Strategy for the Global Ex Situ Conservation of Sorghum Genetic Diversity

Online Consultation


Sorghum [Sorghum bicolor (L.) Moench] is the fifth most important cereal crop in the world. It is traditionally grown in marginal agricultural lands, primarily because of its adaptation to drought. However, this does not mean that it cannot produce tremendous yields under optimal growing conditions. For example, the Bahia region of Mexico consistently reports sorghum yields of more than 13,000 kg ha-1 on an annual basis. As the population of the world continues to grow, the demand for greater and more reliable food and feed sources will expand agricultural lands into environments that will be challenged by limited water supplies and high temperatures. Because of sorghum’s wide adaptation, its role in feeding the world will increase in importance. It is unique in that it can be used in human food production systems, as a reliable animal feed, as building material, converted in to both high and low value alcohols, and most recently as an important source of feedstock for biofuels production using starch, sugar, and biomass as feedstocks. Preserving this variability has been a primary goal of the sorghum research community. Sorghum’s relative drought and heat resistance may also increase its importance world wide if the predicted effects of global warming come to pass.

The development of an effective strategy for the conservation and use of sorghum’s immense genetic diversity is therefore essential for its long term maintenance and improvement of its utility.

Purpose of the global conservation strategy for sorghum

To contribute to an efficient and effective conservation system for sorghum genetic resources.


Development of a comprehensive International Sorghum Germplasm Collection through:

  • Identification and assessment of the global, regional and national collections of sorghum genetic resources meeting the international standards for conservation and playing a key role in a global conservation system
  • Identification of critical gaps in existing world collections of sorghum genetic resources and development of strategies to fill these
  • Development of a model for collaboration, cost sharing, and international responsibilities for the effective and efficient management of key sorghum genetic resource collections which will become the International Sorghum Germplasm Collection (ISGC).
  • Identification of information needs for a comprehensive integrated global database network that enhances the maintenance, sharing, and utilization of the
  • Capacity building in order to upgrade and enhance various collection repositories to ensure the maintenance, regeneration, and sharing of the

Focal person coordinating the strategy development process

Dr Robert G Henzell, Retired sorghum plant breeder and consultant 129 Allens Road, Warwick Q4370, Australia.

Email: and

Contributors to the strategy development process

  • Advisors: Dr CLL Gowda and Dr HD Upadhyaya (ICRISAT)
  • Nineteen sorghum collection curators completed the survey
  • Representatives of ten sorghum collections and other experts (Annex 2 – participants) participated in the Expert Consultation Meeting for Developing a Strategy for the Global Conservation of Sorghum Genetic Resources, 12 -14 March 2007, ICRISAT – Patancheru, Andhra Pradesh, India. The results of this meeting are incorporated into this sorghum

Process for Developing the Strategy

This strategy is based on an inventory of basic information and relevant data on the collections from the Germplasm Holdings Database maintained by Bioversity International (formerly know as IPGRI, International Plant Genetic Resources Institute) as at January 2006 with some additional information supplied by Ms Brigitte Laliberte. It listed (Annex 1) 122 different collections containing 194,250 “accessions”. This huge number of accessions no doubt highlights the major issue of duplication of accessions across collections. The 122 collections were prioritised based on size and likely contribution to sampling the world population of land-race-type sorghum genetic diversity. A questionnaire (see format in Annex 3) was sent to the 57 institutes so chosen in September 2006 and 19 responses were received by February 2007. The strategy is based mainly on the information supplied by the 19 respondents and complemented by the outcomes of the Expert Consultation Meeting at ICRISAT in March 2007 (Annex 2) and additional key stakeholders consultations with respect to the key elements of an effective Global Strategy. This is a representative sample containing a total of 86% of the total listed on the Bioversity International data base on January 2006.

Origin and taxonomy of Sorghum bicolor (L.) Moench

Centres of diversity

De Wet and Harlan (1972) have an excellent discussion on the origin and domestication of Sorghum bicolor (L.) Moench. In short, it is generally agreed that S. bicolor (L.) Moench originated and was domesticated in the Sub-Saharan region of Africa and spread to India and China. It probably follows that the Sub-Saharan and north east regions of Africa are the primary centres of diversity and that India and China are secondary centres. A tertiary pool of diversity is considered to be the nineteen wild species indigenous primarily to Australia, but also to South East Asia and Africa (Lazarides et al 1992).

The literature cited in the Taxonomy section below, suggest that S. bicolor (L.) Moench (or Sorghum bicolor subs. bicolor) was derived from Sorghum verticilliflorum and S. drummondii (or Sorghum bicolor subs. verticilliflorum and Sorghum bicolor subs. drummondii) in Africa and from S. halepense and S. propinquum in Asia.

Taxonomy of Sorghum bicolor (L.) Moench

Doggett (1988) and Dahlberg (2000) have comprehensive discussions on the classification of sorghum. Their papers form the basis of the following discussion along with some key publications for further reading. It is common that there are variable interpretations of taxonomic literature and that for sorghum is no exception.

Sorghum (described by Linnaeus in 1773 and named by Moench in 1794) belongs to the Family Poaceae, Tribe Andropogonaea which consists of 16 sub-tribes, one of which is Sorghastrae (Stapf 1917 and Garber 1950). Garber (1950) considered this sub-tribe comprised two main genera, Cleistachne (Hackel 1889) and Sorghum, the latter having a basic chromosome number of n=5 (see also Celarier 1956a). Recent DNA evidence suggests that n=10 is a possibility (Sprangler et al 1999). Snowden (1935 and 1936) and Garber 1950) suggested that the genus Sorghum comprises six sub- genera (later named “sections”):

  1. Eu-Sorghum – is considered to be the same as Snowden’s (1935 and 1936) section Eu- sorghum. The term Eu-sorghum has been discarded (de Wet 1978) and is now known as Sorghum. It is the likely progenitor of bicolor (n=10) in Africa, of the n=20 chromosome S. halepense in India and of S. propinquum (n=10) in South East Asia. It has been suggested that crosses between S. propinquum and S. bicolor, when the latter reached China, gave rise to the distinct Chinese landraces. S. propinquum is rhizomatous so is considered to be a progenitor of the strongly rhizomatous and geographically neighbouring, S. halepense
  2. Sorghastrum (Nash) – Sorghum and Sorghastrum probably had common ancestors

Sorghastrums were found in Africa and the Americas

  1. Chaetosorghum – (Snowden 1936) found only in Australia
  2. Stiposorghum– (Snowden 1936) found only in Australia
  3. Heterosorghum – (Snowden 1936) found in South East Asia, Phillipines and Australia
  4. Para-sorghum (Snowden 1936) – widely spread in South and East Africa, India, South East Asia, Australia, and as trichocladum in western Mexico and Guatemala

Doggett (1988) presents a discussion on how the locality of these ancient types can be related to continental shifts during the breakup of Pangaea, the super continent.

For further reading see Celarier 1956 and 1959, Lazaredes et al 1991, and Dillon et al 2004. See Sun et al (1994), Spangler (2003) and Dillon et al (2004) for a discussion of the use and results of using molecular technology to elucidate the taxonomy of Sorghum. Sprangler’s suggested changes are not included here.

The six sections (sub-genera) are distinct types and until very recently (Price et al 2006) no crosses between Sorghum and the five other sections were reported. While Prices’s work needs to be developed it seems that the genetic diversity in this tertiary centre of diversity will be available for S. bicolor grain or forage sorghum breeding.

The basis for the unique genetic diversity within the Sorghum section is derived from thousands of years’ of natural and farmer/user selection and the grain and forage sorghum breeding programs that have occurred internationally during the last century. To assist breeders, this large genetic diversity represented in the world collection of sorghums has been grouped on the basis of their similarities and this forms the basis of the following discussion, (Doggett 1988).

De Wet (1978) considered the Sorghum subgenus comprise three species, Sorghum halepense ((Linn.) Pers.)), S. propinquum ((Kunth) Hitchc.)), and Sorghum bicolor ((L.) Moench)). The cultivated taxa of the subgenus Sorghum were first grouped into 28 species by Snowden (1936). Classification schemes since then have all been based on his historic work. Later all named S. bicolor (L.) Moench were grouped into working groups (WG) by Murty and Govil (1967) and later into five races and ten

intermediate groups by Harlan and de Wet (1972). The five races are Bicolor, Kafir, Caudatum, Durra and Guinea. Dahlberg (2000) integrated the WG and race/intermediates grouping system for use by researchers who require a more detailed description of groups. Most sorghum workers recognize and use the race/intermediate method of grouping. Historically, the Kafir, Caudatum and Durra races have contributed most to grain breeding. For example, the cytoplasmic-genetic male-sterility system almost exclusively used to produce F1 hybrid cultivars is based on the Milo cytoplasm (A1) from the Durra race and to the non-restorer genes from Kafir. A1 cytoplasm male-fertility restoration genes are ubiquitous across the races.

It is evident from the surveys that the currently accepted taxonomy of sorghum has not been used. This is a point that was discussed at the Experts Meeting and it was agreed that getting the taxonomy section in this strategy correct and agreed upon is therefore very important.