Printer Friendly

Evolution and classification of Cannabis sativa (marijuana, hemp) in relation to human utilization.


This review reports on recent agricultural, industrial and medicinal advances concerning Cannabis sativa, stressing how, to meet various utilitarian needs, humans have guided the evolution of the plant into a range of diverse domesticated kinds. These have been recognized as "land races," "cultivars," "strains" and "biotypes," which have been grouped taxonomically as species, subspecies and varieties. Two comprehensive categories of domesticated plants are evident, one kind selected for stem fiber and rarely usable for narcotic purposes and the other kind selected for narcotic content. Within the narcotic category, two sets of plants have been recognized, but the distinction between these has been obscured by extensive hybridization. In parallel, within the non-narcotic category, two sets of plants have been recognized, but the distinction between these has also been obscured by extensive hybridization. Further complicating the overall variation pattern, cultivated plants regularly escape to the wild, abandon the domesticated features that shackled them to servitude in cultivation, and establish colonies throughout the world that often interbreed by long-distance pollination with their domesticated relatives. This review first summarizes the history and ecology of C. sativa, then examines phenotypic and physiological aspects of the plant that have been moulded by human selection, and concludes with an examination of how to classify the confusing variation patterns that have been generated.

The subject of this review is how humans have domesticated Cannabis, causing it to evolve in divergent ways to supply different products. Virtually all major crops have undergone domestication, although the degree of divergence among the different kinds of cannabis plant is more extreme than in most other plants. "The Classification and Nomenclatural Issues" section provides background for the evolutionary nature of domestication, stressing that so-called "artificial selection" (selection by humans) is by nature quite comparable to natural selection, although classifying domesticated plants requires special considerations. There is an intriguing symbiotic aspect of the evolutionary relationship between Cannabis and humans: cultural evolution of cannabis use over millennia (i.e., how people have discovered new uses and created new technologies) is the cause of biological evolution of the plant, so that one can discern with exceptional insight how human preferences have resulted in morphological, anatomical, chemical and physiological transformations of the plant.

Cannabis sativa, best known as the source of marijuana, is probably the world's most recognizable, notorious and controversial plant. Because of its criminal association, almost all research and economic development--both narcotic and non-narcotic aspects--were suppressed for most of the 20th century. Most investigations authorized in Western countries were either forensic studies to aid law enforcement, or medical and social research specifically intended to document and reduce harmful effects. By the last decade of the 20th century, however, several developments contributed to a surge of scientific and technological development of C. sativa (reviewed in Small & Marcus, 2002; Small, 2007, 2014b). First, in many countries (with the conspicuous exception of the United States), after a half century of prohibition of cultivation, there was a resurrection of production of the plant for non-narcotic purposes. Second, non-narcotic hemp has acquired a reputation for being phenomenally beneficial for the environment, and has become a leading symbol of sustainable agriculture (Montford & Small, 1999a, b; Small, 2012). Third, there has been a substantial and increasing usage of marijuana prescribed for medical purposes. Fourth, in much of Western society there has been a growing tolerance of the extremely widespread recreational use of marijuana, as reflected by a romantic, idealized image in the media, less enthusiastic law enforcement, and even decriminalization in some jurisdictions. Although this article mentions recent potential medical applications, the intent is not to assess the physiological harm or benefit of marijuana, and indeed most countries (even those with provisions for usage of medical marijuana) have officially adopted the position that there are no legitimate medical benefits. The current sociological, philosophical, political and legal debates concerning cannabis drugs are also outside the boundaries of this review.

Until recently, the genera Cannabis and Humulus (best known for H. Iupulus, the hop plant) were considered to constitute the Cannabaceae (Small, 1978a). Recent phylogenetic studies have considerably expanded the family (Sytsma et al., 2002; Yang et al., 2013), but it is clear that Cannabis and Humulus are well separated from the eight or so other genera that are now included, and constitute a coherent phylad. Grudzinskaya (1988) added the fossil genus Humulopsis and split Humulus into two genera (although only Humulus is currently accepted). Humulus species are vines, and easily distinguished from Cannabis. However, the fruits (achenes) are very similar, and could be confused. Older texts commonly use the obsolete orthography Cannabinaceae and Cannabiaceae for the family (Miller, 1970). The Cannabaceae are closely related to the Urticaceae and Moraceae, and have sometimes been put in the latter families. A conclusion of this review is that only one species, C. sativa L., merits recognition, and the" Classification and Nomenclatural Issues" section of this contribution is concerned with classification issues and the various taxonomic groups of Cannabis that have been recognized to date.

The vernacular word "cannabis" has evolved as a generic abstraction from the genus name Cannabis, conventionally italicised. Non-italicised, cannabis is employed as a noun and adjective, and frequently (often loosely) used both for cannabis plants and/or any or all of the intoxicant preparations made from them. Cannabis sativa is usually called "hemp" when used as a source of fiber, "hempseed" when used as a source of seed oil, and "marijuana" (more commonly spelled "marihuana" in the past) when used for euphoric inebriants and therapeutic drugs. "Industrial hemp" refers to non-narcotic cultivars of the crop grown for fiber or oil, usually licensed for these purposes. The industrial hemp industry is making great efforts to point out that "hemp is not marijuana." Nevertheless, both names have been applied loosely to all forms of C. sativa. Although the term "hemp" mostly indicates C. sativa, it has also been applied to dozens of species representing more than 20 genera, often prominent fiber crops. For examples, Manila hemp (abaca) is Musa textilis, sisal hemp is Agave sisalina, and sunn hemp is Crotolaria juncea. Especially confusing is the phrase "Indian hemp," which has been used both for narcotic Asian varieties of C. sativa (so-called "C. indica Lamarck," a name alluding to the historical narcotic use in India) and Apocynum cannabinum, which was used by North American Indians as a fiber plant. Adding further to the confusion, "Indian hemp" is sometimes applied to jute (Corchorus capsularis), another fiber plant (Ash, 1948). Law enforcement personnel in the U.S. commonly call ruderal Cannabis plants (i.e., those growing as established weeds) "ditch weed" (a reflection of its weedy propensities and adaptation to moist soils as found in drainage channels). There are dozens of species with an epithet like cannabinus in the scientific name, indicative of similarity with C. sativa, but the resemblance is generally superficial (Small, 1975e).

Cannabis sativa is an annual plant, growing vegetatively in the early part of its life cycle, and induced to flower by photoperiod, the timing of induction being one of many adaptive features of the plant, discussed in this review (in the "Evolution of Photoperiodism Under Domestication" section). The plants are predominantly dioecious, with pistillate plants bearing only female flowers and staminate plants developing only male flowers (Figs. 1 and 2). Male (staminate) plants die after anthesis while female (pistillate) plants persist until frost. Female plants grown in a greenhouse or in climates lacking a cold winter can remain alive for years, although declining steadily in vigor. This potential longevity has led some to term the plants "annual or perennial depending on climate," but it is clear that the species is normally an annual. Sex expression has been remarkably manipulated in domesticated plants, and this topic is dealt with in the "Evolution of Sex Expression Under Domestication" section. The main stalk is erect, furrowed (especially when large), with a somewhat woody interior, and may be hollow in the internodes. Although the stem is more or less woody, the species is frequently referred to as a herb or forb. Plants vary enormously in height depending on environment and whether selected for fiber (the tallest kind), but are typically 1-5 m (heights of 12 m or more in cultivation have been claimed).

The leaf of Cannabis is probably more widely recognized than the foliage of any other plant. The leaves tend to be decussate on the lower stem (with opposite pairs, the succeeding pairs turned 180[degrees]), usually alternate near the stem apex, petiolate, palmately compound (except for small unifoliolate leaves at branch apices), with an odd number (3-13) of coarsely serrate, lanceolate leaflets. The foliage and stems of some populations are sometimes anthocyanin-streaked, and frost often causes plants to become suffused with purple; as discussed in the "Evolution of Color Under Domestication" section, this represents one of the kinds of coloration that has been preferentially selected.



Cannabis has been employed for numerous purposes, primarily fiber from the main stalk, narcotic drugs from the flowering parts (used mostly illicitly for recreation and more or less legally as medicinals), and oilseed (employed for human food, livestock feed, nutritional supplements, industrial oils, and occasionally as a biofuel). Historically, the same plants were often used simultaneously for different purposes. However, this review is particularly concerned with how humans have selected kinds of Cannabis that are especially productive for just one of the three commodities. The stem is an important source of bast (phloem) fiber, and the extensive modification of the anatomy of strains domesticated for fiber is discussed in detail in the "Evolution of Stem Fiber Production Under Domestication" section. Much of the above-ground parts of the plant are pubescent with stiff, pointed cystolithic trichomes, which are a mechanical defense against herbivores. By definition, such hairs contain a basal concentration of calcium carbonate, which presumably is unpleasant to chew, so protecting the plants from being eaten. Small, secretory, resin-producing glands are also present on the shoot epidermis. The unique chemicals (cannabinoids) in these glands have undergone considerable evolution under the hand of mankind, and are discussed in detail in the "Evolution of Narcotic Drug Production Under Domestication" section. Cannabis is also employed as a source of a multi-purpose fixed (i.e., non-volatile vegetable) oil in the achenes ("seeds"), and this additional dimension of variability caused by human selection will be examined in the "Evolution of Seed Oil Production Under Domestication" section.

Geography, Ecology and Ancient Domestication

Cannabis saliva is widely regarded as indigenous to temperate, western or central Asia. However, no precise area has been identified where the species occurred before it began its association with humans. De Candolle (1885) speculated that the ancestral area was the southern Caspian region, and other authors (e.g., Walter, 1938; Sharma, 1979) have suggested that the plant is native to Siberia, China or the Himalayas. Certainly, the plant is of Old World origin. For at least the last 6000 years, C. sativa has been transported widely, providing extensive opportunities for establishment outside of its original range (Abel, 1980; Clarke & Merlin, 2013). Because the species has been spread and modified by humans for millennia, there does not seem to be a reliable means of accurately determining its original geographical range, or even whether a plant collected in nature represents a primeval wild type or has been influenced by domestication (Schultes, 1970). The seeds of some wild-growing populations in India are remarkably small, unlike those collected from any other area of the Old World, but whether this is indicative of an ancient wild form is unclear. As discussed in this review, whatever ecological constraints once limited C. sativa to its ancestral home range, over the millennia it has become adapted to grow in much of the world.

Of the well over 100 informal or vulgar names that have been recorded for the marijuana form of C. sativa, "weed" is the most frequent, and accurately reflects the nature of the species. Cannabis growing outside of cultivation is indeed basically a weed, growing mostly in habitats created or modified by humans (Fig. 3). The extremely wide range of circumstances in which one finds weedy growth includes: the borders of fields, on rubbish heaps near settlements or habitations, in farmyards, waste places, vacant lots, in disturbed areas of pastures, in fallow fields (but not those that are sod-bound), along or beside roadsides, railways, ditches, creeks, fence rows, borders of cultivated fields, bridge embankments, lowland drainage tributaries and open woods (Haney & Bazzaz, 1970; Haney & Kutscheid, 1975). The species seems very poorly adapted to penetrating into established stands of perennial vegetation, and generally invades such areas only after the soil is freshly disturbed. As a colonizer, weedy hemp spreads slowly, except in drainage channels, a habitat to which it is very well adapted.

The circumstances and adaptations of extant wild-growing populations of C. sativa provide a basis for judging its ecology before human influence. The species thrives in mammalian-manured, continuously moist but well-drained soil, in open areas with limited competition from other plants. This suggests that ancestral C. sativa grew on the alluvial soils near streams and other water bodies, and depended on herds of wild, large, mammalian grazers to deposit excrement (Fig. 4).


Cannabis sativa is the most widely cited botanical example of a crop that is postulated to have evolved initially as a "camp follower" (Anderson, 1954; Schultes, 1970). Humans at the hunter-gatherer stage are thought to have been nomadic, often traveling among temporary camps, and creating trails among these. Abandoned campsites and paths would tend to be open (unshaded), located frequently near lakes or streams, and the soils would be enriched by deposition of organic materials (excrement and unused remains of harvested animals and plants). Seeds and roots from gathered plants that humans would have selected for their usefulness would also be deposited in these open, fertilized areas. This amounts to selective planting of desirable plants in protected situations where they will receive excellent light and soil--a precursor of cultivation. Inevitably, people would have noticed and eagerly harvested materials from the plants that were growing along their routes and former homesteads, especially in garbage dumps, and such plants would have been among the first that would have been considered for deliberate planting. As described by Anderson (1954), this explanation is variously known as the "rubbish heap" or "dump heap" hypothesis (in archaeology, rubbish heaps are referred to as "kitchen middens"). It is interesting that, in parallel, some monkeys have been shown to create "monkey gardens"--concentrations of preferred food plants in areas where they have discarded seeds (Rindos, 1984). Uncultivated, colonizing plants that grow vigorously in human-cleared areas are known as weeds. It is no accident that many, probably the majority of the world's major domesticated crops are related to, or are known to have originated from such plants. The ability to be weedy clearly pre-adapts plants to being domesticated. Cannabis sativa is superbly adapted for the role of camp follower. It is very weedy by nature. It is also a nitrophile, and would have grown exceptionally well in the manured soils around early settlements. Its propagules are thought to be distributed by streams, which as noted above are often near campsites, as well as by people and animals, including domesticates. Because Cannabis has products (stem fiber, edible seeds, intoxicating tissues) that could have been easily harvested and utilized by prehistoric peoples, it was almost certainly associated with humans in very early times (Fig. 5). Indeed, hemp may have been harvested by the Chinese 8500 years ago (Schultes & Hofmann, 1980), and has probably been grown for at least 6000 years, making it one of the world's oldest crops. For most of its history, C. sativa was most valued as a source of stem fiber, considerably less so as an intoxicant, and only to a very limited extent as an oilseed crop. Hemp is one of the oldest sources of textile fibers, with extant remains of hempen cloth trailing back 6 millennia. Hemp grown for fiber was introduced to western Asia and Egypt, and subsequently to Europe somewhere between 1000 and 2000 BC. Cultivation in Europe became widespread after 500 AD. A superb documentation of historical usage and cultural diffusion of Cannabis is provided by Clarke and Merlin (2013).


For most plants, nitrogen is the most critical limiting nutritional element, and most wild plants are adapted to substrates in which nitrogen is in short supply. Most annual domesticated crops, however, have been bred to be nitrophiles, with the capacity to utilize large amounts of nitrogen for productive growth (Emerich & Krishnan, 2009).


Modern agriculture in fact is to a considerable degree based on the creation of crops that can utilize nitrogen fertilizers. The "Green Revolution" of the middle of the last century greatly increased agriculture production, especially in the Developing World, by selecting new cultivars that are especially capable and efficient at using nitrogen fertilizers (Borlaug, 2000). Wild C. sativa is a natural nitrophile, thriving in well-manured substrates, and stripping soils readily of nitrogen. Notes accompanying herbarium specimen collections of the species commonly mention the presence of nearby manure. Vavilov (1926) observed that wild hemp in Russia thrives in low places and ravines into which wild animal excrement is washed, and on soils manured by grazing cattle. Manure not only supplies nutrients, but the humus is important in retaining moisture that hemp demands (Dewey, 1914). Weedy hemp in the U.S. has been collected on sandy soils very low in nitrogen, but the plants are dwarfed (Haney & Bazzaz, 1970). Cultivars are typically fertilized with nitrogen at a rate of 100 kg/ha/ season (Bocsa & Karus, 1998), which is higher than the recommended rates for some modern high-yielding field crops.

Several terms are used to denote plants of different degrees of "wildness" growing outside of cultivation, and it is critical to be aware of their ambiguity in discussing C. sativa as a weed. Plants that develop as a result of seeds unintentionally scattered from cultivated plants are said to be "volunteers," a label used in agriculture. For the most part volunteers appear on or very near the field where the maternal plants were grown. The word "spontaneous" is used in floristics to denote plants that appear locally as a result of human activities, but do not spread. Such plants can be domesticates (e.g., tomatoes growing only on refuse heaps where tomato seeds were discarded; cereals growing only near mills where the seeds were processed), or wild (e.g., seeds of foreign plants transported in ship ballast and appearing only where the ballast has been discarded). The term ruderal (applied both to plants and their habitat) means growing in waste places or rubbish, and is descriptive of the habitat of perhaps the majority of weeds. One also encounters "feral" applied to hemp (and other weeds), although mostly the word is used for escaped domesticated animals (such as dogs and horses) that are living outside of human control. Both the words feral and ruderal are ambiguous, since they are applied to a) those escaped domesticates that basically retain all of their domesticated charateristics but nevertheless establish and spread vigorously outside of cultivation, and to b) types of plants that differ dramatically from domesticates, with adaptations specifically suited to wild existence. The term "wild" is also ambiguous. It has been used in a narrow sense to refer to populations of a species that are essentially uninfluenced genetically by domestication, and in a broad sense to include all populations growing outside of cultivation. The distinctions discussed in this paragraph are examined additionally in the "Classification and Nomenclatural Issues" section.

Weedy hemp is particularly widespread in southeast and central Asia, common in many European countries and, less frequent in South America, Australia and Africa (Davidyan, 1972). According to Haney and Kutscheid (1975), C. sativa seldom becomes naturalized as a result of escapes from cultivated hemp in subtropical and tropical areas. In North America, the species is best established in the American Midwest and Northeast, and in southern Ontario and southern Quebec, all areas where hemp cultivation was concentrated historically in recent centuries. Cannabis sativa has been collected growing outside of cultivation from Canadian provinces from British Columbia to New Brunswick (Small, 1972b; Small et al., 2003). Naturalized hemp is uncommon in the western U.S., rare in the U.S. south of 37[degrees]N latitude, and very rare in Mexico (Haney & Kutscheid, 1975). In most of the world, wild-growing Cannabis is of limited concern, but there have been long-continued efforts by law-enforcement to eradicate ruderal plants in North America. In contrast to the huge social costs, the deleterious effects of Cannabis as a weed in North America are relatively minor. Discovery of extensive growth of ruderal hemp on a farm often invites unwelcome attention, from the legal authorities as well as from delinquents who mistakenly believe that ruderal hemp in North America is as intoxicating as high-quality marijuana. As an agricultural weed, however, ruderal hemp is of limited importance (Small et al., 2003).

Reflective of its extensive geographical distribution, ruderal C. sativa occurs in a wide range of climates. Domesticated forms of the plant have narrower tolerances than the wild-growing counterparts (Small et al., 2003). Both domesticated and wild plants of Cannabis sativa develop best in full sun, and weedy plants thrive in open areas. However, some wild plants have been observed growing well in shaded habitats in Europe (Janischevsky, 1924) and Canada (Small et al., 2003). Both domesticated and wild plants of C. sativa are tolerant of hot, arid conditions provided that the roots are adequately supplied with water, but ruderal plants in Europe have been observed to be much more drought resistant than cultivars (Janischevsky, 1924). In North America, Haney and Bazzaz (1970) noted that wild hemp in sandy soils in Illinois survives dry conditions in deep, loose-textured soils by virtue of the roots growing to gain access to deep water sources. Cannabis sativa does not tolerate cold temperatures well, but once again the weedy forms are more stress-tolerant; in northern areas, the seeds germinate at lower temperatures and the seedlings survive frost better than do cultivars (Haney & Kutscheid, 1975). Compared to most fiber cultivars (which tend to have hollow stems), wild plants are also relatively wind resistant, due to low stature and woodier, flexible stems. Wild plants in the Old World have adapted to various habitats for thousands of years, while those in North America have a history of only a few hundred years. Not surprisingly, in Eurasia the species grows wild over an enormous range of climates and altitudes, much greater than in North America. Vavilov (1926) observed vast stands of wild hemp in Eurasia. In the Himalayas, C. saliva occurs at altitudes of thousands of meters.

Humans, animals, water, and insects have been proposed as disseminating agents for wild hemp. Since wild C. sativa is dioecious, the most effective dispersal agents should distribute at least a seed of each sex to a given site, although pollen is distributed so widely that even isolated plants may participate in reproduction. Since birds are strongly attracted to the seeds, Haney and Bazzaz (1970) suggested they are likely the most important wild animals distributing them in North America. Virtually no wild hemp seeds fed to upland game birds (quail and doves) survived (Small et al., 2003), but it is possible that some seeds are transmitted by adhesion to claws or bills (Merlin, 1972). Weedy hemp in North America is often found in alluvial sites disturbed by flooding, and flood waters may serve to distribute the seeds (Haney & Bazzaz, 1970). Ruderal hemp clearly depends heavily on human activities for dispersal. Because large wild herds of mammalian grazers probably were important to providing manured habitats for Cannabis, and the species characteristically grows in moist areas, the mammals may have distributed seeds caught up in mud on their hooves. In more recent times, domestic livestock may similarly serve as distribution vectors. Seed weight in C. sativa varies enormously, from more than 1000 seeds to the gram in some wild Asian plants to less than 15 seeds to the gram for some cultivated plants (Vavilov, 1926; Watson & Clarke, 1997). The ecology of the species may differ considerably according to the size of the seeds, and this remains to be studied.

Both wild and cultivation plants that grow for many generations in a particular location tend to evolve adaptations to their local climates, and these adaptations may make a given biotype quite unsuitable for a foreign location. In the "Evolution of Narcotic Drug Production Under Domestication" and "Classification and Nomenclatural Issues" sections, the narcotic "Group 4" (so-called "indica type") is discussed. This is established in the arid area of Afghanistan and western Turkmenistan, and when strains from this region are grown in high-humidity climates their dense flowering tops retain moisture and succumb to "bud mold" caused by Botrytis cinerea and Trichothecium roseum (McPartland et al., 2000). In the "Evolution of Photoperiodism Under Domestication" section, photoperiodic adaptation to latitude is discussed, and it is pointed out that when strains adapted to the season of one area are grown in an unsuitable latitude they may fail to develop seeds.

Phenotypic Plasticity: a Key to Success

Phenotypic plasticity is "the ability of individual genotypes to alter their growth and development in response to changes in environmental factors" (Barrett, 1982). It is flexibility of response, and allows a population to survive in a broad range of environments, especially marginal conditions. It is a key component of the genetic system of weeds (Bradshaw, 1965; Baker, 1974), and is often critical to the ability of species to diversify and adapt in response to natural and human-caused selection (West-Eberhard, 2003). Most of the ecological adaptations of Cannabis discussed in the previous section contribute to its exceptional adaptive phenotypic plasticity. Some particular aspects related to this topic are dealt with in the following.

Nature vs. Nurture in the Determination of Characteristics

In the early 20th century, a sort of Lamarckian conception of semi-permanent induction of characteristics by the environment was sometimes applied to explain why C. sativa strains suited for fiber production in temperate climates, when transplanted to hot, dry climates, would apparently transform into narcotic cultivars, and vice-versa (for a proponent of this viewpoint, see Bouquet, 1950). As explained in the "Classification and Nomenclatural Issues" section, maintaining the purity of a strain of Cannabis requires stabilizing selection and protection from contaminating pollen, and the absence of these probably accounts for observations that Cannabis grown in a foreign location seemed to transform remarkably in a few generations. It is clear that although environment does influence the development of the characteristics of Cannabis, indeed of all organisms, strains selected for fiber or narcotic characteristics retain their capacities for such production so long as their gene frequencies are maintained. Of course, the ability of Cannabis to change genetically as a result of hybridization and selection should not be confused with the concept of phenotypic plasticity.

Surviving Soil Infertility

Like many weeds, C. sativa is very plastic in a range of edaphic conditions, responding with dramatically increased growth to a good supply of soil nutrients, but able to produce dwarfed plants in very infertile conditions and still produce a few seeds. Given that the species is a nitrophile, it has a remarkable ability to survive in soils deficient in this element.

Root Flexibility in Relation to Ground Water Level

Cannabis develops a laterally branched taproot. The root system provides another example of flexible response in relation to environmental circumstances. Haney and Bazzaz (1970) noted that wild hemp in sandy soils in Illinois seemed able to tolerate dry conditions because the roots penetrated to deep water sources. In coarse textured, well-drained soils the primary root of wild hemp can extend more than 2 m down, allowing access to a low water table. In medium-textured, moderately water-retentive soils the primary root develops to a depth of about 1 m, with extensive laterals concentrated in two locations: near the surface and at about 1 m, a bet-hedging strategy enabling acquisition of both surface and moderately deep water. If the water table is near the surface (generally undesirable for good growth of C. sativa), the root system is shallow.

Resistance to Catastrophic Stem Damage

Cannabis sativa normally has a dominant leader stem which produces a central stalk. As discussed in the following paragraphs, the species has an amazing capacity to recover from catastrophic damage to the main stem.

The European com borer (Ostrinia nubilalis) or ECB (Fig. 6a), is a major Lepidopteran pest of C. sativa. Young ECB larvae eat hemp leaves until half-grown, then bore holes into the stems. A typical entrance hole resulting from an attack on the main stem is shown in Fig. 6b. The insect is indigenous to the Old World, where it apparently once reproduced mainly in association with Cannabis and its close relative Humulus (although also attacking many other plant species). It was not exposed to com (i.e., maize, Zea mays), which is indigenous to the Americas, until post-Columbian times ("European hemp borer" would have been a better choice of name). In a study of ECB infestation of a large experimental field, Small et al. (2007) discovered that ECB damage to Cannabis increased the shoot weight of the plant by 20 %, concomitantly enlarging seed production, indicating that Cannabis is adapted to the insect. The expanded productivity observed was due to branch proliferation at the site of the attack (see Fig. 6c and d). Figure 6e shows silhouettes of a normal and an ECB-damaged plant, and it is evident that the increased number of branches resulting from the damage has produced more biomass and more seeds. (The insect preferred larger stems, but was unaffected by THC content.)


There is controversy whether insect damage may, at least in a limited sense, be good for plant productivity. McNaughton's (1983) classic paper in this regard proposed that in some circumstances plants can respond to herbivory by just growing faster ("compensation" or "overcompensation"). Verkaar (1986) surveyed papers purporting to support the hypothesis that grazing can have positive effects on plant growth and fitness, and concluded that "the hypothesis may only be tenable under very particular circumstances." Additional literature on the topic is reviewed in Small et al. (2007).

Horticulturally, it is well known that destroying leader buds to induce proliferation of flowers or fruits in a range of plants can increase productivity, so it is logical that insects that carry out this activity might also be beneficial to crop production. Moreover, humans have engaged in the practice of damaging stems to increase productivity of Cannabis. Pate (1998b) noted that when growing hemp for seed, the number of flowers per plant and the number of seeds produced can be increased by "topping" the plants when 30 to 50 cm high. Dewey (1902) observed that hemp grown in North America at the turn of the century was sometimes topped to make it spread and produce more seed. Clandestine growers of narcotic strains also sometimes remove the tops of their plants to produce more of the desired high-TFIC inflorescences, the "buds" (see the "Evolution of Narcotic Drug Production Under Domestication" section).

Evolution of Sex Expression Under Domestication

Sexual selection is often recognized as a special kind of natural selection (Darwin, 1859). It involves competition within a gender for the opposite sex, and is important in evolution. In nature, males often are especially important in sexual selection. Human selection of the sexual characteristics of domesticated species is also a powerful evolutionary force but, by contrast, the males of domesticates have lost much of their importance. Farmers often favor females of livestock (bulls are much harder to manage than cows, do not produce milk or calves, and only a limited number are needed for reproduction). As discussed in the following paragraphs, male Cannabis plants have also suffered significantly under domestication: (a) humans have created many cultivars that are monoecious (the plants bearing both male and female flowers), but a preponderance of female flowers has been favored; (b) cultivars have been created by hybridization that are entirely female; (c) for narcotics production, male plants are usually eliminated; (d) clones maintained for narcotics production are female. A curious aspect of sexual evolution of Cannabis under domestication has to do with the fact that humans have turned a normally dioecious species into forms that are monoecious. This constitutes reversing the normal pattern thought to exist in nature that dioecious species have evolved from monoecious ones (Lewis, 1942).

The wild plants of C. sativa are among the small minority (4 % according to Yampolsky and Yampolsky (1922), 6 % according to Renner and Ricklefs (1995), or some undetermined higher figure according to Bawa (1980)), of flowering plants with male reproductive organs (stamens) and female reproductive organs (carpels) confined to separate plants (i.e., the populations are dioecious, with unisexual flowers, those on a given plant either entirely male or entirely female). Staminate plants, with male flowers only, are routinely called males, and pistillate (carpellate) plants, with female flowers only, are called females, and this standard terminology (albeit technically incorrect, since the sporophytic phase of plants is asexual) is followed here.

Floral primordia are normally initiated in mid-summer, with development proceeding from the base upwards to the top of the inflorescence. The flowers of Cannabis are small but very numerous. The staminate inflorescences are large, showy, loose, axillary, cymose panicles (thyrses), while the pistillate ones are small, obscure, congested, axillary, spicate cymes. Male flowers are pedicellate, with five greenish or whitish tepals and five stamens with flaccid filaments opposite the tepals. The male flowers fall away after anthesis. The female flowers consist of a superior, unilocular ovary and a short apical style with two long filiform stigmatic branches. Unlike the male flowers, the female flowers are essentially sessile. A perigonal bract (sometimes called a floral bract) subtends each female flower, and grows to envelop the fruit (this is important in narcotic resin production, and additional detail is given in the "Evolution of Narcotic Drug Production Under Domestication" section). In contrast to the male flowers, the female perianth is not at all recognizable as tepals, consisting of a thin undivided layer adhering to the ovary (this unusual anatomical feature is very important ecologically as discussed in the "Evolution of Propagules Under Domestication" section, and for classification purposes, as discussed in the "Classification and Nomenclatural Issues" section).

The sexes are dimorphic not just with respect to reproductive organs: male plants tend to be 10-15 % taller, although less robust than the female plants, with slimmer stems, less branching, smaller leaves, and a more delicate appearance, and they die after shedding their pollen. Female plants protected from frost can remain alive for years (gradually losing vitality), although the species is normally an annual. However, cloned biotypes of female marijuana plants are often regenerated for many years by repeated cuttings, which does maintain plant vigor. Before sex in plants was widely understood, many 18th century European botanists (males at the time, reflecting their perception of masculine superiority) often referred to the vigorous females as males, and the wimpy males as females (Bouquet, 1950). (However, by that period some European botanists appreciated the true nature of sex in flowering plants (Anonymous, 1933).)

The sole purpose of the males is to produce pollen, and they excel at this task: a single flower can produce about 350,000 pollen grains (Faegri et al., 1989), and there are hundreds of flowers on larger plants. The stigma is densely covered with receptive trichomes to receive pollen. Cannabis is wind-pollinated, and the pollen can be blown long distances. It has been claimed that crossing has occurred at a span of over 300 km (Clarke, 1977). Cabezudo et al. (1997) noted that C. saliva pollen, apparently from marijuana cultivated in North Africa, was transported by wind currents to southwestern Europe. Hemp pollen is a significant allergen for some people (Lindemayr & Jager, 1980), so its presence is often monitored. Stokes et al. (2000) recorded that in August in the Midwestern United States (where cultivation of hemp is not permitted, but weedy hemp is common) hemp pollen represented up to 36 % of total airborne pollen counts! Because the pollen of Cannabis spreads remarkably, an isolation distance of about 5 km is usually recommended for generating pure-bred seed, exceeding the distance for virtually every other crop (Small & Antle, 2003). Because of widespread clandestine cultivation, the pollen can be found, at least in small concentrations, over most of the planet. While the inverse square law dictates that the probability of pollen distribution decreases rapidly with distance, it is likely that there is frequent genetic interchange among populations.

Although self-fertilization is possible in C. saliva, inbreeding depression is pronounced (conversely, so is hybrid vigor). To promote outcrossing, male plants of a given population tend to come into flower 1-3 weeks before female plants have receptive stigmas. Male flowers at anthesis are very attractive to bees, including bumble bees and honey bees, which collect substantial amounts of pollen. Pollen-collecting flies are also often present. However, these insects do not visit the female flowers and so do not play a role in pollination.

Inheritance of sexual expression in Cannabis has been studied extensively (Hoffmann, 1970). Sexual differentiation in dioecious strains is based on a pair of sex chromosomes, the male being heterogametic (XY, the Y chromosome allegedly larger (Sakamoto et al., 1998), unlike mammals, but like some other plants), producing an approximately 50:50 sex ratio. However, sex expression appears to be somewhat determined autosomally, with an X/autosome dosage type chromosome system (Ainsworth, 2000). Sex development is labile, modifiable by a wide range of environmental factors and hormonal treatments (Heslop-Harrison & Heslop-Harrison, 1969). The application of auxins or ethylene feminizes Cannabis (Heslop-Harrison, 1956; Mohan Ram & Jaiswal, 1970), whereas gibberellins are masculinizing (Atal, 1959; Chailakhan, 1979). The proportion of female plants has been reported to be increased after exposure of seeds to ultraviolet light, and decreased by shorter day-length during the growing season, and higher nitrogen concentrations in the soil (see Haney & Kutscheid, 1975 for references). Such factors can result in sex reversal, and indeed the aberrant production of plants with male, female, and intergradient flowers. In a survey of over 1400 U.S. herbarium specimens, 55 % were male, but only 41 % of the plants collected along streets and highways were male; Haney and Bazzaz (1970) speculated that this could be due to the higher carbon monoxide levels near roadways. This is intriguing as carbon monoxide has been shown to favor the development of female flowers (Heslop-Harrison & Heslop-Harrison, 1957).

There have been numerous studies of male-associated and female-associated DNA markers (e.g., Mandolino et al., 1999, 2002; Sakamoto et al., 2000; Flachowsky et al., 2001; Peil et al., 2003; Shao et al., 2003; Cristiana Molitemi et al., 2004; Rode et al., 2005; Sakamoto et al., 2005).

Many cultivars, especially those selected for stem fiber production, are monoecious (with both male flowers and female flowers, and often with sexually intergradient flowers, on the same plants), or at least substantially so (i.e., some plants may also be entirely or mostly male, some may also be entirely or mostly female). In monoecious forms, staminate flowers, if present (frequently on the upper part of flower-bearing stems) are produced before the pistillate flowers (frequently on the lower parts of stems); staminate flowers, if present, are also produced before transitional hermaphroditic flowers (some of which are sometimes sterile), which are also often encountered. In some populations, one finds plants that are 100 % male, 100 % female, and a spectrum of plants with intermediate sexuality (a population structure that has been termed "subdioecy"). While male plants almost always die after shedding pollen, the presence of even a few female flowers on hermaphroditic plants seems to protect them against dying before seed set (personal observation). However, in a plantation setting there is a much reduced need for the prodigious pollen production that is normal in the wild plants, so hermaphroditic plants tend to be bred that are predominantly female.

Recently escaped plants are occasionally monoecious, but monoecy is associated with inbreeding depression, and is therefore very rare in wild C. saliva, which is naturally strongly outcrossing (Heslop-Harrison & Heslop-Harrison, 1969). Monoecy is also associated with smaller, less vigorous pollen grains. Migalj (1969) found that the acetolyzed pollen grains of dioecious strains tended to have a diameter averaging about 33 pm, while the grains of monoecious strains were smaller, with a diameter averaging about 27 pm; and the pollen of dioecious plants was also more uniform, while that of monoecious plants were more variable in size and in number of pores. Zhatov (1983) reported that pollen viability in monoecious strains tends to be lower than in dioecious strains.

Some artificial hybrids obtained by pollinating females of dioecious lines with pollen from monoecious plants are predominantly female (so-called "all-female," these generally also produce some hermaphrodites and occasional males). All-female lines are productive for some purposes (e.g., they are very uniform, and with very few males to take up space they can produce considerable grain), but the hybrid seed is expensive to produce. So-called "feminized" seeds are often offered in the marijuana trade, these producing plants with female flowers only (as noted below, only female plants are normally used for narcotics production).

For production of narcotic resin, male plants are eliminated before they can shed pollen to fertilize the females, as unfertilized female inflorescences are highly valued (see the "Evolution of Narcotic Drug Production Under Domestication" section). Female narcotic plants have as much as 20 times the concentration of THC as corresponding males (Clarke & Merlin, 2013). By contrast, male fiber plants, although also less productive than corresponding females, produce a higher quality of fiber, and before the 20th century were often harvested separately by hand, when labor was cheap. Today, males are considered undesirable for fiber, because they senesce earlier and degenerate, thus decreasing the overall quality of fiber harvested. In former, labor-intensive times when the plants were hand-harvested separately, selection pressures were probably more or less equal for the sexes, or perhaps there was some preference for male plants. Monoecious varieties are commonly utilized today for fiber, so that all plants mature simultaneously and their quality is uniform. For production of oilseed, dioecious varieties are frequently employed, although at present there are very few varieties exclusively used for oilseed production. Several "dual-purpose" varieties are grown for simultaneous production of fiber and oilseed, and these may be monoecious or dioecious. Because female plants are more valued for oilseed and narcotics, selection has been much more directed to the females than the males.

Humans propagate many crops vegetatively (e.g., apples, potatoes, strawberries) as clones, a tactic to avoid the variability produced by sexual reproduction, in order to maintain a uniform genotype that is especially desirable. This is the method increasingly being used to propagate (female) strains of narcotic Cannabis, particularly the most desirable biotypes (Chandra et al., 2010b). In perhaps an ultimate departure from normal plant sexual reproduction, propagules of narcotic strains, generated by tissue culture, have been encapsulated to form "synthetic" or "artificial" seeds (Chandra et al., 2010a; Lata et al., 2011).

Evolution of Propagules Under Domestication

In nature, plants reproduce mainly by distributing propagules, mostly seeds and fruits (occasionally vegetative tissues), commonly by wind, water, gravity, and cooperating wild animals. Humans have domesticated many wild plants, frequently specifically to harvest the seeds or fruits. Many wild plants cast off their seeds or fruits as soon as they mature, by various mechanisms. This has two undesirable consequences from the human perspective: when a seed or fruit drops away it is more difficult to collect; and when seeds or fruits do not remain attached to the plant at maturity, it necessitates repeated collection of propagules from each plant over the weeks that they sequentially mature. Selecting mutations that inactivate the separation mechanisms (abscission of fruits, dehiscence of fruits to release seeds) so that the mature seeds or fruits remain on the plant greatly facilitates harvest. This is the most important way that humans have domesticated the majority of crops (Harlan, 1995; Fuller & Allaby, 2009). Cereals currently supply more than half of the calories consumed by humans (Small, 2009), and in all of them a "domesticated syndrome" of characteristics is recognizable whereby the edible fruits (caryopses) have lost the features in their wild ancestors that cause the grains to detach and scatter away (see, for example, Sakuma et al., 2011). Although the precise anatomical and morphological changes that keep cereal grains attached differ between domesticated cereals and domesticated C. sativa, one can recognize a comparable domesticated syndrome of propagule characteristics in all strains of fiber hemp, oilseed hemp and narcotic hemp.

Cannabis plants domesticated for fiber, oilseed, or narcotics tend to differ from plants adapted to wild (ruderal) existence, most characteristically in the achenes (Small, 1975a). For many crops and their wild progenitors, propagule characters are an excellent index or gauge of the relative state of domestication, and this is the case in Cannabis (Small, 1975a). In contrast to the achenes of domesticated forms of Cannabis, wild achenes are smaller (generally less than 3.8 mm long), disarticulate more readily (facilitated by an attenuated base), are covered by a tightly adhering camouflagic mottled layer (homologous with the perianth), have relatively thick pericarp walls, are relatively long-lived, and do not all germinate more or less simultaneously (Vavilov, 1926; Small, 1975a). By examining the relative development of these achene features, one can often evaluate whether a Cannabis plant is merely recently escaped from cultivation or derived from plants that have lived in the wild for a considerable period and consequently evolved wild characteristics. The morphological differences between the achenes of wild and domesticated C. sativa are shown in Fig. 7.

The anatomy, morphology and germination behavior of the achenes is key to the survival of wild hemp. The attenuated base and well-developed abscission zone of the wild achenes facilitate disarticulation as soon as the fruits are ripe, and this minimizes the period that they are available for predation by birds. Additionally, the camouflagic mottled layer covering the achenes of wild plants keeps the fallen ones hidden from mammalian and insect herbivores. Janischevsky (1924), working on the ecology of ruderal Russian hemp, noted that birds are very infrequently seen on the ground in pursuit of fallen seeds. By contrast, the achenes of domesticated plants mostly remain on the plant, and birds perch on the infructescences gorging on the seeds. In contrast to the thin wall of domesticated achenes, the comparatively thick wall of wild achenes provides mechanical protection. Additionally, apparently because of the presence of a water-soluble inhibitor (Small et al., 2003), achenes of wild races remain dormant in the soil at least until the spring, and germinate irregularly for several years, providing protection against the entire population being subjected to a catastrophe (Scholz, 1957; Haney & Kutscheid, 1975; Small & Brookes, 2012).


In contrast to the adaptive characters of the achenes of wild Cannabis, the features facilitating disarticulation have been greatly weakened in domesticated forms. The achenes tend to remain on the plant for easy harvest, the development of a thick pericarp to protect the seed has been lessened, the camouflagic perianth attached to the pericarp tends to slough off since it is no longer needed, larger seeds have been selected to give the seedlings a better start, and dormancy has been eliminated so that the seeds germinate immediately and produce a dependably uniform crop.

Janischevsky (1924) alleged that he had discovered a symbiotic relationship between wild hemp and the red bug Pyrrhocoris apterus. He observed it apparently feeding on the attenuated base (the attachment area) of the achene, and concluded that the base was an elaiosome, i.e., a fleshy edible appendage of the achene serving to attract dispersal vectors, constituting an adaptation for distribution of the seeds. However, the insect is a generalized feeder that has no fidelity to Cannabis, and the base of wild achenes do not develop a genuine elaiosome, although the detachment zone is a weak area of the protective pericarp, and might offer some limited nutrition to insects.

Evolution of Photoperiodism Under Domestication

Photoperiodism is a physiological reaction of organisms to length of day or night. In the following discussion the term is used with specific reference to induction of flowering. Toumois (1912) is credited with the first discovery of photoperiodism in plants (see Jarillo et al. (2008) for a review of the subject). Based on studies of hemp and its relative Japanese hop, Toumois observed that flowering was promoted by short daylight (giving rise to the expression short-day plants) and delayed by long days. Today, Cannabis has been evaluated to be a quantitative (facultative) short-day plant that is, flowering is normally induced by a required duration of days with a minimum uninterrupted period of darkness (10-12 h for most cultivars), but at least in some cases flowering may occur regardless of light regime. Some strains of Cannabis can produce flower buds under continuous illumination (Borthwick & Scully, 1954; Heslop-Harrison & Heslop-Harrison, 1969); however, before these open, some cultivars require short days, while others will flower in continuous light, but only after a long period of growth (Schaffner, 1926; Borthwick & Scully, 1954; Heslop-Harrison & Heslop-Harrison, 1969). The critical daylength may be longer for male plants than for female plants in a given population, which is consistent with the fact that males normally come into flower faster (Borthwick & Scully, 1954). Flowering is induced in Cannabis mainly by shortening daylight hours in late summer, but also to some extent by intrinsic, genetic factors. However, environmental stresses also have some effect on flowering time, especially drought, which is the most important factor in speeding up maturation. As noted below, hybridization may also play a role in inducing flowering.

Latitudinal Photoperiodic Adaptation

A world-wide, north-south pattern of clinal (geographically-graduated and genetically fixed) photoperiodic adaptation correlated with stature has evolved in Cannabis. Bergmann's Rule states that within a taxonomic group of animals, individuals are larger in colder environments (an ecogeographic generalization with mixed validity).

For plants, the reverse is often the case: the shorter, colder season at higher latitudes (or altitudes) limits growth and accordingly stature. Annual plants like Cannabis are designed to maximize propagule production, achieved in part by growing as large as possible within the limitations of the length of their season and the cultural conditions of their growth sites. It seems clear that the historical migration of Cannabis throughout much of the world for purposes of cultivation was accompanied by strong selection for local photoperiodic regime. During domestication, some populations could have been selected for photoperiodic insensitivity (like some cultivars of strawberry and other crops), but this has not been important for Cannabis. Wild plants and cultivars are photoperiodically adapted to their local climate; plants adapted to growth in northern areas tend to come into flower readily with shortening days, allowing time for seeds to mature before a killing frost; and conversely plants adapted to areas closer to the equator tend to come into flower slowly with shortening days, in order to grow for a longer period in the milder environment. Russian (U.S.S.R.) agronomists classified hemp into four eco-geographical maturation groups, respectively adapted to a longer season: Northern, Middle-Russian, Southern, and Far Eastern (Serebriakova & Sizov, 1940; Davidyan, 1972), and noted that races of Cannabis are available to meet the local photoperiodic requirements of most regions of the country.

When plants adapted to the photoperiod of semi-tropical climates are grown in north-temperate climates, they may mature so late that they succumb to cold weather before they can produce seeds (Heslop-Harrison & Heslop-Harrison, 1969). Such photoperiodic differences are apparent when Cannabis populations obtained from different latitudes are grown together in a northern experimental garden. In Ottawa, Canada, where I have grown over 1,000 accessions outdoors, those from the northernmost locations (Siberia) sometimes produced seeds in less than a month after planting, while some from near-equatorial locations (India, Africa) sometimes remained vegetative after 5 months (and were killed by frost). When hemp cultivation was authorized in Canada in 1998 (after more than a half century of prohibition), the only source of cultivars with reliably low TFIC (a requirement) was the European Union; embarrassingly, most of the cultivars were so late-maturing that they were unsuitable for Canadian locations. (It is possible to harvest vegetative plants of hemp for fiber, but Canadian plants are chiefly grown for oilseed.)

Most drug forms have historically been cultivated in areas south of the north-temperate zone, sometimes close to the equator, where they may be photoperiodically adapted to near-12-h days and an associated long season. (In the "Evolution of Narcotic Drug Production Under Domestication" and "Classification and Nomenclatural Issues" sections, two groups of narcotic plants are discussed; many strains of the less common Group 4 ("indica-type") are able to mature in relatively northern locations. Although Group 4 strains originate from relatively southern areas of the Northern Hemisphere, they seem to mature earlier than Group 3 ("sativa-type") strains because of adaptation to a shorter season due to drought.) By contrast, non-narcotic plants (both wild and legally cultivated) are mostly found in north-temperate climates, and are photoperiodically adapted to mature by the fall season in such locations. When drug strains are grown in north-temperate climates maturation is much-delayed until late autumn, or the plants die from cold weather before they are able to produce seeds. Before illicit marijuana growers became acquainted with the fact that most narcotic strains are very late-maturing, they often found that their clandestine outdoor plants remained vegetative, not producing the congested flowering tops ("buds") that are most valued. Particularly in California, hybridization and selection produced narcotic strains that are capable of flowering outdoors (Clarke & Merlin, 2013). Of course, photoperiod can easily be controlled indoors by varying light (or dark) period, which is one of the reasons why marijuana is commonly grown in buildings.

In addition to photoperiodic adaptation, climate adaptation determines the success of Cannabis crops selected in one part of the world but grown in a quite foreign location. Most hemp cultivars (mostly fiber strains) were developed for relatively cool northern regions, and do not perform well when moved closer to the equator (Watson & Clarke, 1997).

Autoflowering (day-neutral) Strains

So-called "autoflowering" strains are genotypes that are indifferent to length of day, flowering when the plants reach a certain age or size. Some forms of C. saliva growing naturally in the extreme north appear programmed to come into flower quite early irrespective of daylength, and since the season is short, such indifference to daylenth is adaptive. At the equator, on the other hand, seasonal photoperiodic cycles are insignificant and indeed the seasons are often longer than required for full development. Some forms of C. saliva growing naturally near the equator appear programmed to come into flower only after a lengthy perior of growth, which is also adaptive in maximizing propagule production in a climate that permits large plants to develop. Autoflowering strains have been claimed in the underground marijuana literature to have been generated by hybridization of short-season and long-season plants. It does seem that hybridization can produce odd effects on photoperiodic response; I have observed hybrid-generated seedlings come into flower in less than 2 weeks, at a height of only 5 cm! Autoflowering plants can be grown in continuous light (since dark periods are not necessary for induction of flowering) and so autoflowering strains are becoming more common in the marijuana trade (Potter, 2014).

Evolution of Leaflet Size Under Domestication

The evolution and ecology of leaf size is a complex subject, and is related to the total number of leaves, their turnover rate and their orientation (e.g., Whitman & Aarssen, 2010). Nevertheless, there is a trend, exhibited in numerous annual plants, whereby the leaves of domesticated forms are larger than is the case in related wild species. This is likely due to the greater photosynthetic capacity of larger leaves, the result of selection by humans to be more productive in a given limited area. This pattern seems to be true for the three classes of domesticated Cannabis (fiber, narcotic drugs and oilseed) all of which tend to have larger leaves than do wild Cannabis plants. In Cannabis, the photosynthetic area of individual leaves is often larger in domesticated plants by virtue of (1) having more leaflets and (2) having leaflets that are larger, especially wider. This pattern of larger leaves with wider leaflets in domesticates compared to wild relatives is frequently encountered in other crops with compound leaves, for example in carrot (Daucus carota; Small, 1978b), and in alfalfa (Medicago sativa; Small, 2011). Based on modelling considerations for tomato leaves, Sarlikioti et al. (2011) concluded that for a given leaf area, bigger but fewer leaflets were better at intercepting light than more but smaller leaflets.

Environment can modify leaf size. The leaves of wild plants growing in the wild are often small simply because of environmental modification--from the more stressful conditions encountered in the wild. In Cannabis, however, the leaflets of wild plants are typically relatively small even in excellent growth conditions. When grown closely together as done conventionally, the branching of fiber cultivars is suppressed and they lose most of their lower leaves. The fewer leaves that survive near the top of the plants are larger, partly as a matter of physiological compensation, but also as a genetically controlled tendency to produce larger leaves.

Some kinds of Chinese fiber land races (Group 2, discussed in the "Classification and Nomenclatural Issues" section) and southern Asian narcotic races (Group 4 ("indica type"), discussed in the "Evolution of Narcotic Drug Production Under Domestication" and "Classification and Nomenclatural Issues" sections) are noted for their large leaves with wide leaflets--a clear reflection that they are the products of considerable domestication. As discussed in Clarke and Merlin (2013), these groups are ancient and have undergone long periods of selection.

Larger leaves (and larger leaflets) in domesticated Cannabis may be the result of greater photosynthetic demand, but there are also reasons why leaflets should be narrower and smaller in related wild plants. Brown et al. (1991) examined the hypothesis that the feeding efficiency of leaf-eating insects is lowered on leaves that are small, dissected, or needle-like, all patterns that make insects work harder to reach the edible lamina. It seems plausible that the smaller, narrower leaflets in wild plants of C. saliva are adaptive in making their foliage less accessible to herbivores, and the reduced need for such protection in domesticated plants has allowed them to develop bigger, wider leaflets. It is also possible that smaller and narrower leaflets are more resistant to wind damage, another advantage in wild plants.

As pointed out in the "Evolution of Narcotic Drug Production Under Domestication" section, the two fundamental classes of narcotic plants differ in leaflet width, Group 3 ("sativa type") plants having narrower leaflets than Group 4 ("indica type") plants. (The underground marijuana literature sometimes also contends that the leaves of Group 4 tend to have fewer leaflets than those of Group 3.) Coincidentally, Group 4 plants have much shorter internodes, resulting in pronounced crowding of the foliage, and darker green foliage. These variables seem to be correlated in the same ways that shade leaves differ from sun leaves. Many plants develop smaller, lighter-green leaves in the sun, and larger, darker-green leaves in the shade (e.g., Nobel, 1976; Givnish, 1988), and the crowded (therefore shaded) leaves of Group 4 seem to reflect this observation.

Evolution of Color Under Domestication

This section examines colors of parts of the plant that appear to have been selected in domesticated Cannabis as a result of human preferences.

Propagules that are edible and therefore attractive to various herbivores need to be inconspicuous, and the "Evolution of Propagules Under Domestication" section discussed how a camouflagic mottled layer covering the achenes of wild C. sativa serves to hide them from herbivores. Also pointed out in that section is that this layer tends to be sloughed off in domesticated strains, because it is no longer needed since humans protect the plants against herbivores. Figure 8a contrasts the quite dark achenes of a domesticated narcotic strain (typical of the "seeds" of numerous criminal confiscations I have observed in Canada) and the much lighter achenes of a fiber strain (most European strains have seeds that tend to be lighter shades of brown or gray). In these samples, the camouflagic perianth layer is absent and the color pigmentation resides in the pericarp (achene wall, surrounding the true seed). (It should be noted that achenes exposed to sunlight for long periods may become bleached.) Larger achenes are appropriately planted deeper, and this may be related to their color. Kluyver et al. (2013) proposed that ancient agricultural practices buried seeds quite deeply, leading to an increase in seed size under domestication so that seedlings would have the energy to grow out of the soil. Deeply buried seeds are probably more protected against herbivores, and may therefore be more tolerant of light coloration, which would tend to attract herbivores. However, darkness of the pericarp of domesticated achenes does not seem to be correlated with their size.


Differences in darkness of pericarps among domesticated strains of C. sativa may be the result of random fixation, but they may also reflect a frequently observed preference for light-colored achenes, as exemplified in Fig. 8b and c (for additional examples of similar color selection of fruits and seeds, see Heiser (1988) and Small (2013). The presence of lighter-colored Cannabis achenes in European fiber hemp cultivars (Group 1, discussed in the "Classification and Nomenclatural Issues" section) has been recorded by Vavilov (1931) and Serebriakova (1940). Lighter-colored achenes also are present in Chinese fiber strains, and indeed Clarke and Merlin (2013) hypothesized that Chinese fiber strains (Group 2, discussed in the "Classification and Nomenclatural Issues" section) imported into Europe in the 19th century contributed genes to European land races, and were responsible for the origin of lighter-colored achenes in European cultivars. However, human preference for lighter-colored propagules seems to be so universal that probably such selection occurred independently in Europe and China. It is possible that lighter-colored achenes arose in Cannabis not because of a human preference for lighter color, but because lighter color is associated with some other aspect of the achenes that is of value. Diederichsen and Raney (2006) found that in a large collection of oilseed flax (Linum usitatissimum) lighter-colored (yellow) seeds were heavier and had a higher oil content than darker-colored (brown) seeds, and it seems possible that the lighter color of the flax seeds is the result of correlation with selection for larger, more nutritious seeds.

Another example of human preference for light hues is provided by the inflorescences of narcotic cultivars that have been selected by clandestine breeders in the last several decades. The stigmas of the female flowers are whitish, although becoming brown with age. High concentrations of female flowers in the inflorescence of narcotic strains is extremely desirable, since this increases potency (see the "Evolution of Narcotic Drug Production Under Domestication" section). The secretory glands responsible for producing narcotic compounds are present in high density on the perigonal bracts, and these often glisten under strong light, also contributing to a whitish appearance of the female inflorescence. There appears to have been selection for strains developing whitish inflorescences. So-called "white strains" are very popular, as reflected by such names as White Diesel, White Fire, White Gold, White Haze, White Ice, White Label, White Queen, White Rhino, White Russian, White Skunk, White Widow, Early Pearl, Silver Haze and X-Haze.

Humans are fond of mutations of domesticated plants that develop purplish foliage, due to prominence of anthocyanin pigments (e.g., 'Crimson King', a very popular variant of Norway maple; "purple" (red) cabbage). When C. sativa is exposed to significant frost, it tends to become quite purple (or less green, since chlorophyll tends to degrade, revealing the anthocyanins), and sometimes the same effect is noticed at high altitudes (perhaps related to high, damaging insolation), demonstrating a propensity for violet coloration. Purple coloration of the inflorescences of narcotic strains became quite attractive to consumers in the second half of the 1970s (Clarke & Merlin, 2013; note Fig. 9), many expressing the belief that such strains are qualitatively superior. Examples of purplish strain names include Purple Bubba Kush, Purple Butter, Purple Cheese, Purple Diesel, Purple Dogg, Purple Erkle, Purple Haze, Purple Kush, Purple Maroc, Purple Monkey Balls, Purple Nepal, Purple Passion, Purple Pine, Purple Pineberry, Purple Power, Purple Pussy, Purple Snow, Purple Urkle, Purple Wreck, Grand Daddy Purple, Blackberry, Blueberry, Grape Ape and Mendocino Purple. [Article 2.2 of the current nomenclatural code for cultivated plants (Brickell et al. 2009) forbids the use of the term "strain" as equivalent to "cultivar" for the purpose of formal recognition. Nevertheless, Snoeijer (2002) treated Cannabis strain names as equivalent to cultivar names. Although Cannabis strains are conceptually identical to Cannabis cultivars, in this review the strain names are not denoted in single quotes, the convention for cultivar names. In fact, very few Cannabis strains satisfy the descriptive requirements for cultivar recognition.]


Evolution of Shoot Architecture Under Domestication

This section is concerned with how human selection of C. sativa for different purposes (fiber from the stem, drugs from the inflorescence, or oilseeds) has altered the shoot by comparison with that of wild-growing plants. Shoot features that are of particular adaptive importance to C. sativa include its main stem ("stalk"), and patterns of branching with respect to the disposition of the foliage and reproductive organs. As noted earlier, male plants are less robust than females, and die after flowering. The comments in this section pertain mostly to female plants.

Wild-growing plants of C. sativa normally develop a dominant central stem, from which, under good growth conditions, spreading side branches arise. Figure 10 shows the appearance of well-developed wild plants. As with numerous annual herbaceous plants, ultimate size depends on availability of nutrients, water and light; and crowding from competition tends to suppress lower branching and promote vertical growth. In a given wild population, one may find plants that are less than 30 cm in height, and other that exceed 2 m. The widespread assertion on the internet that there is a unique wild species, "Cannabis ruderalis," that is quite short, is rubbish--very short plants growing outside of cultivation have simply developed in a stressful environment, or are photo-periodically adapted to short-seasons and so do not have time to become large. (Janischevsky (1924), the author of C. ruderalis, noted that well-manured plants of his alleged species grow to heights of 2 m or more.)

The stature and branching pattern of C. sativa have been altered in domesticated plants in ways that maximize production of the desired product (stem fiber, drugs from the inflorescence, or oilseed). These differences have become genetically fixed by selection, but are accentuated by density of planting. The various field configuration patterns that are encountered are shown in Fig. 11, and are discussed in the following paragraphs.


The two top illustrations in Fig. 11 show shoot configurations typical of narcotic C. sativa. As discussed in the "Evolution of Narcotic Drug Production Under Domestication" and "Classification and Nomenclatural Issues" sections, there are two basic classes of narcotic plants, Group 3 ("sativa type," taller ones, at top right) and Group 4 ("indica type," shorter ones, at top left). All of these plants are naturally (genetically) very well branched (like wild plants), but the internodes in Group 4 are much shorter than in Group 3. Narcotic forms are planted at relatively low density, leaving room for the branches to develop well and produce abundant flowers. Maximizing branch production is desirable as this maximizes flower production, the perigonal bracts around the female flowers producing most of the narcotic chemicals that are desired. As discussed in the "Evolution of Narcotic Drug Production Under Domestication" section, male narcotic plants are removed to prevent production of seeds, which are not the desired product.

The short internodes of Group 4 result in quite crowded leaves, and this in turn results in several microhabitat and associated physiological features: higher humidity and lower water loss (likely adaptive, as Group 4 occurs in quite arid areas), and intra-crown shading the "Evolution of Leaflet Size Under Domestication" section discusses the resulting development of shade leaves). Moreover, as pointed out by Clarke and Merlin (2013), when allowed to go to seed the infructescences are so crowded that natural seed dispersal is very limited (the seeds mostly remain within the infructescence), and consequently Group 4 seems to have very limited capacity to escape to the wild. The short stature of narcotic Group 4 minimizes production of stem tissues (in contrast to fiber strains) while maximizing production of floral tissues, and represents a parallel strategy to advanced oilseed cultivars (discussed in the "Evolution of Seed Oil Production Under Domestication" section), which similarly have very short stature and very compact inflorescences, also minimizing stem tissues while maximizing desired reproductive tissues.


As pointed out in the "Geography, Ecology and Ancient Domestication" section, Cannabis requires fertile soil and good water availability. As discussed in the "Evolution of Stem Fiber Production Under Domestication" section, tall Cannabis for fiber has been traditionally produced near rivers, streams or ponds which not only furnish irrigation water but also provide the water in which the stems are immersed to extract (by "retting") the fiber from the stems. By contrast, areas of southern Asia, where narcotic Cannabis developed historically, are often arid. Also, soils may be rather infertile. The dwarf nature of narcotic plants of Group 4 appears to suit them to such areas where soil nutrients and water are limited.

As discussed in the "Evolution of Narcotic Drug Production Under Domestication" section, for the last half-century narcotic plants have frequently been grown clandestinely indoors to avoid detection by law enforcement, a situation in which tall plants are frequently too large (once overhead lighting and ventilation are installed in a room). Legitimate, authorized medicinal marijuana growers also often find tall plants to be too awkward to raise in greenhouses and specially fitted secure rooms. It is possible to adjust height by controlling the photoperiod. Alternatively, indoor growers sometimes resort to removing the tops, pinching stem buds to promote branching, trellising, and other techniques to limit the height of plants (Clarke, 1981). However, plants that are naturally shorter are often grown in these circumstances. "Breeders continue to develop early-maturing and high-yielding varieties that are short and compact for indoor grow room use and to avoid detection outdoors" (Clarke & Merlin, 2013).

As discussed in the "Evolution of Narcotic Drug Production Under Domestication" section, in Asia one method of preparing hashish involved using hands or leather to collect (by adherence) sticky resin from the inflorescences at the top of the plants (alternatively and more conventionally today, hashish is prepared by filtering techniques, described in the "Evolution of Narcotic Drug Production Under Domestication" section). Accordingly, strains suitable for hashish collection based on stickiness should not be too tall. As Bouquet (1950) recorded: "The cultivators, dressed in leather, moved about through the plantations. The resin sticks to their clothes, which are scraped from time to time with a blunt curved knife. This method of collection shows clearly that in those regions the plant does not grow to any great height." In a similar vein today, dwarf varieties of tree fruits have been bred to facilitate collection. An added benefit of low stature is greater wind resistance.

As discussed in the "Evolution of Essential Oil Production Under Domestication" section, a very recent market has developed for the production of essential (volatile) oil, a product substantially from the perigonal bracts, exactly the same source for narcotic drugs. Accordingly, plants of the same architecture as narcotic plants (especially as shown at top right, Fig. 11) are often used as sources of essential oil. Indeed, as discussed in the "Evolution of Essential Oil Production Under Domestication" section, narcotic strains are often pleasant-smelling and therefore suitable for harvest of essential oil, although they pose security problems.

As discussed in the "Evolution of Seed Oil Production Under Domestication" section, there has been comparatively limited selection of strains of C. saliva in historical times specifically for oilseed production. Since plants that are big and well-branched produce many flowers (such as those shown at top right in Fig. 11), when allowed to produce seeds they do so very well. Such plants were occasionally used as sources of oilseeds, but more often as sources of seed to reproduce the following season's plants. In more recent times, as discussed in the "Evolution of Seed Oil Production Under Domestication" section, short plants with flowers (and hence seeds) congested on short branches (as shown at bottom right in Fig. 11) have been grown at moderate densities to produce oilseeds, a strategy that reduces the production of stem tissue in a given area while maximizing the production of seeds on a given acreage. Plants with limited (or at least short) branching are naturally superior than irregularly branching plants for the purpose of fully and uniformly occupying a field, and maximally utilizing solar irradiation.

As discussed in the "Evolution of Stem Fiber Production Under Domestication" section, strains of Cannabis selected for harvest of stem fiber are tall and have limited branching, both characteristics accentuated by growing the plants at extremely high densities. As detailed in the "Evolution of Stem Fiber Production Under Domestication" section, these traits maximize quantity and quality of fiber. Woody tissues in the stem have been suppressed so that the stems are much hollower than in any other category of C. sativa. This makes the stems weaker and less flexible, but the high density of planting protects the plants from being lodged (blown over) by wind. Because of the limited branching, seed production is much more limited than in strains used for oilseed. However, sometimes "dual purpose" cultivars are grown (see bottom center, Fig. 11) with intermediate characteristics between fiber and oilseed strains, so that both products can be harvested, albeit in relatively modest amounts.

As noted above, different densities of planting are used to increase or suppress branching. The different classes of strains have been genetically selected to grown well at either high or low concentrations. Unlike fiber strains that have been selected to grow well at extremely high densities, drug strains tend to be less tolerant of high population densities (de Meijer, 1994).

The different architectures selected by humans are advantageous in production of particular desired products (stem fiber, seeds, or narcotics from the inflorescences), but there are associated susceptibilities to herbivores and pathogens. The long stalks of fiber strains makes them attractive to stalk-boring insects and stalk-canker fungi (McPartland, 1998). Congested inflorescences, as found in many superior narcotic and oilseed strains, makes the plants attractive to budworms and gray mold, Botrytis cinerea (McPartland, 1998). Susceptibility to pests and diseases also differs according to density of cultivation. The very dense plantations in which fiber crops are grown raises the humidity around the stalks and increases infections by fungal diseases. On the other hand, the dense canopy may be protective against many insects. By contrast, both drug and many oilseed crops are grown in open rows, and the increased sunlight is attractive to flea beetles and birds (McPartland, 1998).

Evolution of Stem Fiber Production Under Domestication

A Brief History of Fiber Production and Usage

Hemp is one of the oldest sources of textile fiber. It was harvested by the Chinese 8500 years ago (Schultes & Hofmann, 1980), and to this day China remains the world's chief producer. Hemp cultivation was introduced to western Asia and Egypt, and subsequently to Europe somewhere between 1000 and 2000 B.C. Cultivation in Europe became widespread after 500 A.D. Hemp was first grown in South America in 1545 (in Chile), and in North America in Port Royal, Acadia in 1606 (Small, 1979b). The hemp industry flourished in the U.S., particularly in Kentucky. Hemp was widely grown in North America until the early part of the 20th century, followed by a brief revival during World War II after supplies of tropical fibers were cut off (for the same reason, substantial renewed cultivation also occurred in Germany at the same time). Hemp was one of the leading fiber crops of temperate regions from the 16th through the 18th centuries. It was widely used for rot-resistant, coarse fabrics, such as sailcloth, as well as for paper, and was the world's leading cordage fiber (used for rope and similar purposes) until the beginning of the 19th century. The majority of all twine, rope, ship sails, rigging and nets up to the late 19th century was made from hemp fiber. During the age of sailing ships, Cannabis was considered to provide the very best canvas, and indeed this word, as well as the genus name Cannabis, are derived from an Arabic word for hemp. Until the middle of the 19th century, hemp rivalled flax (Linum usitatissimum) as the chief textile fiber of vegetable origin, and was described as "the king of fiber-bearing plants--the standard by which all other fibers are measured" (Boyce, 1900).

The extraordinarily labor-intensive technology traditionally employed to extract fiber from hemp stems and prepare it for weaving is shown in Fig. 12. After hand-harvesting, fiber was crudely separated by "retting," (discussed in this section), hand-stripping and/or beating, scutching (removal of smaller bits of adhering woody tissues from the phloem fiber, accomplished in the past with mechanical tools), hackling ("hackles" were the steel "brushes" traditionally used to separate the fibers), and perhaps additional combing to remove the remaining pieces of stalks, broken fibers and extraneous material.

Today, several dozen European fiber hemp cultivars make up the bulk of modern registered Cannabis cultivars; most of these originate from European land races dating back at least hundreds of years (de Meijer, 1995, 1998), and are described in the "Classification and Nomenclatural Issues" section as constituting "Group 1." Land races and cultivated selections from China are of much older origin, often more variable, and are described in the "Classification and Nomenclatural Issues" section as "Group 2."

Several developments in the late 19th and early 20th centuries, listed in decreasing order of importance in the following, combined to drastically curtail the importance of hemp fiber outside of Asia. (1) Ships once used enormous amounts of hemp for sails, because hemp fabric is very water- and rot-resistant. The "age of sailing ships" is usually defined for Western countries as lasting from the 16th century to the mid-19th century (peaking in importance in the 19th century, the "Golden Age of Sailing"). The development of motorized ships drastically reduced the need for hemp fiber. (2) Sailing ships also used enormous amounts of hemp for rope (a single ship could require 60 tonnes of hemp rope, 30 km for rigging alone; an anchor cable could exceed 60 cm in diameter). Hemp rope tended to hold water in the interior and to prevent internal rotting they were tarred, a laborious process that was made unnecessary when abaca was substituted. (3) The Industrial Revolution (approximately 1760-1840 in Britain) initiated sustained economic growth and living standards in the Western world, but also accentuated differences for the cost of fiber production between rich temperate regions and poor tropical and semi-tropical regions. As a fiber crop, hemp (like flax) is best adapted to temperate regions, in contrast to other leading fiber crops such as cotton (Gossypium spp.), jute (Corchorus capsularis), and sisal hemp. Outside of Asia, production costs (largely determined by labor) in recent centuries have been much cheaper for tropical and semi-tropical fiber crops, and this contributed to making hemp substantially less competitive. (4) Hemp fiber was once important for production of coarse but durable clothing fabric. In the 19th century softer fabrics took over this market. As the world has judged, cotton is a remarkably more attractive choice for apparel. The invention of the modern cotton gin by Eli Whitney in 1793 enormously increased the efficiency of cotton production, and has been claimed to have contributed to the demise of hemp fiber, which is relatively difficult to separate cleanly from other parts of the plant. Increasing limitation of cheap labor for traditional production in Europe and the New World led to the creation of some mechanical inventions for preparing hemp fiber, but too late to counter growing interest in competitive crops. (5) Human-made fibers began influencing the marketplace with the development of rayon from wood cellulose in the 1890s. Largely during the 20th century, commercial synthetic fiber technology increasingly became dominant (acetate in 1924, nylon in 1936, acrylic in 1944, polyester in the 1950s), providing competition for all natural fibers, not just hemp. (6) Hemp rag had been much used for paper, but the 19th century introduction of the chemical woodpulping process considerably lowered demand for hemp. (7) A variety of other, minor usages of hemp became obsolete. For example, the use of hemp as packing (oakum), once desirable because of resistance to water and decay, became antiquated. (8) The growing use of the cannabis plant as a source of marijuana drugs in the Western world in the early 20th century gave hemp a very bad image, and led to legislation prohibiting cultivation of hemp. By the end of the Second World War hemp cultivation essentially ceased in North America, most of Western Europe, and most non-Asian countries, but continued at a diminished level in Asia, Eastern Europe, France and the Soviet Union.


The Recent Resurrection of Fiber Usage

A surge of interest in re-establishing the hemp industry began in the 1990s, particularly in Europe and British Commonwealth countries. At the time, governments generally were hostile to growing any form of Cannabis sativa for fear that this was a subterfuge for making marijuana more acceptable. Nevertheless, cultivation resumed in the temperate-climate regions of many Western countries. For example, the first crops were established in Australia (Tasmania) in 1990, in England in 1993, in Germany in 1995, and in Canada in 1998. The impetus for growing hemp in the West, despite the hurdle of overcoming governmental reluctance, was economic, motivated by the general need to find new profitable crops. Today, the United States is the only notable Western nation to persist in prohibiting hemp cultivation, at least at the federal level. (Most U.S. states have enacted resolutions or legislation favoring the resumption of hemp cultivation, although federal laws have precedence.) China has always been the predominant producer of hemp, primarily for fiber, but about 3 dozen other countries currently grow significant commercial crops. Security regulations for cultivating hemp in most Western countries are usually stringent, and represent a significant cost. Such requirements may involve the use of approved cultivars obtained from authorized sources, secure fencing and storage facilities, careful maintenance of records, governmental inspections, sampling to ensure material has insignificant levels of THC (the chief intoxicating chemical), and personnel free of recent criminal records. The legislative burden that accompanies hemp puts the crop at a unique disadvantage.
COPYRIGHT 2015 New York Botanical Garden
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2015 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:p. 189-221
Author:Small, Ernest
Publication:The Botanical Review
Article Type:Report
Date:Sep 1, 2015
Previous Article:Inflorescence structure in Koyamaeae and its relationship with Sclerieae, Bisboeckelereae, Cryptagieae and Trilepideae tribes...
Next Article:Evolution and classification of Cannabis sativa (marijuana, hemp) in relation to human utilization.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters