Genetic diversity within and between populations is essential for adaptation and survival of any species. Plant breeders use this breadth of genetic diversity to generate new cultivars. The current need for genetic diversity is critical because of the predicted threats to agricultural production from climate change combined with the ever-increasing human population (FAO, 2008). The Intergovernmental Panel on Climate Change (IPCC) (2007) concluded that the poorest countries would be hardest hit, with increased food insecurity impacting most tropical and subtropical regions due to decreased water availability and new or changed insect pest incidence. Although there are many approaches to improving food security (FAO, 2012a, b), one option that is currently underdeveloped but which could potentially make a significant contribution is a more systematic and targeted use of crop wild relatives (CWR) in crop improvement programmes. CWR have the potential to contribute beneficial traits to crops, such as biotic and abiotic stress resistance, leading to improved yield and stability (Prescott-Allen and Prescott-Allen, Reference Prescott-Allen and Prescott-Allen1988; Feuillet et al., Reference Feuillet, Langridge and Waugh2008; Maxted et al., Reference Maxted, Ford-Lloyd, Jury, Kell and Scholten2006; Maxted et al., 2008a; Guarino and Lobell, Reference Guarino and Lobell2011). While climate change will directly impact CWR diversity itself, it will also undoubtedly alter the conditions under which our crops grow. It is likely that many current crop varieties will need to be replaced if the crop is to be grown in the same places in the future (see Jones et al., Reference Jones, Lister, Jaggard and Pidgeon2003; Duveiller et al., Reference Duveiller, Singh and Nicol2007; Deryng et al., Reference Deryng, Sacks, Barford and Ramankutty2011; Li et al., Reference Li, Takahashi, Suzuki and Kaiser2011; Luck et al., Reference Luck, Spackmand, Freemand, Trębicki, Griffiths, Finlay and Chakraborty2011; Yadav et al., Reference Yadav, Redden, Hatfield, Lotze-Campen and Hall2011). Failure to meet this challenge could have a devastating impact not only on food security but also on the global economy and social stability.
Yet CWR, like other wild plant species, are highly threatened. It is estimated that 20% of all plants globally are threatened with extinction (Brummitt and Bachman, Reference Brummitt and Bachman2010) and specifically for European CWR species, 11.5% (66) of the species are considered as threatened using the IUCN (2001) Red List Criteria, with 3.3% (19) of them being Critically Endangered, 4.4% (22) Endangered and 3.8% (25) Vulnerable – a further 4.5% (26) of the species are classified as Near Threatened (19) (Kell et al., Reference Kell, Maxted and Bitz2011). Furthermore, Jarvis et al. (Reference Jarvis, Lane and Hijmans2008) found that 16–22% of CWR species of three target gene pools would go extinct by 2055 with 50% loss of distributional range for the majority of species. The potential threat and the possible exploitation benefit of sustainably using CWR have been placed firmly on the international conservation agenda with explicit mention of the need for conservation in the Global Plan of Action for the Conservation and Sustainable Utilization of Plant Genetic Resources for Food and Agriculture (FAO, 1996), the Global Strategy for Plant Conservation (CBD, 2010a) and the Conference of the Parties to the Convention on Biological Diversity (CBD) in their Strategic Plan agreed in Nagoya (CBD, 2010b).
To help ensure the effective conservation of crop gene pools, a methodology for genetic gap analysis (Maxted et al., Reference Maxted, Dulloo, Ford-Lloyd, Iriondo and Jarvis2008b) has been proposed and has been applied for several crops (Maxted et al., Reference Maxted, Mabuza-Dlamini, Moss, Padulosi, Jarvis and Guarino2004, Reference Maxted, White, Valkoun, Konopka and Hargreaves2008c; Ramírez-Villegas et al., Reference Ramírez-Villegas, Khoury, Jarvis, Debouck and Guarino2010). The methodology utilizes the GIS to compare in situ CWR diversity with diversity that has been sampled and conserved either in situ or ex situ. The aim of this paper is to undertake gap analysis of wild taxa of the genus Hordeum as represented in genebank collections and to propose conservation strategies to conserve Hordeum genetic diversity.
The genus Hordeum L. belongs to the Triticeae tribe of the family Poaceae (Gramineae) and consists of 32 species and, in total, 45 taxa (Bothmer et al., Reference Bothmer von, Jacobsen, Baden, Jørgensen and Linde-Laursen1995). The primary gene pool of cultivated barley (Hordeum vulgare subsp. vulgare) consists of the wild relative H. vulgare subsp. spontaneum (C. Koch) Thell., the secondary gene pool consists of Hordeum bulbosum L. and the tertiary gene pool consists of all the remaining Hordeum species (Bothmer et al., Reference Bothmer von, Jacobsen, Baden, Jørgensen and Linde-Laursen1995). The species are annual or perennial grasses and are found in the Mediterranean, western North America, southern South America and Central Asia, but the crop barley was domesticated in the Israeli–Jordan part of the Fertile Crescent about 10,000 years ago (Badr et al., Reference Badr, Müller, Schäfer-Pregl, El Rabey, Effgen, Ibrahim, Pozzi, Rohde and Salamini2000). Hordeum species inhabit a wide range of habitats such as sandy beaches, grasslands, fresh and saltwater areas, steppes and alpine meadows (Bothmer et al., Reference Bothmer von, Jacobsen, Baden, Jørgensen and Linde-Laursen1995). Cultivated barley is used widely in brewing alcoholic beverages, or as human foods and livestock feeds.
Materials and methods
Data preparation
Data were gathered from seven sources selected among the largest collections of wild Hordeum accessions (Knüpffer, Reference Knüpffer2009). A total of 16,716 wild Hordeum accessions were taken from the Global Inventory of Barley Plant Genetic Resources, which was compiled at the International Centre for Agricultural Research in the Dry Areas (ICARDA, Aleppo, Syria); Leibniz Institute of Plant Genetics and Crop Plant Research (Gatersleben, Germany) provided data for 1280 accessions of Hordeum wild taxa, Nordic Genetic Resource Centre (Alnarp, Sweden) provided data for 1333 Hordeum accessions, EURISCO (Web-based catalogue of ex situ plant collections maintained in Europe – http://eurisco.ecpgr.org/static/index.html) provided data for 5646 Hordeum accessions, Plant Gene Resources of Canada (Saskatoon Research Centre, Saskatchewan, Canada) provided data for 56,309 Hordeum accessions,GRIN (Genetic Resources Information Network, United States Department of Agriculture – Agricultural Research Service, Beltsville, MA, USA) provided data for 30,617 Hordeum accessions, and R. von Bothmer compiled 2763 Hordeum accessions from various sources. Taxonomic data from each source were standardized using the Bothmer et al. (Reference Bothmer von, Jacobsen, Baden, Jørgensen and Linde-Laursen1995) nomenclature (hybrid accession data were retained in all datasets where there were adequate location identifiers). For each dataset, the countries listed were converted to the three-letter ISO ALPHA-3 standard using the website http://unstats.un.org/unsd/methods/m49/m49alpha.htm (accessed February 2012). The longitude and latitude coordinates in each dataset were converted to decimal degrees to enable the data to be processed by GIS (Geographic Information System) tools.
The seven separate data sources contained duplicate accessions, and these were identified and removed to ensure uniqueness and maximum accuracy. Accessions that had neither locality nor coordinate data in the database were deleted. All accessions without latitude and longitude were georeferenced using an online global gazetteer (Falling Rain, 2012) with the locality data provided. Coordinates of localities which could not be found using the gazetteer were found using either ‘Google Earth’ (2012) or the National Geospatial-Intelligence Agency website (2012); however, the accessions for which no coordinates could be found were deleted from the datasets.
All accessions were mapped using DIVA-GIS (Hijmans et al., Reference Hijmans, Guarino, Jarvis, O'Brien, Mathur, Bussink, Cruz, Barrantes and Rojas2005) and any accessions that obviously had wrong coordinates (for example, were in the sea) were clearly noticeable and were recorded. Each of these accessions was then examined and reassigned coordinates using Google Earth (2012) or the Falling Rain (2012) gazetteer. Google Earth (2012) was used to update accession details held in countries which are no longer recognized, such as the former USSR and former Yugoslavia (UN World Macro Regions and components, 2009). The database was searched for all accessions associated with countries which are no longer recognized, and these entries were updated with the current country name. The final version of the Hordeum database contained the data of 17,131 unique georeferenced germplasm accessions and represents all of the 45 wild taxa in the genus Hordeum. This database is available on request from the Genetic Resources Section, ICARDA, PO Box 5466, Aleppo, Syrian Arab Republic.
Spatial analysis
DIVA-GIS version 5.2 is a free GIS computer program that was downloaded from ‘www.diva-gis.org’ along with current WorldClim climatic data at the 10 min resolution and a world map with administrative boundaries. Data on current and proposed global protected areas were downloaded from the IUCN and UNEP-WCMC (2010) website as a polygon shapefile. This contained both IUCN reserves and other conservation bodies such as the World Heritage Convention.
Ex situ conservation gap analysis
Distribution maps of all the species and subspecies were created by importing the georeferenced data into DIVA-GIS and placing them on a world map. Predicted distribution maps were then created for each species and subspecies using the Bioclim method in DIVA-GIS. This algorithm uses climate data and the known localities of a species to suggest areas where that species is likely to exist. Each distribution map was then compared with its predicted distribution map to identify regions where further ex situ collecting missions might be beneficial (depending on whether missions are permitted in that region or country). The priorities for ex situ conservation were ranked based upon perceived threat (inferred from narrow geographical distribution).
In situ conservation species richness and complementary analysis
DIVA-GIS version 5.2 was used to produce a global map of species richness for the Hordeum genus. A species-rich area in this case will be one where there are many different species of Hordeum present. The default cell size was used (1°) and the output variable was defined as ‘Richness’ and ‘Number of different classes’. This method was also used to produce a map of observation richness (areas with the highest number of accessions) but the output variables were defined here as ‘Richness’ and ‘Number of observations’. The reserve selection algorithm was also used in DIVA-GIS to identify which regions would be best for conserving all of the diversity of the genus Hordeum using a minimum number of 100 km × 100 km grid cells. All of these maps were then compared with the global protected areas dataset to discover where species exist in established reserves. This is an important procedure, as often it costs less to incorporate a species into an existing reserve management plan than to set up a whole new reserve.
Results
Database content
The data for the accessions listed in the Hordeum database were rarely complete, and many accessions contained just taxonomic data and collection location details. The taxonomic data in the Hordeum database were 100% complete and the majority of the geographic data were also complete; however, there were a few accessions that had missing location descriptions but had georeferenced coordinates. The amount of ecological and curatorial data was relatively poorly represented with only 19.7% of accessions having altitude data. This highlights the importance of recording complete and detailed data during germplasm collection expeditions.
H. vulgare subsp. spontaneum was the most frequently collected taxon within the database with 13,258 accessions (77.4% of total accessions). In the database, there were 21 species that have fewer than 100 accessions; these accounted for only 4.4% of the total accessions (Table 1). Hordeum guatemalense Bothmer, Jacobsen & Jørgensen and Hordeum erectifolium Bothmer, Jacobsen & Jørgensen are highlighted as being the most poorly represented species quantitatively in the database with one accession each. It is also likely that duplicates exist in the database as when accessions could not be verified as a complete match they were left in. One source of error is the rather frequent wrong determinations of the wild taxa. This could skew the results of the analysis, particularly for those taxa with few accessions, causing a false impression of the current state of conservation for the taxon.
Number of database seed accessions and priority ranking for ex situ collection for Hordeum species
Geographical distribution
The genus Hordeum is widely distributed from subtropical areas in South America to arctic areas in Central Asia and North America and is found from sea level up to more than 4500 m altitude. Hordeum species can be found in temperate regions in both the northern and southern hemispheres.
The primary gene pool (H. vulgare subsp. spontaneum) is centred in the Middle East, spreading from around Libya and Egypt in the west, Greece across the Fertile Crescent as far east as Kyrgyzstan and as far south-east as Pakistan. The secondary gene pool (H. bulbosum) is centred in the Mediterranean (mainly Greece and Turkey) spreading west to southern Spain and eastwards through Iran to Kyrgyzstan and Uzbekistan. Western Asia has most accessions (13,226) constituting 77% of the total accessions, although it is not a particularly species-rich area. The region with the highest number of species (21) is South America; other species-rich regions include North America (11) and Eastern Europe (10). The majority of the accessions in the database were collected in Israel (62%) where only four of the 32 species of Hordeum occur; however, many of these collected accessions are from the same populations and so could be considered duplicates. The countries in which the least number of accessions were collected include Austria (1), Brazil (1), Albania (2) and Guatemala (2). The analysis of the database shows that the countries with the highest species richness include Afghanistan (7), Argentina (20), Chile (10), Russia (7) and the USA (11). Conversely, there are 26 countries in which two or less Hordeum species have been collected. This suggests that these countries are species-poor for Hordeum; however, this may not be an accurate assessment of the situation, as Maxted et al. (Reference Maxted, Mabuza-Dlamini, Moss, Padulosi, Jarvis and Guarino2004) state that unequal sampling techniques can often lead to underestimation of species richness in under-sampled areas. However, the distribution of Hordeum species is already well known due to previous thorough surveying of herbarium material, so we can assume that these countries are species-poor for Hordeum rather than under-sampled.
Ex situ conservation gap analysis
All 32 species of Hordeum are represented as ex situ germplasm accessions in the survey; however, some are represented very poorly in terms of frequency, with the number of accessions ranging from 1 for H. erectifolium and H. guatemalense to 13,258 for H. vulgare subsp. spontaneum. To assess which species require further germplasm collection, the distribution maps of the Hordeum database accessions need to be studied along with the predicted distribution maps produced in DIVA-GIS. The results of the gap analysis are documented in Supplementary Table S1 (available online only at http://journals.cambridge.org).
In situ species richness and complementary analysis
The most species-rich area is in southern South America, mainly Argentina, as the area with the highest species richness. The main hotspot is in western Argentina, in the Mendoza Province, and contains 11 Hordeum species. The analysis highlights California, USA, as being the most species-rich area in North America with six species in the richest hotspot; however, two of these species, Hordeum marinum and Hordeum murinum are introduced species from Europe and have become invasive weeds in several parts of the world, not least North America. In Asia and Europe, the richest areas in terms of the number of species are northern Iran, western Tajikistan, eastern Azerbaijan, central Turkey and Armenia. The observation richness map shows that Israel, Jordan, Syria and Lebanon are the countries where most collections have occurred. The main observation hotspot covering central to western Israel, the West Bank, and a very small part of Jordan and Egypt contains 3,828 accessions which constitutes 22.3% of the total accessions in the database.
To designate the optimal location of in situ reserves globally based upon the accessions in the Hordeum database, DIVA-GIS employs the Rebelo algorithm which attempts to conserve maximum species diversity in a minimum number of map grid cells. As a result, a total of 12 grid cells were required to represent all 32 species of the Hordeum genus. This number is rather high due to the global spread of Hordeum species and the narrow distribution of endemic species. Five of the grid cells produced by the reserve selection algorithm were found in South America where 21 species are located according to the data in the Hordeum database. The priority site (iteration 1 of the ‘reserve selection’ algorithm) for a genetic reserve also matches with the species richness results, indicating that the Mendoza Province of Argentina is the ideal area for a reserve. Two grid cells were located in south-western USA and one in Guatemala. Three grid cells are present in Asia with one in Turkey, one in central China and one in Pakistan. The final grid cell is located in South Africa, south-west of Lesotho (Figs 1 and 2).
Complementary analysis of 100 km × 100 km grid cells in order of priority for in situ conservation, highlighting South America and Southern Africa (where 1 = high priority and 12 = low priority).
Complementary analysis of 100 km × 100 km grid cells in order of priority for in situ conservation, highlighting North America and Asia (where 1 = high priority and 12 = low priority).
The taxa in gene pools 1B (H. vulgare subsp. spontaneum) and 2 (H. bulbosum) were analysed separately due to their importance in barley breeding. Figures 3 and 4 show the richness of observations for H. vulgare subsp. spontaneum and H. bulbosum, respectively, along with known existing sites where CWR are being conserved. The central regions of Israel are highlighted (Fig. 3) as being the area with most observations of H. vulgare subsp. spontaneum. The other areas of high observations are as follows: the West Bank from Jerusalem to the border of Jordan, the northern region of Israel and across the borders into southern Lebanon, the south-western point of Syria and the north-western tip of Jordan. The Izmir area of Turkey has the most observations of H. bulbosum (Fig. 4) with Gaziantep and Urfa, also in Turkey, having high richness.
Observation richness for Hordeum vulgare subsp. spontaneum using a 50 km × 50 km resolution, also showing known crop wild relative (CWR) reserves.
Observation richness for Hordeum bulbosum using a 50 km × 50 km resolution, also showing known crop wild relative (CWR) reserves.
National protected areas
To compare the reserve sites identified previously with current in situ conservation activities, the World Database on Protected Areas (WDPA) (IUCN and UNEP-WCMC, 2010) that contains polygon shapefiles of the world's different protected areas was used. Supplementary Table S2 (available online only at http://journals.cambridge.org) shows that 11 out of 12 hotspots identified have national protected areas within them. Further analysis was conducted, listing every country where an occurrence of Hordeum species was present and when any accession was found within a protected site; the reserve name and the species were recorded too. This highlighted that the USA and Israel have the most accessions within national protected areas and, globally, 25 out of 32 species are represented at least once in a national protected area.
Discussion
Assessment of current conservation strategies
In situ conservation assessment
After reviewing online material and literature, several established genetic reserves were found that actively conserve Hordeum species, centred near the Fertile Crescent. There is a reserve in Ammiad, Galilee that includes wild Hordeum species in its management plan (Anikster et al., Reference Anikster, Feldman and Horovitz1997), and there are three genetic reserves in Turkey, one in the Kaz Dag area, another in Ceylanpinar and the last in the Amanos region, that include wild Hordeum species in their management plans (Ertug and Tan, Reference Ertug Firat and Tan1997). The reserves in Haffeh, Syria and Ham, Lebanon also conserve wild Hordeum species. The Erebuni reserve in Armenia is located 8–10 km from Yerevan and holds several species of wild Hordeum (Avagyan, Reference Avagyan2008). In addition to this, the distribution of the accessions in the Hordeum database was compared with the WDPA, and it was found that there are over 100 national protected areas in which Hordeum species potentially exist. This suggests that passive in situ conservation (where there is no active population management of the target species) is occurring for 25 species of Hordeum across the geographic range of the genus. The species not covered even passively are H. cordobense, H. erectifolium, H. euclaston, H. flexuosum, H. guatemalense, H. procerum and H. stenostachys. It is important, however, that target sites receive active in situ conservation and Hordeum species are incorporated into active management plans to ensure the species' future protection.
Ex situ conservation assessment
The Results section details where each species of Hordeum could be further collected. Overall, South America contains a wealth of Hordeum diversity which may prove to hold important genes for the future of barley breeding and the region has not been conserved sufficiently in ex situ collections; the 21 species (66% of species) found in this area account only for 7.2% of the accessions in the Hordeum database. Of the wild taxa, H. vulgare subsp. spontaneum has the highest number of accessions in the primary gene pool of cultivated barley; however, further collection of this taxon would still be advised but in varied ecological niches to ensure that the full diversity of the species is conserved. H. bulbosum is also reasonably well represented in ex situ collections with 792 accessions; this is most likely due to the relative ease of breeding useful traits from H. bulbosum into cultivated barley compared with other wild relatives (excluding H. vulgare subsp. spontaneum). Additionally, some perennial and outcrossing species are more difficult to conserve ex situ, therefore their conservation is limited in genebanks. The gap analysis indicates that the majority of Hordeum species are under-represented and require further collection.
Formulating conservation strategies
In situ conservation priorities
Hordeum species with highest priority for in situ conservation are the ones with the highest socio-economic value, these being H. brevisubulatum, H. bulbosum, H. chilense, H. cordobense, H. jubatum, H. marinum, H. murinum and H. vulgare subsp. spontaneum. This is due to the fact that these species have a particularly good prospect for utilization in wheat and barley breeding (Xu and Snape, Reference Xu and Snape1988; Kindler and Springer, Reference Kindler and Springer1991; Vaz Patto et al., Reference Vaz Patto, Aardse, Buntjer, Rubiales, Martín and Niks2001; Martín et al., Reference Martín, Cabrera, Hernández, Ramírez, Rubiales and Ballesteros2000; Knüpffer, Reference Knüpffer2009; Bothmer and Komatsuda, Reference Bothmer von and Komatsuda2010).
A genetic reserve in South America, either in Chile or Argentina, is required as a large amount of Hordeum diversity is native to this region. The species richness analysis and reserve selection identify Mendoza near the Chilean border as the best place to establish a reserve. However, further investigation is required to identify the three distinct ecotypes of H. chilense to ensure that all of them are conserved adequately. H. chilense does appear in five national protected areas; so, as an interim measure before a reserve is set up for this species, these protected areas could alter their management plans to include H. chilense conservation.
In the Middle East, several areas seem likely places to establish further reserves, the first being near Izmir in Turkey (for H. bulbosum) and the second (for H. vulgare subsp. spontaneum) being the Israel/Jordan area (DIVA-GIS suggests east of Jerusalem) where the crop was first domesticated; however, a network of several reserves in the Israel/Jordan region would be more desirable, to more effectively conserve the genetic diversity of this important taxon. Using Fig. 3, the following places would be good sites to establish reserves for H. vulgare subsp. spontaneum: near Be'er Sheva, Israel, between Hebron and Jerusalem in the West Bank, and finally the area around Irbid, Jordan. Although there is already a genetic reserve conserving wild Hordeum in Ammiad, Israel, this region is the centre of diversity for the primary gene pool of cultivated barley and could possibly benefit from another genetic reserve being established or at least incorporating wild Hordeum into the management plans of the 12 protected areas in Israel where species have a georeferenced accession. However, the genetic diversity of populations of both priority taxa should be examined to further determine that the proposed sites are conserving as much genetic diversity as possible. Additional sites should be set up in Europe and Asia to conserve H. bulbosum and H. vulgare subsp. spontaneum (covering the range of habitats H. brevisubulatum, H. bulbosum, H. chilense, H. cordobense, H. jubatum, H. marinum, H. murinum and H. vulgare subsp. spontaneum in which these species occur); however, this requires further ecological analysis and, ideally, the use of molecular marker techniques to ensure the conservation of genetically distinct populations.
A further genetic reserve should also be established (or existing protected area management plans amended) in California, USA to ensure protection of the endemic species found there, as they are threatened by anthropogenic disturbance more than other Hordeum species (Bothmer et al., Reference Bothmer von, Jacobsen, Baden, Jørgensen and Linde-Laursen1995). The gap analysis suggests that south of Bootjack, next to the Sierra National forest, would be a good place to establish such a genetic reserve.
Genetic reserves should also be created to conserve populations of H. guatemalense and H. erectifolium in Guatemala: Departamento de Huehuetenango, Chiantla, Llano de Pacquix and Argentina: Buenos Aires, Buenos Aires Provincia, Bahía Blanca Partido, respectively. As these are endemic and have small populations, it is better to set up in situ reserves than collect further seed for ex situ storage, as over-collecting could damage the populations and the existing collections in genebanks most probably already contain the majority of the genetic diversity of the species. Also, being endemic and such small populations, there is a greater threat of extinction, so setting up in situ reserves would hopefully offer better protection.
All prospective sites for reserves should be checked to actually establish that the target taxa are present and ensure viable population size before any steps to establish reserves or alter management plans are taken.
Ex situ conservation priorities
It has been argued by Ford-Lloyd et al. (Reference Ford-Lloyd, Kell and Maxted2008) that: ‘… species must be prioritised so that those which require the most urgent attention and conservation planning can be dealt with as soon as possible…’. The gap analysis used a simple prioritization technique based upon perceived threat (inferred from narrow geographical distribution) and relative numbers of ex situ conserved accessions. The species are ranked as follows:
High priority – Hordeum species represented by less than 50 ex situ accessions and/or species poorly represented in their known and predicted geographical range.
Medium priority – species that are well represented across the known and predicted geographical range with less than 100 accessions stored ex situ.
Low priority – species well represented throughout their geographical range with more than 100 accessions stored ex situ.
See Table 1 for the priority ranking for ex situ collection for Hordeum species.
For a more effective ex situ and in situ conservation of population diversity (genetic diversity), the gap analysis on the basis of species richness, geographical distribution, should be complemented with genetic diversity analysis of existing accessions using mainly molecular marker techniques. An example of such a studies is the series of papers on H. marinum (Jakob et al., Reference Jakob, Ihlow and Blattner2007), Patagonian species (Jakob et al., Reference Jakob, Heibl, Rödder and Blattner2010) and the (South) American diploids (Jakob and Blattner, Reference Jakob and Blattner2010), but these studies require extension to the entire genepool. Targeting valuable traits for ex situ and in situ conservation can also add another important dimension to concentrate the conservation efforts on populations' adaptive traits such as heat, drought and salt tolerance.
Conclusion
In conclusion, the majority of wild Hordeum taxa are under-represented in genebanks globally with 14 wild species having less than 50 accessions stored ex situ. There is an additional need for further investigation into the ecological environments that Hordeum species occupy to ensure that any adaptive traits expressed are fully conserved, particularly in the case of H. vulgare subsp. spontaneum and H. bulbosum.
Genetic reserves should ideally be established in Argentina, Guatemala and the USA to cover both the threatened endemics and high species diversity in the Americas. An important network of reserves should also be set up across the Fertile Crescent in Israel, Palestine, Syria, Jordan, Lebanon and Turkey to provide effective conservation within the centres of diversity for gene pools 1B and 2. Although there are several existing genetic reserves that claim to be actively conserving in situ Hordeum diversity, it would be expedient to review the conservation procedures in place and ensure that they conform to current international standards (Iriondo et al., Reference Iriondo, Maxted and Dulloo2008, Reference Iriondo, Maxted, Kell, Ford-Lloyd, Lara-Romero, Labokas and Magos Brehm2012).
Finally, studies are required to genetically characterize existing collections and to test the viability of rare species accessions held in genebanks to determine whether further ex situ collections are required alongside the proposed in situ conservation.
