need not be considered a phenomenon of its own. Clinically, the best-established effect is the modulation of the THC effect by CBD: several studies show that CBD weakens some of the psychoactive and anxiety-inducing effects of THC. For the broader terpene hypothesis, on the other hand, large-scale clinical studies with standardised single-terpene doses are still lacking.

The practical value of the entourage concept nevertheless lies in putting the sole fixation on THC percentages into perspective. Two strains with identical THC content can feel clearly different subjectively if their terpene profiles diverge. For medical applications – for example in pain therapy or for sleep disorders – the chemovar view of the overall profile is therefore more informative than the sativa/indica label.

8. The taxonomic dispute: monotypic or polytypic?

Parallel to the criticism of the vernacular sativa/indica division, academic botany has argued for decades about how many species the genus Cannabis contains at all. Three positions stand opposite each other:

Monotypic classification (today’s mainstream): Cannabis comprises a single, phenotypically highly variable species, Cannabis sativa L. Differences at a lower taxonomic level are described as subspecies (subsp. sativa, subsp. indica) and varieties. This position is supported by DNA barcoding, allozyme studies and chloroplast analyses.

Polytypic classification with two species: This position – represented by botanists such as Ernest Small in combination with Cronquist – distinguishes C. sativa and C. indica at species level. In practical application, Small defines the difference pragmatically via a THC limit of 0.3 %, which has been adopted in numerous countries for the legal distinction between industrial hemp and drug hemp.

Polytypic classification with three species: Authors such as Karl Hillig argue, on the basis of allozyme and morphology studies, for recognising a third species, C. ruderalis – a small-growing, wild form from Central Asia with a very low THC content, which plays a role as a crossing partner in the commercial autoflower segment. This position is, however, a minority view; the molecular data of the last ten years mostly support the monotypic view.

For practical purposes – for example in pharmacology or in the trade – the dispute hardly matters, because the chemovar classification works independently of formal taxonomy anyway.

9. Practical consequences

For consumers, patients and growers, the state of research yields a few rules of thumb.

What is informative is the lab certificate, not the label. For regulated products there is a Certificate of Analysis that documents THC and CBD content, the ratio of the two, the concentration of minor cannabinoids (CBG, CBN, THCV, CBC) and – for high-quality strains – the terpene profile. These figures are the actual information content of the strain; the word “sativa” or “indica” on the packaging carries only a weak, at best aromatic, additional statement.

Terpene profiles are the most useful compass in practice. Myrcene-dominant strains usually feel heavier and more sedating, limonene-dominant ones lighter and more mood-lifting, terpinolene-dominant ones fresher and more stimulating. These assignments are, however, probabilistic and interact with the individual biochemistry of the person consuming and with the dose.

In growing, genetics is not destiny. Growth habit, yield and active compound content are considerably influenced by light, temperature, nutrients, harvest time and drying. Two plants of the same strain from different growing situations can develop clearly different chemovar profiles.

Landraces are worth protecting culturally and genetically. McPartland and Small explicitly warn that the original landraces of Central and South Asia are threatened with extinction through introgression of commercial genetics. For breeder projects and genetic reserves, access to unhybridised landraces is an important concern.

10. Summary

The division of cannabis into sativa, indica and hybrid rests on a historical mixture of botanical observation (Linnaeus 1753, Lamarck 1785), breeder tradition (Schultes 1974, Anderson 1980) and marketing. The classic traits – growth habit, flowering time, geographical origin – are real and traceable in landraces, even if they are overlaid even there by natural introgressive hybridisation between neighbouring varieties. In modern commercial breeding the separation has largely dissolved: practically all available strains are genetic hybrids. Chemical and genetic analyses consistently show that the labels sativa, indica and hybrid do not reliably predict the actual active compound composition of a plant. A small core of reality remains in terpene synthase genes that encode the characteristic aroma – so the labels are not meaningless, but they are not a reliable classification of effects.

Classification by chemotype (THC or CBD dominance) and chemovar (complete cannabinoid and terpene profile) has become established as the scientifically sound alternative. The analytical gold standard for this is HPLC.

In practice: the binary division remains as a marketing shortcut and will persist for the foreseeable future; scientifically it is outdated. Anyone who really wants to assess the effect of a strain looks at the lab certificate – cannabinoids, dominant terpenes and overall potency.

Sources

Primary scientific literature

Historical primary literature (no stable online sources)

  • Linné, C. von (1753). Species Plantarum.
  • Lamarck, J.-B. (1785). Encyclopédie Méthodique, Botanique, volume 1.
  • Anderson, L. C. (1980). Leaf variation among Cannabis species from a controlled garden. Botanical Museum Leaflets, Harvard University, 28(1): 61–69.
  • Small, E. & Cronquist, A. (1976). A practical and natural taxonomy for Cannabis. Taxon 25(4): 405–435.
  • McPartland, J. M. & Guy, G. W. (2014). The evolution of Cannabis and coevolution with the cannabinoid receptor – a hypothesis. In: Pertwee, R. (ed.), Handbook of Cannabis. Oxford University Press.

Supplementary specialist and industry sources (explicitly marked)

  • Seedfinder database – strain genealogies and breeder information; not a peer-reviewed source, but the most comprehensive publicly accessible database of commercial cannabis strains.
  • Sensi Seeds – retrospectives and company information on the history of Dutch breeding; a primary source for breeder tradition, not scientifically validated.