Optimisation of breeding for agronomic traits in fibre hemp (Cannabis sativa L,) by study of parent-offspring relationships

 Euphytica 78: 69-76, 1994 

Sebastiaan Hennink  

CPRO, Wageningen, The Netherlands 

Summary

In the framework of a comprehensive Dutch research project the feasibility of hemp cultivation for paper pulp production is being investigated. This project also includes a hemp breeding programme which was initiated at CPRO-DLO in 1990. Hemp breeding is primarily aimed at improving bast fibre production since bast fibre is qualitatively superior to woody core fibre for paper pulp production. The progress of the ongoing breeding programme is hampered by the lack of knowledge on the inheritance of agronomic traits.

The following traits were examined on 252 parental plants in 1991 and subsequently on their progenies in 1992; fibre content, plant height, date of flowering, stem diameter and cannabinoid contents. For each trait heritabilities were estimated and direct and indirect effects of artificial selection were studied.

The heritability of bast fibre content was high and mass selection proved to be an efficient method causing no undesirable influence on other characters. Characters not directly related to bast fibre yield such as date of flowering, plant height and stem diameter were shown to have disadvantages as selection criteria for the improvement of bast fibre yield.

The cannabinoids THC and CBD were studied, as the acceptance of hemp cultivation requires a low level. The content of THC, the psychoactive component can be successfully reduced by mass selection, but it is not certain that mass selection is the most efficient method.  

Introduction 

In the framework of a comprehensive Dutch research project the feasibility of hemp cultivation (Cannabis sativa L.) for paper pulp production is being investigated. This project also includes a hemp breeding programme which was initiated at CPRO-DLO in 1990. The progress of the ongoing breeding programme is hampered by the lack of knowledge on the inheritance of agronomic traits. 

Traditionally only the bast fibre was used for rope and cloth production, while the woody core was considered as waste. When hemp is grown for fibre production, seed is sometimes considered a valuable by-product.

For paper production both bast fibre and woody core fibres can be utilized. The quality of paper made from bast fibre is superior to that from woody core fibre. Therefore, hemp breeding for paper production has primarily aimed at improved bast fibre production. The yielding (fibre) capacity can be improved directly by selection for yield components, i.e. stem yield and bast fibre content, or indirectly by selection for characters such as date of flowering, plant height or stem diameter (Hoffman, 1957). Late flowering cultivars produce greater fibre yields than early flowering cultivars due to a longer vegetative growing period (van der Werf, 1993). For annual fibre crops a strong relationship between plant height and stem yield is often reported, for instance, in kenaf total dry matter yield is closely associated with plant height (Muchow, 1979).  

Breeding of fibre hemp has been subject of many studies (Bredemann, 1938; Hoffman, 1957; Bocsa,1958; Bocsa, 1969). Since the 1920's bast fibre content has more than doubled from approximately 12-15% to 25-35% (Bredemann, 1922; Bredemann, 1938; Heimann, 1990). Breeders in Hungary and France discontinued selection for increase of bast fibre content, since further increase of bast fibre content is accompanied by an increase of secondary bast fibre. This would decrease quality for cloth and rope production (Bocsa, personal communication, 1991). In contrast high bast fibre levels have no negative effect on paper quality. 

For political and social acceptance of large scale hemp cultivation, the content of cannabinoids, especially delta-9-tetrahydrocannabinol (THC), should be minimal. Plants containing, at initial seed maturity less than 0.3% THC in the dry matter of the female inflorescense are considered to have no psychoactive potency (De Meijer et al., 1992), individual plants can have THC levels above the tolerated maximum, although the overall population has an acceptable level.  

A successful breeding programme for the reduction of cannabinoid contents was initiated in the former USSR in the 1970's. Cultivars completely lacking THC were obtained. In the 1980's breeders in France and Poland also started breeding for low cannabinoid contents. In France new selections of existing cultivars were released having a THC content less than 0.07% instead of the original content of 0.1-0.3% (van der Werf, 1992).

Mass selection is a common breeding method in hemp, indicating that the heritability of traits under selection are high. No studies have reported any estimates of realized or narrow-sense heritabilities.

Evidently the maximisation of the response to selection is never considered.

 The use of knowledge on relations among characters is also seldom considered. The relationship between narcotic components and other characteristics has been subject of some studies. De Meijer et al. (1992) found no apparent relation except for a weak correlation between leaflet width and THC content. Small et al. (1976) found only a weak association between a large set of characters and narcotic components using multivariate analyses. It is never reported that hemp breeding was hampered by unwanted relationships between characters.

In the present paper parent-offspring (HS-families) relationships for relevant agronomic traits are used to estimate genetic parameters such as narrow-sense heritability. With the aid of artificial selection the estimated genetic parameters are verified and relationships between characters assessed. The response to selection for the observed traits will be discussed in relation with the optimisation of yield capacity.