American law and the species problem in Cannabis: Science and semantics
The treatment of cannabis in the literature
Taxonomy and the "species"
Semantic questions concerning the use of terms
The scientific question concerning taxonomic structure in Cannabis
Author: Ernest SMALL
Pages: 1 to 20
Creation Date: 1975/01/01
Biosystematics Research Institute, Research Branch, Agriculture Canada Ottawa, Canada, KIA 0C6
Cannabis, the source of the psychotomimetic drug preparations marihuana and hashish, is widely proscribed in law. In the United States, in federal and most state legislation governing cannabis drugs, specific reference is made to " C. sativa L." as the source plant. In recent court cases defendants have argued that their material was obtained from non-proscribed species of Cannabis, C. indica Lam. or C. ruderalis Janisch. The eminent plant taxonomist A. Cronquist of the New York Botanical Garden and I have testified for the state on numerous occasions recently. In no case known to the writer where the state was adequately informed on the botanical aspects of the issue has the ploy resulted in circumvention of cannabis legislation. However in a few instances where the state was not adequately prepared, cases have been dismissed on the basis of the issue. At present in both the United States and Canada, an explosion is occurring in the number of cases involving the question (appendix I).
The well-known student of hallucinogenic plants, R. E. Schultes of Harvard University, has testified for the defence in dozens of court cases since 1972, following a notable reversal of opinion regarding the taxonomy of Cannabis. Schultes and a number of colleagues have recently published their viewpoint (1974), that Cannabis comprises three species, essentially as found in some of the literature. Schultes' claims have been supported in court and print by Emboden (1974) and additionally in court recently by a number of reputable botanists. Fullerton and Kurzman (1974) have recently presented an extensive evaluation of the merits of the cases of myself, and the supporters of the defence, and Fullerton has vigorously defended the position of the defence on this and other matters connected with narcotics trials, in the courtroom. Their extensive analysis of the botanical issue is unwarranted, since neither is a botanist, and so neither has the competence necessary to evaluate the issue. The presentations advanced to this point by taxonomists and others who have testified on behalf of the defence are, in my opinion, seriously deficient both in terms of providing adequate orientation to the nature of taxonomy, and in analysing Cannabis scientifically. The purpose of this paper is to provide in abbreviated form the essential information necessary to a satisfactory understanding of the problem, and to present the basic considerations which serve to invalidate the contention that present legislation governing cannabis drugs is seriously defective in terminology.
1 Based in part on a lecture for which the author received the George M. Cooley award for 1974 of the American Society of Plant Taxonomists.
Although a number of names have been proposed for ostensibly distinctive "species" of Cannabis, only three have been given any credence as reflecting distinctive entities. These are C. sativa, C. indica, and C. ruderalis.
The father of the modern system of naming plants, the Swede Linnaeus, in 1753 recognized only one species of Cannabis, C. sativa L.
In 1785, the famous French biologist Lamarck described a putatively distinctive second species of Cannabis, C. indica. 2 The critical passage from Lamarck's work translates as follows:
"This plant, of which Mr. Sonnerat has sent us some samples which he collected in India, appears to us a species very distinct from the preceding. It is smaller, more branched, with a firmer, nearly cylindrical stem, and it particularly is distinguished in that the leaves are all constantly alternate. The leaflets are very narrow, linear-lanceolate, and very acuminate. Male individuals have five or seven leaflets; but those which are female commonly carry only three on each petiole and the upper leaves themselves are quite simple. The calyces of the female flowers are velvety, the long styles are similarly velvety. This plant grows in the East Indies. Its firm stem and thin bark make it incapable of furnishing similar fibres to the preceding species ( C. sativa L.) of which so much use is made."
Lamarck recognized C. indica as a distinct species after relatively little study, and his original species concept is not clear to us today. This statement is not meant to criticize Lamarck. In the "exploratory age" of plant taxonomy scientists often were forced to come to conclusions on the basis of very limited material.
In 1924 the Russian Janischevsky recognized the wild cannabis plants of southeastern central Russia as the species C. ruderalis. Janischevsky conducted meticulous studies of the differences between wild hemp of southeastern central Russia, and the hemp cultivars known to him. He noted that the fruits (usually termed "seeds") of the wild hemp were smaller, had a drawn-out base, a hard protective pericarp (casing), and were attached by a joint by which the fruits are easily disarticulated from the plant.
Janischevsky discovered a curious symbiotic relationship between certain red bugs ( Pyrrhocorius apterus L.) and wild hemp fruits. He observed that oily material at the base of the fruits (sometimes termed "caruncle" or "eliasome") attracted the bugs who fed on this basal fatty tissue. As the fruits were disseminated by the insects, the plants benefited in return. Finally, he noted that in the development of the wild hemp fruits that he studied, a cup-like tissue proliferated from the base of the ovary and grew over the whole fruit, adhering tightly to it. This coating integument was termed "perianth" (indicating that it was botanically equivalent to the petals of a flower) and patches of brown cells in the perianth were found to be responsible for the "marbled" appearance of the fruits. Janischevsky's diagnosis (the essential defining part of the description of a species, required to be in Latin since 1935) translates, loosely: "fruits rather small, rather hard, and narrowly ovate, the base attenuate, the perianth marbled with spots and persistent; the mature achene is included in the perianth and at maturity quickly drops off. The plants are distributed throughout the major part of southeastern Russia."
2 The date that Lamarck described C. indica is commonly reported as 1783. Breistroffer (1948) and Stafieu (1967) have noted that in fact Lamarck's description was published in 1785.
A careful reading of Janischevsky's article reveals that he had reservations concerning the advisability of recognizing wild hemp of central Russia as a separate species, and that he regarded his studies as incomplete and his conclusions as very tentative. Janischevsky stated: "Despite the notable polymorphism of hemp, and despite attempts of individual flora researchers to isolate special species of it, for the entire genus Cannabis only one species remains generally recognized in the systematic classification - Cannabis sativa. All other species described for this genus have at best been reduced to varieties, and more often are regarded as identical with the species C. sativa". Janischevsky wrote with regard to the wild form of hemp in Russia: "In supplying it with a new name, I leave for future study its rank as a systematic unit. However even now I am inclined to consider it a well marked variety". It is noteworthy that he gave as an alternative to the species name Cannabis ruderalis the varietal name C. sativa L. var. ruderalis. (The giving of alternative ranks today is unacceptable.)
In the English-speaking world until very recently there was almost universal recognition of only one species of Cannabis, C. sativa L. This is epitomized by the statement of the hemp specialist Dewey (1914): "Hemp, cultivated for three different products - fiber from the bast, oil from the seeds, and resinous drugs from flowers and leaves has developed into three rather distinct types or groups of forms. The extreme, or more typical, forms of each group have been described as different species, but the presence of intergrading forms and the fact that the types do not remain distinct when cultivated under new conditions make it impossible to regard them as valid species." A similar sentiment is expressed in the Dispensatory of the United States (Wood 1937, p. 275): "The hemp plant of India has been considered by some as a distinct species, and named Cannabis indica; but the most observant botanists, upon comparing it with our cultivated plant, have been unable to discover any specific differences. It is now, therefore, regarded merely as a geographical variety." That this viewpoint persisted down to the present decade is best evidenced by the writing of Schultes and Emboden, the men who now most vigorously attack it. Schultes wrote in 1970: "Botanists now generally agree that Cannabis is a monotypic genus, a genus with one species, C. sativa, that there cannot be recognized any true botanical varieties within this species, and that this one species has diversified into a great number of ecotypes and cultivated races. Modern taxonomists, thus, are in agreement with Linnaeus' treatment of the genus." Emboden wrote in 1972: "Despite much writing to the contrary there is but a single species of Cannabis - popularly called marihuana and that is Cannabis sativa L., the cultivated hemp, so named by Linnaeus in 1753"; and: "There is only one species of Cannabis, but there is evidence to suggest that the plant has undergone natural selection and selection by man for perhaps 6000 years, and this fact, coupled with migration, has led to recognition of three varieties within the single species: mexicana, americana ( gigantea), and indica."
Occasionally one finds the concession in recent English-language writings that there may be more than one species of Cannabis worthy of recognition. For example, Stearn (1974) noted: "the possibility that the genus Cannabis (Cannabaceae) comprises more than one species as believed by Zhukovsky (1962) and other Russian botanists and as noted by Tutin et al. in Flora Europaea 1:67 (1964), or consists of one variable species divisible on fruit characters into several subspecies with differing chemical properties...". Miller (1970) with regard to C. ruderalis wrote: "a possible second species occurring in Central Asia, southeastern Russia, and as a weed in eastern and Central Europe, is reported to differ from C. sativa in size and achene characters. Most authors accept only one species, and whether the diagnostic characters of C. ruderalis are distinct and do not overlap the extensive variability of C. sativa remains to be worked out carefully." Despite such occasional qualifications, the majority of anglophone botanists have labelled all plants of Cannabis as C. sativa L.
It is mainly in the non-English literature that one finds the opinion that Cannabis contains more than one species. Even here, however, the majority of non-anglophone authors regard Cannabis as comprising only one species (e.g. Alefeld 1866, De Candolle 1869, Heuser 1924, Kirchner et al. 1938, Hegi 1957, Kitamura 1960, Hegnauer 1964, Melchior 1964, Senchenko 1965, Takhtajan 1966, Davidyan 1972). It should also be noted that considerable disagreement and uncertainty have been expressed by various non-anglophone botanists who have recognized a number of species within Cannabis. For example in Russia, Yarmolenko (1936) included C. indica within C. sativa and recognized C. ruderalis as a separate species, whereas Serebriakova and Sizov (1940) included C. ruderalis within C. sativa but afforded distinct specific recognition to C. indica. In 1926, the famous Russian geneticist Vavilov presented an extensive article on cultivated hemp, and on wild hemp including C. ruderalis. With respect to wild hemp, Vavilov noted: "this hemp is characterized by biological features which are common to many wild plants as weeds, and sharply distinguish them from cultivated plants", and he proposed to recognize this fact (contra Janischevsky), by giving the wild hemp status as a variety of Cannabis sativa, i.e. var. spontanea. Vavilov's article includes a discussion of the fact that many forms may be recognized in Europe, Afghanistan and India, all of which he considered as Cannabis sativa. Vavilov recognized, as Nikiforov (1963) noted subsequently, that "series of races of the typical wild and typical cultivated forms of hemp show many overlapping characters", and that all transitional forms from wild hemp to cultivated hemp can be found, and that for many forms "the morphological limits between weed hemp and true wild hemp are difficult to draw, and in regard to the characters of the fruits, the flowers and the leaves, they are indistinguishable from each other". On the other hand, in a brief treatment dealing with material collected in Afghanistan, Vavilov and Buchinich (1929) recognized C. indica as distinct from C. sativa by small leaves, small fruit, and low growth.
Zhukovsky (1962, 1964), in a very brief treatment of Cannabis, recognized three species, C. sativa, C. indica and C. ruderalis. There is no evidence suggesting careful taxonomic analysis of Cannabis by Zhukovsky. Janischevsky's concept of C. ruderalis was adopted. Zhukovsky noted Nikiforov's conclusion that some small-seeded forms of wild hemp belong to C. ruderalis rather than to C. sativa, and that C. sativa contains both wild and cultivated races.
Apparently for Zhukovsky the concepts of C. sativa and C. ruderalis merge. " Indian hemp" ( C. indica) is also recognized but it is not at all apparent how the three species are to be distinguished morphologically. Cannabis indica is described as being a narcotic plant occurring wild in Pakistan and Afghanistan, and cultivated for narcotics in India, Iran, Turkey, Syria and northern Africa. Zhukovsky notes that wild varieties of narcotic hemp, like fibre hemp, possess disarticulating seeds.
American law and the species problem in Cannabis: Science and semantics 5
Kirchner et al. (1938), like many taxonomists, recognized as varieties the groups of plants which Zhukovsky termed species. "Indian hemp", the form of Cannabis cultivated for drug purposes, is characterized by heavy branching, and dark, small leaves.
The most intensive attempt at classification of Cannabis was by Serebriakova and Sizov (1940), who recognized the genus Cannabis as composed of two species, C. sativa and C. indica. This treatment was adopted by Ceapoiu (1958), Mansfeld (1959), and Kender (1962). Hoffman (1961) recognized only one species of Cannabis and, noting Ceapoiu's adoption of the scheme (Hoffman 1970), he points out that the two "species" are completely interfertile.
Serebriakova and Sizov recognized Cannabis as comprising two species with numerous subspecies, groups of varieties, varieties, and forms. "Indian hemp" ( C. indica Lam.) is used for the extraction of narcotic substances, and is distributed in India, Afghanistan, Iran, Turkey, Syria, Israel, and North Africa. In comparison to "common hemp" ( C. sativa), Indian hemp is described as being usually relatively branched, and having narrower, smaller leaves, and shinier seeds. Serebriakova and Sizov recognize that there are wild and cultivated phases in both species. "Common hemp" is divided into two subspecies on this basis: subsp. culta (cultivated hemp) and subsp. spontanea (wild hemp). This scheme agrees in general concept with my own evaluation of Cannabis although as discussed later, the morphological distinctions are quite insufficient for specific delimitation. Serebriakova and Sizov recognized four groups within the cultivated phase of C. sativa, one of which encompasses a group of narcotic varieties, which might more logically have been placed in their "Indian hemp". It appears that Serebriakova and Sizov concentrated on evaluating the wild and cultivated hemps of Russia, and spent relatively little effort assessing the differences between their "common hemp" and "Indian hemp". Fifteen varieties of common hemp were recognized, whereas no attempt was made to assess variation of "Indian hemp", and thereby evaluate whether the two units deserve separate recognition.
It should be noted that a tradition of "splitting" of species (by Western standards) has characterized Russian plant taxonomy (Heywood and Bobrov 1965). It is most unlikely that a Western botanist would have ever recognized C. ruderalis as a species. The main support for the recognition of more than one species of Cannabis has come from Russia. The Russian Davidyan (1972) has, however, surveyed the Russian literature and concluded categorically that in Cannabis only one species deserves recognition. And recently the most authoritative Russian plant taxonomist, A. Takhtajan, has written (pers. com.) that "there is only one species of Cannabis, namely C. sativa."
Taxonomy is not readily comprehensible to the layman and it has been my impression that its nature has been obfuscated in the courtroom and in recent articles dealing with the present forensic issue. In particular the concept of "species" has been clouded by seemingly abstruse rhetoric. A recent compilation of 21 divergent definitions of "species" (Radford et al. 1972, pp. 28-29) makes it clear that there is not universal agreement on this concept. However plant taxonomists generally insist that a number of minimal conditions be satisfied before groups be recognized as species, and by examining these requirements with respect to Cannabis I believe it becomes evident that the case for recognizing more than one species is unwarranted at present. A text by Davis and Heywood (1965) is widely considered as the best available exposition of the nature and principles of plant taxonomy, and may be consulted for details of aspects discussed in this paper. Good briefer treatments of the principles of plant taxonomy have been published by Heywood (1967) and Jeffrey (1968). Useful discussions of theory are found regularly in the journals Taxon and Systematic Zoology. A discussion of pertinent aspects of the nature of species follows.
Some biologists believe that in nature organisms are substantially organized into a characteristic kind of population aggregate, which is sufficiently discrete to merit special recognition. Such biologists frequently believe that the principal criterion for recognizing such units, which they term "species", is the demonstrated existence of substantial physiological barriers to breeding. Thus species are distinguishable on the basis that individuals interbreed only within the confines of the species of which they are constituents. Some biologists relax this criterion, and would consider populations which are isolated either geographically or ecologically to be worthy of species recognition, for although they may be potentially interfertile, their isolation prevents interbreeding in practice. However, in such cases, there must be accompanying morphological distinction. This point is consistently overlooked by the defence. Stocks of Cannabis from widely divergent sources have been found to be completely interfertile (Small 1972a, Davidyan 1972) and the extensive morphological intergradation of Cannabis in all regions of the world indicates lack of reproductive isolation on a geographical or ecological basis.
Virtually all plant taxonomists give principal consideration to external appearance (morphology) in assessing taxonomic relationships. "Hidden" features, such as fine structure revealed by high-resolution electron microscopy, internal structure (anatomy), chemical constitution, and physiological and ecological behaviour also provide important information, but unless accompanied by breeding and/or morphological distinctiveness almost no plant taxonomist would recognize more than one species in a group. Just how different in appearance do kinds of organism have to be before one calls them distinct species? There is no standard level of divergence which biologists require before recognizing species. Thus species are in fact at least partly arbitrary units. This subjectiveness is epitomized by the terms "splitters" and "lumpers" sometimes used to indicate the dispositions of taxonomists to species delimitation. However, the recognition of species is by no means an entirely subjective exercise. Simpson (1961, p. 107) phrased the situation as follows: "Taxonomy is a science, but its application to classification involves a great deal of human contrivance and ingenuity, in short, of art. In this art there is leeway for personal taste, even foibles, but there are also canons that help to make some classifications better, more meaningful, more useful than others".
Botanical taxonomists generally use several criteria, often unconsciously, in deciding whether entities are sufficiently distinctive to merit species designation. One guide to the desirability of recognizing entities as species is the degree of difference of appearance. When individuals are dramatically different, and as easy to distinguish as black from white, there is at least the possibility of designating the groups as species. However one must also consider the critically important minimal criterion of discontinuity of variation. Thus one may recognize black and white as different "species" only if most individuals are either black or white. When the limits of the groups are substantially obscured by grays, and many individuals can not be reliably identified, most taxonomists would not designate the groups as separate species. Variants of lesser degree of discreteness are sometimes placed in infraspecific categories such as "subspecies" and "variety".
Geographical isolation is an additional characteristic of taxonomic importance. Often geographical isolation is due to the different ecological requirements of closely related species. There are many examples of closely related species which are ecologically and geographically isolated except for zones of contact, where hybridization occurs, which sometimes obscures the primary morphological discontinuity between the otherwise separate populations. In one courtroom, this situation was represented as an indication that taxonomists do not require discontinuity of variation before species are recognized. However, the situation really shows that taxonomists may accept species without a complete discontinuity of variation in the special situation where very different kinds of plants occupy adjacent differing ecological zones. If the contact zones where hybridization occurs are very extensive, it is more appropriate to designate the overlapping groups as subspecies of one species rather than as distinct species. Another confusing contention of the defence is that species frequently show "introgression" (the acquisition of a degree of similarity by one group of organisms during a limited amount of interbreeding with another group of organisms), and that therefore it is acceptable to recognize species of Cannabis even if many populations are more or less intermediate. This is putting the cart before the horse, for it is mandatory to first demonstrate that there is sufficient non-overlap in the variation pattern to speak of species, before one can speak of overlap (as implied by introgression). In general one is justified in recognizing groups as separate species only when there is substantial discontinuity of variation. It should be emphasized that the groups represented as species in the genus Cannabis (i.e. the wild phase, the phase cultivated for fibre, and the phase cultivated for drugs) are interspersed to a considerable degree, and cannot be considered to represent species on the basis of the arguments presented in this paragraph.
It has been claimed by proponents for the defence that variation in Cannabis is simply too extensive to be contained within the confines of one species. This position simply ignores the minimal requirement of discreteness of variation for species delimitation, and appeals to the inherent, justified desire of people to have labels for different kinds of creatures. The crux of the matter, however, is not the assignment of labels, but their assignment at the species level. Human beings provide an instructive example. Homo sapiens encompasses an enormous range of variants, and some of the races of man are strikingly different in appearance. The geographical isolation and sociologically-based restrictions of interbreeding between some of the races should be noted as an example of this kind of situation within the confines of one species. Wide-ranging species such as Homo sapiens frequently exhibit extremely distinctive local variants, which when considered by themselves might be thought worthy of recognition as species, but when viewed in the light of the entire world variational pattern are seen to be insufficiently demarcated to warrant species designation, in the particular way in which biologists use the term species. The differences between the races of man are adequately labelled by such epithets as "Caucasian", "Negro", and "Mongol", and there are good reasons for using similar non-latinized labels for the races of Cannabis.
Although discontinuity of variation provides a minimal criterion for species delimitation, it does not necessarily provide a sufficient criterion. It should be appreciated that rather spectacular non-overlapping variation may be included within the confines of one species, especially (as in Cannabis) when domestication has been responsible for the evolution of highly divergent types of organisms. This problem has been essentially ignored by the proponents for the recognition of several species in Cannabis. Its importance is indicated by the following excerpt from Charles Darwin's discussion, in The Origin of Species, of pigeon breeding (Darwin 1859; 1959 ed., p. 97): "Altogether at least a score of pigeons might be chosen, which if shown to an ornithologist, and he were told that they were wild birds, would certainly be ranked by him as well-defined species. Moreover, I do not believe that any ornithologist would in this case place the English carrier, the short-faced tumbler, the runt, the barb, pouter, and fantail in the same genus -more especially as in each of these breeds several truly-inherited sub-breeds, or species, as he would call them, could be shown him. Great as are the differences between the breeds of the pigeon, I am fully convinced that the common opinion of naturalists is correct, namely, that all are descended from the rock-pigeon ( Columbia livia)..."
An equally impressive example, in the plant kingdom, of the inclusion of a wild progenitor and several highly divergent domesticated derivatives within one species, is provided by Brassica oleracea. This species includes inedible wild plants, as well as brussels sprouts, cabbage, kale, broccoli, kohlrabi, and savoy. Dog breeds provide another familiar example of a species containing many divergent kinds of creatures which are different simply because of selection by man. A frequently employed tactic by the defence in the present debate has been to present as evidence carefully chosen divergent specimens of wild Cannabis, and Cannabis cultivars selected for fibre purposes and for drug purposes, representing these three kinds of plant respectively as examples of C. ruderalis, C. sativa, and C. indica. This manoeuvre has proven extremely effective, since the particular specimens exhibited have invariably been tangibly distinctive. As will be noted later, the three groups mentioned are not nearly as divergent as has been claimed, and indeed much confusion has existed in the minds of the expert witnesses for the defence concerning the nature of the entities they have represented as species. In particular the proponents for the recognition of several species in the genus Cannabis have not seemed to realize that the characters which have been used to delimit three species by some botanists are fundamentally the result of domestication. Agronomists have selected tall, sparsely branched, hollow-stemmed plants, since this makes for long, easily accessible fibres. It is hardly surprising then that there are tall, sparsely branched plants with solid stems, and short, highly branched plants with solid stems. These differences are comparable in nature to those between race horses and draught horses, the former possessing slim bodies, and the latter stocky, massive frames-each kind ideally suited for its purpose. Of course race horses and plough horses belong to one species. Even if one were to concede that the entities represented as species in Cannabis were highly distinctive, in view of the fact the differences are the result of domestication, there is no more reason for the different kinds of these Cannabis plants to be labelled different species than there is for the kinds of pigeon, horse, dog, or Brassica oleracea.
The assignment of different species names to the wild and cultivated phases of a plant is justified only when there is appreciable divergence between the two. Taxonomists may adopt highly individualistic species concepts, and those who wish to do so are free to label as species cultivated derivatives which scarcely differ from their wild progenitors. However, it should be clearly understood that such treatment represents a radical departure from the philosophy of most plant taxonomists. Most botanists have regarded the minor differences which distinguish plants like Cannabis which have basically similar co-existent wild and domesticated phases as too trivial to warrant recognition as distinctive species, in view of the fact that the differences are artificially generated and maintained (see for example Harlan 1969, Baker 1970, Hawkes 1970, Harlan and de Wet 1971). The International Code of Botanical Nomenclature (Stafleu et al. 1972) provides that: "Plants brought from the wild into cultivation retain the names that are applied to the same taxa growing in nature", and: "Variants of infraspecific rank, which arise in cultivation through hybridization, mutation, selection or other processes, and which are of sufficient interest to cultivators to be distinguished by a name, receive cultivar epithets preferably in common language (i.e. fancy epithets) markedly different from the Latin epithets of species and varieties". The separate recognition of the wild and various cultivated phases of Cannabis as species is therefore questionable.
Taxonomists have adopted a complex set of rules governing the application of names to plants (Stafleu et al. 1972, Jeffrey 1973). These rules are operable under specified circumstances, have particular but limited implications, and employ terms in restricted senses. Unaware of the limitations of botanical nomenclature, some courts have lacked the basis necessary for arriving at an intelligent decision.
Because of incomplete knowledge or lack of competence, different botanists have frequently assigned different names to the same species. To promote stability of nomenclature taxonomists have adopted regulations specifying the single name which must be adopted when more than one name is available for a given species. (If no name is available, a taxonomist creates a new name.) The "type method" is of central importance to nomenclature. Botanical taxonomists since 1958 have been obliged to designate a single specimen known as the "type" When they name a species. This specimen is permanently deposited in a herbarium and serves as a reference point for the name. It would not seem unreasonable to suppose that type specimens necessarily represent the best possible exemplar of a species, and this has frequently been implied and inferred in courtrooms. This is absolutely untrue, however, and this belief represents a fundamental misunderstanding of the purpose of the type method. Type specimens may be extremely atypical (although aberrancy in the sense of a monstrosity may be grounds for rejecting a specimen as a type). Type specimens are not primarily intended to demonstrate variation, but rather are a device for promoting nomenclatural stability, since names are permanently attached to specimens, and are carried with the specimen should a biologist change the concept to which the type specimen is assigned. It is incorrect by definition to speak of the type of a taxonomic group (as has been done repeatedly in courts) since only names have types. Indeed the phrase "type specimen" is highly misleading, and for purposes of comprehension might better be discarded and replaced by a term such as "name-bearing specimen".
Type specimens of C. sativa and C. indica were not designated by Linnaeus and Lamarck, respectively, since the type method was not employed in the early days of taxonomy. However the rules of nomenclature are retroactive, and a type specimen of C. sativa was recently designated by Steam (1974), and the choice for C. indica is essentially narrowed down to a single specimen in Lamarck's preserved herbarium. The type specimen of C. ruderalis is in the author's office as of this writing and has not yet been seen by any of the other participants in the present forensic debate, who have been, and are, accordingly, simply not in a good position to evaluate the applicability of the name C. ruderalis.
Before a name becomes available for consideration, a number of requirements must be fulfilled. With adherence to an increasing number of these requirements, a species name is progressively described as "effectively published", "validly published", and "legitimate". The emotive content of these terms has been fully exploited in the courtroom. However these terms have been assigned very restricted meanings in the International Code of Botanical Nomenclature, article 6 (Stafleu et al. 1972), and merely indicate that there has been adherence to a number of codified conventions. By following these conventions, one could re-describe any species of plant as a species of Cannabis. For example, one could describe tomatoes, or carrots, or marigolds as new species of Cannabis, and by following the appropriate conventions specified in the botanical code, one could thereby produce new species names for Cannabis, which are effectively published, validly published and legitimate. By the same token, one could split C. sativa into a million species, restricting the original term to only one of the million. Of course, no one would take such names seriously. However they would have to be listed in Index Kewensis (Jackson et al. 1893--present), the continuously-updated, multi-volume botanical repository of species names. It has been claimed that the mere existence of several legitimate names of Cannabis produces sufficient doubt to make the use of the name C. sativa L. unworkable in legislation, but this proposition is clearly absurd. A peculiar definition of "murder" by a small segment of society does not alter the import of the term in current legislation. The most important technical term, "correct", has been conveniently omitted from the presentations of the defence. Whereas there may be many "effectively published", "validly published", and "legitimate" names for a taxon, for any particular taxonomic treatment, there can be only one "correct name".
How do taxonomists decide on correct names? This is done in two stages. First taxonomists decide, on the basis of scientific investigation, how many taxa deserve recognition. The point cannot be overemphasized that prior to the satisfactory exhibition of discrete entities the number of names in the literature has no bearing on how many species deserve recognition. An analogy can be made with the Unicorn, Loch Ness Monster and Abominable Snowman. The names exist, but are irrelevant to taxonomist until the creatures are scientifically demonstrated to exist also.
Once a taxonomist has decided on the existence and boundaries of a species, he determines its correct name by ascertaining whether and how many names exist whose types fall within the species as he has defined it. When, as is frequently the case, more than one legitimate name is available, the earliest name takes precedence by the principle of priority. In the case of Cannabis, if the decision is made that all plants belong to one species, the name of the species is C. sativa L. If on the other hand the decision is made that there are more than one species of Cannabis, the name C. sativa L. acquires a more restricted meaning, and refers only to some plants of Cannabis. Occasionally when there is dispute concerning the comprehensiveness of a name, taxonomists add the phrase sensu lato ("in the wide sense") or sensu stricto ("in the narrow sense") to make the meaning of their name clearer. Usually, however, it is necessary to establish the comprehensiveness intended in using a taxonomic name by examining the total context in which it is used. This situation is so pervasive in taxonomy that similar forensic debates could arise for virtually all living creatures, and substances derived from these, which have been subjected to legislative controls. It would be tragic if the present debate has opened a Pandora's Box. However the solution would seem to be obvious. Clearly names of organisms are just words whose precise meaning is highly mutable, and can only be understood by referring explicitly to the concept of the users of the term. And clearly, because names of organisms can be used in different senses than the originators of the names intended, and the comprehensiveness of a name is governed by what the term means to the user, it is imperative to determine what the meaning was when used. This is precisely the question which most judges have seen fit to interpret as the crux of the matter. It would be impertinent of me to comment on the appropriateness of their decision with respect to the canons of the American judicial system, but I would agree that logically this is the only tenable position.
What then is the meaning of the name "C. sativa" as used in legislation and how does it relate to the terms C. indica and "C. ruderalis"? Congressional hearings were held in 1937 to aid in the drafting of the federal marihuana legislation and are most pertinent to an understanding of that legislation, and by extension, since most state legislation is patterned on federal legislation, to legislation generally in the United States.
In hearings before the Committee of Ways and Means, House of Representatives, 75th Congress, 1st Session (H.R. 6385), H. J. Anslinger, Commissioner of Narcotics, testified:
"Marihuana is the same as Indian hemp, hashish... It is known as cannabis, cannabis America, or cannabis Sativa. Marihuana is the Mexican term of cannabis Indica... It is known in various countries by a variety of names" (p. 29).
Eugene Stanley, District Attorney of New Orleans, testified:
"The plant or drug known as cannabis indica, or marihuana, has as its parent the plant known as cannabis sativa. It is popularly known in India as cannabis indica; in America as cannabis americana; in Mexico as cannabis mexicana, or marihuana. It is all the same drug, and is known in different countries by different names. It is scientifically known as cannabis sativa, and is popularly called cannabis americana, cannabis indica, or cannabis mexicana, in accordance with the geographical origin of the particular plant" (p. 37).
"Although the different forms of the plant have been described under different botanical names, there are no essential differences in any of the specific characteristics, and all cultivated or wild hemp is now recognized as belonging to one species, cannabis sativa" (p. 38).
A botanist, L. H. Dewey, author of an agricultural monograph on hemp (Dewey 1914) testified:
"There are different botanists who have seen the different forms of the plant and have given it different names, so far as its identity is concerned. For instance, the plant in India was very excellently described by Lamarck in 1788 under the name of "cannabis sativa" 3 (sic). That was the name they gave to Indian hemp in the primitive days and down to comparatively recent time. Until I got some seed from India and grew them I had supposed that was a different species. I grew them in comparison with the Kentucky hemp with which you are all familiar, but in a year or two they were just alike. There was no specific difference; so it is with the other forms that have been described as different species. So there is only one species known as "hemp" (p. 55).
In delimiting C. indica, Lamarck (1785) went into some detail concerning its use as an intoxicant. Although he also referred to the "narcotic" properties of C. sativa, his relatively greater emphasis on the inebriating potential of C. indica may have given rise to the practice in the literature of referring to C. indica rather than C. sativa when the writer wished to indicate intoxicating properties. Pharmacologists came to use the term "C. indica" in "official" or "officinal" nomenclature, as denoting a drug concoction prepared from C. sativa. Even this usually unappreciated ambiguity between biological and pharmacological nomenclature was clarified before a Congressional committee. In a hearing held before a subcommittee of the Committee of Finance, U.S. Senate, 75th Congress, 1st Session (H.R. 6906) Matt Rens of the Rens Hemp Company testified:
3 A transcription or oral error. Substitute indica for sativa.
"As used in the bill (H.R. 6906) the term "marihuana" is synonymous with true hemp, the scientific name of which is Cannabis sativa L. The chemical substance found in hemp which produces the narcotic effect has been officially termed "cannabis indica", and is known throughout the world as cannabis indica. Since botanists now recognize hemp as consisting of only one species, the term "indica" should be discontinued; thus in referring to hemp in a narcotic sense, the term "cannabis" is most appropriate and more universally understood" (p. 23).
This record makes it clear that the legislators of the time appreciated that there was a minor problem over the use of the name C. sativa, but that they accepted the name as comprehensive. The question of the competence of legislators and past scientists to have arrived at correct decisions regarding biological nomenclature has prompted considerable discussion by the defence. However the topic is irrelevant, the point being that whatever the scientific advisability of so doing, at least at the time of the establishment of the laws regulating narcotics the term C. sativa was almost universally used and understood in the United States as inclusive of all plants of Cannabis.
Without the backing of so eminent a botanist as Schultes it is most unlikely that the current controversy would have arisen, or would be maintained. And so the debate in American courts with regard to the species question has focussed on the research of Schultes and his associates, and on my research. The relative merits of our disparate conclusions concerning the scientific facts must be judged ultimately by our peers in the botanical community. Unfortunately, it has now become urgent that the lay community acquired some understanding of the weaknesses and strengths of our scientific positions, since the Schultes interpretation results in the possible circumvention of legislation when courts choose to disregard the semantic aspects of the question.
Schultes radically reversed his position on the taxonomy of Cannabis, as evidenced by his contrasting and contradictory papers of 1970 and 1974. Although he had written (1970, p. 25): "One cause of the chaos in taxonomic and nomenclatural recognition of polymorphism in Cannabis sativa has been misunderstanding of species delimitation in cultivated plants. No botanist can draw a sharp line between a cultivated and a wild or spontaneous plant", in his 1974 paper he contends that the wild phase of Cannabis in the USSR, which has been called C. ruderalis, deserves specific recognition. Further, it is now contended that one should ignore cultivated and spontaneous plants entirely, and only consider what are termed "wild" plants. Although he had written (1970, p. 25): "The recognition of races, ecotypes and other kinds of convivia with Latin designations in the rank of species for convenience confuses classification, since the minor concepts-e.g. Cannabis indica-are not the equivalent of a Linnean species" he now believes that C. indica deserved specific recognition. Although he had written with respect to the views of Zhukovsky and Vavilov on C. indica and C. ruderalis (1970, p. 17): "Obviously neither Vavilov nor Zhukovsky envisage genetically stable "varieties" in these concepts", he now contends that the views of these two Russian botanists provide the best support in the literature for the notion that these entities are full-fledged species. Schultes has stated the reasons for this reversal of view (Schultes et al. 1974, p. 341) as follows: "Subsequent critical studies of the literature; examination of material from many areas preserved in several of the world's largest herbaria; preliminary field work in Afghanistan; and a survey of the plantings of Cannabis in Mississippi from seed imported from many localities around the world under the auspices of the National Institute of Health." I cannot help but observe that I can perceive little literature cited in his paper of 1974 that was not fully appreciated in his masterful synthesis of 1970. Inadequacies of the Mississippi plantation material, which I have studied, are pointed out later.
Schultes et al. (1974) and Emboden (1974) have stressed that a distinction should be drawn between "wild" plants (those, growing in native habitats, which have not been modified by domestication) and plants variously described as "spontaneous" or "escaped", which have escaped from cultivation and which have perhaps re-evolved adaptations for living in nature. My work has been criticized as not addressing this problem. However, this criticism is not only specious, but irrelevant. In the first place, there is simply no reliable non-circular way of identifying a plant of cannabis collected in nature as "wild" or "escaped". Further, only speculative evidence exists as to where the original range of Cannabis was located prior to domestication. Given the extraordinary expansion of territory occupied by Cannabis as it became domesticated, the frequent escape of cultigens back to nature, and the ability of all populations to hybridize freely, it is doubtful that unaltered ancestral stocks of Cannabis exist anywhere, quite apart from the impossibility of being able to identify these as such. 4 In any case, the question of exclusively "wild" species does not really even enter into the evaluation of the three "species" which have been discussed in courts, since C. sativa includes fibre cultigens, C. indica includes drug cultigens, and C. ruderalis includes "ruderal" plants. The defence has also alleged that the existence of as yet undescribed wild species of Cannabis so complicates interpretation of the name C. sativa as to make it impossible to identify this species. However referring to undiscovered hypothetical species constitutes speculation which is unwarranted in this context.
In their recent paper Schultes et al. (1974) beg the question of how many species should be recognized in Cannabis by simply referring to "the species of Cannabis", presenting a key to the "species" of Cannabis as they understand them, and discussing character variation between these "species". Features used in the key include plant height, branching pattern, and a number of seed characters. I have found that the character combinations proposed are poor for delimiting the taxa purportedly described, and I can only strongly disagree with the statement by Schultes et al. (p. 358): "the species can easily be distinguished".
The reader should clearly understand that a detailed description of kinds of individuals characterized by clusters of attributes, as that of Schultes et al., does not constitute evidence that the groups are necessarily of any taxonomic significance.
4 Anderson (1952) provides good discussions of the difficulty, and in some cases, impossibility, of ascertaining that a plant is "wild".
For example, consider men who are short, bald and fat, or tall, blond and lean, or blue-eyed, buck-toothed and bow-legged. Such kinds of men could be recognized as groups, but as the characters in question are known to occur randomly in all combinations, such groups are considered too trivial for taxonomic consideration. For given constellations of characteristics to actually represent significant, distinctive, taxonomic groupings, there must be a substantial absence of intermediate types.
The reader may well ask how such differing evaluations could be entertained. Taxonomy as practised by many individuals and as practised to date with Cannabis by those who have testified for the position of the defence consists mainly of simply looking at organisms under study. This is quite adequate when the organisms constitute distinctively different non-overlapping groups, and it is possible for careful observers like Schultes to perceive minor differences. Such methodology by itself is doubtfully adequate for the recognition of new species in a genus such as Cannabis, where there are innumerable minor variants, and where most taxonomists have not previously perceived differences worthy of specific designation. Scientists are no more immune than laymen to mentally restructuring reality so as to justify their preconceptions. Armed with the prejudice that there are certain distinct kinds of plants, it is difficult to avoid perceiving just these when in fact there are numerous intermediate plants. I believe those who are currently advocating the recognition of several species in Cannabis, have fallen into this trap, and are only comparing selected extreme variants. Consider the statement (Schultes et al. 1974): "Preliminary examination of the wood anatomy of material which we collected in Afghanistan and which we believe to represent Cannabis indica discloses differences from that of material of C. sativa grown in the United States. This research, being carried out by Dr. Loran C. Anderson of Kansas State University, is in its preliminary stages and will be the subject of a later paper. The anatomical differences between these two species are very substantial, and Dr. Anderson feels that some comparable differences in other groups of plants might be given even generic status". 5
Anderson, who has testified for the defence in recent cases based his conclusion on the examination of only one sample of Schultes' Afghanistan material, and one sample of Kansas wild hemp (Anderson, 1974). Since the Schultes Afghanistan material, which I have studied, is likely a variant selected for drug properties, and since the Kansas hemp is derived from fibre strains, it would be surprising if internal differences did not exist. In fact, as discussed later, I have found profound internal differences (as large as those between some genera!) between such races with respect to chemical constitution. As I have noted, a proposal to split Cannabis into several species must be based on an examination of the entire spectrum of variants of Cannabis. It is a simple matter to "prove" that there are several "species" of Cannabis by comparing only selected variants, but this disregard for representative sampling and for adequate sample size cannot result in acceptable conclusions regarding the taxonomic structure of Cannabis.
Research employing two procedures can help resolve a contentious taxonomic situation such as described here.
5 This statement is misleading. Jansonnius (1950) has pointed out that variability in wood anatomy is often much greater within some species than between other species. For example, profound anatomical differences exist between the succulent roots of the cultivated carrot and the woody roots of wild carrot, but these different kinds of plants are placed in the same species. Anderson's conclusions are contradicted by the more extensive studies of Nassonov (1940).
American law and the species problem in Cannabis: Science and semantics 15
First of all, it is necessary to grow material representing the diversity of possible kinds of Cannabis in a standard garden, to be sure that the features of the plants reflect fundamental inherited traits rather than environmentally induced modification. For example, I have found that plant height and branching pattern are strongly influenced by environment, so much so that these features are of limited usefulness in separating populations of Cannabis. To my knowledge the only large plantation intended primarily for taxonomic study yet established is my own. Additionally, at the University of Mississippi, M. W. Quimby grew several dozen seed stocks. Unfortunately, in subsequent years many plants were generated from stocks which were allowed to cross-pollinate, obscuring differences between the populations. Quimby, an excellent scientists who had the opportunity to closely examine the original unhybridized material, has concluded that Cannabis is best considered monotypic (Quimby et al. 1974; cf. Small 1975a). The Mississippi material was limited compared to our own plantation in which some 350 stocks were grown outdoors. In addition, we have grown stocks in greenhouses, enabling us to bring into flower populations which failed to mature outdoors.
The second requirement, as any good scientist would insist, is the use of appropriate statistical procedures. Taxonomy has recently adopted mathematical techniques designed to clarify how variation in organisms is structured (Sneath and Sokal 1973). The use of these techniques may well be considered imperative in Cannabis, where strong prejudices are clearly biasing judgement. We have very recently concluded extensive statistical evaluation of about 2,500 plants scored for 47 characteristics. All plants were cultivated simultaneously in a standard garden to ensure that the traits we observed were not simply the results of environmental modification. The large sample included plants originating from wild and cultivated populations from all parts of the world. We have also carried out quantitative studies of the variation of herbarium specimens collected throughout the world (Small 1975b) and these support the work described below. (Small, Jul and Lefkovitch 1975.)
Through our use of a spectrum of "clustering'' 6 techniques (Sneath and Sokal 1973), it has become evident that variation in Cannabis is not dearly definable into different clusters of populations which could be reasonably considered candidates for specific designation. However, a powerful set of statistical techniques, sometimes termed "ordination analysis" (Blackith and Reyment 1971), has revealed that it is possible to recognize groupings within Cannabis on the basis of the intuitive concepts which have been entertained by the supporters of those who have chosen and would choose to designate entities within Cannabis as species.
In particular, we employed an ordination technique known as canonical analysis. In simplest terms, this methodology answers the question: to what extent are we able to distinguish potential groupings (defined on an a priori basis, such as wild v. cultivated origin), based on the most discriminating set of characters. In effect, this methodology establishes whether or not there exists the minimal discontinuity of morphological variation necessary for species delimitation.
We found that there was too much intergradation between wild and cultivated plants to reliably distinguish two taxa at the species level on the basis of this criterion.
6 Clustering techniques in contrast with canonical analysis described later, do not presume on an a priori basis the constitution of potential taxonomic grouping (clusters). Both approaches are valid methods of assessing taxonomic structure.
We also examined the relationships between plants of three different intoxicant potentials. "Intoxicant" plants almost always originate from fairly southern countries, especially from Africa, southern Asia, and the central countries of the New World-areas which traditionally have been associated with the use of Cannabis for "narcotic" purposes (Small, Beckstead and Chan 1975). In such plants the bulk of the resin is composed of the intoxicating cannabinoid, Δ 9-tetrahydrocannabinol (THC). We found that these types of plant required a long growing season before they would mature into males and females, reflecting their origin from southern climates with long growing seasons. Very curiously (in view of the well-known practice of roguing male plants out of plantations), male and female plants both contained large amounts of resin. "Intoxicant" plants contrast with "non-intoxicant" and "semi-intoxicant" plants. The latter two closely-related categories of plant both usually originated from comparatively northern countries. Such plants tended to reach maturity relatively quickly. Oddly, in these groups female plants usually have much higher resin content than male plants. Both non-intoxicant plants and semi-intoxicant plants have limited THC content, but semi-intoxicant plants have sufficient THC in the females (but not in the males) to be utilizable for intoxicant preparations (Small and Beckstead 1973a, 1973b). "Semi-intoxicant" plants may represent the products of hybridization between noni-ntoxicant and intoxicant types of plants. We found that, using morphological characteristics alone, we were able by canonical analysis to distinguish intoxicant plants from less intoxicant types (both semi-intoxicant and non-intoxicant), with 94 per cent accuracy. Normally, the ability to distinguish groups on morphological grounds with this level of confidence might be a reasonable basis for recognizing them as separate species. In this case, however, the differences between the groups are only made apparent by cultivating them under the uniform growth conditions of a standard garden. Plants grown in nature are too variable to be distinguished. In addition, the mathematical function used to discriminate the groups is rather cumbersome, and very few plant taxonomists would be willing to use such a difficult way to distinguish species. For these reasons, there is no practical basis at present for distinguishing intoxicant plants as a separate species, quite apart from the theoretical reasons already discussed for not following this course.
It should be noted that there are wild and cultivated plants in both the intoxicant and less intoxicant (non-intoxicant, semi-intoxicant) categories of plant. The relationships of the four basic groups which may be perceived are shown in the figure below, and are labelled (1) to (4). An interpretation of the relationships of these four groupings to the concepts of Janischevsky and Lamarck is given in this diagram. It should be appreciated that the particular delimitations of groups that I have proposed within Cannabis are not coincident with those of previous or contemporary students, but in general represent an expansion and interpretation of vague concepts which have resulted from failure to survey the entire variational pattern in Cannabis. As has been stressed throughout this paper, all of these groups are best treated as constituent elements of Cannabis sativa L.
I wish to express my gratitude to Drs. A. Cronquist, J. McNeill, L. Weresub, and B. Baum for criticism of the manuscript. I alone am responsible for the interpretations and opinions expressed.