As indicated in Section 2.3, ICRAF’s Agroforestry Tree Genetic Resources Strategy states that criteria used to determine key species for conservation should include a strong use focus. There are a number of potential sources of information on agroforestry tree user’s needs that ICRAF may wish to use to prioritise which collections it should acquire and/or retain. These include ICRAF’s own user surveys, the ICRAF GRU’s species-specific records on requests for material, and useful wild genotypes and domesticated cultivars in other genebanks. Each of these is described below.
4.1. ICRAF user’s surveys
Analysis of seed distributed from ICRAF HQ Seed Bank by user sector (derived from 2012-2016 ICRAF Genebank Annual Reports; Table 4.1) show most of the seeds from the GRU going to ICRAF research projects and individual farmer/farmer organisation/CBO categories but fewer distribution to research institutions and the commercial sector. These users also have the ability to add value to and/or scale up the use of material. Similarly, distribution analysis of seedlings from Field Genebanks 2014 to 2016 ICRAF Genebank Annual Reports data show the same pattern (Table 4.2). In addition, the geographic distribution of seed samples from ICRAF HQ Seed Bank in 2012-2016 (Table 4.3) showed that 82% of the samples distributed by the GRU went to Kenya. Although some of the ‘Kenyan’ recipient seed may have been distributed more widely through ICRAF projects, it wasn’t possible to quantify this, and the conclusion was that the international reach of ICRAF’s GRU was quite limited. Notably, quantifying dispatch by samples distributed may not be a true reflection of seeds distributed. Depending on the species, the seed samples may contain several seed numbers; for example in 2014 of the 309 samples distributed HQ Seed Bank, the database records show that these were indeed 3,195,555 seeds of 46 species; if all these seeds developed to maturity, this is a great number of long-lived tree species added to agroforestry systems.
Table 4.1. Seed distribution from ICRAF HQ Seed Bank by user sector between 2012 and 2016
Table 4.2. Seed distribution from ICRAF Field Banks by user sector between 2014 and 2016
Number of samples
% of total
Agricultural Research Institutes or Universities
NARS or national genebanks
Individuals, unknown or other users
Table 4.3. Seed distribution from ICRAF HQ Seed Bank by country in 2012-2016
Number of samples
Number of species
Unites States of America
The ICRAF GRU’s 2016 user survey results confirmed these patterns (see Section 2.2.1) but also included a request for information on which species users wanted to plant. ICRAF also carried out consultations with users in 2013 (Africa) and in 2015 (Africa and SE Asia). A combined list of 65 preferred agroforestry tree species identified by users is reproduced in Annex 1.
4.2. ICRAF GRU’s species-specific records of requests for material A further useful indicator of demand is ICRAF’s seed supply records for requests for individual species. Annex 2 lists the top 50 taxa requested from ICRAF over the period 2013 to 2016.
Assorted agroforestry tree seed samples packed for distribution
4.3. Useful wild genotypes and domesticated cultivars in other genebanks
The tree collections held in the world’s agricultural, horticultural, forestry and botanical garden sectors are highly diverse and, sometimes, are accompanied by passport data that includes their utility and extent of use. These genebanks are a source of material and data encompassing wild taxa and cultivars.
The most diverse ex situ wild tree collections are held in the living collections and seed banks of the world’s 3,000 botanic gardens and arboreta (https://www.bgci.org/garden_search.php). Collectively, these manage at least one third of known plant diversity in their living collections. For example, wild tree seed collections are well represented in Kew’s Millennium Seed Bank (MSB) that holds more than 5,000 tree taxa, many with multiple accessions. The MSB’s seed list (http://apps.kew.org/seedlist/SeedlistServlet) indicates that a proportion of these collections are freely available to users, with many more taxa available under specific conditions of use. Duplicates of all MSB collections are held in national genebanks in the countries of origin, and in cases where material is not directly available from the MSB, it is often available from the counterpart national genebank. Other, large, specialist wild species genebanks include Kunming Institute of Botany’s Genebank of Wild Species (GBOWS) and the Royal Botanic Garden, Sydney’s PlantBank. The most comprehensive listing of the names of tree accessions held in the world’s botanic gardens and arboreta is Botanic Garden Conservation International’s PlantSearch database (https://www.bgci.org/plant_search.php ). PlantSearch lists more than 1.3 milllion collection records, representing more than 500 thousand taxa (including cultivars), at over a thousand contributing institutions. However, PlantSearch does not yet provide individual accession level data. This information is currently only available, where it exists, from the institutions themselves.
In the agricultural/horticultural sector, the most comprehensive list of accessions is provided by the GENESYS database (https://www.genesys-pgr.org/welcome). GENESYS currently lists more than 3.5 million accessions from almost 500 institutes, and includes accession names and some passport data. GENESYS is primarily devoted to food crops but includes some tree collections from ICRAF, ILRI, CIAT and a number of national genebanks, including those of the USDA.
Characterisation and/or passport data for wild collections in genebanks is often very limited. While collections from large, well-resourced, modern wild seed banks like the MSB and GBOWS come with detailed passport data, including latitude/longitude coordinates (from which environmental variables can be derived), soils and geology, most wild collections in national forestry genebanks may simply be accompanied by a rough locality and a date of collection. Similarly, for cultivars, characterisation data although more likely to be recorded is often of variable quality. For this reason, a key question in identifying priority collections for acquisition will be “How good is the accompanying data?” Ideally, material will have both good passport data (including GPS geo-referenced location data and other detailed information from collection sites) and good characterisation data (phenotypic and/or genotypic – especially on key traits important for breeding purposes). Accurate sample location data is increasingly important because of the ability to extrapolate from it to environment based on global climate overlays and other global geo-referenced datasets.
In addition, of course, material will need to be available for acquisition, preferably under the SMTA of the International Treaty (see Section 6.3.3.).
5. Gap analysis
A key criterion identified in ICRAF’s Agroforestry Tree Genetic Resources Strategy in terms of goals for germplasm collection, acquisition and rationalisation is to assess and act on gaps in collections. Below we examine some of the gaps in ICRAF’s collections against the global importance and user’s needs criteria discussed in Sections 3 and 4 of this document.
5.1. ICRAF GRU’s global coverage
A comparison of ICRAF’s current accessions with the top 137 useful tree species used in five or more developing/emerging countries according to the SOWFGR Review is detailed in Table 5.1.
Table 5.1. Comparison of ICRAF GRU species collections with widely used species in developing and emerging countries according to the SOWFGR Review (FAO, 2014)
5.2. ICRAF GRU’s coverage of user priority species
When compared against the species identified by user surveys as preferred or priority taxa (Annex 1), ICRAF’s GRU fares much better in terms of representation, with 48% of these species stored by the GRU and available to users. Species that are however not currently available from the ICRAF GRU are listed in Table 5.2.
Table 5.2. User priority species not currently available from the ICRAF GRU
Again, there are species in Table 5.2 that are readily available from other sources (e.g. Azadirachta indica, Mangifera indica). Some species that are widely used in important ICRAF mandate areas such as tropical Africa are however not currently available from the GRU. The ICRAF GRU should thus consider acquiring these (e.g. Khaya anthotheca, Lannea humilis, Pterocarpus angolensis and Syzygium guineensis).
5.3. ICRAF GRU’s coverage of wild genotypes and cultivars for domestication
Species under domestication by ICRAF are listed in Table 5.3.
Table 5.3. Species covered by ICRAF’s domestication programme
ICRAF’s current collections of the species that are priorities in its domestication programme (all fruit trees) are listed in Table 5.4. These collections are currently being characterised, which will enable ICRAF to formulate and develop collections acquisition, retention and discard policy for these taxa.