Industrial Crops and Products 68 (2015) 32–41

Contents lists available at ScienceDirect

Industrial Crops and Products

journal homepage: www.elsevier.com/locate/indcrop

New developments in fiber hemp (Cannabis sativa L.) breeding

Elma M.J. Salentijn a , Qingying Zhang b , Stefano Amaducci c ,
Ming Yang b , Luisa M. Trindade a,∗
a
Wageningen UR Plant Breeding, Wageningen University and Research Centre, P.O. Box 386, 6700 AJ Wageningen, The Netherlands
b
Industrial Crops Research Institute, Yunnan Academy of Agricultural Sciences, Kunming, PR China
c
Istituto di Agronomia, Genetica e Coltivazioni erbacee, Facoltà di Scienze Agrarie, Alimentari e Ambientali, Università Cattolica del Sacro Cuore, Via Emilia
Parmense, 84, 29122 Piacenza, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Fiber hemp (Cannabis sativa L.) is a sustainable and high yielding industrial crop that can help to meet
Received 22 April 2014 the high global demand for fibers. Hemp can be grown for fiber, seeds, and/or for dual purpose in a
Received in revised form 24 July 2014 wide range of geographic zones and climates. Currently the main hemp producing regions in the world
Accepted 8 August 2014
are China, Europe, and Canada. The number of new cultivars developed for each of these regions has
Available online 2 September 2014
gradually increased, with each region producing its own typical hemp cultivars for different purposes.
In this article, the state of the art of fiber hemp breeding programs in Europe, China, and Canada are
Keywords:
reviewed. The breeding strategies and tools used in the breeding of hemp cultivars are discussed. We also
Hemp
Fiber quality
provide an overview of genetic diversity in hemp for different traits. In addition, the current knowledge
Breeding of the main breeding goals for fiber hemp, which are an improvement of fiber quality and fiber yield,
Genetics breeding for specific cannabinoid profiles, control of flowering behavior, male flowering control, and
breeding of cultivars for specific environments are evaluated. Lastly, we discuss the inestimable value of
next generation technologies to breed new hemp cultivars that are suitable for a biobased economy.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction globally on 61,318 ha, of which 11,400 ha in China, 14,344 ha in The

European Union, and 15,720 ha in Canada (Source: Health Canada).
Hemp (Cannabis sativa L, 2n = 20) is one of the world’s oldest cul-
tivated annual crops (C3 annual), traditionally grown for its long
and strong bast fibers and seeds. In most Western countries the 1.1. Products
cultivation of hemp vanished or was interrupted for decades as
a result of competition with other feedstock’s such as cotton and Hemp is involved in a diverse range of products, and has inte-
synthetic fibers, high labor costs, and the prohibition of cultiva- grated many agro-industrial fields such as agriculture, textile,
tion due to the use of cannabis (C. indica) as a narcotic. Only in bio composite, paper-making, automotive, construction, bio-fuel,
Eastern Europe, the former Soviet-Union and China has a substan- functional food, oil, cosmetics, personal care, and pharmaceutical
tial hemp industry survived (De Meijer, 1995). In the late 1970s industry (Fig. 1). Traditionally, hemp bast fiber is used in textiles,
and the early 1980s, the average hemp planting area once reached paper pulp, and materials for building and insulation. Hemp hurds
about 160,000 ha in China. After that, the cultivation area declined (also termed ‘shives’), the woody and lignified core tissues of the
because of the above-mentioned reasons. The crop can be grown stems, are used as horse-bedding, pulping, and concreting (Elfordy
in a wide range of geographic zones of climate, and is well adapted et al., 2008; Karus and Vogt, 2004). Besides the traditional uses,
to most regions of the world. China, Europe, and Canada are the novel applications for fiber hemp (fibers/biomass) are being devel-
three most important hemp planting regions in the world. Accord- oped. The high cellulose content of hemp cell walls (Amaducci
ing to FAOSTAT (excluding Canada), in 2011 hemp was cultivated et al., 2000) together with the relatively high productivity make
hemp biomass an interesting renewable feedstock for energy pro-
duction (Hanegraaf et al., 1998; Prade et al., 2011, 2012a,b; Ragit
et al., 2012), for the production of second generation bio-ethanol
∗ Corresponding author. Tel.: +31 317 482127. (Gonzalez-Garcia et al., 2012) and as a reinforcement in ‘green com-
E-mail address: [email protected] (L.M. Trindade). posite’ materials (Khalil et al., 2012; Shahzad, 2012) and concrete

http://dx.doi.org/10.1016/j.indcrop.2014.08.011
0926-6690/© 2014 Elsevier B.V. All rights reserved.
 E.M.J. Salentijn et al. / Industrial Crops and Products 68 (2015) 32–41 33

Fig. 1. Flowchart of multi-purpose hemp utilization.

(Elfordy et al., 2008). It is estimated that the global market for hemp cellulose (∼55%), hemi-cellulose (∼16%), pectin (∼18%), and lignin
consists of more than 25,000 products. (∼4%) (Garcia-Jaldon et al., 1998).
Although hemp has the potential to produce fiber of excellent
quality, the cultivars currently available deliver fiber of variable
1.2. Yield
quality. A better understanding of the key factors determining fiber
quality, and how these can be regulated, is of crucial importance
Hemp may potentially yield 25 t ha−1 above ground dry mat-
for breeding fiber hemp. This article provides an overview of the
ter, 20 t ha−1 stem dry matter and 12 t ha−1 cellulose (Struik et al.,
state of the art with regard to hemp breeding and the possibilities
2000) but in many cases the yield varies. For instance, in North-
of novel molecular breeding approaches to increase the value of
ern Italy total dry matter yield (cv. Futura 77) ranged from 18.7 to
industrial hemp.
8.3 t ha−1 over years and locations (Amaducci et al., 2000). In gen-
eral cellulose yield is 7–10 t ha−1 (Zatta et al., 2012). The yield of the
dry bast fiber varies from 1.2 to 3.0 t ha−1 , and seed yield from 0.7
2. Genetic variation in hemp
to 1.8 t ha−1 . Industrial hemp production statistics for Canada indi-
cate that average yield of seeds is about 0.78 t ha−1 , and an average
Hemp (C. sativa L.) can be classified according to different
5.9 t ha−1 of straw, which can be transformed into about 1.45 t ha−1
attributes including: (i) population type such as wild and natural-
of fiber (Johnson, 2013).
ized populations, landraces, and cultivars; (ii) plant-use as fiber
cultivars (for long fibers or for pulp), seed cultivars, drug strains,
1.3. Hemp fiber and ornamentals (De Meijer, 1995; De Meijer and Keizer, 1996);
(iii) flowering time, which includes early ripening, intermediate-
At present, hemp has been re-discovered as an interesting sus- ripening, late-ripening cultivars; (iv) gender, whether they are
tainable (Amaducci et al., 2000; Struik et al., 2000; Van der Werf dioecious or monoecious cultivars, and (v) geographic origin, e.g.
and Turunen, 2008) high yielding industrial fiber crop (Van der North-type and South-type cultivars in China. Hemp is believed
Werf et al., 1996) that can help to meet the high global demand to have originated in Central Asia, and several advocate for two
for fibers (Shui and Plastina, 2013). Hemp stems can be divided centers of diversity, Hindustani and European–Siberian (Zeven and
in a ‘bark’ or ‘bast’ section, corresponding to the tissue located in Zhukovsky, 1975). There is still debate over the taxonomic orga-
the outer part of the stem outside the vascular cambium, and the nization of the genus Cannabis. Some authors have proposed a
‘woody core’ which is located inside the ring of vascular cambium monotypic genus, C. sativa, while others state that two species
and consists of lignin rich xylem tissue. The primary bast fibers (ele- can be distinguished, C. sativa and C. indica, and maybe even a
mentary fiber about 20 mm. to 50 mm. long) and secondary bast third species C. ruderalis. On the basis of allozyme data for 157
fibers (about 2 mm long) are derived from the vascular bundles accessions from diverse geographic origin, Hillig (2005) recog-
in the bark of stems whereas the core fibers are located inside of nized three gene pools, C. sativa, C. indica and C. ruderalis and
the vascular cambium in the woody core (0.5–0.6 mm long) (De suggested a polytypic concept (=having several variant forms,
Meijer and Keizer, 1994; De Meijer, 1994; Mediavilla et al., 2001; especially subspecies or cultivars) of Cannabis, with seven puta-
Van der Werf et al., 1994). The primary bast fibers of hemp are tive taxa. The characteristics attributed to such subspecies have
made up of bundles of pericyclic elementary fibers that are char- generally evolved as a result of geographical distribution or isola-
acterized by thick and lignified cell walls. They are composed of tion. Russian botanists recognized four ‘eco-geographical’ groups
34 E.M.J. Salentijn et al. / Industrial Crops and Products 68 (2015) 32–41

(ecotypes) of hemp: Northern (Northern Russian, Finland), Central geographic zones of climate in China (latitude range about
ecotypes (Central Russian, Ukraine), Southern ecotypes (Mediter- 23–50◦ N), hundreds of hemp landraces have been established.
ranean region, Balkan, Turkey, Caucasus), and Far Eastern ecotypes Examples are ‘Liuan HuoMa’ and ‘Liuan HangMa’ from Anhui
(China, Japan and Korea). The genus is now distributed worldwide province, ‘Laiwu DaMa’ and ‘Laiyang DaMa’ from ShanDong
from the equator to about 60◦ N latitude, and throughout much of province, ‘Gushi KuiMa’ in Henan province, ‘Wenxian DaBaiPi’ in
the southern hemisphere (Hillig, 2005; Mukherjee et al., 2008 and Hebei province, ‘Liuzhi DaMa’ in Guizhou province, and ‘Dayao
references therein). In order to distinguish drug types from fiber DaMa’ and ‘Weishan DaMa’ in Yunnan province. All these landraces
types, Gilmore et al. (2007) used chloroplast and mitochondrial are valuable resources of starting material for further hemp breed-
markers to study the phylogeny of a panel of 188 plants derived ing and some of these are still being cultivated in certain areas
from 76 populations, representing plant-use groups and geograph- nowadays. Chinese landraces were also used to breed the now
ical regions. Despite the uni-parental (maternal) inheritance and extinct ‘Kentucky hemp’ cultivars (Dewey, 1913, 1927) that were
low mutation rate in organelles, six minor haplotypes and three cultivated in the United States until the mid-1950s, when the cul-
major haplotypes were recognized, that were highly suggestive for tivation of hemp was prohibited. Large collections of germplasm
both crop-use and geographic region. Two organelle haplotypes resources have been collected and maintained in the Yunnan
(I and II) were observed in populations from Europe, mostly fiber Academy of Agricultural Sciences, which comprise approximately
types, whereas the other haplotypes were more indicative C. indica 350 accessions with a good representation of fiber/seed hemp
drug types from southern Europe and Asia and haplotypes V and VI groups. In 1970s, several cultivars (e.g. 882, 812, 333, 112, 544, 370,
for drug types from Africa, India, SE Asia, and Mexico. Central Rus- and hybrid strain-2) were developed and, although rarely, some are
sian and Mediterranean fiber hemp landraces and cross-progenies still used in production now. From the 70s till the end of the 20th
of these two groups are the ancestors of the present European and century limited research on hemp breeding was carried out. In the
west Asian cultivars (De Meijer, 1995). Fiber strains from China past decade, many new applications for hemp biomass have arisen
(Far Eastern hemp) may be somewhat distinct. In an analysis of and they have been accompanied with the development of related
the genetic variation in ribosomal and cannabis specific chloroplast industries and an increase in hemp cultivation area in China. Since
DNA regions, ancient DNA from an extinct Chinese hemp showed 2007, hemp breeding research has continuously received financial
a close relationship to present day C. sativa hemp material from support from China Agriculture Research System, and five indus-
China as well as to Humulus japonicus (‘Japanese hop’), whereas C. trial hemp cultivars (YunMa 1, YunMa 2, YunMa 3, YunMa 4, YunMa
sativa subspecies indica formed a separate clade (Mukherjee et al., 5) have been bred and widely cultivated in China. Other hemp culti-
2008). vars (LongDaMa 1, JinMa 1, WangDaMa 1, WangDaMa 2) have been
registered and used in certain provinces.

3. Hemp breeding in Europe, China, and Canada

3.3. Hemp breeding in Canada

3.1. Hemp breeding in Europe

Hemp was banned in North America from the late 1930s until
March 1998. Since the first commercial licenses for hemp grow-
The historical importance of hemp cultivation in Europe is well
ing were issued in 1998, the Canadian hemp market has gradually
reflected by the abundance of cultivars, traditional landraces, and
developed and Canada has become the main supplier of hemp seed
populations that were selected in the main areas of hemp culti-
and oil-cake for the United States (Hanks, 2008; Johnson, 2013).
vation throughout Europe. Mass selection was used in the past to
Hence, the Canadian hemp oil processing value chain is well estab-
select the most important cultivars, such as Carmagnola in Italy
lished by comparison to the fiber processing chain. Government
(Ranalli, 2004) or Novosasdka konoplia in Yugoslavia (Berenji et al.,
subsidies were granted to sustain the market, resulting in the devel-
2013). In mass selection pollination cannot be controlled and any
opment of novel hemp enterprises, improvements in the processing
improvement in fiber content is very slow. A large contribution to
technologies, and development of new hemp cultivars for Canadian
the increase of stem fiber content was obtained by the application
environments. To increase the interest for hemp, efforts were made
of the Bredemann method (Bredemann, 1942), that consisted in the
to produce hemp as a certified bio-based crop and in 2012 a food
individual selection of male plants on the basis of the fiber content,
safety accreditation for hemp food was issued to Hemp Oil Canada
measured on a longitudinal section of the stem.
Inc. (Laate, 2012; Robbins et al., 2013).
A great breeding work carried out in Europe by the German
Field production was dominated by the cultivars Finola (orig-
researcher Von Segenbuch has yielded the first monoecious cul-
inating from Finland), Crag (Canada), and USO 14 (Ukrainian)
tivars and the first hybrids bred by Professor Bocsa in Hungary had
for a long time. Several breeding programs included developing
a high fiber content. In Hungary a Chinese accession has been used
commercial strains from feral Canadian stocks, creating superior
as a heterosis breeding parent (hybrid breeding) (for references see
oilseed cultivars by increasing seed yield and optimizing fiber use
De Meijer, 1995). In 1995, De Meijer reported that 12 hemp culti-
for a variety of regions (Hanks, 2008). Together with the Bast Insti-
vars were registered in the EU of which only seven French cultivars
tute (Summy region, Ukraine) a program to evaluate hemp cultivars
were readily available (De Meijer, 1995). In 2004, the number of
for Canada was started, and it included the Ukrainian cultivars
registered hemp cultivars increased to 45 (Ranalli, 2004), in 2008
USO 14, USO 31 Zolotonoshskaya 11 (Zolo 11), and two Canadian
the list contained 46 industrial hemp cultivars (Jankauskienè and
cultivars, Anka, and Carmen (Watson et al., 2012). Since then a
Gruzdevienè, 2009) and currently the number of cultivars regis-
number of high yielding cultivars suited to a wide range of growing
tered for the EU is 51 (Table 1) reflecting the increased interest in
conditions, including both monoecious and dioecious cultivars
the crop.
grown for seed, fiber or for dual purpose, have been developed
and tested. The most common cultivars that are presently being
3.2. Hemp breeding in China contracted and grown in Canada are Alyssa, Anka, CRS-1, CFX-1,
CFX-2, Delores, and Finola (Laate, 2012). The 2013 list of approved
According to archeological finds and ancient records, it has been cultivars for Canada contains 39 cultivars, of which 24 are kept
more than 6000 years since China started cultivating hemp for in Canada and the others are in European countries (http://hc-sc.
fiber and seed (Yang, 1991). Due to the long history of cultiva- gc.ca/hc-ps/pubs/precurs/list cultivars-liste2013/index-eng.php).
tion and the wide spread of this fiber crop throughout different THC levels are not a problem since they can be kept low and well
 E.M.J. Salentijn et al. / Industrial Crops and Products 68 (2015) 32–41 35

Table 1
Hemp cultivars registered in the EU on 25-09-2013, 51 entries.

Denomination Type Maintainers Country Admission Last modification Deletion Market extension

1 Armanca RO 1002 RO 24.11.2010

2 Asso IT 15 IT 23.02.2004
3 Beniko PL 893 PL 23.12.1985 16.02.2009
NL 613 NL 04.11.2002
CH 168 PL 07.06.1999 01.06.2009 01.06.2011
AT 567 PL
4 Bialobrzeskie AT 567 AT
CZ 1067 CZ 31.10.2008
PL 893 PL 31.12.1967 16.02.2009
5 Cannakomp HU 149424 HU 19.02.2004 06.01.2012
6 Carma IT 15 IT 16.01.2006
7 Carmagnola IT 15 IT
8 Chamaeleon NL 391 NL 25.03.2002
9 Codimono IT 15 IT 23.02.2004
10 CS IT 15 IT
11 Dacia Secuieni RO 1018 RO 19.12.2011
12 Delta-405 ES 275 ES 18.01.2012
13 Delta-llosa ES 275 ES 18.01.2012
14 Denise RO 1018 RO 25.10.2007
15 Diana RO 1018 RO 14.07.2009
16 Dioica 88 FR 8194 FR
17 Epsilon 68 FR 8194 FR
18 Fedora 17 FR 8194 FR
CH 170 07.06.2002 01.06.2012 01.06.2014
19 Felina 32 FR 8194 FR
20 Férimon FR 8194 FR
21 Ferimon DE 4668 DE, FR
22 Fibranova IT 15 IT
23 Fibrimor IT 15 IT 17.10.2003
24 Fibrol HU 149424 HU 05.05.2006 06.01.2012
25 Finola FI 6157 FI 05.02.2003
26 Futura 75 FR 8194 FR
27 Ivory NL 722 NL 04.06.2012
28 KC Dora HU 149424 HU 10.03.2009; 06.01.2012
29 KC Virtus HU 149424 HU 12.03.2013
30 KC Zuzana HU 149424 HU 12.03.2013
31 Kompolti hibrid TC Hybrid HU 149424 HU 19.05.1983
32 Kompolti HU 151322 HU 25.04.1954 06.01.2012
AT 625 AT 30.06.2010
CH 173 CH 01.06.2010
NL 612 NL 04.11.2002
33 Lipko HU 151322 HU 19.02.2004
34 Lovrin 110 RO 1002 RO 22.11.2007
CH 172 CH 30.06.2011
35 Marcello NL 722 NL 04.06.2012
36 Markant NL 722 NL 04.05.2012
37 Monoica HU 149424 HU 05.05.2006 06.01.2012
CZ 666 CZ 16.07.2009
38 Santhica 23 FR 8194 FR
39 Santhica 27 FR 8194 FR 04.05.2002
40 Santhica 70 FR 8194 FR 30.03.2007
41 Secuieni Jubileu RO 1018 RO 29.02.2012
42 Silvana RO 1002 RO 22.11.2007
43 Szarvasi HU 108887 HU 12.03.2007 12.03.2007
44 Tiborszállási HU 105303, HU 19.02.2004
IT 1229 IT 16.01.2006
45 Tisza HU 105303 HU 10.03.2010
46 Tygra PL 893 PL 12.03.2007 16.02.2009
47 Uniko B Hybrid HU 151322 HU 29.04.1965
CH 173 CH 07.06.1998 27.04.2009 01.06.2010
48 Uso-31 NL 647 NL 31.01.2005
CH174 CH 07.06.1999 01.06.2009 30.06.2011
49 Wielkopolskie PL 893 PL 06.03.2009
50 Wojko PL 893 PL 08.03.2011
51 Zenit RO 1018 RO 14.07.2009

Market extension: Certification and marketing of seed of the variety is allowed until the indicated. http://ec.europa.eu/food/plant/propagation/catalogs/database/public/index.
cfm?event=RunSearch.

below the Canadian standards of 0.3% of the dry weight. Breeding component and important for the (healthy) food market. Cultivars
for stable production of either seed, fiber or for dual purpose in with high contents of this fatty acid in the oil have been obtained,
specific environments is the main target for breeding. Another e.g. Canda >3.5%, Debby >5%, and Joey >4%.
important trait for Canada is increase in gamma linolenic acid In 2013, a large decortication plant (parkland industrial hemp
(GLA) in the seed oil. This is a highly desirable essential fatty acid processing PIHP) opened in Gilbert Plains, Manitoba. In 2012
36 E.M.J. Salentijn et al. / Industrial Crops and Products 68 (2015) 32–41

several hemp cultivars that may produce a high biomass with a 1994). One example of a high heritable trait is bast fiber content and
high fiber yield for this new production plant were evaluated. Both mass selection proved to be efficient to breed for this parameter.
mono- and dioecious cultivars were tested for grain and fiber yield, Characteristics such as date of flowering, plant height, and stem
as well as for quality evaluation of oil profiles and percent fiber con- diameter, which are not directly related to bast fiber yield, were
tent. The cultivars were evaluated for their potential as grain-only, shown to have disadvantages as selection criteria for the improve-
fiber-only or dual grain-fiber crop in environments across Canada ment of bast fiber yield (Hennink, 1994).
(Watson et al., 2012).
4.2. Cross-breeding
4. Breeding methods in hemp
In later stages, more efficient crossing methods and selection
criteria were employed. Dedicated crosses were made to breed
Methods for breeding hemp have also changed throughout the
for improved characteristics such as fiber content, stem and seed
years (reviewed for Europe by De Meijer, 1995; Ranalli, 2004;
yield, gender, phenological development, stem quantity, low 9-
Ranalli and Venturi, 2004).
tetrahydrocannabinol (THC) content, resistance to pathogens (such
Hemp is open pollinated (wind-pollinated) and is usually a dioe-
as root-knot nematodes) and lodging, as well as suitability for
cious annual crop, where female and male flowers are on different
different cultivation regions taking into account the effects of geno-
individuals, indicating that hemp is naturally outcrossing (cross-
type and environment (GxE) on the yield and quality of hemp.
pollinator). All cannabis strains can inter-cross creating, in some
Half-sib family selection was employed, based on the evaluation
cases, a continuous pattern of variation. In hemp the control of
of progeny from each mother plant. The latter are open pollinated
pollination is therefore an important issue. In the case of a dioe-
and fertilized by more than one pollen male parent. The selection is
cious hemp population the male and female plants are intermixed
based on general combining ability (GCA; the average performance
and the female plants are always cross-pollinated. Entirely female
of an individual in a particular series of test crosses), with the entire
populations exist that can be used to produce hybrid cultivars by
population used as a tester. Furthermore, hybridization (between
crossing with a selected pollen donor. In the case of monoecious
cultivars) was performed to create new variability (Ranalli, 2004
hemp cultivars the male and female flowers are on the same indi-
and references therein). The cultivars YunMa 2 and YunMa 4 are
vidual, which enables selfing. The breeding of a cross-pollinator
good examples of the application of this method in China.
such as hemp requires a specific breeding approach (Posselt, 2010)
that comprises three breeding phases: (1) search for the natural
4.3. Inbreeding and hybrid breeding
variation in the material and create a base population, (2) gener-
ate varietal parents through selection and improve the population
Dioecious hemp populations also contain monoecious or inter-
through recurrent selection steps to create a breeding population,
sexual plants. These populations were used to develop monoecious
and (3) develop and test experimental cultivars. The hemp cultivars
hemp. Monoecious plants and subdioecious plants (female plant
available are mainly population cultivars, such as ‘open pollinated
with induced male flowers) enabled self-fertilization and inbreed-
cultivars’ that are the result of recurrent selection and ‘synthetic
ing. Often the progeny of such plants is exclusively female, which
cultivars’ that are advanced generations of a population initiated
enables heterosis and hybrid breeding. The unisexual female char-
by crosses among a restricted number of selected parents and mul-
acter can be considered an analog for male sterility and allows for
tiplied by a number of random out-crossings in isolation.
large scale hybrid seed production with limited labor. Tetraploid
The methods commonly used in hemp breeding are ‘mass selec-
forms of hemp were developed (2n = 40) that were completely fer-
tion’, ‘cross-breeding’, ‘inbreeding’, and ‘hybrid breeding’, and more
tile but showed reduced fiber quality.
recently there are a few examples of the use of molecular markers
Heterosis breeding of hemp has been used in Hungary and China
to assisted breeding (reviewed in Ranalli, 2004).
and has resulted in several F1 hybrid cultivars, including Uniko-
B, Kompolti hybrid TC, and YunMa 3. Uniko-B is a single cross
4.1. Mass selection hybrid (Kompolti (female) × Fibrimon 21 (male)). The F1 of this
cross is almost unisexual female and is used to produce an F2
Mass selection is performed by selecting seeds after harvest containing 30% males that is cultivated for fiber production. Kom-
(method I) and selecting seeds from the best plants in the field polti hybrid TC is a three way-cross hybrid (1◦ Kinai dioecious,
(method II). Initially selections from old naturalized (weedy) popu- female × Kinai monoecious, male = Kinai uniszex, female; 2◦ Kinai
lations resulted in the establishment of landraces. In the case of uniszex (female) × Kompolti (male) = Kompolti hybrid TC, sex ratio
fiber hemp mass selections or single plant selections of such lan- 50/50). The Kinai material is of Chinese origin (Ranalli, 2004 and
draces selections were performed to produce more genetically references therein). Finally the YunMa 3 from China is one of the F1
uniform breeding material, cultivars and ecotypes. For example, hybrid cultivars. Interestingly, the single cross hybrid can only be
in the first quarter of the 20th century cultivars were selected produced in winter in the low latitude region like Yunnan Province
from landraces using selection characteristics such as the length because its parents cannot cross in the normal growth season (Sum-
of the vegetative period length, height, diameter, weight, and in mer) due to distinctly different flowering times.
some cases seed weight. The predominant approach to improve
the material was continuous mass selection, a cyclic procedure 4.4. Marker assisted breeding
that attempted to upgrade whole populations by directed selection
(Ranalli, 2004). This type of ‘family breeding’ led to the devel- Next generation sequencing technologies opened the gate
opment of cultivars in Italy, Hungary and Romania, all of them toward genotyping by sequencing (Elshire et al., 2011) and defining
coming from the famous Northern Italian landrace Carmagnola. In causal relationships between genetic polymorphism’s and pheno-
Italy, Carmagnola was further improved for specific regions by mass typic differences on a large scale. The expansion of the genetic
selection, leading to the development of ‘Bolognese’, ‘Toscana’, and and phenotypic data and the development of molecular markers
‘Ferrarese’ ecotypes (the names define the places of cultivation). In are of inestimable value for plant-breeding. Genome wide genetic
China some cultivars (e.g. YunMa 1 and YunMa 5) were bred by polymorphisms can be used to explore the genetic diversity of the
this method from the local landraces. Mass selection is a common available germplasm or breeding populations and genetic maps
breeding method in hemp for traits with high heritability (Hennink, with markers linked to a trait of interest can assist the selection
 E.M.J. Salentijn et al. / Industrial Crops and Products 68 (2015) 32–41 37

for complex traits. The history of marker-assisted selection in properties (Hänninen et al., 2012). Breeding for genotypes that
hemp was reviewed by Mandolino and Carboni (2004). Molecular are easier to process would provide an advantage in terms of fiber
markers have mostly been used to exploit hemp for forensic appli- quality and a reduction of costs related to fiber extraction (Van den
cations, to trace back the origin of illegal hemp material and Broeck et al., 2008). These objectives seem to be met by cultivars
recognize the presence of cannabis in materials. Five main conclu- recently released in The Netherlands: Chamaeleon (Toonen et al.,
sions have been made regarding the genetic structure of hemp: (1) 2004), Markant, Ivory, and Marcello.
It has a high genetic variation and a high level of heterozygosity due From studies in Arabidopsis and other model plants it is known
to open-pollination and the out-crossing character of hemp. The that many genes are involved in cell wall related processes. Using
majority of the alleles occurred in low frequency in the population microarrays a random set of hemp cDNAs (n = 3414 cDNAs) derived
(<0.30) with only a few major alleles being observed in cultivars from fiber material was tested for the differential expression in
with a high level of inbreeding, such as cultivar Fibrimon; (2) it has between bast fiber tissue and core tissue (Van den Broeck et al.,
a low discriminative power of marker loci to distinguish among 2008) across developmental stages of two hemp cultivars, cv.
cultivars or accessions; a high degree of variation within cultivars Chamaeleon and cv. Felina 34. Hemp genes that were highly
or accessions was observed, even in female inbred lines; (3) There expressed in core tissue were coding for proteins involved in lignin
is no clear split between drug types and fiber types; (4) Hemp has a biosynthesis and C1 metabolism, a process closely connected to
widely shared gene pool, with limited cultivar boundaries and lit- lignin biosynthesis. Genes coding for arabinogalactan related pro-
tle segregation between populations; and finally, (5) The practical teins, lipid transfer proteins, lipoxygenase, and endoxyloglucan
use of a genetic maps and molecular markers is limited in cannabis transferases were highly expressed in the bast fiber tissue (Van den
because of the high level of variability. Broeck et al., 2008).

5. Breeding goals
5.2. Cannabinoid profiles

Traits that are important in fiber hemp breeding comprise: high

Breeding for low delta-9-tetrahydrocannabidol (THC) has been
fiber yield and fiber quality, cannabinoid content and composition,
a main target in fiber hemp breeding and levels below <0.2% THC
degree of monoecy, length of vegetative cycle, and resistance to
have been reached for some cultivars. Government regulations that
diseases and pests. A great deal of attention is given to high fiber
allowed a THC content of only 0.2% were implemented in the Euro-
yield and quality together with low THC content. Some traits show
pean Union in 2001. Since then, a further and stable reduction of
a high plasticity, especially cannabinoid content and phenologi-
THC gained importance as a breeding goal. It was necessary to start
cal development. Because hemp is very sensitive to environmental
the difficult task of further decreasing the THC content while keep-
conditions, such as day length and temperature, cultivars are
ing the productivity and other positive characters intact. In the
typically developed for specific environments and cropping condi-
former USSR a successful breeding program for the reduction of
tions. However, the ranking of cultivars for most traits in different
cannabinoids was initiated in the 1970s. Cultivars completely lack-
environments are expected to be fairly stable. Furthermore, the
ing THC were obtained (Hennink, 1994; Mandolino and Carboni,
suitability of certain cultivars to a given environment also depends
2004). In a joint effort between scientists in France and Ukraine
on the purpose for which they are cultivated (Ranalli, 2004; Ranalli
several new monoecious cultivars were developed. These culti-
and Venturi, 2004).
vars have very low THC levels (THC < 0.07%) and lack the typical
hemp aroma (e.g. USO-45) (Holoborodko et al., 2014 report at:
5.1. Fiber quality and fiber yield www.interchanvre.com/docs/article-Laiko.pdf).
The distinction between fiber and drug accessions can only
Fiber quality strongly depends on the morphology of the fiber be made on the basis of the cannabinoid profile (chemotype).
bundles and on the chemical composition of the cell wall of the Three major ‘chemotypes’ are recognized in hemp based on the
elementary fiber (Rowell et al., 2000). In polymer reinforcement or ratio in the inflorescence dry matter between the two major
biocomposites, surface characteristics and finesses are important cannabinoids of hemp, delta-9-tetrahydrocannabidol (THC) and
fiber traits that influence the interfacial bond strength between the cannabidiol (CBD): (i) drugs type with THC prevalent, (ii) inter-
fibers and the matrix (Gamelas, 2013), as well as the fiber tensile mediate type with similar amounts of both THC and CBD, and (iii)
strength (Placet, 2009). The variability of natural fiber properties, fiber type with CBD prevalent. Two alleles at the B locus (BT and BD )
moisture absorption, and processing costs are weak factors of nat- are controlling the trait (Fig. 2, De Meijer et al., 2003). The fourth
ural fibers for composite applications (Deyholos and Potter, 2014). and fifth chemotype are minor chemotypes that are not frequently
The variability of fiber characteristics is a consequence of the high found. In the fourth chemotype, cannabigerol (CBG), the precursor
heterogeneity of hemp cultivars, but it is also due to an unavoidable of THC and CBD is the major cannabinoid. This chemotype is most
interplant heterogeneity. In fact, fiber maturity decreases from bot- likely to be controlled by a B0 allele, a mutant form of the BD locus
tom to top of the stem (Amaducci et al., 2008b) and from the outer (De Meijer and Hammond, 2005). The fifth chemotype has unde-
to the inner fiber layers in the same internode (Amaducci et al., tectable amounts of cannabinoids (zero cannabinoid). This trait is
2005), and the fiber yield and quality changes with time during controlled by is single locus (O) that operates upstream of the B
plant development; for instance cellulose increased up to 56–65% locus. This zero cannabinoid chemotype is most likely due to a block
until late flowering. Furthermore, cultural techniques such as plant in the metabolic pathway leading to the production of cannabinoids
density, nitrogen fertilization, and harvest time are also important and not to an alteration in the glandular trichomes (De Meijer et al.,
factors that affect fiber yield and quality (Amaducci et al., this issue). 2009a,b).
A high cellulose content, a low degree of lignification and a A sixth chemotype is found in plants with a specific morpho-
reduced number of cross links between the pectins and the struc- logical phenotype ‘prolonged juvenile chemotype’ that produce
tural components of the cell wall are important characteristics for cannabichromene (CBC). The genetic factors controlling this trait
a suitable extractable fiber for both paper- and textile industries (termed locus C or BC ) are independent from the B loci that encode
(Mandolino and Carboni, 2004). Fiber extraction by mechani- THC- and CBD synthase (De Meijer et al., 2009a,b) (Fig. 2). Cannabis
cal decortication and scutching has a strong influence on fiber chemotypes having no cannabinoids, or only CBG or CBC are inter-
quality and it is often the main cause of damage to the fiber esting because of their pharmaceutical value. CBC dominates the
38 E.M.J. Salentijn et al. / Industrial Crops and Products 68 (2015) 32–41

cannabinoid genes. For instance, CBD-synthase [genbank:

AB292682] is predominantly expressed in Finola (BT :BD ) (Van
Bakel et al., 2011).

5.3. Flowering behavior

Hemp, C. sativa L., is a dioecious annual with separate male

and female plants, and occasionally monoecious individuals can
be found within a population. As a “short day” plant hemp has a
critical photoperiod, under which flowering is induced, at approx-
imately 14 h (Amaducci et al., 2008a, 2012; Lisson et al., 2000).
Variation for the timing of the transition from vegetative to repro-
ductive growth (=flowering time) is present among cultivars and
Fig. 2. The biosynthetic pathway of the most common cannabinoids in cannabis.
heterogeneity in hemp stands is partly due to the differences in
9-tetrahydrocannabinol (THC), cannabidiol (CBD) and cannabichromene (CBC).
The first step in the cannabinoid biosynthesis pathway, the condensation of ger- flowering time and the dioecious nature of hemp (Amaducci et al.,
anylgeraniol diphosphate (GPP) with olivetolic acid (OA) is catalyzed by the enzyme 2008c). These characteristics of hemp are reviewed in Hall et al.
geranyldiphosphate:olivetolate geranyltransferase (GOT), leading to the production (2012). Three developmental stages are recognized in hemp; the
of (CBG). The next step is the action of THC-, and CBD-synthases that are converting
juvenile-, photosensitive-, and flowering-stage. Based on the flow-
CBG into THC (BT locus), CBD (BD locus), and in some cases CBC-synthase which
catalyzes the conversion of CBG into the cannabinoid cannabichromene (CBC). BC ,
ering time, three main groups of genotypes can be distinguished;
BT , BD and Bo are alleles of the B locus that account for the different chemotypes. early (flowering after 40–60 days), intermediate (60–90 days), and
late (90–120 days) cultivars (Zatta et al., 2012). Early and interme-
diate groups are selected under Northern conditions (short growing
cannabinoid fraction of juvenile cannabis plants and declines with seasons with long day-lengths in the summer). If grown under
maturation. Southern conditions such cultivars start flowering even earlier as
Chemotype-associated molecular markers could well be used the critical day-length for hemp flowering in the South comes ear-
to assist the selection of THC producing plants for elimination. To lier than in the North and the biomass yield is therefore lower. On
define such markers a cross was made between inbred lines with the contrary, late ripening cultivars should be selected in the low
contrasting chemotypes (THC x CBD). The F1 was almost entirely latitude regions in order to get high fiber yield (e.g. the south of
hybrid (CBD/THC; chemotype II), while the F2 segregated in a China, Thailand, Australia, Southern Europe). The fiber yield can
Mendelian way 1:2:1 for the chemotypes, THC [BT ,BT ]: CBD/THC increase if the cultivar from a low latitude location is grown in a
[BD ,BT ]: CBD [BD ,BD ], which was in agreement with one gene with higher latitude location. This effect of increase in yield becomes
two codominant alleles (BT and BD ). The F2 population was used to obvious when the value of latitude change is above 2 degrees (Guo
identify RAPD markers that resulted in a SCAR marker (B190/B200) et al., 2013). There are differences among hemp cultivars for sen-
that was useful for chemotype identification in that particular sitivity to changes in photoperiod: ‘Felina 34’ and ‘Futura’, and
population, but unfortunately was not linked to chemotype in Chinese landrace ‘Huoqiuzi’ are regarded low sensitive, whereas
other populations (Mandolino and Carboni, 2004). The genes ‘Tiborszallasi’ and most Chinese landraces are very sensitive. The
for THC and CBD synthases have been isolated (1635 bp, 89.3% Italian landrace ‘Carmagnola’ and the cross-bred variety ‘Fibranova’
identity, deletion of a Serine coding triplet at position 757–759 are intermediate sensitive. The flowering behavior is, besides the
in the nucleotide sequence for THC synthase) and allele specific length of the photoperiod, depending on temperature and light
primers have been developed that allow the identification of the quality. Amaducci et al. (2008a, 2008b, 2012) generated a model
allelic composition at the B locus (Mandolino and Carboni, 2004). to estimate the flowering time of hemp that is useful to make deci-
Based on a genetic model with two codominant alleles BT and BD sions on the sowing time and harvesting time and cultivar choice.
on the same locus, drug types are homozygous for BT (BT/BT) and The basis for a good fiber quality and yield is the use of a cultivar in
expected to harbor only the THC synthase. With the publication combination with good environmental conditions and crop man-
of the draft genome of the drug strain ‘Purple Kush’ in 2011 (Van agement (Struik et al., 2000) whereby each application also has its
Bakel et al., 2011) analysis of the genes involved in cannabinoids own demand on fiber characteristics and processing of fibers.
biosynthesis was facilitates. In this hemp genome only a single
THC synthase gene (THCAS) with a full open reading frame was 5.4. Control of gender (male flowering control)
identified (with 99% nucleotide identity to the published sequence
PK29242.1, genbank: JP450547) whereas no intact CBD synthase Monoecious hemp cultivars have a higher seed yield and higher
gene (CBDAS) was found. However, the genomic organization of the uniformity compared to dioecious cultivars and therefore mechan-
locus is more complex because pseudogenes with pre-mature stop ical harvesting of such cultivars is easier. Drawbacks are the
codons and frame-shift mutations (THCAS-like pseudogenes and narrower genetic base, necessity to maintain the monoecious trait
CBDAS-like pseudogenes, genbank) for both genes were observed (including the selfing of a monoecious plant and elimination of male
indicating that the cannabinoid synthase genes are part of a small plants), strict isolation of propagations and seed batch control for
gene family (Van Bakel et al., 2011). In a study based on chemotype male plants. In dioecious cultivars, selection of males before polli-
and genotype of cannabinoids in hemp landrace, Chen et al. nation and pollination only with the best-scoring male is a common
(2013) speculated that there is probably another (BT /BT )-like locus practice in breeding. Selection for sex is therefore important in
involved. RNAseq analysis (Van Bakel et al., 2011) of the hemp hemp breeding. The determination of the gender in C. sativa L. is
cultivars ‘Finola’ (fiber type), Purple Kush (drug type), and ‘USO-31’ influenced by both genetic and environmental factors. In cannabis,
(fiber type) showed that besides cannabinoid synthase other genes two sex chromosomes, X and Y are present, whereby the Y chro-
involved in the cannabinoid pathway were up-regulated in the drug mosome is much larger than X chromosomes and autosomes. True
type Purple Kush and that the differences between drugs and fiber male plants have one X and one Y chromosome, females have two
strains are not only due to changes in the coding DNA-sequence of X chromosomes resulting in a difference in genome size between
the genes themselves but are either due to cis-or trans-regulatory male and female plants as determined by flow cytometry (respec-
transcription factors that regulate the expression from the tively 1683 Mbp and 1636 Mbp for the diploid genome). The long
 E.M.J. Salentijn et al. / Industrial Crops and Products 68 (2015) 32–41 39

arm of the Y chromosome differed from the other chromosomal Currently, hemp-based products are increasingly being diversi-
regions by showing early condensation at the metaphase stage fied and developed on industry-specific goals, which is a striking
(Sakamoto et al., 1998). However, as is the case in many other plant feature for this crop. Breeding of cultivars for specific end-uses
species, in C. sativa L., determination of gender is not only controlled will become critical. For making functional foods, the seed yield,
by sex-chromosomes (Van der Werf and Van den Berg, 1995) and nutrient composition and content will be the most important tar-
can be altered by chemicals such as growth regulating hormones gets of breeding. For the oil-type cultivar breeding, the major
or silver thiosulfate; Gibberellic acid induces male characters while objectives include high seed yield and the content and composi-
auxins, ethylene and cytokines induce femaleness, silver thiosul- tion in fatty acids. For paper making, both bast fiber and woody
fate promotes male flowers on female (XX) plants and is a useful core can be used but the quality of paper made from the bast
tool for producing seeds that give rise to only female plants (Ram fiber is higher. Therefore, breeding for paper production has pri-
and Sett, 1982; Kaushal, 2012). In dioecious hemp cultivars differ- marily aimed at improved bast fiber production (Hennink, 1994).
ences in growth rate and development between male and female The recent increasing demand in the market for cannabinoid-CBD
plants are evident whereby the male plants tend to flower and will result in a rapid growth of production for medicine-type hemp
senesce earlier (Meijer et al., 1995; Struik et al., 2000). This gener- (high CBD content but less than 0.2% THC in Europe Union or 0.3% in
ates variation in the crop and also the competition between plants other countries). Thus new cultivars need to be developed. In 2013,
may lead to the suppression of smaller ones which may result in about 800 ha hemp with cultivation licenses were grown for fiber
self-thinning of the field (Van der Werf et al., 1995). As such, com- and by-product CBD in Yunnan province of China, which greatly
pounds that control gender are useful tools to achieve uni-sexuality benefitted the hemp farmers economically. With this trend, two
and uniformity in hemp (Hall et al., 2012; Ram and Sett, 1982; new industrial hemp cultivars with high CBD content are expected
Kaushal, 2012). to be registered and released by Yunnan Academy of Agricultural
Several markers for maleness have been identified. SCAR mark- Sciences by the end of 2015.
ers fragments were present in both female, male and monoecious Studies were done on the improvement of hemp for specific
plants but a single band was specific for male plants. These are industries, such as textile (Amaducci, 2003, 2005), papermaking
thus not the primers themselves but feature in the region amplified (Italian paper and pulp organization), biobuilding compounds, and
that is male specific. A specific marker (SCAR, OPA08; developed plywood. Such initiatives paved the way for further industry-driven
on the MADC2 region; 391 bp fragment in male plants and two studies and the selection of new genotypes for specific uses (Zatta
larger fragments in females and monoecious plants) allows the et al., 2012). For example, new genotypes especially for textile uses
safe identification of male plants in dioecious and monoecious cul- (Toonen et al., 2004; Di Candilo et al., 2010).
tivars at all developmental stages with a quick and easy, direct
PCR method (Mandolino and Carboni, 2004 and references therein).
Three other male specific markers were developed on the MADC5, 6. Perspectives of ‘Genomics’ for hemp breeding
MADC6, and MADC400-S208 loci (Törjék et al., 2002; Li et al., 2012).
A marker for the monoecious trait is still required to fully char- In 2011 the result of the first genome sequence of C. sativa
acterize the sexuality in hemp. Obstacles are the environmental was published (Van Bakel et al., 2011). The size of the haploid
influences altering the expression of male flowers in monoecious hemp genome is 818 Mbp (female) to 843 Mbp (male). A major
plants that can change the female:monoecious rate. part of the hemp genome, 534 Mb, encompassing the vast major-
ity of the non-repetitive genome and the individual genes was
5.5. Hemp cultivars for specific environments and end-uses sequenced and an assembled draft genome of hemp became avail-
able including DNA sequences of three cultivars, the drug type
Given the strong influence of the environment on hemp biomass Purple Kush (clonal propagated) and the industrial types Finola
yield and quality, hemp cultivars were developed for specific envi- and USO-31. The genomic sequences from Finola, Purple Kush, and
ronments and end-uses. As result of the efforts in hemp breeding USO-31 were compared and elucidated many SNPs that supported
specific cultivars were designed for cultivation in Italy (Ranalli, the separation of drug strains from the fiber types. Whole genome
2004), France, Hungary, Poland, Romania, Bulgaria, Russia (former re-sequencing and genome comparisons can provide information
USSR), former Yugoslavia, Spain, former Czechoslovakia, Germany, about the amount of genetic variation that is present in the hemp
The Netherlands, Finland (De Meijer, 1995; Ranalli, 2004), Canada germplasm. A physical map of the hemp genome saturated with
(Watson et al., 2012), and China. For example, Ukrainian cultivars DNA polymorphisms will provide targets for the development of
and French cultivars differ in the length of their vegetative period. molecular markers for breeding. In hemp however, due to the high
For seed production, flowering, and seed ripening is required. level of genetic variation within and among accessions, the high
Therefore early ripening cultivars are more suitable for seed pro- number of minor alleles and the plasticity of traits in different envi-
duction in Northern Europe, where the growing season is short, ronments, such genomics studies require a specific approach in
and late ripening cultivars are suitable for the same use in the which the effect of the environment on the phenotype is included,
South of Europe. As the fiber formation finishes already a month and in which a correction is made for variation in the genetic
before ripening, late cultivars are often grown in Northern regions, background of populations (false positive correlations in specific
for the production of stem and high quality fibers. In the Northern populations). In the ‘Multihemp’ project (www.multihemp.eu) a
regions late flowering cultivars have a prolonged vegetative phase large mapping panel including wild material, landraces, and culti-
and a higher stem yield, in situations where flowering and seed vars are being genotyped by whole genome re-sequencing. Upon
ripening is not required. The harvest can be performed at different phenotyping of the mapping panel in three different environ-
developmental stages depending on the use. ments, genome wide association mapping (GWAS; reviewed in
Hemp is a promising phytoremediation crop for clean-up of Korte and Farlow, 2013) is being performed to provide informa-
heavy metal pollutions, different genotypes demonstrate distinct tion about important alleles, quantitative trait loci, and linked DNA
differences in their abilities to tolerate and accumulate heavy polymorphisms in the underlying genes. Next generation sequenc-
metals (Zeng et al., 2013). Besides, different hemp cultivars have ing technologies increase the possibilities of mutation detection
variable tolerance abilities under low temperature, drought and by facilitating the screening of large numbers of plants for rare,
saline stress, and resistance for lodging and diseases (Guo et al., induced or natural genetic variation in specific target genes known
2010, 2011). to be involved in important traits. The targeted mutations eliminate
40 E.M.J. Salentijn et al. / Industrial Crops and Products 68 (2015) 32–41

or alter the functionality of genes and selected mutant genotypes Gamelas, J.A.F., 2013. The surface properties of cellulose and lignocellulosic mate-
are directly useful as breeding material (Comai et al., 2004; Elshire rials assessed by inverse gas chromatography: a review. Cellulose 20, 2675–
2693.
et al., 2011; McCallum et al., 2000; Metzker, 2010). Garcia-Jaldon, C., Dupeyre, D., Vignon, M.R., 1998. Fibers from semi-retted hemp
bundles by steam explosion treatment. Biomass Bioenergy 14, 251–260.
Gilmore, S., Peakall, R., Robertson, J., 2007. Organelle DNA haplotypes reflect crop-
Acknowledgments use characteristics and geographic origins of Cannabis sativa. Forensic Sci. Int.
172, 179–190.
The authors gratefully acknowledge funding from the European Gonzalez-Garcia, S., Luo, L., Moreira, M.T., Feijoo, G., Huppes, G., 2012. Life cycle
assessment of hemp hurds use in second generation ethanol production.
Union consortia FIBRA (project ID 311965) and MultiHemp (project Biomass Bioenergy 36, 268–279.
ID 311849). Guo, H.Y., Yang, M., Xu, Y.P., Guo, M.B., Zhang, Q.Y., Chen, X., Wang, H.H., Wu, J.X.,
2013. Cultivation Techniques for Hemp in Dryland. The Nationalities Publishing
House of Yunnan, Kunming, pp. 1–98, in Chinese.
References Guo, L., Wang, D.K., Wang, M.Z., Ren, C.M., Li, Y.F., Jie, X., Zheng, W., 2010. Adaptation
analysis on different hemp varieties in drought and saline area. Plant Fiber Sci.
Amaducci, S., Amaducci, M.T., Benati, R., Venturi, G., 2000. Crop yield and quality China 32 (4), 202–205, in Chinese.
parameters of four annual fiber crops (hemp, kenaf, maize and sorghum) in the Guo, Y., Wang, Y.F., Qiu, C.S., Long, S.H., Deng, X., Hao, D.M., 2011. Preliminary study
North of Italy. Ind. Crops Prod. 11, 179–186. on effects of drought stress on physiological characteristics and growth of dif-
Amaducci, S., 2003. HEMP-SYS: design, development and up-scaling of a sustainable ferent hemp cultivars (Cannabis sativa L.). Plant Fiber Sci. in China 33, 235–239,
production system for hemp textiles – an integrated quality system approach. in Chinese.
J. Ind. Hemp 8, 79–83. Hall, J., Bhattarai, S.P., Midmore, D.J., 2012. Review of flowering control in industrial
Amaducci, S., 2005. Hemp production in Italy. J. Ind. Hemp 10, 109–115. hemp. J. Nat. Fibers 9, 23–36.
Amaducci, S., Pelatti, F., Medeghini Bonatti, P., 2005. Fiber development in hemp Hanks, A., 2008. Canadian hemp update 2007. J. Ind. Hemp 13, 49–57.
(Cannabis sativa L.) as affected by agrotechnique: preliminary results of a micro- Hennink, S., 1994. Optimization of breeding for agronomic traits in fiber hemp
scopic study. J. Ind. Hemp 10, 31–48. (Cannabis sativa L.) by study of parent–offspring relationships. Euphytica 78,
Amaducci, S., Colauzzi, M., Bellocchi, G., Venturi, G., 2008a. Modelling post-emergent 69–76.
hemp phenology (Cannabis sativa L.): theory and evaluation. Eur. J. Agron 28, Hanegraaf, M.C., Biewinga, E.E., Van der Bijl, G., 1998. Assessing the ecological and
90–102. economic sustainability of energy crops. Biomass and Bioenergy 15, 345–355.
Amaducci, S., Zatta, A., Pelatti, F., Venturi, G., 2008b. Influence of agronomic factors Hänninen, T., Thygesen, A., Mehmood, S., Madsen, B., Hughes, M., 2012. Mechani-
on yield and quality of hemp (Cannabis sativa L.) fiber and implication for an cal processing of bast fibers: the occurrence of damage and its effect on fiber
innovative production system. Field Crops Res. 107, 161–169. structure. Ind. Crops Prod. 39, 7–11.
Amaducci, S., Colauzzi, M., Zatta, A., Venturi, G., 2008c. Flowering dynamics in Hillig, K.W., 2005. Genetic evidence for speciation in Cannabis (Cannabaceae). Genet.
monoecious and dioecious hemp genotypes. J. Ind. Hemp 13, 5–19. Resour. Crop Evol. 52, 161–180.
Amaducci, S., Colauzzi, M., Bellocchi, G., Cosentino, S.L., Pahkala, K., Stomph, T.J., Holoborodko, P., Virovets, V., Laiko, I., Bertucelli, S., Beherec, O., Fournier,
Westerhuis, W., Zatta, A., Venturi, G., 2012. Evaluation of a phenological model G., 2014. Results of Efforts by French and Ukranian Breeders to
for strategic decisions for hemp (Cannabis Sativa L.) biomass production across Reduce Cannabinoid Levels in Industrial Hemp (Cannabis sativa L.).
European sites. Ind. Crops Prod. 37, 100–110. www.interchanvre.com/docs/article-Laiko.pdf
Berenji, J., Sikora, V., Fournier, G., Beherec, O., 2013. Genetics and selection of hemp. Jankauskienè, Z., Gruzdevienè, E., 2009. Beniko and Bialobrezskie – industrial
In: Bouloc, P., Allegret, S., Arnaud, L. (Eds.), Hemp: Industrial Production and hemp varieties in Lithuania. Proceedings of the 7th International Scientific and
Uses. CABI, Wallingford, UK, pp. 48–71. Practical Conference. Environment Technology Resources 1, 228–234, ISBN 978-
Bredemann, G., 1942. Die Bestimmung des Fasergehaltes bei Massenunter- 9984-44-027-9.
suchungen von Hanf, Flachs, Fasernesseln und anderen Bastfaserpflanzen. Johnson, R., 2013. Hemp as an Agricultural Commodity. CRS Report for Congress.
Faserforschung 16, 14–39, in German. Congressional Research Service https://www.fas.org/sgp/crs/misc/RL32725.pdf
Chen, X., Guo, M.B., Zhang, Q.Y., Xu, Y.P., Guo, H.Y., Yang, M., Yang, Q.H., 2013. Chemo- Karus, M., Vogt, D., 2004. European hemp industry: cultivation, processing and prod-
type and genotype of Cannabinoids in hemp landrace from southern Yunnan. uct lines. Euphytica 140, 7–12.
Acta Bot. Boreal. Occid. Sin. 33, 1817–1822, in Chinese. Kaushal, S., 2012. Impact of physical and chemical mutagens on sex expression in
Comai, L., Young, K., Till, B.J., Reynolds, S.H., Greene, E.A., Codomo, C.A., Enns, L.C., Cannabis sativa. Indian J. Fundam. Appl. Life Sci. 2, 97–103.
Johnson, J.E., Burtner, C., Odden, A.R., Henikoff, S., 2004. Efficient discovery of Khalil, H.P.S.A., Bhat, A.H., Yusra, A.F.I., 2012. Green composites from sustainable
DNA polymorphisms in natural populations by Ecotilling. Plant J. 37, 778–786. cellulose nanofibrils: a review. Carbohydr. Polym. 87, 963–979.
De Meijer, E.P.M., Keizer, L.C.P., 1994. Variation of cannabis for phenological devel- Korte, A., Farlow, A., 2013. The advantages and limitations of trait
opment and stem elongation in relation to stem production. Field Crops Res. 38, analysis with GWAS: a review. Plant Methods 9, 29 http://www.
37–46. plantmethods.com/content/9/1/29
De Meijer, E.P.M., 1994. Variation of cannabis with reference to stem quality for Laate, E.A., Report of the Government of Alberta, Agriculture and Rural Development
paper pulp production. Ind. Crops Prod. 3, 201–211. Department 2012. Industrial hemp production in Canada., pp. 1–7.
De Meijer, E.P.M., 1995. Fiber hemp cultivars: a survey of origin, ancestry, availability Li, S.J., Xin, P.Y., Guo, H.Y., Yang, M., 2012. Study of male-specific RAPD and SCAR
and brief agronomic characteristics. J. Int. Hemp Assoc. 2, 66–73. marker in Cannabis sativa L. Guangdong Agric. Sci. 24, 151–154, in Chinese.
De Meijer, E.P.M., Keizer, L.C.P., 1996. Patterns of diversity in cannabis. Gen. Resour. Lisson, S.N., Mendham, N.J., Carberry, P.S., 2000. Development of a hemp (Cannabis
Crop Evol. 43, 41–52. sativa L.) simulation model 2. The flowering response of two hemp cultivars to
De Meijer, E.P.M., Bagatta, M., Carboni, A., Crucitti, P., Moliterni, V.M., Ranalli, P., photoperiod. Aust. J. Exp. Agric. 40, 413–417.
Mandolino, G., 2003. The inheritance of chemical phenotype in Cannabis sativa Mandolino, G., Carboni, A., 2004. Potential of marker-assisted selection in hemp
L. Genetics 163, 335–346. genetic improvement. Euphytica 140, 107–120.
De Meijer, E.P.M., Hammond, K.M., 2005. The inheritance of chemical phenotype McCallum, C.M., Comai, L., Greene, E.A., Henikoff, S., 2000. Targeted screening for
in Cannabis sativa L. (II): cannabigerol predominant plants. Euphytica 145, induced mutations. Nat. Biotechnol. 18, 455–457.
189–198. Mediavilla, V., Leupin, M., Keller, A., 2001. Influence of the growth stage of industrial
De Meijer, E.P.M., Hammond, K.M., Sutton, A., 2009a. The inheritance of chemical hemp on the yield formation in relation to certain fiber quality traits. Ind. Crops
phenotype in Cannabis sativa L. (IV): cannabinoid-free plants. Euphytica 168, Prod. 13, 49–56.
95–112. Meijer, W.J.M., Van der Werf, H.M.G., Mathijssen, E.W.J.M., Van den Brink, P.W.M.,
De Meijer, E.P.M., Hammond, K.M., Micheler, M., 2009b. The inheritance of chemical 1995. Constraints to dry matter production in fiber hemp (Cannabis sativa L.).
phenotype in Cannabis sativa L. (III): variation in cannabichromene proportion. Eur. J. Agron. 4, 109–117.
Euphytica 165, 293–311. Metzker, M.L., 2010. Sequencing technologies – the next generation. Nat. Rev. Genet.
Dewey, L.H., 1913. Hemp. In: Yearbook of the USDA., pp. 283–316. 11, 31–46, http://dx.doi.org/10.1038/nrg2626.
Dewey, L.H., 1927. Hemp varieties of improved type are result of selection. In: Year- Mukherjee, A., Roy, S.C., De Bera, S., Jiang, H.E., Li, X., Li, C.S., Bera, S., 2008. Results of
book of the USDA., pp. 358–361. molecular analysis of an archaeological hemp (Cannabis sativa L.) DNA sample
Deyholos, M.K., Potter, S., 2014. Engineering bast fiber feedstocks for from North West China. Genet. Resour. Crop Evol. 55, 481–485.
use in composite materials. Biocatal. Agric. Biotechnol. 3, 53–57, Placet, V., 2009. Characterization of the thermo-mechanical behaviour of Hemp
http://dx.doi.org/10.1016/j.bcab.2013.09.001. fibers intended for the manufacturing of high performance composites. Compos.
Di Candilo, M., Bonatti, P.M., Guidetti, C., Focher, B., Grippo, C., Tamburini, E., A: Appl. Sci. Manuf. 40, 1111–1118.
Mastromei, G., 2010. Effects of selected pectinolytic bacterial strains on water- Prade, T., Svensson, S.E., Andersson, A., Mattsson, J.E., 2011. Biomass and energy
retting of hemp and fiber properties. J. Appl. Microbiol. 108, 194–203. yield of industrial hemp grown for biogas and solid fuel. Biomass Bioenergy 35,
Elfordy, S., Lucas, F., Tancret, F., Scudeller, Y., Goudet, L., 2008. Mechanical and 3040–3049.
thermal properties of lime and hemp concrete (hempcrete) manufactured by Prade, T., Finell, M., Svensson, S.E., Mattsson, J.E., 2012a. Effect of harvest date on
a projection process. Construct. Build. Mater. 22, 2116–2123. combustion related fuel properties of industrial hemp (Cannabis sativa L.). Fuel
Elshire, R.J., Glaubitz, J.C., Sun, Q., Poland, J.A., Kawamoto, K., Buckler, E.S., 2011. 102, 592–604.
A robust, simple genotyping-by-sequencing (GBS) approach for high diversity Prade, T., Svensson, S.E., Mattsson, J.E., 2012b. Energy balances for biogas and solid
species. PLoS ONE 6, e19379, http://dx.doi.org/10.1371/journal.pone.0019379. biofuel production from industrial hemp. Biomass Bioenergy 40, 36–52.
 E.M.J. Salentijn et al. / Industrial Crops and Products 68 (2015) 32–41 41

Posselt, U.K., 2010. Breeding methods in cross-pollinated species. In: Boller, B. (Ed.), Törjék, O., Kiss, E., Bucherna, N., Homoki, H., Finta-Korpelova, Z., Bocsa, I., Nagy, I.,
Fodder Crops and Amenity Grasses. Handbook of Plant Breeding, vol. 5. Springer Heszky, L., 2002. Identification and characterization of sex-specific molecular
Science and Business Media, LLC. markers in hemp. Novenytermeles 51, 639–655.
Ragit, S.S., Mohapatra, S.K., Gill, P., Kundu, K., 2012. Brown hemp methyl ester: trans- Van Bakel, H., Stout, J.M., Cote, A.G., Tallon, C.M., Sharpe, A.G., Hughes, T.R., Page, J.E.,
esterification process and evaluation of fuel properties. Biomass Bioenergy 41, 2011. The draft genome and transcriptome of Cannabis sativa. Genome Biol. 12,
14–20. R102.
Ram, H.Y.M., Sett, R., 1982. Induction of fertile male flowers in genetically female Van den Broeck, H.C., Maliepaard, C., Ebskamp, M.J.M., Toonen, M.A.J., Koops, A.J.,
Cannabis sativa plants by silver-nitrate and silver thiosulfate anionic complex. 2008. Differential expression of genes involved in C-1 metabolism and lignin
Theor. Appl. Genet. 62, 369–375. biosynthesis in wooden core and bast tissues of fiber hemp (Cannabis sativa L.).
Ranalli, P., 2004. Current status and future scenarios of hemp breeding. Euphytica Plant Sci. 174, 205–220.
140, 121–131. Van der Werf, H.M.G., Harsveld van der Veen, J.E., Bouma, A.T.M., ten Cate, M., 1994.
Ranalli, P., Venturi, G., 2004. Hemp as a raw material for industrial applications. Quality of hemp (Cannabis sativa L.), stems as a raw material for paper. Ind. Crops
Euphytica 140, 1–6. Prod. 2, 219–227.
Robbins, L., Snell, W., Halich, G., Maynard, L., Dillon, C., Spalding, D., 2013. Van der Werf, H.M.G., Wijlhuizen, M., De Schutter, J.A.A., 1995. Plant-density and
Economic considerations for growing industrial hemp: Implications for Ken- self-thinning affect yield and quality of fiber hemp (Cannabis sativa L.). Field
tucky’s farmers and agricultural economy department of agricultural economics Crops Res. 40, 153–164.
(AEC). University of Kentucky http://www2.ca.uky.edu/cmspubsclass/files/ Van der Werf, H.M.G., Van den Berg, W., 1995. Nitrogen fertilization and sex expres-
EconomicConsiderationsforGrowingIndustrialHemp.pdf sion affect size and variability of fiber hemp (Cannabis sativa L.). Oecologia 103,
Rowell, R.M., Han, J.S., Rowell, J.S., 2000. Characterization and factors effecting fiber 462–470.
properties. In: Frollini, E., Leao, A.L., Mattoso, L.H.C. (Eds.), Natural Polymers Van der Werf, H.M.G., Mathijssen, E.W.J.M., Haverkort, A.J., 1996. The potential of
and Agrofibers based Composites: Preparation, Properties and Applications. hemp (Cannabis sativa L.) for sustainable fiber production: a crop physiological
Embrapa Instrumentacao Agropecuaria, Sao Carlos, Brazil. appraisal. Ann. Appl. Biol. 129, 109–123.
Sakamoto, K., Akiyama, Y., Fukui, K., Kamada, H., Satoh, S., 1998. Characterization, Van der Werf, H.M.G., Turunen, L., 2008. The environmental impacts of the produc-
genome sizes and morphology of sex chromosome in hemp (Cannabis sativa L.). tion of hemp and flax textile yarn. Ind. Crops Prod. 27, 1–10.
Cytologia (Tokyo) 63, 459–464. Watson, K., Kostuik, J., McEachern, S., Melnychenko, A., 2012. Industrial hemp
Shahzad, A., 2012. Hemp fiber and its composites: a review. J. Compos. Mater. 46, fiber variety trial. In: Parkland Crop Diversification Annual Report., pp. 107–
973–986. 124.
Shui, S., Plastina, A., 2013. FAO/ICAC World Apparel Fiber Consumption Survey. Inter- Yang, X.Y., 1991. History of cultivation on hemp, sesame and flax. Agric. Archaeol.
national Cotton Advisory Committee, Washington, DC, ISBN 9780979390395. 03, 267–274, in Chinese.
Struik, P.C., Amaducci, S., Bullard, M.J., Stutterheim, N.C., Venturi, G., Cromack, H.T.H., Zatta, A., Monti, A., Venturi, G., 2012. Eighty years of studies on industrial hemp in
2000. Agronomy of fiber hemp (Cannabis sativa L.) in Europe. Ind. Crops Prod. the Po Valley (1930–2010). J. Nat. Fibers 9, 180–196.
11, 107–118. Zeng, M., Guo, H.Y., Guo, R., Yang, M., Mao, K.M., 2013. A study on phytoremediation
Toonen, M.A.J., Maliepaard, C., Reijmers, T.H., van der Voet, H., Mastebroek, D., van of Cannabis sativa L. in heavy metals polluted soil. Chin. J. Soil Sci. 44, 472–476,
den Broeck, H.C., Ebskamp, M.J.M., Kessler, W., Kessler, R.W., 2004. Predicting in Chinese.
the chemical composition of fiber and core fraction of hemp (Cannabis sativa L.). Zeven, A.C., Zhukovsky, P.M., 1975. Dictionary of Cultivated Plants and Their Centres
Euphytica 140, 39–45. of Diversity. Pudoc, Wageningen, The Netherlands, ISBN 9022005496.