Biopiracy by another name?

A critique of the FAO-CGIAR trusteeship system

GRAIN

Print version

Nearly ten years ago, back in 1994, the world's most important genebank collections for plant breeding of our major food crops were placed under the auspices of the UN Food and Agriculture Organisation (FAO). The idea was to protect this germplasm from misappropriation – or biopiracy – by establishing legal and political protection called “trusteeship”. While few people are aware of the arrangement, it is a very important one. The trusteeship agreement is about to undergo a major revision, to bring the system in line with the new FAO Treaty on Plant Genetic Resources. This presents an opportunity to critically examine the trusteeship system and asks ourselves if we're on the right boat.

Building up the collections

Since the start of the ‘green revolution'[1] in the 1960s, international agricultural research centres have been building up vast collections of plant genetic diversity called “genebanks.” They did this almost as fast as they replaced that diversity in the farmers' field with their new so-called “high-yielding varieties” which were supposed to boost productivity in Third World countries and end hunger. These genebanks are something like seed libraries: vast collections of different types of rice, corn, wheat or potatoes, held under special storage conditions to hopefully outlive us all. But they are not museums. They are supposed to serve as a reservoir of genetic diversity – all the existing genetic variation amongst our food crops that the scientists can get hold of – for breeding new and supposedly better plant types year after year.

In the early 1970s, these research centres were regrouped under the umbrella of the Consultative Group on International Agricultural Research (CGIAR), a freshly launched consortium of donors. The CGIAR was a shady operation from the start. With no legal charter and a posh secretariat inside the World Bank, the CGIAR has operated freely on its own, unhampered by democratic decision-making processes or UN oversight. And yet its impact has been significant, as the CGIAR's ensemble of research centres were very successful in convincing Third World governments that they should replace their ‘backward peasant farming systems' with modern agricultural technology and thinking – courtesy of the CGIAR.

Genebanks have been a core component of this enterprise. The green revolution scientists needed a pool of farmers' varieties from which to produce their own ‘high-yielding' ones. That's how breeding has worked for 12,000 years now: breeders don't create new genes, they rearrange old ones. Over the years, the genebank strategy for collecting and conserving this crop diversity was pumped and primed, especially through the work of one CGIAR Centre, the International Board for Plant Genetic Resources.[2] Today, there are over six million accessions – or entries – in the world's crop genebanks. About 10% of them are held by the CGIAR Centres. While that may seems like a small portion, it's actually the most important part. The CGIAR collections are reputed to be the most complete, best documented and best preserved.

Who controls the CGIAR genebanks? For whose benefit? And under what terms? The trusteeship deal that the CGIAR sealed with FAO was meant to provide something of an answer to these long nagging questions.

The purpose of the FAO-CGIAR arrangement

The essence of the arrangement agreed between the FAO and the CGIAR in 1994 was to bring the CGIAR's germplasm collections under intergovernmental authority. That was certainly the FAO's purpose, with the added benefit that the authority would be exercised through the FAO itself. But the CGIAR had an additional motive.

Placing the germplasm collections under some more official auspices than their own would help shield the CGIAR Centres from pressures they were facing in terms of intellectual property rights (IPR). The Centres collaborate with all forms of scientific research outfits: from small universities in the Pacific islands to DuPont Corporation in the US. Increasingly, intellectual property concerns were creeping into these relationships and plaguing the kind of work that could be undertaken. Plant breeders' rights – a kind of patent for plant varieties – had been available in the developed countries since the 1960s. But more restrictive, industrial patents were now being pushed worldwide to control the commercial use of genes and biotechnology. These systems give legal monopoly rights over plant genetic resources in a way that restricts the flow of information and materials, threatening plant breeding itself. As They also taint the mandate of public research institutes, since they encourage the siphoning off of publicly-funded work to private gains and compromised agendas.

Beyond day-to-day problems of research partners coming in with IPR-related restrictions on scientific cooperation, international policies were also shifting all around the CGIAR at the time. The Convention on Biological Diversity (CBD), which asserts national sovereignty over genetic resources, had just been signed. The World Trade Organisation, with its highly prescriptive Trade Related Intellectual Property Rights (TRIPS) Agreement, was just months away from being launched. Fences were being built around genetic resources left and right.

All of this was making life increasingly complicated for the CGIAR, which just wanted to produce what it considers to be “international public goods”. For the CGIAR, the arrangement with FAO was a kind of high-level political ‘hands off' sign – a shroud of protection in a world getting more greedy and individualistic about who owns what. This is not to say that going into this scheme was an altruistic move on the part of the CGIAR. Its scientific lifeline was up for grabs.

The components of the system

There are three components to the trusteeship system: the FAO-CGIAR Agreement, the joint FAO-CGIAR Statement and a model Material Transfer Agreement (MTA).

The Agreement was signed on 24 October 1994 between FAO and 12 of the 16 CGIAR Centres (those that are supposed to hold plant germplasm collections). The Agreement is identical in all 12 cases. It basically says that:

·        designated germplasm will be held “in trust” by the Centres for the benefit of humanity;

·        germplasm will be maintained properly and shared freely for the purpose of conservation, research and plant breeding; and

·        no one will be allowed to take out IPR on the germplasm.

Under the terms of the Agreement, each Centre has to decide which accessions in its genebank will be “designated” to the FAO. This selective approach is supposed to respect the different conditions under which the germplasm may have been acquired by the Centres. In some cases, there may be restrictions imposed by the donor, but the overall aim is to designate as much germplasm as possible. Currently, about 70% of the CGIAR's germplasm holdings are designated to the trusteeship system (see table). The Agreement is subject to renewal every four years. It was renewed without modification in October 1998 and is up for renewal again in October 2002 – when it will start undergoing important changes to fall into line with the new FAO International Treaty on Plant Genetic Resources.

Attached to the Agreement are joint Statements signed by the FAO and the CGIAR (one drawn up at the first signing of the Agreement, and the second accompanying the 1998 renewal.) These Statements serve to clarify a few matters, especially in terms of how trusteeship will honour the principles of the CBD. Among other points, they make two significant “clarifications”. The first is that the intention of the Agreement, and the function of any tools used to implement it, is to ensure that the germplasm “remains in the public domain”. In other words, both the FAO and the CGIAR Centres apparently equate “trusteeship” – protection of the germplasm from IPR claims – with a “public domain” status of the material. The second one is that the Centres explicitly reject any responsibility for enforcement of the Agreement. When shipping out germplasm to people who request samples, or to institutes doing safety backups, “in neither case will the source Centre be under an obligation to monitor the compliance of the recipient with these undertakings [i.e. to refrain from claiming IPR]; the obligation of the source Centre will be limited to obtaining such undertakings on the part of the recipient.”

Finally, the FAO and the CGIAR drew up a standard MTA. This is a contract that is supposed to accompany all transfers of designated germplasm from the Centres to whoever requests samples. The main feature of the MTA is that it tells the germplasm recipient that he or she has “no rights” to claim intellectual property over the material – and that this provision extends to anyone with whom he or she shares it further (third parties). No signature is required to express acceptance of these terms on the part of the recipient: merely opening the package signifies agreement. The MTA also makes it clear that direct commercialisation of the germplasm – through, for example, multiplication and distribution of seeds to farmers – is perfectly within bounds of the agreement.

The problems

On the surface, a hands-off policy to prevent ‘misappropriation' of the crop genetic resources held by the CGIAR seems like an excellent thing. However, if you look closely at how the system is set up, what it implies for farmers, and how well it's being implemented, there is a lot that appears fundamentally wrong with it.

1. Farmers' rights denied

The FAO-CGIAR trusteeship system does not recognise, much less support or help enforce, farmers' rights. This is the most profound problem with the whole arrangement, not least since it involves two institutions which claim to be all for the rights of farmers with respect to the management of agricultural biodiversity.

Why are farmers' rights so important? More than 70% of the designated germplasm is composed of landraces and wild materials. That translates to half a million different plant varieties that were taken from farmers and indigenous peoples across the developing world. (Many wild plants are actively cared for by local communities, even if they are not directly cultivated.) No one is saying that farmers or indigenous peoples should be considered “owners” of those materials. But they developed them, conserved them, and ended up giving them to all of us through the genebank system set up by the CGIAR. And yet the trusteeship system ignores this history and walks right over any relationship that the farmers had – or continue to have – with the seeds.

No farmers were involved in drafting the Agreement or asked to approve it. Neither do farmers have any authority or practical means to oversee its implementation or express any grievances about it. If they don't like it, tough luck. This is not a theoretical problem. In 1998, RAFI[3] discovered that a whole number of chickpea and lentil varieties taken from farmers in South Asia and ‘entrusted' to one of the CGIAR Centres (the International Crops Research Institute for the Semi-Arid Tropics in India) had been sent to Australia where some local breeders obtained plant breeders' rights on them. The materials had not been modified in any way through plant breeding, but they were ‘new' to Australia and were legally appropriated. The same happened with cereals from another CGIAR institute, the International Centre for Agricultural Research in the Dry Areas, based in Syria. The same could easily be happening with more seeds from more farmers in other places. In other words, genetic resources taken from farming communities are being claimed as other people's intellectual property through or despite this trusteeship system.

The problem of farmers being locked out of the system goes further still. In 2001, Thai NGOs and social movements learned that the country's cherished Jasmine rice, which was being conserved under the trusteeship arrangement at the CGIAR's International Rice Research Institute (IRRI), in the Philippines, was shipped to the United States without an MTA. This triggered off huge protests because of concerns about the Jasmine rice being genetically engineered, patented, and/or grown in the US to undercut Thailand's export market. Not only did IRRI fail to honour its responsibilities, but important economic and political insecurities were generated among the farming communities that the rice came from.

These materials come from farmers. Yet farmers are denied any rights or role to control what happens to them under this system. In the case of the Jasmine rice, the farmers had no immediate recourse but to go down to Bangkok and protest in front of the Prime Minister's office. Some even flew over to the Philippines and protested in front of IRRI. What else were they to do when they are not part of the arrangement? They wouldn't even have known about this trusteeship deal – and how it turned against them – had local NGOs and newspapers not blown the whistle. It's not as if the FAO or the CGIAR had informed them or kept them updated with new developments. The CGIAR has even denied that the Jasmine affair was of any importance, despite the point of view of the Thai farmers. This is an extreme and concrete denial of farmer's rights, no matter how you define it.

More fundamentally, by unilaterally declaring all 500,000 designated accessions as “public domain”, the CGIAR in particular has severed whatever relationship farmers and indigenous peoples may consider that they have towards those materials. This raises a whole range of questions about due legal process and the substance of the rights involved here. In terms of process, what legal basis is this “public domain” status grounded in? What tradition does it emerge from and is it properly framed in international law? Where does the CGIAR, whose history of collecting and distributing germplasm “on behalf of the international community” it is supposed to protect, get the right to operate it? With regard to substance, what does “public domain” mean here? Free access? Public property? Is the CGIAR a public entity? What if the farmers or indigenous groups consider the seeds part of their heritage and want to have a certain kind of control over who accesses them, how they are used and what happens to them? What if they have their own proactive understanding of sovereignty – not necessarily vested in national governments, the FAO or the CGIAR for that matter – which applies? Are the sovereign rights of indigenous peoples, which are increasingly getting crystallised in national and international law, properly accounted for through this “public domain” tag? What if they see these seeds, these plant varieties taken from their lands and farms, as subject to collective rights – historically, today and for the future? Are these collective rights compatible with “public domain”?

The CGIAR, and this especially reflects the position of the CGIAR in many documents, does not seem to care about any of this. Its main concern is to have these materials effortlessly declared public property so that its scientists can freely continue their work. It's like trying to construct a kind of ‘no man's land', so that no one can come and pester the CGIAR genebanks with conflicting claims and competing rights. The centres are like ostriches sticking their heads in the sand. Below the surface, maybe they see a world free of IPR pressure, a peaceful “public domain” – but there are definitely no farmers' rights down there either. The links have conveniently been cut. Which leaves us with free access to the CGIAR-held ‘raw materials' (below ground) and IPR on the rest (above ground).

2. An unenforceable system

Another major problem is that the system cannot be enforced – and those of in charge of it do not even seem serious about wanting to enforce it. This is borne out by a number of key facts already mentioned:

·        No signature indicating acceptance of the terms of the MTA is required;

·        The Centres officially renounce responsibility to monitor compliance with the Agreement;

·        Centres have been caught distributing designated germplasm without an MTA; and

·        Designated material distributed from the Centres has been directly appropriated under intellectual property regimes.

It is clear from all the legal documents that the system relies on good intentions and nothing more. The CGIAR Directors have reiterated many times their desire to see things work well, but they are not committed to ensuring that the policy works. Yes, they have steps to deal with violations. But they cannot and have to no desire to police, say, the 60,000 shipments of germplasm that left their Centres in the year 2000. Nor can the FAO. Violation-spotting is left to public interest NGOs instead. But tracking seed packets is not the heart of the enforcement problem. The real problem is that the people running this system want everyone to believe that it's a sound and credible one, when this does not appear to be true.

3. An empty ‘hands off' policy

When you look closely, the ‘no IPR' policy of the trusteeship system only applies to the germplasm accessions in the form that they are sitting on the genebank shelves. Despite all the trumpeting, components of the germplasm accessions (such as individual genes) and derivatives (selections made from the plants or descendents of crosses) may be legally appropriated by anyone through IPR. This is explicitly upheld by the system in several documents. And it happens all the time.

This makes for a very deceptive system in many ways. First, despite all the assurances, the protection from IPR that the system is supposed to provide is extremely limited. Okay, no one should be able to take a seed sample from a genebank and patent it. But under most any IPR law, that would be impossible anyway because the seed is not new. So who needs trusteeship? Another kind of deception is that the MTA is not very up front about this. The implicit policy is that no IPR should be applied to the materials “in the form received”. Nowhere in the current setup will you find those four little words.

But those four little words are fast becoming the de facto policy governing the distribution of genetic resources all around the world. It means that if you get hold of two rice accessions from IRRI, you cannot patent them as such – in the form you received them. But you can cross them together and patent the offspring. The final deception is neglecting to remind anyone that the Centres themselves may also get IPR protection over designated materials in this devious way. Under the same unspoken principle of “in the form received”, the Centres would never try to get IPR on the materials as contained in the genebanks. But their components and derivatives are another story. Many CGIAR Centres have pro-IPR policies which state that the Centres are prepared to seek IPR over the results of their own research and breeding work, i.e. modifications of the genebank materials.

The hands-off sign provided by trusteeship is largely irrelevant, since the material would not qualify for IPR protection anyway. Any patent office should turn a tempted biopirate – someone trying to claim that what already exists is actually new – away.

4. Commercialisation, but no benefits

The FAO-CGIAR trusteeship Agreement states that designated germplasm is to be distributed for the purpose of “scientific research, plant breeding or genetic resources conservation”. Full stop. The MTA, which is supposed to accompany every mail-out of material, adds a new dimension. It opens the doors to commercial release of the material. “This [MTA] does not prevent the recipient from making the germplasm directly available to farmers or consumers for cultivation, use or consumption,” it says. The Centres themselves clearly understand this to mean commercial release. For example, the Director General of the International Centre for Tropical Agriculture in Colombia has written to her staff stipulating that, “Commercialisation of the material, as well as use for research and breeding, is allowed under the agreement.”

For people concerned about faithfulness to the CBD, this is a bit odd. Because nowhere in the trusteeship system up to now – and it has been around for as long as the CBD – has there been any provision for sharing benefits from the direct commercial use of the material with the country it came from.

Where will trusteeship go?

In October 2002, the FAO member states will sit down to revisit the nuts and bolts of the trusteeship system. There are two reasons for this. On the one hand, the current four-year agreements with the CGIAR are about to terminate and have to be renewed. On the other hand, the whole system has to be shifted to the auspices of the new FAO Treaty on Plant Genetic Resources, which will now govern the “in trust” germplasm under its own provisions.

It is expected that the current agreements with each of the 12 Centres will be renewed for now without much fuss. In the background documents for this procedure, the CGIAR has reported some problems with IPR being taken out on designated germplasm. But these problems are said to be under control. (Nothing is said about the Jasmine rice fiasco, however.) Rehauling the system towards compliance with the new FAO Treaty is where things get complicated. And the proposals on the table should generate a host of questions and debates.

The revised trusteeship system, to be administered under the FAO Treaty, is basically the old one plus deeper problems.

One deeper problem is that “in the form received” is now being entrenched as the legal framework for distributing farmers' materials from genebanks. Under the FAO Treaty – which still needs to be ratified by more than 30 governments before it comes into force – these four little words will become the letter of international law. “Recipients shall not claim any intellectual property or other rights that limit the facilitated access to the plant genetic resources for food and agriculture, or their genetic parts or components, in the form received from the Multilateral System,” the Treaty says. The revised trusteeship system will have to implement this rule. As we've seen, “in the form received” is no shield from biopiracy. You cannot access Jasmine rice from IRRI and patent it. But you can access Jasmine rice, add one gene, and patent the result. Never mind that 99% of what gets patented this way was in the “public domain” yesterday, or fell under the national sovereignty of the Thai government or was developed by Thai farmers. The policy now is that the slightest modification of so-called public germplasm, or its “parts or components”, qualifies it for private monopoly control. With patent offices eager to call the isolation of a gene a feat of invention, “in the form received” amounts to a very poor defense against the privatisation of genetic resources and the new trusteeship system will have to implement it to the letter.

Another deepening problem is enforcement. The old system had weak legal bones and half-hearted commitments; its implementing tools, like the MTA, even deviated from the basic policy. The new trusteeship system could get more problematic. Without necessarily solving those previous problems – with the exception of the legal bones, since the FAO Treaty will be a legally binding instrument with its own governing body – it now has to respect different categories of germplasm with different strings attached. The Treaty only covers about 35 crops, itemised in its Annex 1. The CGIAR Centres will handle those crops in accordance with the Treaty. Crops outside the coverage of the Treaty but held by the Centres will be handled according to whatever specific terms were agreed to either with the country of origin of the material or with the country the Centres got it from. Add to this the fact that not all countries from which the materials came (one way or another) are party to the CBD, plus the possibility that not all countries which contributed to or wish to access the CGIAR genebanks will join the FAO Treaty, and it gets quite complicated.

For example, there will probably be fights over how commercial benefits from the use of “in trust” germplasm will be shared under the revised system. This fight will have to play itself out no matter what when the Treaty comes into force, since there are no details on this with respect to any kind of Treaty-covered material. But access to CGIAR-held germplasm has been generating a lot of commercial benefits up to now which have never been shared through the trusteeship system – and some governments may raise noise about that. Similarly, the FAO Treaty provides for benefit sharing upon commercialisation of modified “in trust” material, but not unmodified “in trust” material. Further, there are still mixed messages being put across in terms of the purpose of distributing the “in trust” material. The new draft trusteeship Agreement talks about research, breeding and training. (“Conservation” has been booted out.) The new draft MTA once again adds a footnote allowing direct commercialisation (but no benefit sharing).

Most importantly, the new trusteeship system – like the old one – risks being founded on wrong principles. Farmers have rights in relation to those seeds, but nowhere in the new construction do we see those rights being upheld. In fact, just where the “trust” to be implemented through the system arises from, in legal terms, is not at all clear. For the FAO, it arises from “a historical process”: a series of agreements, one superseding the other with very different legal weights attached. When and where did the farmers and indigenous peoples from whom the seeds come delegate authority over all this to the CGIAR? When did they give it to the FAO or to the governments for that matter? Where did the right to designate all these materials to the “public domain” come from, and who is this supposed to serve, and how?

The CGIAR and the FAO ought to start from scratch and redesign the whole system rather than simply update it to bring it into line with the new FAO Treaty. That means anchoring it on a full and unquestionable commitment to farmers' and indigneous peoples' rights. Not the way Northern governments – like the European Union, for whom farmers' rights are a “retrospective reward” for having given us the genetic diversity now subject to trusteeship – might want to reduce them. Trusteeship only makes sense if it is properly negotiated with those who have rights vested in the materials. That means providing the proper forum and using an appropriate legal basis  – such as that of collective rights – so that farmers and indigenous peoples can decide themselves what the principles and policies should be, who does what with the seeds and for whose benefit. If these agencies can't do that, then all this talk about farmers' rights at FAO or within the CGIAR itself is plain hot air. In this case, farmers and indigenous peoples will be left to take care of their rights in defiance of these systems.

 

GRAIN would like to thank Antonio Onorati of the Italian NGO Crocevia for his important contributions to this article.

 

Sources and further reading

van Caenegem, William, “The Public Domain: Scientia Nullia?”, European Intellectual Property Review, Vol 24, Issue 6, June 2002.

Commission on Plant Genetic Resources for Food and Agriculture, Consultations with the IARCs of the CGIAR and Other Relevant International Institutions on the Agreements to be Signed with the Governing Body of the ITPGRFA, First Meeting Acting as the Interim Committee for the International Treaty on PGRFA, 9-11 October 2002, Document CGRFA/MIC-1/02/08, FAO, Rome.
http://www.fao.org/ag/cgrfa/docsic1.htm

 

Commission on Plant Genetic Resources for Food and Agriculture, Report on the International Network of Ex Situ Collections Under the Auspices of FAO and Renewal of the Agreements Between the IARCs of the CGIAR and FAO and Draft Revised MTA, Ninth Regular Session, 14-18 October 2002, Documents CGRFA-9/02/11 and CGRFA-9/02/20, FAO, Rome.
http://www.fao.org/ag/cgrfa/docs9.htm

 

ETC Group (formerly RAFI), “Pinning the Tail on the Ostrich: The Australian PBR Scandal”, Communique, 30 January 1998.
http://www.rafi.org/article.asp?newsid=189

 

European Commission, Communication by the European Communities and their Member States to the TRIPs Council on the Review of Article 27.3(b) of the TRIPs Agreement, and the Relationship between the TRIPs Agreement and the Convention on Biological Diversity (CBD) and the Protection of Traditional Knowledge and Folklore: A Concept Paper, Directorate-General for Trade, Brussels, 12 September 2002.
http://trade-info.cec.eu.int/europa/2001newround/comnr_trips.pdf

 

Huamán, Zosimo (formerly with CIP), Changes in MTA for the Distribution of FAO-Designated Germplasm Held in Trust at CIP, PROBIOANDES, Lima, March 2001. http://www.geocities.com/probioandes/changes_in_mta_for_the_distribut.htm

 

Koo, Bonwoon, Philip G Pardey and Brian D Wright, Endowing Future Harvests: The Long-Term Costs of Conserving Genetic Resources at the CGIAR Centres, SGRP/IFPRI, Rome, 2002.
http://www.sgrp.cgiar.org/SGRP_GenebankEndowment.pdf

 

Lianchamroon, Witoon (BIOTHAI), Bioprospecting: The Case of Jasmine Rice, paper presented to the “Second South-South Biopiracy Summit”, Johannesburg, 22-23 August 2002.
http://www.biowatch.org.za/wliancham.htm

 

System-Wide Genetic Resources Program and CGIAR Genetic Resources Policy Committee (producers), Booklet of the CGIAR Centre Policy Instruments, Guidelines and Statements on Genetic Resources, Biotechnology and Intellectual Property Rights, Version 1, SGRP, Rome, September 2001.
http://www.sgrp.cgiar.org/Policy_Booklet_Version1.doc

 

Table: Trusteeship germplasm at the CGIAR, 2001
Â	Â	number of accessions
Centre*	crop	 held “in trust”	remainder	total
CIAT	Cassava	5,728	2,332	8,060
	Common bean	30,590	810	31,400
	Forages	16,339	7,845	24,184
	Total	52,657	10,987	63,644
CIMMYT	Wheat	79,912	75,000	154,912
	Maize	20,411	4,675	25,086
	Total	100,323	79,675	179,998
CIP	Potato	5,057	2,582	7,639
	Sweet potato	6,413	1,246	7,659
	Andean roots & tubers	1,112	383	1,495
	Total	12,582	4,211	16,793
ICARDA	Cereals	54,218	5,795	60,013
	Forages	24,581	5,947	30,528
	Chickpea	9,116	2,103	11,219
	Lentil	7,827	2,135	9,962
	Faba bean	9,074	1,671	10,745
	Total	104,816	17,651	122,467
ICRAF	Agroforestry trees	25	10,000	10,025
ICRISAT	Sorghum	35,780	941	36,721
	Pearl millet	21,250	142	21,392
	Pigeon pea	12,698	846	13,544
	Chickpea	16,961	289	17,250
	Groundnut	14,357	985	15,342
	Minor millets	9,050	202	9,252
	Total	110,096	3,405	113,501
IITA	Bambara groundnut	2,029		2,029
	Banana		400	400
	Cassava	2,158	1,371	3,529
	Cowpea	15,001	1,000	16,001
	Soybean	1,909	1,144	3,053
	Wild Vigna	1,634	50	1,684
	Miscellaneous legumes		400	400
	Yam	2,878	822	3,700
	Total	25,609	5,187	30,796
ILRI	Forages	11,537	1,667	13,204
IPGRI/INIBAP	Musa	914	229	1,143
IRRI	Rice, cultivated	77,827	16,737	94,564
	Rice, wild	2,790	1,778	4,568
	Total	80,617	18,515	99,132
WARDA	Rice	14,917	460	15,377
CGIAR total		514,093	151,987	666,080
* Sixteen CGIAR Centres have signed the Agreement with FAO, but one of them, the Centre for International Forestry Research, does not hold germplasm collections.   Source: Adapted from Bonwoo Koo, Philip G Pardey and Brian D Wright, "Endowing Future Harvests: The Long Term Costs of Conserving Genetic Resources at the CGIAR Centres", SGRP/IFPRI, 2002, p. 13.