The greatest service which can be rendered to any country is to add a useful plant to its culture.
Although it is possible to breed Cannabis with limited success without any knowledge of the laws of
inheritance, the full potential of diligent breeding, and the line of action most likely to lead to success, is realized by breeders who have mastered a working knowledge of genetics.
As we know already, all information transmitted from generation to generation must be contained in
the pollen of the staminate parent and the ovule of the pistillate parent. Fertilization unites these two sets of genetic infor- mation, a seed forms, and a new generation is begun. Both pollen and ovules are known as gametes, and the trans- mitted units determining the expression of a character are known as genes. Individual plants have two identical sets of genes (2n) in every cell except the gametes, which through reduction division have only one set of genes (in). Upon fertilization one set from each parent combines to form a seed (2n). In Cannabis, the haploid (in) number of chromo- somes is 10 and the diploid (2n) number of chromosomes is 20. Each chromosome contains hundreds of genes, influ- encing every phase of the growth and development of the plant.
If cross-pollination of two plants with a shared genetic trait (or self-pollination of a hermaphrodite)
results in off- spring that all exhibit the same trait, and if all subsequent (inbred) generations also exhibit it, then we say that the strain (i.e., the line of offspring derived from common an- cestors) is true-breeding, or breeds true, for that trait. A strain may breed true for one or more traits while varying in other characteristics. For example, the traits of sweet aroma and early maturation may breed true, while off- spring vary in size and shape. For a strain to breed true for some trait, both of the gametes forming the offspring must have an identical complement of the genes that influence the expression of that trait. For example, in a strain that breeds true for webbed leaves, any gamete from any parent in that population will contain the gene for webbed leaves, which we will signify with the letter w. Since each gamete carries one-half (in) of the genetic complement of the offspring, it follows that upon fertilization both "leaf- shape" genes of the (2n) offspring will be w. That is, the offspring, like both parents, are ww. In turn, the offspring also breed true for webbed leaves because they have only w genes to pass on in their gametes.
On the other hand, when a cross produces offspring that do not breed true (i.e., the offspring do not all
re- semble their parents) we say the parents have genes that segregate or are hybrid. Just as a strain can breed true for one or more traits, it can also segregate for one or more traits; this is often seen. For example, consider a cross where some of the offspring have webbed leaves and some have normal
compound-pinnate leaves. (To continue our system of notation we will refer to the gametes of plants
with compound-pinnate leaves as W for that trait. Since these two genes both influence leaf shape, we
assume that they are related genes, hence the lower-case w and upper- case W notation instead of w for
webbed and possibly P for pinnate.) Since the gametes of a true-breeding strain must each have the same
genes for the given trait, it seems logi- cal that gametes which produce two types of offspring must have
genetically different parents.
Observation of many populations in which offspring differed in appearance from their parents led
Mendel to his theory of genetics. If like only sometimes produces like, then what are the rules which
govern the outcome of these crosses? Can we use these rules to predict the outcome of future crosses?
Assume that we separate two true-breeding popula- tions of Cannabis, one with webbed and one with
compound-pinnate leaf shapes. We know that all the gametes produced by the webbed-leaf parents will
contain genes for leaf-shape w and all gametes produced by the compound-pinnate individuals will have
W genes for leaf shape. (The offspring may differ in other characteristics, of course.)
If we make a cross with one parent from each of the true-breeding strains, we will find that 100% of the
off- apring are of the compound-pinnate leaf phenotype. (The expression of a trait in a plant or strain is known as the phenotype.) What happened to the genes for webbed leaves contained in the webbed leaf
parent? Since we know that there were just as many w genes as W genes combined in the offspring, the
W gene must mask the expression of the w gene. We term the W gene the dominant gene and say that
the trait of compound-pinnate leaves is dominant over the recessive trait of webbed leaves. This seems
logical since the normal phenotype in Cannabis has compound- pinnate leaves. It must be remembered,
however, that many useful traits that breed true are recessive. The true-breeding dominant or recessive
condition, WW or ww, is termed the homozygous condition; the segregating hybrid condition wW or
Ww is called heterozygous. When we cross two of the F1 (first filial generation) offspring resulting from
the initial cross of the ~1 (parental generation) we observe two types of offspring. The F2 generation
shows a ratio of approximately 3:1, three compound pinnate type-to-one webbed type. It should be
remembered that phenotype ratios are theoretical. The real results may vary from the expected ratios,
especially in small samples.
In this case, compound-pinnate leaf is dominant over webbed leaf, so whenever the genes w and W are
combined, the dominant trait W will be expressed in the phenotype. In the F2 generation only 25% of
the offspring are homo- zygous for W so only 25% are fixed for W. The w trait is only expressed in the
F2 generation and only when two w genes are combined to form a double-recessive, fixing the recessive
trait in 25% of the offspring. If compound-pinnate showed incomplete dominance over webbed, the
geno- types in this example would remain the same, but the phenotypes in the F1 generation would all
be intermediate types resembling both parents and the F2 phenotype ratio would be 1 compound-
pinnate :2 intermediate :1 webbed.
The explanation for the predictable ratios of offspring is simple and brings us to Mendel's first law, the first of the basic rules of heredity:
I. Each of the genes in a related pair segregate from each other during gamete formation.
A common technique used to deduce the genotype of the parents is the back-cross. This is done by
crossing one of the F1 progeny back to one of the true-breeding P1 parents. If the resulting ratio of
phenotypes is 1:1 (one heterozygous to one homozygous) it proves that the parents were indeed
homozygous dominant WW and homozygous-recessive ww.
The 1:1 ratio observed when back-crossing F1 to P1 and the 1:2:1 ratio observed in F1 to F1 crosses
are the two basic Mendelian ratios for the inheritance of one character controlled by one pair of genes.
The astute breeder uses these ratios to determine the genotype of the parental plants and the relevance of genotype to further breeding.
This simple example may be extended to include the inheritance of two or more unrelated pairs of
genes at a time. For instance we might consider the simultaneous inheritance of the gene pairs T (tall)/t
(short) and M (early maturation)/m (late maturation). This is termed a poly- hybrid instead of
monohybrid cross. Mendel's second law allows us to predict the outcome of polyhybrid crosses also:
II. Unrelated pairs of genes are inherited indepen- dently of each other.
If complete dominance is assumed for both pairs of genes, then the 16 possible F2 genotype
combinations will form 4 F2 phenotypes in a 9:3:3:1 ratio, the most frequent of which is the double-
dominant tall/early condition. In- complete dominance for both gene pairs would result in 9 F2
phenotypes in a 1:2:1:2:4:2:1:2:1 ratio, directly re- flecting the genotype ratio. A mixed dominance
condition would result in 6 F2 phenotypes in a 6:3:3:2:1:1 ratio. Thus, we see that a cross involving two
independently assorting pairs of genes results in a 9:3:3:1 Mendelian phenotype ratio only if dominance
is complete. This ratio may differ, depending on the dominance conditions present in the original gene
pairs. Also, two new phenotypes, tall/late and short/early, have been created in the F2 genera- tion; these phenotypes differ from both parents and grand- parents. This phenomenon is termed recombination and
explains the frequent observation that like begets like, but not exactly like.
A polyhybrid back-cross with two unrelated gene pairs exhibits a 1:1 ratio of phenotypes as in the
mono- hybrid back-cross. It should be noted that despite domi- nance influence, an F1 back-cross with
the P1 homozygous- recessive yields the homozygous-recessive phenotype short/late 25% of the time,
and by the same logic, a back- cross with the homozygous-dominant parent will yield the homozygous
dominant phenotype tall/early 25% of the time. Again, the back-cross proves invaluable in determin- ing
the F1 and P1 genotypes. Since all four phenotypes of the back-cross progeny contain at least one each
of both recessive genes or one each of both dominant genes, the back-cross phenotype is a direct
representation of the four possible gametes produced by the F1 hybrid.
So far we have discussed inheritance of traits con- trolled by discrete pairs of unrelated genes. Gene
inter- action is the control of a trait by two or more gene pairs. In this case genotype ratios will remain the same but phenotype ratios may be altered. Consider a hypothetical example where 2 dominant gene pairs Pp and Cc control late-season anthocyanin pigmentation (purple color) in Cannabis. If P is present alone, only the leaves of the plant (under the proper environmental stimulus) will exhibit accumulated anthocyanin pigment and turn a purple color. If C is present alone, the plant will remain green through-out its life cycle despite environmental conditions. If both are present, however, the calyxes of the plant will also ex- hibit accumulated anthocyanin and turn purple as the leaves do. Let us assume for now that this may be a desir- able trait in Cannabis flowers. What breeding techniques can be used to produce this trait?
First, two homozygous true-breeding ~1 types are crossed and the phenotype ratio of the F1 offspring
The phenotypes of the F2 progeny show a slightly altered phenotype ratio of 9:3:4 instead of the
expected 9:3:3:1 for independently assorting traits. If P and C must both be present for any anthocyanin
pigmentation in leaves or calyxes, then an even more distorted phenotype ratio of 9:7 will appear.
Two gene pairs may interact in varying ways to pro- duce varying phenotype ratios. Suddenly, the
simple laws of inheritance have become more complex, but the data may still be interpreted.
Summary of Essential Points of Breeding
1 - The genotypes of plants are controlled by genes which are passed on unchanged from generation to
2 - Genes occur in pairs, one from the gamete of the staminate parent and one from the gamete of the
3 - When the members of a gene pair differ in their effect upon phenotype, the plant is termed hybrid or
4 - When the members of a pair of genes are equal in their effect upon phenotype, then they are termed
true- breeding or homozygous.
5 - Pairs of genes controlling different phenotypic traits are (usually) inherited independently.
6 - Dominance relations and gene interaction can alter the phenotypic ratios of the F1, F2, and
Polyploidy is the condition of multiple sets of chro- mosomes within one cell. Cannabis has 20
chromosomes in the vegetative diploid (2n) condition. Triploid (3n) and tetraploid (4n) individuals have
three or four sets of chro- mosomes and are termed polyploids. It is believed that the haploid condition
of 10 chromosomes was likely derived by reduction from a higher (polyploid) ancestral number (Lewis,
W. H. 1980). Polyploidy has not been shown to occur naturally in Cannabis; however, it may be induced
artificially with colchicine treatments. Colchicine is a poi- sonous compound extracted from the roots of
certain Colchicum species; it inhibits chromosome segregation to daughter cells and cell wall formation,
resulting in larger than average daughter cells with multiple chromosome sets. The studies of H. E.
Warmke et al. (1942-1944) seem to indicate that colchicine raised drug levels in Cannabis. It is
unfortunate that Warmke was unaware of the actual psychoactive ingredients of Cannabis and was
therefore unable to extract THC. His crude acetone extract and archaic techniques of bioassay using
killifish and small freshwater crustaceans are far from conclusive. He was, however, able to produce
both triploid and tetraploid strains of Cannabis with up to twice the potency of dip- bid strains (in their ability to kill small aquatic organisms). The aim of his research was to "produce a strain of hemp with materially reduced marijuana content" and his results indicated that polyploidy raised the potency of Cannabis without any apparent increase in fiber quality or yield.
Warmke's work with polyploids shed light on the nature of sexual determination in Cannabis. He also
illus- trated that potency is genetically determined by creating a lower potency strain of hemp through
selective breeding with low potency parents.
More recent research by A. I. Zhatov (1979) with fiber Cannabis showed that some economically
valuable traits such as fiber quantity may be improved through polyploidy. Polyploids require more
water and are usually more sensitive to changes in environment. Vegetative growth cycles are extended
by up to 30-40% in polyploids. An extended vegetative period could delay the flowering of polyploid
drug strains and interfere with the formation of floral clusters. It would be difficult to determine if
canna- binoid levels had been raised by polyploidy if polyploid plants were not able to mature fully in
the favorable part of the season when cannabinoid production is promoted by plentiful light and warm
temperatures. Greenhouses and artificial lighting can be used to extend the season and test polyploid
The height of tetraploid (4n) Cannabis in these exper- iments often exceeded the height of the original
diploid plants by 25-30%. Tetraploids were intensely colored, with dark green leaves and stems and a
well developed gross phenotype. Increased height and vigorous growth, as a rule, vanish in subsequent
generations. Tetraploid plants often revert back to the diploid condition, making it diffi- cult to support tetraploid populations. Frequent tests are performed to determine if ploidy is changing.
Triploid (3n) strains were formed with great difficulty by crossing artificially created tetraploids (4n)
with dip- bids (2n). Triploids proved to be inferior to both diploids and tetraploids in many cases.
De Pasquale et al. (1979) conducted experiments with Cannabis which was treated with 0.25% and
0.50% solu- tions of colchicine at the primary meristem seven days after generation. Treated plants were
slightly taller and possessed slightly larger leaves than the controls, Anoma- lies in leaf growth occurred in 20% and 39%, respectively, of the surviving treated plants. In the first group (0.25%) cannabinoid levels were highest in the plants without anomalies, and in the second group (0.50%) cannabinoid levels were highest in plants with anomalies, Overall, treated plants showed a 166-250% increase in THC with respect to controls and a decrease of CBD (30-33%) and CBN (39-65%). CBD (cannabidiol) and CBN (cannabinol) are cannabinoids involved in the biosynthesis and degrada- tion of THC. THC levels in the control plants were very low (less than 1%).
Possibly colchicine or the resulting polyploidy interferes with cannabinoid biogenesis to favor THC. In treated plants with deformed leaf lamina, 90% of the cells are tetraploid (4n 40) and 10% diploid (2n 20). In treated plants without deformed lamina a few cells are tetraploid and the remainder are triploid or diploid.
The transformation of diploid plant The transformation of diploid plants to the tetraploid level
inevitably results in the formation of a few plants with an unbalanced set of chromosomes (2n + 1, 2n -
1, etc.). These plants are called aneuploids. Aneuploids are inferior to polyploids in every economic
respect. Aneu- ploid Cannabis is characterized by extremely small seeds. The weight of 1,000 seeds
ranges from 7 to 9 grams (1/4 to 1/3 ounce). Under natural conditions diploid plants do not have such
small seeds and average 14-19 grams (1/2- 2/3 ounce) per 1,000 (Zhatov 1979).
Once again, little emphasis has been placed on the relationship between flower or resin production and
poly- ploidy. Further research to determine the effect of poly- ploidy on these and other economically
valuable traits of Cannabis is needed.
Colchicine is sold by laboratory supply houses, and breeders have used it to induce polyploldy in
Cannabis. However, colchicine is poisonous, so special care is exer- cised by the breeder in any use of it.
Many clandestine cultivators have started polyploid strains with colchicine. Except for changes in leaf
shape and phyllotaxy, no out- standing characteristics have developed in these strains and potency seems
unaffected. However, none of the strains have been examined to determine if they are actually poly ploid
or if they were merely treated with colchicine to no effect. Seed treatment is the most effective and
safest way to apply colchicine. * In this way, the entire plant growing from a colchicine-treated seed
could be polyploid and if any colchicine exists at the end of the growing season the amount would be
infinitesimal. Colchicine is nearly always lethal to Cannabis seeds, and in the treatment there is a very
fine line between polyploidy and death. In other words, if 100 viable seeds are treated with colchicine
and 40 of them germinate it is unlikely that the treatment in- duced polyploidy in any of the survivors.
On the other hand, if 1,000 viable treated seeds give rise to 3 seedlings, the chances are better that they are polyploid since the treatment killed all of the seeds but those three. It is still necessary to determine if the offspring are actually poly- ploid by microscopic examination.
The work of Menzel (1964) presents us with a crude map of the chromosomes of Cannabis,
Chromosomes 2-6 and 9 are distinguished by the length of each arm. Chromo- some 1 is distinguished
by a large knob on one end and a dark chromomere 1 micron from the knob. Chromosome 7 is
extremely short and dense, and chromosome 8 is assumed to be the sex chromosome. In the future,
chromosome *The word "safest" is used here as a relative term. Coichicine has received recent media
attention as a dangerous poison and while these accounts are probably a bit too lurid, the real dangers of expo- iure to coichicine have not been fully researched.
The possibility of bodily harm exists and this is multiplied when breeders inexperi- enced in handling toxins use colchicine. Seed treatment might be safer than spraying a grown plant but the safest method of all is to not use colchicine. mapping will enable us to picture the location of the genes influencing the phenotype of Cannabis. This will enable geneticists to determine and manipulate the important characteristics contained in the gene pool. For each trait the number of genes in control will be known, which chromosomes carry them, and where they are located along those chromosomes.
All of the Cannabis grown in North America today originated in foreign lands. The diligence of our
ancestors in their collection and sowing of seeds from superior plants, together with the forces of natural selection, have worked to create native strains with localized characteris- tics of resistance to pests, diseases, and weather conditions. In other words, they are adapted to particular niches in the ecosystem. This genetic diversity is nature's way of pro- tecting a species. There is hardly a plant more flexible than Cannabis. As climate, diseases, and pests change, the strain evolves and selects new defenses, programmed into the genetic orders contained in each generation of seeds. Through the importation in recent times of fiber and drug Cannabis, a vast pool of genetic material has appeared in North America. Original fiber strains have escaped and become acclimatized (adapted to the environment), while domestic drug strains (from imported seeds) have, unfortunately, hybrid- ized and acclimatized
randomly, until many of the fine gene combinations of imported Cannabis have been lost.
Changes in agricultural techniques brought on by technological pressure, greed, and full-scale
eradication programs have altered the selective pressures influencing Cannabis genetics. Large
shipments of inferior Cannabis containing poorly selected seeds are appearing in North America and
elsewhere, the result of attempts by growers and smugglers to supply an ever increasing market for mari-
juana. Older varieties of Cannabis, associated with long- standing cultural patterns, may contain genes
not found in the newer commercial varieties. As these older varieties and their corresponding cultures
become extinct, this genetic information could be lost forever. The increasing popular- ity of Cannabis
and the requirements of agricultural technology will call for uniform hybrid races that are likely to displace primitive populations worldwide.
Limitation of genetic diversity is certain to result from concerted inbreeding for uniformity. Should
inbred Cannabis be attacked by some previously unknown pest or disease, this genetic uniformity could
prove disastrous due to potentially resistant diverse genotypes having been dropped from the population.
If this genetic complement of resistance cannot be reclaimed from primitive parental material, resistance
cannot be introduced into the ravaged population. There may also be currently unrecognized favorable
traits which could be irretrievably dropped from the Cannabis gene pool. Human intervention can create
new phenotypes by selecting and recombining existing genetic variety, but only nature can create variety
in the gene pool itself, through the slow process of random mutation.
This does not mean that importation of seed and selective hybridization are always detrimental. Indeed
these principles are often the key to crop improvement, but only when applied knowledgeably and
cautiously. The rapid search for improvements must not jeopardize the pool of original genetic
information on which adaptation relies. At this time, the future of Cannabis lies in govern- ment and
clandestine collections. These collections are often inadequate, poorly selected and badly maintained.
Indeed, the United Nations Cannabis collection used as the primary seed stock for worldwide governmental research is depleted and spoiled.
Several steps must be taken to preserve our vanishing genetic resources, and action must be immediate:
• Seeds and pollen should be collected directly from reliable and knowledgeable sources. Government
seizures and smuggled shipments are seldom reliable seed sources. The characteristics of both parents
must be known; conse- quently, mixed bales of randomly pollinated marijuana are not suitable seed
sources, even if the exact origin of the sample is certain. Direct contact should be made with the farmer-
breeder responsible for carrying on the breeding traditions that have produced the sample. Accurate
records of every possible parameter of growth must be kept with carefully stored triplicate sets of seeds.
• Since Cannabis seeds do not remain viable forever, even under the best storage conditions, seed
samples should he replenished every third year. Collections should be planted in conditions as similar as
possible to their original niche and allowed to reproduce freely to minimize natural and artificial
selection of genes and ensure the preservation of the entire gene pool. Half of the original seed
collection should be retained until the viability of further generations is confirmed, and to provide
parental material for compari- son and back-crossing. Phenotypic data about these subse- quent
generations should be carefully recorded to aid in understanding the genotypes contained in the
collection. Favorable traits of each strain should be characterized and catalogued.
• It is possible that in the future, Cannabis cultiva- tion for resale, or even personal use, may be legal
but only for approved, patented strains. Special caution would be needed to preserve variety in the gene
pool should the patenting of Cannabis strains become a reality.
• Favorable traits must be carefully integrated into existing strains.
The task outlined above is not an easy one, given the current legal restrictions on the collection of
Cannabis seed. In spite of this, the conscientious cultivator is making a contribution toward preserving
and improving the genet- ics of this interesting plant.
Even if a grower has no desire to attempt crop im- provement, successful strains have to be protected
so they do not degenerate and can be reproduced if lost. Left to the selective pressures of an introduced
environment, most drug strains will degenerate and lose potency as they accli- matize to the new
conditions. Let me cite an example of a typical grower with good intentions.
A grower in northern latitudes selected an ideal spot to grow a crop and prepared the soil well. Seeds
were selected from the best floral clusters of several strains avail- able over the past few years, both
imported and domestic. Nearly all of the staminate plants were removed as they matured and a nearly
seedless crop of beautiful plants re- sulted. After careful consideration, the few seeds from accidental
pollination of the best flowers were kept for the following season, These seeds produced even bigger and
better plants than the year before and seed collection was performed as before. The third season, most of
the plants were not as large or desirable as the second season, but there were many good individuals.
Seed collection and cul- tivation the fourth season resulted in plants inferior even to the first crop, and this trend continued year after year. What went wrong? The grower collected seed from the best plants
each year and grew them under the same conditions. The crop improved the first year. Why did the
This example illustrates the unconscious selection for undesirable traits. The hypothetical cultivator
began well by selecting the best seeds available and growing them properly. The seeds selected for the
second season resulted from random hybrid pollinations by early-flowering or overlooked staminate
plants and by hermaphrodite pistil- late plants. Many of these random pollen-parents may be undesirable
for breeding since they may pass on tendencies toward premature maturation, retarded maturation, or
hermaphrodism. However, the collected hybrid seeds pro- duce, on the average, larger and more
desirable offspring than the first season. This condition is called hybrid vigor and results from the hybrid crossing of two diverse gene pools. The tendency is for many of the dominant characteristics from both parents to be transmitted to the F1 off- spring, resulting in particularly large and vigorous plants.
This increased vigor due to recombination of dominant genes often raises the cannabinoid level of the F1
offspring, but hybridization also opens up the possibility that unde- sirable (usually recessive) genes may form pairs and express their characteristics in the F2 offspring. Hybrid vigor may also mask inferior
qualities due to abnormally rapid growth. During the second season, random pollinations again
accounted for a few seeds and these were collected. This selection draws on a huge gene pool and the
possible F2 combinations are tremendous.
By the third season the gene pool is tending toward early-maturing plants that are acclimatized to their new conditions instead of the drug- producing conditions of their native environment. These acclimatized members of the third crop have a higher chance of maturing viable seeds than the parental types, and random pollinations will again increase the numbers of acclimatized individuals, and thereby increase the chance that undesirable characteristics associated with acclimatization will be transmitted to the next F2 generation.
This effect is compounded from generation to generation and finally results in a fully acclimatized weed strain of little drug value. With some care the breeder can avoid these hidden dangers of unconscious selection. Definite goals are vital to progress in breeding Cannabis. What qualities are desired in a strain that it does not already exhibit? What character- istics does a strain exhibit that are unfavorable and should be bred out?
Answers to these questions suggest goals for breeding. In addition to a basic knowledge of Cannabis
botany, propagation, and genetics, the successful breeder also becomes aware of the most minute
differences and similarities in phenotype. A sensitive rapport is established between breeder and plants
and at the same time strict guidelines are followed. A simplified explanation of the time-tested principles of plant breeding shows how this works in practice.
Selection is the first and most important step in the breeding of any plant. The work of the great
breeder and plant wizard Luther Burbank stands as a beacon to breeders of exotic strains. His success in
improving hundreds of flower, fruit, and vegetable crops was the result of his meticulous selection of
parents from hundreds of thou- sands of seedlings and adults from the world over.
Bear in mind that in the production of any new plant, selection plays the all-important part. First, one
must get clearly in mind the kind of plant he wants, then breed and select to that end, always choosing
through a series of years the plants which are approaching nearest the ideal, and rejecting all others. -
Luther Burbank (in James, 1964)
Proper selection of prospective parents is only possible if the breeder is familiar with the variable
characteristics of Cannabis that may be genetically controlled, has a way to accurately measure these
variations, and has established goals for improving these characteristics by selective breed- ing. A
detailed list of variable traits of Cannabis, including parameters of variation for each trait and comments pertaining to selective breeding for or against it, are found at the end of this chapter. By selecting against unfavorable traits while selecting for favorable ones, the unconscious breeding of poor strains is avoided.
The most important part of Burbank's message on selection tells breeders to choose the plants "which
are ap- proaching nearest the ideal," and REJECT ALL OTHERS! Random pollinations do not allow the
control needed to reject the undesirable parents. Any staminate plant that survives detection and roguing
(removal from the popula- tion), or any stray staminate branch on a pistillate her- maphrodite may
become a pollen parent for the next gen- eration. Pollination must be controlled so that only the pollen-
and seed-parents that have been carefully selected for favorable traits will give rise to the next
Selection is greatly improved if one has a large sample to choose from! The best plant picked from a
group of 10 has far less chance of being significantly different from its fellow seedlings than the best
plant selected from a sample of 100,000. Burbank often made his initial selections of parents from
samples of up to 500,000 seedlings. Difficul- ties arise for many breeders because they lack the space to
keep enough examples of each strain to allow a significant selection.
A Cannabis breeder's goals are restricted by the amount of space available. Formulating a well defined goal lowers the number of individuals needed to perform effective crosses. Another technique used by breeders since the time of Burbank is to make early selections. Seedling plants take up much less space than adults. Thousands of seeds can be germinated in a flat. A flat takes up the same space as a hundred 10-centimeter (4-inch) sprouts or six- teen 30-centimeter (12-inch) seedlings or one 60-centimeter (24-inch) juvenile. An adult plant can easily take up as much space as a hundred flats. Simple arithmetic shows that as many as 10,000 sprouts can be screened in the space required by each mature plant, provided enough seeds are available. Seeds of rare strains are quite valuable and exotic; however, careful selection applied to thousands of individuals, even of such common strains as those from Colombia or Mexico, may produce better offspring than plants from a rare strain where there is little or no oppor- tunity for selection after germination. This does not mean that rare strains are not valuable, but careful selection is even more important to successful breeding. The random pollinations that produce the seeds in most imported marijuana assure a hybrid condition which results in great seed- ling diversity. Distinctive plants are not hard to discover if the seedling sample is large enough.
Traits considered desirable when breeding Cannabis often involve the yield and quality of the final
product, but these characteristics can only be accurately measured after the plant has been harvested and
long after it is possible to select or breed it. Early seedling selection, therefore, only works for the most basic traits. These are selected first, and later selections focus on the most desirable characteristics exhibited by juvenile or adult plants. Early traits often give clues to mature phenotypic expression, and criteria for effective early seedling selection are easy to establish. As an example, particularly tall and thin seedlings might prove to be good parents for pulp or fiber production, while seedlings of short internode length and compound branching may be more suitable for flower production. However, many important traits to be selected for in Cannabis floral clusters cannot be judged until long after the parents are gone, so many crosses are made early and selection of seeds made at a later date.
Hybridization is the process of mixing differing gene pools to produce offspring of great genetic
variation from which distinctive individuals can be selected. The wind performs random hybridization in
nature. Under cultivation, breeders take over to produce specific, controlled hybrids. This process is
also known as cross-pollination, cross-fertilization, or simply crossing. If seeds result, they will produce hybrid offspring exhibiting some characteristics from each parent.
Large amounts of hybrid seed are most easily produced by planting two strains side by side, removing
the staininate plants of the seed strain, and allowing nature to take its course. Pollen- or seed-sterile
strains could be devel oped for the production of large amounts of hybrid seed without the labor of
thinning; however, genes for sterility are rare. It is important to remember that parental weak- nesses are transmitted to offspring as well as strengths. Because of this, the most vigorous, healthy plants are al
ways used for hybrid crosses.
Also, sports (plants or parts of plants carrying and expressing spontaneous mutations) most easily
transmit mutant genes to the offspring if they are used as pollen parents. If the parents represent diverse
gene pools, hybrid vigor results, because dominant genes tend to carry valu- able traits and the differing
dominant genes inherited from each parent mask recessive traits inherited from the other. This gives rise
to particularly large, healthy individuals. To increase hybrid vigor in offspring, parents of different geo-
graphic origins are selected since they will probably repre- sent more diverse gene pools.
Occasionally hybrid offspring will prove inferior to both parents, but the first generation may still
contain recessive genes for a favorable characteristic seen in a par- ent if the parent was homozygous for
that trait. First gen- eration (F1) hybrids are therefore inbred to allow recessive genes to recombine and
express the desired parental trait. Many breeders stop with the first cross and never realize the genetic
potential of their strain. They fail to produce an F2 generation by crossing or self-pollinating F1
offspring. Since most domestic Cannabis strains are F1 hybrids for many characteristics, great diversity
and recessive recombi- nation can result from inbreeding domestic hybrid strains. In this way the
breeding of the F1 hybrids has afready been accomplished, and a year is saved by going directly to F2
hybrids. These F2 hybrids are more likely to express reces- sive parental traits. From the F2 hybrid
generation selec- tions can be made for parents which are used to start new true-breeding strains. Indeed,
F2 hybrids might appear with more extreme characteristics than either of the P~ parents. (For example,
P1 high-THC X P1 low-THC yields F1 hybrids of intermediate THC content. Selfing the F1 yields F2
hy- brids, of both P1 [high and low THC] phenotypes, inter- mediate F1 phenotypes, and extra-high
THC as well as extra-low THC phenotypes.)
Also, as a result of gene recombination, F1 hybrids are not true-breeding and must be reproduced from
the original parental strains. When breeders create hybrids they try to produce enough seeds to last for
several successive years of cultivation, After initial field tests, undesirable hybrid seeds are destroyed
and desirable hybrid seeds stored for later use. If hybrids are to be reproduced, a clone is saved from
each parental plant to preserve original paren- tal genes.
Back-crossing is another technique used to produce offspring with reinforced parental characteristics.
In this case, a cross is made between one of the F~ or subsequent offspring and either of the parents
expressing the desired trait. Once again this provides a chance for recombination and possible
expression of the selected parental trait. Back- crossing is a valuable way of producing new strains, but it
is often difficult because Cannabis is an annual, so special care is taken to save parental stock for back-
crossing the following year. Indoor lighting or greenhouses can be used to protect breeding stock from
winter weather. In tropical areas plants may live outside all year. In addition to saving particular parents,
a successful breeder always saves many seeds from the original P1 group that produced the valuable
characteristic so that other P1 plants also exhibiting the characteristic can be grown and selected for
back-crossing at a later time.
Several types of breeding are summarized as follows:
1 - Crossing two varieties having outstanding qualities (hybridization).
2 - Crossing individuals from the F1 generation or selfing F1 individuals to realize the possibilities of
the ori- ginal cross (differentiation).
3 - Back crossing to establish original parental types.
4 - Crossing two similar true-breeding (homozygous) varieties to preserve a mutual trait and restore
It should be noted that a hybrid plant is not usually hybrid for all characteristics nor does a true-
breeding strain breed true for all characteristics. When discussing crosses, we are talking about the
inheritance of one or a few traits only. The strain may be true-breeding for only a few traits, hybrid for
the rest. Monohybrid crosses involve one trait, dihybrid crosses involve two traits, and so forth. Plants
have certain limits of growth, and breeding can only pro- duce a plant that is an expression of some gene
already present in the total gene pool. Nothing is actually created by breeding; it is merely the
recombination of existing genes into new genotypes. But the possibilities of recombi- nation are nearly
The most common use of hybridization is to cross two outstanding varieties. Hybrids can be produced
by crossing selected individuals from different high-potency strains of different origins, such as
Thailand and Mexico. These two parents may share only the characteristic of high psycho- activity and
differ in nearly every other respect. From this great exchange of genes many phenotypes may appear in
the F2 generation. From these offspring the breeder selects individuals that express the best
characteristics of the par- ents. As an example, consider some of the offspring from the P1 (parental)
cross: Mexico X Thailand. In this case, genes for high drug content are selected from both parents while
other desirable characteristics can be selected from either one. Genes for large stature and early
maturation are selected from the Mexican seed-parent, and genes for large calyx size and sweet floral
aroma are selected from the Thai pollen parent. Many of the F1 offspring exhibit several of the desired
characteristics. To further promote gene segregation, the plants most nearly approaching the ideal are
crossed among themselves. The F2 generation is a great source of variation and recessive expression. In
the F2 generation there are several individuals out of many that exhibit all five of the selected
characteristics. Now the process of inbreeding begins, using the desirable F2 parents.
If possible, two or more separate lines are started, never allowing them to interbreed. In this case one
accept- able staminate plant is selected along with two pistillate plants (or vice versa). Crosses between
the pollen parent and the two seed parents result in two lines of inheritance with slightly differing
genetics, but each expressing the desired characteristics. Each generation will produce new, more
If two inbred strains are crossed, F1 hybrids will be less variable than if two hybrid strains are crossed.
This comes from limiting the diversity of the gene pools in the two strains to be hybridized through
previous inbreeding. Further independent selection and inbreeding of the best plants for several
generations will establish two strains which are true-breeding for all the originally selected traits. This
means that all the offspring from any parents in the strain will give rise to seedlings which all exhibit the
selected traits. Successive inbreeding may by this time have resulted in steady decline in the vigor of the
When lack of vigor interferes with selecting pheno- types for size and hardiness, the two separately
selected strains can then be interbred to recombine nonselected genes and restore vigor. This will
probably not interfere with breeding for the selected traits unless two different gene systems control the
same trait in the two separate lines, and this is highly unlikely. Now the breeder has pro- duced a hybrid
strain that breeds true for large size, early maturation, large sweet-smelling calyxes, and high THC level.
The goal has been reached!
Wind pollination and dioecious sexuality favor a heter- ozygous gene pool in Cannabis. Through
Anbreeding, hybrids are adapted from a heterozygous gene pool to a homozygous gene pool, providing
the genetic stability needed to create true-breeding strains. Establishing pure strains enables the breeder
to make hybrid crosses with a better chance of predicting the outcome. Hybrids can be created that are
not reproducible in the F2 generation. Commercial strains of seeds could be developed that would have
to be purchased each year, because the F1 hybrids of two pure-bred lines do not breed true. Thus, a seed
breeder can protect the investment in the results of breeding, since it would be nearly impossible to
reproduce the parents from F2 seeds.
At this time it seems unlikely that a plant patent would be awarded for a pure-breeding strain of drug
Cannabis. In the future, however, with the legalization of cultivation, it is a certainty that corporations with the time, space, and money to produce pure and hybrid strains of Cannabis will apply for patents. It may be legal to grow only certain patented strains produced by large seed companies. Will this be how government and industry combine to control the quality and quantity of "drug" Cannabis?
Much of the breeding effort of North American cultivators is concerned with acclimatizing high-THC
strains of equatorial origin to the climate of their growing area while preserving potency. Late-maturing, slow, and irregularly flowering strains like those of Thailand have difficulty maturing in many parts of North America. Even ~:n a green- house, it may not be possible to mature plants to their full native potential.
To develop an early-maturing and rapidly flowering 8train, a breeder may hybridize as in the previous
example. However, if it is important to preserve unique imported genetics, hybridizing may be
inadvisable. Alternatively, a pure cross is made between two or more Thai plants that most closely
approach the ideal in blooming early. At this point the breeder may ignore many other traits and aim at
breeding an earlier-maturing variety of a pure Thai strain. This strain may still mature considerably later than is ideal for the particular location unless selective pressure is ex- erted. If further crosses are made with several individuals that satisfy other criteria such as high THC content, these may be used to develop another pure Thai strain of high THC content.
After these true-breeding lines have been established, a dihybrid pure cross can be made in an attempt to produce an F1 generation containing early- maturing, high-THC strains of pure Thai genetics, in other words, an acclimatized drug strain.
Crosses made without a clear goal in mind lead to strains that acclimatize while losing many favorable
characteristics. A successful breeder is careful not to overlook a characteristic that may prove useful. It
is imperative that original imported Cannabis genetics be preserved intact to protect the species from
loss of genetic variety through excessive hybridization. A currently unrecognized gene may be
responsible for controlling resistance to a pest or disease, and it may only be possible to breed for this gene by back- crossing existing strains to original parental gene pools.
Once pure breeding lines have been established, plant breeders classify and statistically analyze the
offspring to determine the patterns of inheritance for that trait. This is the system used by Gregor
Mendel to formulate the basic laws of inheritance and aid the modern breeder in predict- ing the
outcome of crosses
1 - Two pure lines of Cannabis that differ in a particular trait are located.
2 - These two pure-breeding lines are crossed to produce an F1 generation.
3 - The F1 generation is inbred.
4 - The offspring of the F1 and F2 generations are classified with regard to the trait being studied.
5 - The results are analyzed statistically.
6 - The results are compared to known patterns of inheritance so the nature of the genes being selected
for can be characterized.
Fixing traits (producing homozygous offspring) in Cannabis strains is more difficult than it is in many
other flowering plants. With monoecious strains or hermaphro- dites it is possible to fix traits by self-
pollinating an individ- ual exhibiting favorable traits. In this case one plant acts as both mother and
father. However, most strains of Cannabis are dioecious, and unless hermaphroditic reactions can be
induced, another parent exhibiting the trait is required to fix the trait. If this is not possible, the unique individual may be crossed with a plant not exhibiting the trait, inbred in the F1 generation, and
selections of parents exhibiting the favorable trait made from the F2 generation, but this is very difficult.
If a trait is needed for development of a dioecious strain it might first be discovered in a monoecious
strain and then fixed through selfing and selecting homozygous offspring. Dioecious individuals can
then be selected from the monoecious population and these individuals crossed to breed out monoecism
in subsequent generations.
Galoch (1978) indicated that gibberellic acid (GA3) promoted stamen production while indoleacetic
acid (IAA), ethrel, and kinetin promoted pistil production in prefloral dioecious Cannabis. Sex alteration has several useful applications. Most importantly, if only one parent expressing a desirable trait can be found, it is difficult to perform a cross unless it happens to be a hermaphrodite plant. Hor- mones might be used to change the sex of a cutting from the desirable plant, and this cutting used to mate with it. This is most easily accomplished by changing a pistillate cutting to a staminate (pollen) parent, using a spray of 100 ppm gibberellic acid in water each day for five consecutive days. Within two weeks staminateflowers may appear. Pollen can then be collected for selfing with the original pistillate parent. Offspring from the cross should also be mostly pistillate since the breeder is selfing for pistillate sexuality. Staminate parents reversed to pistillate floral production make inferior seed-parents since few pistillate flowers and seeds are formed.
If entire crops could be manipulated early in life to produce all pistillate or staminate plants, seed
production and seedless drug Cannabis production would be greatly facilitated.
Sex reversal for breeding can also be accomplished by mutilation and by photoperiod alteration. A well-
rooted, flourishing cutting from the parent plant is pruned back to 25% of its original size and stripped of all its remaining flowers. New growth will appear within a few days, and several flowers of reversed sexual type often appear. Flowers of the unwanted sex are removed until the cutting is needed for fertilization. Extremely short light cycles (6-8 hour photoperiod) can also cause sex reversal. How- ever, this process takes longer and is much more difficult to perform in the field.
Genotype and Phenotype Ratios
It must be remembered, in attempting to fix favorable characteristics, that a monohybrid cross gives
rise to four possible recombinant genotypes, a dihybrid cross gives rise to 16 possible recombinant
genotypes, and so forth.
Phenotype and genotype ratios are probabilistic. If recessive genes are desired for three traits it is not effective to raise only 64 offspring and count on getting one homo- zygous recessive individual. To
increase the probability of success it is better to raise hundreds of offspring, choosing only the best
homozygous recessive individuals as future parents. All laws of inheritance are based on chance and
offspring may not approach predicted ratios until many more have been phenotypically characterize
and grouped than the theoretical minimums.
The genotype of each individual is expressed by a mosaic of thousands of subtle overlapping traits. It is
the sum total of these traits that determines the general pheno- type of an individual. It is often difficult to determine if the characteristic being selected is one trait or the blending of several traits and whether these traits are controlled by one or several pairs of genes. It often makes little difference that a breeder does not have plants that are proven to breed true. Breeding goals can still be established.
The selfing of F1 hybrids will often give rise to the variation needed in the F2 generation for selecting parents for subsequent generations, even if the characteristics of the original parents of the F1 hybrid are not known. It is in the following generations that fixed characteristics appear and the breeding of pure strains can begin. By selecting and crossing individuals that most nearly approach the ideal described by the breeding goals, the variety can be continuously improved even if the exact patterns of inheritance are never deter- mined. Complementary traits are eventually combined into one line whose seeds reproduce the favorable parental traits. Inbreeding strains also allows weak recessive traits to express themselves and these abnormalities must be diligently removed from the breeding population. After five or six generations, strains become amazingly uniform. Vigor is occasionally restored by crossing with other lines or by backcrossing.
Parental plants are selected which most nearly approach the ideal. If a desirable trait is not expressed by the parent, it is much less likely to appear in the offspring. It is imperative that desirable characteristics be hereditary and not primarily the result of environment and cultivation. Acquired traits are not hereditary and cannot be made hereditary. Breeding for as few traits as possible at one time greatly increases the chance of success. In addition to the specific traits chosen as the aims of breeding, parents are selected which possess other generally desirable traits such as vigor and size.
Determinations of dominance and recessiveness can only be made by observing the outcome of many crosses, although wild traits often tend to be dominant. This is one of the keys to adaptive survival. However, all the possible combinations will appear in the F2 generation if it is large enough, regardless of dominance.
Now, after further simplifying this wonderful system of inheritance, there are additional exceptions to
the rules which must be explored. In some cases, a pair of genes may control a trait but a second or third pair of genes is needed to express this trait. This is known as gene inter- action. No particular genetic attribute in which we may be interested is totally isolated from other genes and the effects of
environment. Genes are occasionally transferred in groups instead of assorting independently. This is
known as gene linkage, These genes are spaced along the same chromosome and may or may not control
the same trait. The result of linkage might be that one trait cannot be inherited without another.
At times, traits are associated with the X and Y sex chromosomes and they may be limited to expression in
only one sex (sex linkage). Crossing over also interferes with the analysis of crosses. Crossing over is
the exchanging of entire pieces of genetic material between two chromosomes. This can result in two
genes that are nor- mally linked appearing on separate chromosomes where they will be independently
inherited. All of these processes can cause crosses to deviate from the expected Mendelian outcome.
Chance is a major factor in breeding Cannabis, or any introduced plant, and the more crosses a breeder
attempts the higher are the chances of success. Variate, isolate, intermate, evaluate, multiplicate, and disseminate are the key words in plant improvement.
A plant breeder begins by producing or collecting various prospective parents from which
the most desirable ones are selected and isolated. Intermating of the select parents results in offspring
which must be evaluated for favorable characteristics. If evaluation indicates that the offspring are not
improved, then the process is repeated. Improved off- spring are multiplied and disseminated for
commercial use. Further evaluation in the field is necessary to check for uniformity and to choose
parents for further intermating. This cyclic approach provides a balanced system of plant improvement.
The basic nature of Cannabis makes it challenging to breed. Wind pollination and dioecious sexuality,
which account for the great adaptability in Cannabis, cause many problems in breeding, but none of
these are insurmountable.
Developing a knowledge and feel for the plant is more important than memorizing Mendelian ratios. The words of the great Luther Burbank say it well, "Heredity is indelibly fixed by repetition."
The first set of traits concerns Cannabis plants as a whole while the remainder concern the qualities of
seedlings, leaves, fibers, and flowers. Finally a list of various Cannabis strains is provided along with
specific characteristics. Following this order, basic and then specific selections of favorable
characteristics can be made. List of Favorable Traits of Cannabis in Which Variation Occurs:
1. General Traits
a) Size and Yield
e) Disease and Pest Resistance
g) Root Production
2. Seedling Traits
3. Leaf Traits
4. Fiber Traits
5. Floral Traits
c) Calyx Size
e) Cannabinoid Level
f) Taste and Aroma
g) Persistence of Aromatic Principles and Cannabinoids
h) Trichome Type
i) Resin Quantity and Quality
j) Resin Tenacity
k) Drying and Curing Rate
l) Ease of Manicuring
m) Seed Characteristics
q) Cannabinoid Profile
CONTINUED with 6. Gross Phenotypes of Cannabis Strains