On 19-12-2018 we have updated this article.×
Cannabis archaeology It can be difficult to obtain an accurate record of the evolution of a plant species. Cannabis plants do not possess bony skeletons, which are often the sole remnant of an ancient animal species, and are comprised of vegetative matter that decomposes rapidly. Despite this, traces left behind in fossils and soil can yield valuable clues.
Seeds, woody fibres and leaves may be preserved in fossils, if the process of mineralisation commences soon enough after death of the organism to bypass decomposition. These specimens have given us a good overall understanding of plant evolution in general, but there are plenty of gaps in the fossil record. When attempting to trace the lineage of one particular plant species, this can cause difficulties.
Pollen grains, on the other hand, are extremely hardy and may be preserved in various ways. From studying pollen in soil and sediment deposits, researchers have managed to glean abundant information on the early development of plants, including cannabis.
Cannabis in the Fossil Record
No fossils of Cannabis sativa itself have thus far been found, though fossil leaves (1) that are very similar to cannabis have been discovered in central Kazakhstan and dated to late Eocene epoch, around 38 million years ago. Fossils of another cannabis-like plant, cannabis oligocenica, have been discovered in Germany, and take their name from the Oligocene, the geological epoch that followed the Eocene. The Cannabaceae family is a member of the Rosales order, to which Moraceae (mulberry) also belongs.
The timing of the diversification of Cannabaceae is uncertain, however genomic research into the mulberry family (2) has indicated that Moraceae had diversified from its common ancestor by the mid-Cretaceous, between 89.1 – 103.4 MA. Cannabaceae comprises ten genera (3) including Cannabis, Humulus (hops), and Celtis (hackberry).
Of these, hops and cannabis are more closely related, and hackberry is more closely related to the remaining seven. These genera comprise around 180 species that are found across most the globe, from the tropics to the northern temperate regions. Most members are trees and shrubs, but the family also includes herbs such as cannabis and vines such as hops.
Cannabis Phylogenetics in More Detail
The Rosales order is a vast group of related plant families which also includes the Rosaceae family, which in turn includes apples, strawberries, and of course roses. The Rosaceae is the basal (oldest) family of the Rosales phylogenetic tree; as noted above, the Moraceae branch diverged from the common ancestor in the mid-Cretaceous, and it is believed that the Urticaceae (nettle family) branched off in a similar time period.
There has been some controversy over whether cannabis is more closely related to the Moraceae (mulberry) or to the Urticaceae (nettle). Convincing evidence that Cannabaceae and Urticaceae diverged no earlier than 34 MYA, when the nettle family first appeared, comes from a study of the parasites that affect each family. Cannabaceae shares seven parasites (4) in common with Urticaceae, but none with Moraceae, suggesting shared ancestry with the former.
Divergence of the Cannabaceae Family
There is much confusion as to the evolution of the various genera of the Cannabaceae family. It is thought that cannabis is among the more recent divergences, although the time of speciation of the Cannabis and Humulus genera is not clear.
However, research into Humulus indicates that was fully speciated by 6.38 MYA (5) (at this time, H. lupulus (common hop) and H. japonicas (wild or Japanese hop) diverged from their most recent common ancestor), so Cannabis and Humulus are likely to have diverged at some point between 34 and 6.38 MYA.
Evolution of the Cannabis Genus
To understand the more recent evolution of the cannabis plant it is necessary to understand how the earth’s climate has changed over recent geological time. The earth has been in the grip of an Ice Age for the last 2.6 million years. This was not a continuous encasement in ice; rather it was divided into extremely cold glacial periods (during which thick ice sheets marched south, bringing polar conditions to the temperate latitudes), and warmer interglacial periods, during which the earth’s climate was similar to that of today. The glacial periods each lasted many tens of thousands of years, while the warmer interglacial periods each lasted a few thousand years.
Indeed, we are currently in an interglacial, though it is considered highly unlikely that we will see a return to glacial conditions for at least fifty thousand years due to the high levels of greenhouse gases emitted since the industrial revolution (6).
During warmer interglacial periods, the plants and animals of the temperate regions would spread to higher latitudes, only to be pushed back to lower latitudes when glacial conditions returned. Some would manage to survive in isolated regions with more favourable microclimates, known as refugia. As evolution continued its work of natural selection on these isolated populations they began to diverge from one another, creating new species. When conditions once again became more favourable these populations could spread northwards to recolonise their former territory.
Modern humans appeared on the scene comparatively recently, within the last 300,000 years. They were initially confined to Africa, only beginning their spread across the rest of the globe within the last 100,000 years. The last interglacial before the present one was around 125,000 years ago – thus, modern humans have only experienced one glacial cycle since they left Africa.
As they colonised Asia and Europe the climate was steadily cooling. The ice sheets reached their maximum extent during the Last Glacial Maximum (LGM, around 18,000 years ago). Under the harsh conditions of the LGM, plants and animals – including modern humans – retreated to temperate refugia in the foothills of southern and south-eastern Europe and the temperate mountain valleys of southern Asia.
Clarke and Merlin consider it most likely that modern humans first encountered the cannabis plant on the steppe of Central Asia at least 35,000 years ago (1). It is not known what form this cannabis took. It is possible that ancestral cannabis speciated into cannabis indica and cannabis sativa during early glacial cycles, and had already gained distinct and characteristic traits by the time modern humans encountered them. It is, however, also possible that modern humans themselves played a role in this speciation event, carrying ancestral cannabis seeds with them to the refugia, where they altered them by selecting for desirable characteristics.
Regardless of both the timing of speciation and the extent of human involvement, it is considered likely that what we call cannabis sativa evolved in the European refugia, while cannabis indica evolved in the Asian refugia. There is also speculation that C. ruderalis (a cold-adapted, day-neutral biotype) survived the LGM in more northerly refugia(1).
Classifications & Controversies
Considerable confusion surrounds the classification of the genus cannabis. Does it comprise one species with three or more subspecies, or three (or more) species with their own respective subspecies and varieties? It is worth going back to the very beginning to lay out this muddled story of cannabis taxonomy(1).
The name cannabis sativa was first used in 16th century Germany by Leonhart Fuchs. Fuchs was describing cultivated hemp (sativa means ‘cultivated’). In the 18th century the Swedish botanist Carl Linnaeus, who gave us the binomial nomenclature used to describe genera and species, also described a cannabis sativa. Also in the 18th century, Jean-Baptiste Lamarck obtained some cannabis from India. The plant was sufficiently distinct from European hemp in terms of its appearance, aroma and efficacy as an inebriant that it merited description as a new species. He called it cannabis indica.
Soviet botanists made further contributions in the 1920s. In 1924 D. E. Janischevsky described a new variety, cannabis ruderalis, which he found growing wild in Central Asia (a ruderal plant is one that readily colonises disturbed ground). In 1929, Nicolai Vavilov made extensive field trips, during which he observed stocky Afghan cannabis plants that were distinct from European and other Asiatic varieties, and considered them cannabis indica. In the north of Afghanistan he made the surprising discovery of cannabis sativa of the European type, tall with narrow leaves, being cultivated for hashish production. In the east he found a wild variety which he named cannabis indica ssp. kafiristanica.
In the 1970s, the esteemed ethnobotanist Richard Evans Schultes went to Afghanistan, where he encountered short stocky cannabis plants. He stuck with Vavilov’s classification as cannabis indica. However, the Vavilov / Schultes indica bore no relation to the cannabis indica described by Lamarck two centuries earlier! The seeds of an almighty confusion had been sown.
This confusion has been amplified by the common use of the labels ‘indica’ and ‘sativa’ to describe the available cannabis strains. In this common usage an indica is a short, stocky, broad-leaved plant, generally of Afghan origin, which gives a narcotic body high. A sativa is a tall, gangly, narrow-leaved plant, generally of equatorial origin, that gives a psychedelic, cerebral high. In recent years it has become clear that this classification system is of little value in making meaningful correlations between the appearance of different types of cannabis and their effects or chemical compositions.
In ‘Cannabis: Ethnobotany and Evolution’, Robert C. Clarke and Mark Merlin stepped away from simplistic indica / sativa classifications and instead differentiated based on leaf width and psychoactivity, giving the classification of Narrow-leaf hemp (NLH, cannabis sativa ssp. sativa), Narrow-leaf drug (NLD cannabis indica ssp. indica), Broad-leaf hemp (BLH, cannabis indica ssp. chinensis) and Broad-leaf drug (BLD, cannabis indica ssp. afghanica). They also acknowledge Janischevsky’s cannabis ruderalis as the putative ancestor of both c. sativa and c. indica, and Vavilov’s cannabis indica ssp. kafiristanica as either an ancestor of NLD varieties, or, more likely, a feral NLD variety(1).
The finer details of taxonomy are often hotly disputed, and those pertaining to the cannabis plant are no exception. Ernest Small proposes that there is in fact only one species, cannabis sativa, which is divided into two subspecies, the non-psychoactive subspecies cannabis sativa ssp sativa and the psychoactive subspecies cannabis sativa ssp indica.
Under the single species scheme both cannabis sativa ssp sativa and cannabis sativa ssp indica are each classified into two varieties based on the size, shape and pigmentation of their seeds. Cannabis sativa ssp sativa is divided into cultivated or feral hemp (cannabis sativa ssp sativa var sativa) and its wild relative (cannabis sativa ssp sativa var spontanea). Cannabis sativa ssp indica is divided into cultivated drug cannabis, cannabis sativa ssp indica var indica and its wild relative cannabis sativa ssp indica var kafiristanica (7).
Interesting though these taxonomic disputes undoubtedly are, they are of little use to the consumer who wants to buy or grow cannabis that will provide a particular effect. Cannabis contains such a complex mixture of chemical compounds – cannabinoids, terpenoids and flavonoids – that it has been labeled ‘The Plant of the Thousand and One Molecules’ (8). Though the ‘indica’ and ‘sativa’ classification has some descriptive utility, at least in the inbred world of European and North American recreational cannabis, in reality there is a continuum between the two extremes (9), in part due to extensive hybridisation between ‘indica’ and ‘sativa’ strains.
It has been shown recently that the primary differentiator between the effect of ‘indica’ and ‘sativa’ strains may not be where the cannabis originated, or how thin its leaves are. It may in fact be determined by the amount of the sedative terpene myrcene that is present (10), along with the ratio and quantity of other terpenoids and cannabinoids expressed by the variety.
A consumer needs to know which chemical compounds are present in a particular variety, and at what levels. By analysing the chemical variety or “chemovar”, we are beginning to develop an approach that promises to at last provide a classification system for cannabis that everybody can agree upon (11).
Pollen Grain Analysis
Fossils are crucial to gaining insight into the ancient history and evolution of plants, but are less useful when it comes to establishing a more recent history. Due to the relative paucity of preserved plant tissue from more recent times, botanists are thus forced to search elsewhere for a timeline of plant evolution—and once detected, pollen grains can provide crucial evidence to paleobotanists.
For sure, the three families of Moraceae, Cannabaceae and Urticaceae have such similar pollen grains that identification can often be difficult. Within the family, distinguishing one type of pollen from another (e.g. separating Humulus from Cannabis) can be even more difficult. However, careful analysis has enabled researchers to differentiate between the various different species to obtain a clearer picture of the evolution and spread of cannabis.
Cannabis in the Age of Humans
Cannabis pollen grains have been various in various locations throughout Eurasia, typically in sediment found at the bottom of lakes, wells and ponds. As a wind-pollinated plant, cannabis sheds abundant pollen; in some sites, it accounts for over 50% of all pollen grains present. The earliest putative evidence of cannabis pollen dates to around 4,500 BP in China; in Europe, pollen sites indicating the existence of hemp cultivation become relatively abundant from around 3,000 BP onwards.
Scattered examples of ancient achenes (seeds) and fibres have also been found—even trichomes and carbonised cannabis remains are occasionally found; such specimens are rare, but become increasingly abundant subsequent to 5,000 BP.
Fibre and seed impressions left in pottery or clay can be preserved for even longer; in the present-day Czech Republic, impressions dated as early as 26,980 BP have been found, although these have not been confirmed as cannabis.
Fortunately for those studying ancient remains of cannabis, both the fibres and the trichomes are highly resistant to decay. The achenes are also fairly resistant to decay, and unlike pollen and fibre remains, these can be positively identified with little difficulty due to their characteristic shape.
There is substantial evidence that we have existed alongside and potentially utilised cannabis for as long as 50,000 years, or even longer. Understanding more about the evolution of cannabis and the role early humans played in its worldwide spread may yield invaluable clues as to our own development.
This article was updated with the contributions of independent scientist Dr Gavin Macfie, to ensure accuracy and academic rigour.
- Clarke R, Merlin M. Cannabis: Evolution and Ethnobotany. University of California Press; 2013:453.
- Zerega NJ, Clement WL, Datwyler SL, Weiblen GD. Biogeography and divergence times in the mulberry family (Moraceae). Mol Phylogenet Evol. 2005;37:402-416.
- Yang M-Q, van Velzen R, Bakker FT, Sattarian A, Li D-Z, Yi T-S. Molecular phylogenetics and character evolution of Cannabaceae. Taxon. 2013;62:473-485.
- McPartland JM, Nicholson J. Using parasite databases to identify potential nontarget hosts of biological control organisms. New Zealand Journal of Botany. 2003;41:699-706.
- Boutain J. On the origin of hops: genetic variability, phylogenetic relationships, and ecological plasticity of humulus (cannabaceae) [dissertation]. Manoa: University of Hawaii; 2014.
- Ganopolski A, Winkelmann R, Schellnhuber HJ. Critical insolation-CO2 relation for diagnosing past and future glacial inception. Nature. 2016;529:200-203.
- Small E. Evolution and Classification of Cannabis sativa (Marijuana, Hemp) in Relation to Human Utilization. Bot Rev. 2015;81:189-294.
- Andre CM, Hausman JF, Guerriero G. Cannabis sativa: The Plant of the Thousand and One Molecules. Front Plant Sci. 2016;7:1-17.
- Fischedick JE, S. Cannabinoids and Terpenes as Chemotaxonomic Markers in Cannabis. Nat Prod Chem Res. 2015;03
- Hazekamp A, Tejkalová K, Papadimitriou S. Cannabis: From Cultivar to Chemovar II—A Metabolomics Approach to Cannabis Classification. Cannabis and Cannabinoid Research. 2016;1:202-215.
- Piomelli D, Russo EB. The Cannabis sativa Versus Cannabis indica Debate: An Interview with Ethan Russo, MD. Cannabis Cannabinoid Res. 2016;1:44-46.