Food Security

, 1:463

The other, ignored HIV — highly invasive vegetation


    • Plant SciencesWeizmann Institute of Science
  • Bernal E. Valverde
    • Investigación y Desarrollo en Agricultura Tropical (IDEA Tropical)
    • The University of Copenhagen

DOI: 10.1007/s12571-009-0038-7

Cite this article as:
Gressel, J. & Valverde, B.E. Food Sec. (2009) 1: 463. doi:10.1007/s12571-009-0038-7


The greatest cost to the farmer in time and variable costs is control of weeds (Highly Invasive Vegetation = HIV). Despite farmer efforts, weeds still result in the greatest crop losses of all biotic constraints. The costs due to this HIV are greatest in the parts of the developing world where manual labor (more typically “femanual”) is used, and the populations are becoming more aged (due to youth abandoning agriculture), or more feeble due to debilitating diseases such as malaria and HIV-AIDS. Two case studies are presented: the previously intractable problems with Striga (witchweed) species in Africa; and rice cultivation and the emerging problems with the labor-saving shift to direct seeding, which has resulted in outbreaks of a weedy/feral form of rice, as well as herbicide-resistant Echinochloa spp. Biotechnology has much to offer as part of the solution to these major HIV problems, whether as transgenic herbicide resistant crops, or as weed resistant crops or through transgenically enhanced weed-specific biocontrol agents. Where necessary, transgenic tricks will also be needed in many cases to prevent transgene flow from crop to related weedy relatives, and possible technologies are described. Ever since the advent of agriculture, solutions to crop protection problems have been effective but ephemeral, and new solutions will be needed in the future. Using mixed solutions leads to solutions lasting synergistically longer, so efforts should be made to use mixed technologies to extend the lifetime of the next generation of sorely needed solutions to the problems of HIV.


WeedsStrigaWitchweedRiceWeedy riceFeral riceEchinochloaMaizeSorghum


In the developed world, the major variable costs in agriculture are in weed control; plowing, discing, purchase of weed-free seeds, cultivation, herbicide treatments - all necessary to keep weeds at bay. Even with minimum tillage and transgenic resistance to inexpensive herbicides, weed control is still the major variable input. The situation in costs is similar in the developing world; except that much of the input is for manual or animal labor and the losses due to weeds are greater. In the developed world, a few percent of the working population provide the highly successful weed control that allows them to feed the whole population and export. In the developing world 60–80% of the population is engaged in this area. Weed control is so successful in the developed world and such a way of life in the developing world that the highly invasive vegetation being dealt with goes unnoticed by the experts in other areas of plant protection.

In the developed world it has become accepted that the agrichemical companies deal with weeds, and weed science is a minimal component of academic research and teaching; much less than entomology and plant pathology, despite their dealing with economically less damaging biotic constraints. The developed world supports much of developing world agricultural research and development; and the emphasis on weed problems is minimal. This is unfortunate, as the drudgery of weed control was eliminated long ago in the developed world by mechanization and agrichemicals. In the developing world, especially in Africa, weed control is relegated mainly to women, subjugating them to a life in the field if they are willing. One anthropological study claims that during the weeding season 80% of a peasant womans’s time awake is spent weeding (Akobundu 1991). Where there are better jobs available, even if they are menial sweatshop type jobs, women prefer them to the paddy or field. Weeds are a major reason for land becoming unproductive, through direct competition with the crop, as well as by depletion of nutrients and water from the soil. When land becomes unproductive, men leave the land looking for work, leaving crop production in the hands of women, children and the elderly.

Partial abandonment leaves the fields under-weeded and productivity decreases, further exacerbating an already bad situation. The population shifts of men looking for jobs in the city has had the same effect it always had throughout human history on all continents—contagious spread of sexually transmitted disease. HIV-AIDS debilitated men and women do not have the strength for weeding, nor do their orphans in the villages where only children are left. Thus, there is a relationship between the Highly Invasive Vegetation and HIV- AIDS. In Asia, especially where there is more industrialization, the youth leave the farms, and the aged are incapable of keeping the weeds down. Again – weed control is the most labor-intensive part of subsistence agriculture.

It is only by understanding the ecology, physiology and agronomy of weed-crop relationships that the sociology and economics can be understood. Only when all are understood, can workable solutions be proposed, modeled and tested in the field. It is thus that we describe two case histories: Striga (witchweed) in sub-Sahara Africa, and the emerging weed problems of rice throughout the rice-growing world, as macro-issue case studies. There are other macro-issues such as grass weeds in wheat that have evolved resistance to all known herbicides in wheat (Gressel 1988, 2002; Powles and Shaner 2001), which are more a problem of the developed world and thus are not discussed in the case studies. Many other micro-issues of intractable weeds occur in many other places in the world—but one can consider how to approach them from these case studies.

Striga—the bewitching of Africa

Maize was introduced to Africa centuries ago, replacing the more drought tolerant sorghum and millets that are better adapted to drier conditions. Maize predominated for a variety of reasons; some forgotten by the researchers trying to re-introduce sorghum. Maize has higher yields in good years, maize tastes better, and the often forgotten issue that maize has a closed cob. The open panicle seed head of sorghum is susceptible to the red-billed weaver bird (Quelea quelea), which gathers in flocks of millions and can completely destroy a sorghum crop (Doggett 1988). These “feathered locusts” are dealt with, where sorghum is still grown and the weaver migration coincides with early-maturing (“milk stage”) sorghum, by spraying fenthion, an organo-phosphate insecticide on the nesting areas (Mundy 2000). Alternatively, dynamite is exploded above ground at the nesting sites to create a lethal shock wave. Maize culture predominated over sorghum, creating a new niche for weeds. There had been (and still is) a wild species Striga aspera that parasitically attacked cultivated and wild sorghum species as well as wild grasses. This and all Striga species are obligate parasites. They germinate only in the presence of a hapless host due to chemical signals exuded by crop roots. They grow towards the crop root, attach, and suck nutrients, minerals, and water from the crop. Maize was more susceptible than sorghum; it did not co-evolve with Striga, and Striga has had a heyday (Fig. 1a). S. aspera evolved into S. hermonthica, a version that can only attack crops and in East Africa has no wild hosts. In colonial times, S. hermonthica was kept in check by a confluence of reasons: in fertilized soils it is less of a problem; share-croppers were required to remove Striga before it set seed; and crop rotation abetted the problem. As long as remaining individuals were sparse it was not a huge problem because it required cross fertilization by a butterfly. Thus, even though Striga could easily produce 100,000 seeds on a single stalk, the insects had to find the flowers among the crops. Sorghum and millet eventually succumbed to S. hermonthica along with maize, and all the grass crops are also attacked by S. asiatica. Leguminous crops are attacked in West Africa by S. gesnerioides. In north Africa and around the Mediterranean, the Striga-related Orobanche species parasitize all the vegetable and legume crops.
Fig. 1

Damage due to HIV (highly invasive vegetation) in maize and rice. aStriga hermonthica in western Kenya. In foreground, a bad infestation of a local open pollinated variety. In background, an experimental imidazolinone herbicide resistant line, where the seeds had been treated with about 0.25 mg imazapyr herbicide per seed (Kanampiu et al. 2003). Pictured: Dr Fred Kanampiu (right), the agronomist responsible for field development of the material, and (left) Dr. Alfa Diallo, the breeder responsible for the development the newer and better lines now released. Picture courtesy of Dr. Dennis Friesen. b A rice paddy heavily infested with weedy rice (tall plants with upright panicles) in Costa Rica. Mature cultivated rice panicles (lighter color) are seen below the weedy rice canopy. It is estimated that there will be a 40% yield loss from this field and the farmer will be docked 35% of the value due to contamination of the seed. Picture by B.E. Valverde

The situation with Striga became worse after the colonial period ended (Table 1). Governments wanted cheap grain for the city dwellers so prices were controlled down by fiat or by importing free or “dumped” grain from the west. Farmers could no longer afford fertilizer, and Striga thrived best without fertility. Once a high Striga seed bank is attained, restoring fertility is no longer effective (Ransom et al. 2007), fertilizer feeds the Striga as well as the crop, and the Striga becomes more productive. The tiny seeds are long lived in the soil, so crop rotation does not help that much once a threshold has been reached.
Table 1

Seeming paradox: overcoming Striga damage can provide profits that are triple the annual losses due to the weed in east Africa


Striga infested area (ha)

Cereal losses to Striga (MT)a

Losses with no Striga control ($US)

Potential gains from control (MT)b

Potential gains from control ($US)b

























amaize, sorghum, millet combined in Kenya, maize, sorghum, rice in Tanzania and Uganda

bIt is uneconomic for farmers to use good seeds, fertilizers, and pesticides when fields are Striga infested. When Striga is controllable, farmers can afford these inputs, leading to very high profitability

Modified from Alokit (2009)

Striga has a second effect once it becomes massive, with many individuals of the parasite attached to a single crop plant; the crop becomes poisoned while the Striga is still underground attached to the crop root. It is not clear what the poison is or whether it comes from the Striga or from some reaction of the crop. Initially, the farmer thinks the crop is healthy, and then suddenly leaves turn yellow and dry up; the crop becomes “bewitched”, and thus the name witchweed. Why weed the field when this happens and the crop is lost? And thus the emerging witchweed uses the crop roots to supply it with water and minerals, and its own green leaves perform photosynthesis. Thus Africa has been in a downward spiral, no fertilizer to keep Striga at bay, few weeders willing to pull Striga stalks on sick maize or sorghum, males leaving the farms for the city, fewer weeders, HIV-AIDS, malaria, fewer weeders again, and more and more Striga, both in area infested and in the intensity of infestation (Ejeta 2007; Parker 2009).

In some seasons the infestation is worse, in some minimal, for reasons researchers are yet unable to predict. Some hope remains and farmers still plant, but without fertilizer and without adequate control of other weeds. Subsistence agriculture has come to mean that the farmers produce 80% of the calories needed for their under-nourished families.

The situation is sufficiently bad that two recent compendiums have been devoted to updating the community on the status of Striga research, a book edited by Ejeta and Gressel (2007), and issue 5 of volume 65 (2009) of the journal Pest Management Science was wholly devoted to parasitic weeds.

Solutions to the Striga problem

If the Striga problem can be solved, Africa could leave subsistence agriculture behind, and go into production agriculture (Hearne 2009). The situation is so bad that good seed, fertilizer, and pesticides are unaffordable, and yields plummet, even in seasons without Striga (Table 1). If Striga could be controlled, farmers would regularly use these inputs, and yields and profitability would vastly increase (Table 1) (if the markets can absorb the added yields—transport and storage are major infrastructural problems not discussed here). The conventional solution to weed problems is to use herbicides. In the case of Striga, there was no herbicide that could be used in maize or sorghum that would kill Striga while underground. Some contact herbicides could be used after Striga had emerged, but by then the crop had been bewitched and ruined. Farmers in the developing world need a current crop yield to pay for pesticides.

Generations of young agronomists came to Africa with a mandate to deal with Striga. They repeated soil fertility experiments to little avail unless they put massive amounts of manure on small plots. They tried crop rotations that were doomed because they entailed years of agroforestry on land needed to feed mouths. They found insects that ate Striga seeds; too late to help the current crop and not enough to lower the soil seedbank so that infestations would be diminished. The countries that sent the scientists called this “foreign aid” but it neither helped the farmers nor served an educational purpose. The embittered researchers whose unworkable solutions were not adopted wrote bitter articles blaming the conservativeness of the farmers (e.g. Oswald 2005).

Still there have been glimpses of success that have led to local or general solutions to this problem that are workable and economic. Whether they will be sustainable for long periods is discussed after the solutions, as no solution has been forever in agriculture.

Limited scope solutions to Striga

Some solutions, while being excellent, are crop or ecosystem specific. Other solutions can be used on a wider scale. The limited scope solutions are discussed first, along with speculations on how their scope can be enlarged.

Crop breeding

Maize, coming from the Americas might not be expected to have any inherent resistance to African Striga, and researchers have searched for resistance genes in maize and its wild relatives for decades. From this lack of success, there may be one sign of light—a wild relative of maize Zea diploperennis seems to possess a gene, which when crossed into a maize inbred conferred a modicum of resistance at the post-penetration stage (Amusan et al. 2008). What effect this trait will have when hybrids are developed is a yet unanswered question. There have been many other claims for inherited resistance in maize, but they remained valid only on a local level, and the possible reasons for the local effectiveness was never adequately assessed. Still, breeders are optimistic that they can find and stack genes in maize (Rich and Ejeta 2008).

African sorghums would be expected to have some resistance, while the sorghum that wandered to China a few thousand years ago may not; and this seems to be the case as the Chinese sorghum lines produce far more strigolactone Striga germination stimulants than the African lines. Breeders tried conventional breeding to enhance the levels of Striga resistance with little success (Doggett 1988). It took the good eyes and insights of one breeder, Gebisa Ejeta, to change this. He characterized cultivated and wild sorghum lines that had a modicum of resistance at different levels; those that produced very little germination stimulant; those that caused the Striga to produce faulty attachment structures (haustoria); those that blocked Striga entry; and those upon which the Striga grew poorly after attachment and penetration (Ejeta et al. 2007). They found that combining a few of these factors allowed the sorghum to have normal yields, even if some Striga grew and set seed. They could do this relatively easily (for breeding) because they developed molecular markers for each of the traits (Grenier et al. 2007). Because the best results included the most traits, quite an effort is needed to get them into locally adapted varieties because they are inherited on different chromosomes and can segregate from each other during breeding. Each region has different varieties that do not grow well elsewhere, and different consumer groups prefer different types of sorghum. Still, many sorghum breeders throughout Africa took on these new lines to generate locally adapted material, with considerable success (see Ejeta and Gressel (2007) for information on different areas). This insightful breeding resulted in Gebisa Ejeta rightfully being awarded the 2009 World Food Prize. This is the first time the committee recognized control of lowly weeds as an appropriate subject.

Similarly, molecular markers are being used for successful breeding of cowpeas for resistance to S. gesnerioides (Li et al. 2009).

The problem with the breeding approach is that it is species specific. One cannot transfer the genes from cowpeas or sorghum to maize or other crops needing Striga resistance by conventional breeding. This can only be performed if the genes are isolated and transferred by genetic engineering. There has been some recent progress on that front, as described in a later section.

Desmodium intercropping

Farmers in Africa have often intercropped legumes with maize, both as a source of fixed nitrogen for the maize, and as an insurance against total maize loss if Striga infestation was heavy. The intercrops had little direct effect on Striga. Various agroforestry legumes were found to be excellent at reducing Striga but are impractical. It was a surprise to researchers that were using a legume to fend off stem borers as part of a “push-pull” strategy that there was no Striga. Desmodium is a perennial shrub that can be harvested as feed for ruminant livestock. It is a poor seed producer with poor germination and stand. It takes a few years until the rows of Desmodium are thick enough to perform their underground magic of controlling Striga such that maize can be planted between the rows. There are excellent reports of its utility in the specific geographic region in which this species will grow (Khan et al. 2007). The researchers are looking towards finding relatives that are also Striga-toxic so that the technology can be extended out of the very limited geographical range where D. uncinatum can grow. The technology is only appropriate in regions where farmers have enough land to keep livestock. Where maize is cultivated and there is no livestock, the Desmodium will have little value. It is not known whether Desmodium can be effective in West Africa, where legume-attacking S. gesnerioides is indigenous, as it is unknown whether it will also attack Desmodium.

Broad spectrum strategies for Striga control

Use of systemic, undegraded herbicides

There are systemic herbicides that when sprayed on leaves travel systemically through the crop vascular system to the roots, where they enter parasitic weeds, and kill them. These were initially used at very low dose levels so as not to affect the crop, which was rather tricky (Foy et al. 1989). The advent of transgenic and mutant crops resistant to the herbicides opened a new window; spray larger doses of the herbicide and the parasite is dead without damage to the crop (Joel et al. 1995). While it was ascertained that this approach could work with a herbicide-resistant mutant maize (Abayo et al. 1998), it was deemed inappropriate for Africa, where it was doubted that subsistence farmers in Striga stricken areas could afford the amount of herbicide required or the sprayers to apply it. Instead, it was proposed to apply the herbicide directly to the crop seed before planting, to control the Striga alone. The approach worked (Kanampiu et al. 2001, 2003; Gressel 2009), and the mutant gene was back-crossed into elite open pollinated and hybrid varieties appropriate for the Striga-stricken area of western Kenya (Kanampiu et al. 2007).

Large scale farmer acceptance studies were performed with a group of NGOs distributing seed (De Groote et al. 2008). Farmers could receive free seed for a single season only, after which they had to purchase the treated seed. Interestingly they preferred the hybrid. The local seed companies cannot keep up with demand. It is clear that the only way that the full potential of the seed can be achieved is if fertilizer is used with the maize. Thus, the international breeding organization (CIMMYT) that developed the technology, AATF (African Agricultural Technology Foundation) who brokered part of the technology, governmental authorities, and the NGOs together agreed on an initial “tie in” where the seed is sold only with a small bag of fertilizer. The yield differential is immense, especially in seasons when Striga is rampant (Fig. 1a). One advantage of the seed treatment, besides using >10 fold less herbicide than would be sprayed, is that it is an “appropriate technology” insofar as many African farmers intercrop a legume between the maize as an insurance against crop devastation by Striga (Kanampiu et al. 2003). If one were to spray the herbicide, the legume would be killed. The seed treatment keeps the herbicide under the maize plant and away from the root zone of the legume, so intercropping is not precluded.

Like all technologies, this one is not without problems. If there is not enough rain after planting, the herbicide remains near the maize seed, at a very high local concentration. This can cause some damage to the young maize plants. If there is too much rain, the herbicide can be washed away from the maize roots. For this reason a novel slow release formulation was developed to prevent too much unbound herbicide from causing early damage and also to keep it from leaching away. These slow release formulations are based on ion exchangers holding the herbicide (Kanampiu et al. 2009). An interesting off-shoot is that these slow release formulations, developed specifically for the developing world, are finding uses in the developed world—quite a reverse trend.

Because a single gene confers resistance, African breeders can backcross the material into their own locally adapted breeding lines, and this is now being done for all Striga-infested areas in Africa where maize is cultivated. It was initially modeled that Striga would rapidly evolve resistance to the herbicides used that are acetolactate synthase (ALS) inhibitors, the most resistance prone group known to which many other weeds quickly evolved resistance (Gressel et al. 1996). That this has not happened is probably due to the form of application. Sprayed weeds get a much lower dose of herbicide than Striga next to a treated crop seed, where the local dose is very high, but the amount used per field is much less. Weeds need be only heterozygously resistant to survive the sprayed dose, but would have to be homozygously resistant to survive the dose near a seed. The model was ‘only’ off by a factor of a million (Gressel 2005).

The present technology is with a mutant line of maize; the same mutation might eventually evolve in Striga with devastating effects. Thus, one needs other herbicide resistances to replace this technology. One technology that works with Orobanche is the use of transgenic resistance to the herbicide glyphosate (Joel et al. 1995) included with seed treatments (Gressel and Joel 1997). In Africa, transgenic glyphosate resistant maize has only been released in South Africa, the first country in sub-Sahara Africa commercially registering transgenics. South Africa is also without S. hermonthica. Still, the technology has been rapidly adopted by resource poor farmers for general weed control. A woman with a backpack sprayer can now control two hectares of weeds in maize in a day, where she could not provide timely weed control on a single hectare wielding a short-handled hoe throughout a whole season. The much greater yield and much higher productivity does not mean less employment for women—it means that they can do other family jobs, grow other crops and they have additional income from the greater yields. It is surprising that some decry the use of herbicides, claiming that they will take jobs from women—instead of considering the backs it is saving from demeaning manual cultivation.

The mutant conferring imidazolinone herbicide resistance has been found in weedy sorghum (“shattercane”) and has been backcrossed into African material (Tuinstra and Al-Khatib 2008, Tuinstra et al. 2009). Weedy sorghum (shattercane) is the same species as cultivated sorghum, but drops its seeds before harvest.

The technology would also be useful in other crops attacked by Striga, but so far, no one has considered using herbicide resistance.

Biological control of Striga

As noted above, some misconceived methods of using insects that eat Striga seeds were promulgated (Traoré et al. 1995, Anderson and Cox 1997). The simple calculations were ignored that suggested that this approach was doomed (Smith et al. 1993), as the insects cannot substantially reduce the weed seedbank.

Much more success has been achieved with Striga-specific Fusarium species as mycoherbicides. Researchers now mainly focus on strains (forma speciales) of F. oxysporum, initially found on infected Striga (Ciotola et al. 1995; Elzein and Kroschel 2004). The Fusarium spores can be applied to grain hills or directly on the crop seed. Initially it was thought that spore production could become a cottage industry, but it was quickly appreciated that the spore production is a far more sophisticated endeavor than had been thought—and the results far too variable. The technologies now being developed are for central, quality-controlled production of robust spores that farmers or seed companies can apply to seed (Venne et al. 2009).

Because Striga hermonthica is so prolific, and because it is out-crossing it can be expected to mutate and evolve quickly and adapt itself to biocontrol agents, allelochemicals (chemicals naturally produced and released by plants that have harmful effects on other, competing plants), or herbicides. Thus, it might be wise to “stack” (combine) biocontrol with herbicide resistance, delaying this adaptive evolution. Herbicide resistance has also been coupled to the use of Desmodium to provide weed control until Desmodium can close its stand (Kanampiu et al. 2007).

Molecular solutions to Striga

The use of transgenic herbicide resistance was described in a previous section. It has also been proposed to render mycoherbicides more virulent to increase their effectiveness by adding transgenes that should enhance activity. There was a modicum of success with related Orobanche spp. using this approach, but not enough to try field testing (e.g. Meir et al. 2009), and it has not been tried with Striga, due to regulatory constraints. There has even been a suggestion to transgenically insert genes into Striga that when disseminated in a population via pollen would render the females sterile (Rector 2009).

Two other molecular approaches can be considered that bypass the need for a herbicide to control Striga.

Use of RNAi technologies

Interference RNA (RNAi) is a technology where pieces of a gene are put in a special construct that makes shorter pieces of RNAi. These in turn basically suppress that gene from being expressed. This is best known as a method for suppressing viruses.

The short pieces of RNAi can be transported short distances. For example, plants have been transformed to make an RNAi that targets a nematode gene. The nematodes attacking the plants are severely inhibited after ingesting the RNAi (Huang et al. 2006). RNAi can traverse the junction between a crop and a parasitic weed (Westwood et al. 2009). Thus it was posited that an RNAi that specifically affects a metabolic pathway in Striga, which has gene sequences not occurring in the crop should have no effect on the crop, but should kill Striga. The first trials were ineffectual (de Framond et al. 2007), but various groups are still trying this approach, hoping to hit on the right gene (Yoder et al. 2009)

Introducing genes for resistance to

Striga Various species are not affected by Striga. Why they are resistant is not clear as some secrete the stimulants of Striga germination. If the gene(s) were known, they could be used to confer resistance. An easier approach is to take genes from crops that have resistant and susceptible lines, such as the sorghum lines described used for breeding described above. There are differential molecular technologies that allow genes which are expressed in one line, but not the other to be picked out. When the genes separately conferring lack of germination stimulant, poor attachment, poor penetration and poor establishment are isolated, they could be put in a single construct for engineering into any Striga sensitive crop; the closest to “one size fits all” you can get in biology. If they are in a single construct, they will be inherited as a single gene trait, and backcrossing to multiple varieties will be much simpler than backcrossing four genes, each on a different chromosome.

The sequencing of crops and the new chip technologies are rendering it easier to isolate the resistance genes, even if their metabolic function is unknown. Thus, a quantitative trait locus (QTL) in rice conferring resistance has rapidly been whittled down to a 1.5 Mbase pair segment of rice chromosome 4 (Swarbick et al. 2009). Three other levels of resistance, each inherited on separate genes have also been identified in rice (Yoshida and Shirasu 2009). Now that sorghum and rice have been sequenced it is hoped that the four QTLs conferring resistance can soon be isolated as genes for transformation into other species. One of the resistance genes has been isolated and cloned from cowpeas (Timko and Li 2009). Suppressing the gene in resistant cowpea renders it susceptible to Striga. It is now important to demonstrate that adding the gene confers resistance, hopefully in other species as well.

Intractable weed problems in rice

For thousands of years, the major method of weed control in rice was to work a field to kill all the weeds, and then flood it to keep new weeds from germinating. While this was being done, rice seeds were planted in a nursery, hand-weeded, and month old seedlings were planted into the just flooded paddy. With this head start, the transplants could close their canopy and nearly cover a paddy before a new growth of weeds can make it through the water. This and later hand weeding were the methods of weed control. Transplanting and weeding in a rice paddy are not romantic forms of labor: work in a sweatshop is far more appealing and pays more. Thus, as countries industrialized, there was a need to abandon transplanting, and instead depend on herbicides. Therefore, direct seeding of rice seeds into dry or wet soil became the new convention as regions industrialized or where farm workers found more appealing employment elsewhere. Herbicide use was imperative, as the crop no longer had a month’s head start. Nature abhors a weed free vacuum and two major problems appeared: herbicide resistance evolved in different members of the grass-weed genus Echinochloa (different Echinochloa species in different regions), and a feral, weedy strain of rice (Fig. 1b) became a major problem. These are discussed below, along with possible solutions.

Herbicide-resistant Echinochloa species

Several species of Echinochloa are well adapted to compete and grow together with the rice crop under a variety of environments and production systems. Two Echinochloa species are in the top ten worst weeds of the world (Holm et al. 1977). They are also a major problem in other crops, but this will not be discussed here. Echinochloa has so closely associated and evolved with the rice crop for centuries that it represents one of a few cases of plant mimicry. At very early stages, Echinochloa resembles the rice seedlings and it requires some training of the eye to perceive its differences from the crop. Physiologically, Echinochloa also imitates the rice crop. Thus resistance to the most loyal chemical companion of rice, the herbicide propanil, evolved in Echinochloa species by mimicking the selectivity mechanism of rice to the herbicide (Leah et al. 1994; Carey et al. 1997). Soon after propanil was commercialized it was found that rice had arylacylamidase enzymes that are capable of degrading the herbicide to dichloroaniline, which is further inactivated by conjugation with sugars and other plant components, rendering the crop unaffected by the herbicide (Still and Kuzirian 1967; Yih et al. 1968). Nature taught Echinochloa to do the same, and with the help of the farmers who relied on propanil so much, resistance became evident after about 25–30 years of use (Talbert and Burgos 2007; Valverde et al. 2000). Butachlor, a chloroacetamide type herbicide is widely used in China, but there are now millions of hectares of Echinochloa in rice that are resistant to butachlor, and surprisingly to the totally unrelated thiocarbamate herbicide, thiobencarb (Huang and Gressel 1997).

Each Echinochloa plant produces thousands of very small seeds, which is an order of magnitude more seeds than produced by rice. The Echinochloa seeds mature and shatter (drop off) before rice harvest. Sometimes a second generation is produced when late emerging individuals are able to complete their life cycle during a fallow period or between rice crops in the most intensive rice monocultures. These seeds can remain alive in the soil for years waiting for the right moment to germinate and invade a new rice crop (Valverde et al. 2000).

Rice farmers have been very fortunate to have an arsenal of selective herbicides that effectively control Echinochloa without harming the crop. They belong to a wide group of chemistries and modes of action, making it possible to find good combinations to widen the spectrum of weed control and to delay or prevent the selection of resistant populations when properly used in rotations or mixtures. But selectivity to the crop, economics and ease of management has favored the use of some herbicides over others with the inevitable selection of resistant biotypes. Currently, all the economically important Echinochloa species have evolved resistance to the most widely used herbicides in each region or production system (Valverde and Itoh 2001). Some of them have done it so effectively that they can withstand several chemically unrelated herbicides that have different modes of action but share within-the-plant degradation mechanisms, even before they were ever sprayed in the field (Fischer et al. 2000; Yun et al. 2005; Valverde 2007). These multiple-resistant populations have become a nightmare for farmers whose fields are infested, as very few herbicides (sometimes none, depending on the number of active ingredients approved by the registering authorities) remain effective on them.

Weedy feral rice—already resistant to selective herbicides in rice

With weedy rice, the situation is more complicated than with Echinochloa. Here we are dealing with a type of rice that, being the same species as the cultivated form (Oryza sativa), has the characteristics of a noxious weed. The most important character is the shattering of its seed. Weedy rice typically matures and shatters most of its seed before rice harvest, assuring sufficient contamination of the field. Some seeds remain attached to the panicles and combine harvesters gather them with the grain (Fig. 1b). If those seed are planted in another paddy the following season, weedy rice will spread. To make things worse, many weedy rice strains have colored (red to brownish) grain that contaminates the final product. Because of consumer rejection, rice mills have to provide additional polishing to the grain to get rid of the color, reducing nutrients in the entire rice batch and the grain becomes prone to breaking, further reducing the value of the crop (Valverde 2005).

Weedy rice seeds have another trait that distinguishes them from the crop. Humans have selected seed during domestication that germinates rapidly and uniformly. Weedy rice has reverted to the wild feature of extended seed dormancy. The dormant weedy rice seed survives in the soil for long periods until conditions are suitable again for germination. This is a characteristic that disseminates weedy rice in time, precluding crop and herbicide rotation from being effective management practices. The practical implication is that, as it is the case with most weeds, the farmer is sentenced to deal with increasing weed problems from season to season.

Even though weedy rice is extremely difficult to differentiate from the crop early in the season, it typically grows taller than rice plants later in the season. Unfortunately by then damage to the crop through competition for nutrients, light, and water has already occurred. Hand removal is totally ineffective at typical infestation densities, and physically harms the crop. Herbicides that will kill weedy rice will do the same to the crop, unless the farmer performs some sophisticated tricks (delayed planting after a herbicide application or directed chemical applications), usually with undesirable yield penalties.

Solutions to the Echinochloa problem

The book about Echinochloa and weedy rice control has yet to be written. Both weeds have had plenty of time to co-evolve with rice and have become highly adapted to changes in production systems, agronomical management and herbicides. There is not a single control tactic that will be sustainable and we are farther away from finding a vaccine to these HIVs than to HIV-AIDS. Still, using good husbandry and profiting from already available knowledge, several tactics can be integrated into an economically-viable control strategy. As with HIV-AIDS, the best strategy is prevention. After that, there is no cure; all treatments are palliative.

The principle of a clean start is quite applicable to Echinochloa management. In the field, Echinochloa germinates and emerges in flushes, particularly at the beginning of the cropping season. In direct-seeded rice production, soil preparation for planting provides a suitable condition for thousands of Echinochloa seedlings to emerge. Densities of over 1,000 plants m-2 are not uncommon. These seedlings will start competing with the emerging rice crop almost as soon as they appear. By delaying planting and allowing the first generations of Echinochloa to germinate (“stale seed bed preparation”), which can be stimulated by irrigation or timely light rain, a non-selective herbicide can be used to kill them. The product of choice worldwide for this purpose has been glyphosate, especially because it leaves no bioactive residues in the soil that would harm rice when planted a few days later. Other products are also available, including paraquat and glufosinate ammonium. This practice also makes a tremendous difference when the farmer has to deal with Echinochloa resistant to propanil or other herbicides (Valverde et al. 2001). Having depleted these initial populations, rice can be planted with as little soil disturbance as possible, sometimes even broadcasting pre-germinated seed, especially when the rain has left the soil too muddy to use agricultural equipment. Needless to say, certified, weed-free rice seed should be planted, as planting weed seeds overcomes the stale seed bed. The emerging rice seedlings will have a head start and the weeds emerging with the crop or right after can be controlled by additional light tillage. The mantra of some organizations espousing “farmer saved seed” is clearly counter-indicated when it comes to the needs of healthy agriculture, as it exposes crops to HIV. Selective herbicides are readily available, which if carefully chosen can also delay the evolution of resistance (Riches et al. 1997). There are herbicides also available that can be applied late in the season (at a cost that has to be balanced with the impact of those weeds), which allow control of weeds that have escaped previous treatments.

Crop rotation is also a management tactic that complements physical and chemical Echinochloa control. By rotating to other crop species, agricultural practices and herbicides that are not normally used and performed in rice as part of the cropping system can be implemented impacting both the emerging Echinochloa seedlings as well as the seed bank. Still, many farmers are loathe to plant other crops in a paddy. Herbicides that are not selective to rice, belonging to other chemical and mode-of-action groups can be used to eliminate the weed, particularly those individuals that have been selected for herbicide resistance in the rice paddies.

If these practices are carefully assembled and conducted, we will not only decrease the amount of Echinochloa seed going into the soil seed bank but also will delay the evolution of herbicide resistance.

Transgenic and conventional herbicide resistance for Echinochloa management

Experimentally, rice varieties resistant to several herbicides have been already developed. In practice, the only resistance commercially available, which is non-transgenic, is that of imidazolinone-herbicide resistant rice (IMI-R rice), sold as Clearfield®. Although transgenic glufosinate rice has been de-regulated in the USA, it is not commercially grown there or elsewhere. Quite recently, however, it is being discussed whether to allow glufosinate-resistant rice varieties to be grown in Brazil; with some opposition by academic and grower groups, mostly because of the risk of gene flow and fear of negative reactions from possible export markets. Transgenic glyphosate-resistant rice has also been produced but, as with glufosinate-resistant rice, has not been released, because of the high risk of the transgene moving to weedy rice, making weedy rice a worse problem (see discussion on gene flow below).

There is no over-riding need for herbicide-resistant rice to control Echinochloa, except in a few cases where the populations have become resistant to all available (registered) herbicides, as is the case with Echinochloa phyllopogon in California (Fischer et al. 2000; Yasuor et al. 2009). But even so, it has been demonstrated that with ingenuity and a solid knowledge of the biology of the weed, agronomic practices can be modified to cope with the problem (Fischer 2009). Transgenic herbicide resistant rice could be considered where few herbicides remain that can control Echinochloa, especially in areas where weedy rice is also an issue, but this is not without drawbacks. Herbicide resistant rice, on the contrary, can exacerbate already serious problems with Echinochloa and other weeds. Such is the case in Costa Rica, where Echinochloa populations previously subjected to strong selection pressure by ALS-inhibiting graminicides (particularly bispyribac) rapidly became imidazolinone-resistant in fields planted with IMI-R rice and subjected to repeated applications of these herbicides.

Biocontrol of Echinochloa

The ubiquitous invasion of rice fields by Echinochloa at high densities makes it a good candidate for biocontrol. Several organisms have been evaluated as possible biocontrol agents but none has yet been commercially developed. One mycoherbicide, Collego based on Colletotrichum gloeosporioides f. sp. aeschynomene, was commercially used for the control of the leguminous broadleaf weed Aeschynomene virginica in rice. Unfortunately, Collego encountered commercial failure because it required very specific weather conditions to realize its efficacy and controlled a single species that was not as important as Echinochloa. Another fungus, Exserohilum monoceras has been investigated as a possible biocontrol for Echinochloa (Zhang and Watson 1997a, b) but has not achieved commercial status. Other pathogens have also been evaluated as possible biocontrol agents of Echinochloa (Yang et al. 2000).

Solutions to the weedy rice problem

Wherever weedy rice evolves in and/or invades a paddy, it becomes the worst phytoprotection problem, thus the best strategy is to prevent fields from being infested. But this is not easy. One might think that using certified seed will keep weedy rice away. This is true if certified seed does not contain seed of weedy rice. Unfortunately, in many rice-growing countries a few weedy rice seeds are allowed in certified seed, usually one per kilogram; sometimes many more. Considering that the seeding rate can be as high as 150 kg ha-1, it is not difficult to realize how infestations can begin. Assuming conservatively that only 50% of this contaminating weedy rice seed germinates in the current season and that only 50% of the seedlings will survive and reach maturity, the 35 weedy rice plants per hectare will be able to initiate the establishment of a soil seed bank. Under a worst-case scenario, a single weedy rice plant bearing only one panicle would produce about 100 seeds. So for the next season, maintaining our assumptions, more than 900 (recruited from the seed bank and arriving with the new lot of certified seed) weedy rice plants will initiate their struggle with rice for needed resources for growth, lowering crop yield and continuing establishment. We should emphasize that this can be avoided if certified seed is truly weedy-rice free.

The stale seed bed preparation mentioned for Echinochloa control is also an effective means of decreasing the weedy-rice infestation that initiates simultaneously with the crop (Olofsdotter et al. 2000; Rao et al. 2007; Vidotto and Ferrero 2005). Where weedy rice is rampant, several irrigation flushes are given to heavily-infested fields to promote weedy rice germination and emergence that is later killed with a non-selective herbicide. Weedy rice seedling recruitment can be very high as seen when emerged densities are determined after each of these irrigation flushes. It is not surprising to find 600 plants m-2, equivalent to a very desirable density of the planted crop. By continuing with this practice for a few times the densities can be brought down to 20–25 plants m-2, at which time rice can be planted without soil disturbance, and a final application of a non-selective herbicide (e.g. glyphosate) is made before the rice crop germinates. For many years California rice has been weedy-rice free, mostly because of the widespread used of certified seed that allows no contamination with the weed and another agronomic tactic that puts weedy rice at a disadvantage: the water seeding of pre-soaked seed. California now has weedy rice. Whether it evolved from a back mutation to a feral form or was introduced from elsewhere is an open question.

No in-crop selective herbicides are presently available to kill weedy rice. Some herbicides, however, can be used as pre-plant treatments provided that enough time is given for the herbicide to dissipate in the soil before rice is planted. Other practices are also performed by farmers, depending on the region and infestation levels to decrease weedy rice densities before planting. Soil puddling, the preparation of soil under completely water saturated conditions, is practiced in irrigated areas, particularly in Latin America, to kill weeds and to improve water retention (Rao et al. 2007). But as mentioned before, no single tactic is enough. Thus weedy rice in Venezuela remains a serious problem despite the widespread use of soil puddling. With increased fuel and machinery costs, water scarcity and environmental concerns, puddling is becoming less attractive as a weedy rice control option.

Crop rotation substantially impacts weedy rice. As with Echinochloa management, crop rotation allows the implementation of agronomic practices and the use of herbicides that are not possible in the rice crop, particularly if the rotation crop is a row-planted, broadleaf species. There are several effective graminicides that control weedy rice available for crops such as soybeans, cotton, peanuts and others. Rotation with sugar cane is also quite effective as this crop is usually ratooned (recut) a few times, withdrawing the field from rice production for at least 3 years, further depleting weedy rice seed from the soil seed bank. Successful weedy-rice free production has been sustained in Uruguay by rotating with pastures for 3 years and using clean certified seed (Garcia-Prechac et al. 2004). But when heavy investment has been made in field infrastructure for rice production (laser leveling, levees, irrigation technology) farmers are reluctant to rotate because of the cost of restoring those conditions for the rice crop.

As rice grain is heavily penalized when contaminated with colored weedy rice, heavily infested fields are hand rogued (i.e. the weedy rice is removed) or weedy rice panicles are slashed as a last resort to avoid harvesting a highly tainted crop. Roguing does not necessarily prevent shattering and up-rooted plants are seldom removed from the field, leaving seed on the ground for the next cycle. Depending on the timing, slashing can be counterproductive as the decapitated plants are stimulated to mature secondary tillers (shoots), and considerable physical damage is inflicted upon the crop by treading. In Malaysia, special sickles are attached to a long pole that allows a worker (again this is a job conferred on women) to cover a much larger area with less walking and less damage. Weedy rice, however, has adapted to this type of weeding and currently rice paddies are infested with weedy rice biotypes of the same height as the dwarf green-revolution rice varieties.

Biological control is not viable for weedy rice since any natural enemy of the weed will also be detrimental to the rice crop.

Transgenic and conventional herbicide resistance

The introduction of non-transgenic IMI-R rice varieties was very successful in terms of adoption by farmers. The opportunity to selectively control weedy rice with herbicides otherwise lethal to the crop was very attractive and allowed the return of abandoned, heavily infested fields into production. But the problem that had been anticipated by scientists and ignored by industry based on the cross pollination of both rice types soon emerged: the herbicide resistance gene rapidly moved to the weedy forms jeopardizing the sustainability of the technology (Gealy 2005; Madsen et al. 2002; Messeguer et al. 2001; Olofsdotter et al. 2000). “Stewardship programs” that were designed and implemented as part of the licensing to sow IMI-R rice were conceived on the premise that the low outcrossing rate between rice and weedy rice, typically less than 1%, was manageable through a series of agronomic practices. These included repeated applications of imidazolinone herbicides at full doses to eliminate most of the weedy individuals, roguing of plants escaping the treatments (seldom done in large operations), and limiting the number of successive plantings of IMI-R varieties.

Despite stewardship the resistance gene moved and resistant hybrids between weedy and cultivated rice were detected after just two sequential IMI-R cropping seasons (Valverde 2007). In Brazil, widespread “unauthorized” planting of IMI-R rice outside of stewardship programs became out of hand. An ineffective penalization scheme was imposed upon farmers who brought more IMI-R grain to the mill than the expected yield from authorized fields, or by farmers who had not signed contracts allowing them to plant the variety. The herbicide manufacturer considered that its commercial interests were seriously affected and decided to withdraw IMI-R varieties from the market and stop further cooperative research with IRGA, the Rio Grande do Sul Rice Institute (Gressel and Valverde 2009). Just recently, however, the manufacturer announced that they are prepared to re-introduce IMI-R varieties and hybrids under a campaign using the slogan “all united against weedy rice”1. The problem with their stewardship system is that it is not biologically sustainable; it is based on rotation of rice varieties and herbicides. The weedy rice seed bank builds up to such an extent that plenty of dormant seed will emerge two or three seasons later when the IMI-R varieties are again planted, obviating the stewardship strategy promoted by industry. The same mistakes are now being repeated in Italy with 20,000 ha of IMI-R rice2.

The problems faced by transgenic, herbicide resistant rice will not differ from those seen with IMI-R rice. The transgene will rapidly move to weedy rice, aggravating a problem that is already serious enough. If the transgene is one of those conferring glyphosate-resistance, the implications could be devastating. As stated before, glyphosate is one of the most cost-effective herbicides to control weedy rice before planting and during fallow periods. Additionally, the no-till rice production system is dependent on this herbicide for weed control before sowing. Glyphosate resistant weedy rice would substantially jeopardize both conventional and no-till rice production with the likely implication that affected areas could become unproductive or abandoned to rice production.

Imidazolinone herbicides target their action against the key enzyme ALS as do many other rice herbicides used today (sulfonylureas, pyrimidinyl-oxy-benzoates and triazolopyrimidines). Herbicides of these chemical groups are sometimes used in IMI-R rice to supplement weed control and are regularly applied when conventional rice is planted in rotation. The increased selection pressure on a single a common target site enhances the rate of evolution of other weeds to ALS inhibitors. Several broadleaf species and sedges (Cyperaceae) have already evolved resistance to ALS herbicides in rice (Valverde and Itoh 2001). Similarly, increased selection pressure by the use of herbicides to which transgenic resistant varieties have been developed could have similar effects. Echinohcloa is a herbicide-resistance prone genus. E. colona already evolved resistance to glyphosate in cereal production in Australia (Preston et al. 2008) and is suspected in Argentina in soybeans. There are strong indications of glyphosate-resistant Echinochloa biotypes in rice production in California (Fischer AJ, personal communication) and Venezuela (Anzalone A, personal communication).

Preventing transgene flow to weedy rice

The experience with IMI-R rice clearly demonstrates that gene flow matters. It is therefore necessary to find biologically-sound, practical ways to prevent or mitigate gene flow. Agronomic practices must be a part of both types of programs but they are insufficient to guarantee the sustainability of herbicide-resistant rice. Biotechnology offers several tools for rice that have been dealt with in detail elsewhere (Valverde and Gressel 2005; Gressel and Valverde 2009, Chapter 4 in Gressel 2008). They include methods to contain the genes within the rice crop, but these are all leaky and thus it is necessary to mitigate the establishment of the transgene in weedy rice, when the gen(i)e gets out of the bottle! Some possibilities to mitigate gene flow between transgenic rice and weedy rice are briefly discussed here. The strategies can work only with transgenic rice, and not with the non-transgenic IMI-R type rice varieties. Thus there must be extreme caution in introducing herbicide resistant rice varieties that are not transgenic, as well as transgenic varieties that do not contain transgenes that preclude the establishment of the herbicide resistant transgenes in weedy rice.

Rendering weedy rice hybrids unfit

As it is clear that transgenes encoding herbicide resistance will eventually move to weedy rice, a possible way to mitigate gene flow is to find ways to make the hybrids between the crop and weedy rice unfit. Conceptually, this could be achieved by engineering a construct having the desirable gene (herbicide resistance) in tandem with a second gene that would be deleterious to the hybrid but not to the crop (Gressel 1999). Being so tightly linked, both genes would be inherited together. Whenever herbicide resistant rice hybridized with weedy rice, the offspring of the resistant hybrids would also carry the unfit mitigation gene. This unfitness would keep the herbicide resistance transgene at a very low frequency in the population, thus preventing or delaying its establishment and spread. What kind of unfitness genes can be used as mitigators? The concept has been proven with a dwarfing gene in tobacco and oilseed rape (canola) (Al-Ahmad et al. 2005, 2006; Al-Ahmad and Gressel 2006) but a specific rice dwarfing (green revolution) gene could be used for rice. The idea would be to reduce the competitive ability of typical, very tall weedy rice with the dwarfing gene. Other genes, however, would be more suitable, particularly those that could prevent seed shattering and seed dormancy, the two main characteristics of weedy rice that were eliminated from the rice crop during domestication.

Chemical traits to mitigate transgene establishment

An interesting approach, “selectively terminable transgenic rice”, was recently proposed based on using a gene conferring susceptibility to a herbicide that is selective (non phytotoxic) to conventional rice as the mitigator gene (Lin et al. 2008). In other words, a crop (rice) is transgenically modified to make it highly susceptible to a herbicide that otherwise is not harmful to the crop. The gene of choice CYP81A6 encodes a natural resistance to the herbicide bentazon in rice (Zhang et al. 2007). Using the RNAi-type technology described in the section on “Molecular solutions to Striga, rice plants susceptible to bentazon were generated by suppressing the expression of the detoxification gene. Additionally, a gene coding for glyphosate resistance was engineered into rice by placing it in a tandem construct with the antisense form of CYP81A6 to assure that they will be inherited together. The implication of this novel research for gene flow in rice is that hybrids with weedy rice (as well as volunteer transgenic glyphosate-resistant rice) could be controlled in the following conventional-rice cropping season by spraying bentazon. By producing rice varieties with a combination of paired engineered resistances and susceptibilities to herbicides, a meaningful rice variety and herbicide rotational program could be designed for rice monoculture systems so that hybrids with weedy rice and volunteers could be controlled by different herbicides in sequential cropping seasons, preventing the persistence of herbicide-resistant weedy rice hybrids (Gressel and Valverde 2009).

Concluding remarks

Lessons can be learnt from experience in dealing with similar problems in other crops. The multiple herbicide resistance situations with Echinochloa are in many respects comparable to those occurring with multiple-resistant Lolium rigidum in wheat production in Australia (Walsh and Powles 2007; Broster and Pratley 2006), and Alopecurus myosuroides in Europe (Moss et al. 2007). This is typically due to the use of herbicides at lower doses than recommended, or at late growth stages, which for practical purposes has the same implication as applying a low dose (Gressel 2002). It has been primarily the lack of diversification in crops and agricultural practices, with thousands of hectares dedicated to monocultures grown with the same technologies that has resulted in widespread evolution of resistance to herbicides, including the most serious cases of multiple herbicide resistance. As we have illustrated with Echinochloa, the dynamics of the resistant weed populations can be altered to the benefit of the farmer by returning to good crop husbandry although this would most frequently require additional time dedicated to managing the crop and weeds and an increase in production costs. Diversification and integrated crop management are also key to dealing with weedy rice, which thrives only in association with the rice crop and not elsewhere.

Relatives can be your worst enemy

It is clear from the above examples with just three weeds, of how weeds are a major constraint to crop production. An issue not accentuated beyond rice, is that close relatives can be a crops worst enemy. Rice is not the only crop where a closely related weed is chemically uncontrollable in the interbreeding crop. If a transgenic solution is found, gene flow must be precluded. A feral/weedy form of sorghum (shattercane), which is the same species as the crop, is the worst enemy of sorghum. Unlike weedy rice, which is only a major problem in rice, shattercane is a pernicious weed in other crops, especially maize. Sorghum also has an interbreeding relative Sorghum halepense, which is in the top 10 worst weeds of the world (Holm et al. 1977). Likewise one of the worst weeds of beet is a feral form that flowers instead of forming a beet. Even wheat has an interbreeding relative Aegilops cylindrica that is a bad weed in many geographic areas. It is fascinating to observe how many scientists have jumped on the bandwagon of studying introduced and naturalized weeds, after xenophobically re-labeling them as “alien” or “invasive” weeds. There is even a new journal wholly devoted to invasive weeds. So few study the de-domesticated feral black sheep, weeds related to the crops. One wonders if all these introduced weeds together cause more damage than just weedy rice and shattercane, or more than native born and bred weeds such as Amaranthus spp. in the Americas, or Striga and Orobanche in Africa.

Conventional crop-based solutions vs. problem based solutions

Most agricultural research at present is focused on each crop separately. This is to be found in agronomy departments with specialists for each crop and has been extrapolated to the large international institutions devoted to assisting developing world agriculture; one institution deals only with potatoes, another with rice, one with sorghum and millet, another is devoted to legumes, and one deals with wheat and maize. The constraints to agriculture typically cross crop barriers; stem borers and grain weevils have omnivorous, voracious palates and chew most crops, mildews mold many crops, drought dries across species barriers, and Echinochloa and Striga infest a multitude of crops. Gene flow from crop to related weed is another issue that is general to many crops, as discussed above.

Over the last century, with the advent of plant physiology and genetics, we have begun to accrue considerable knowledge about these constraints that cross crop barriers. During the last two decades modern molecular biology has greatly enhanced the acquisition of this knowledge but more importantly, has provided the tools to use this knowledge to overcome constraints. Bt, glyphosate and glufosinate genes have been transformed into a number of crops performing the same task in each. The solutions are across the board. The time has come to realize this and reshape our international agricultural science centers to deal with issues, not crops.

The first green revolution came from breeding dwarf wheat at one institution and dwarf rice at another, independently. Now that we know that dwarfing adds yield at the expense of stems, we can use the same gene to transgenically dwarf other crops and increase yield, as has been done with oilseed rape (canola) (Al-Ahmad et al. 2006). Fragmenting research on constraints by crops does not provide the critical mass of researchers working together to conquer the problems. Even though parasitic weeds attack all crops, less than half the international crop research centers have a single researcher dealing with them. Others have none. The situation is probably not much better for stem borers, grain weevils, or mycotoxins. The research of these far-flung researchers is not even loosely coordinated; and the very best coordination is adjacent laboratories.

The universities and the major research institutions seem too conservative to change the paradigm. Unfortunately, the granting agencies who could force the change from a crop to a problem based approach have stuck to the crop-based paradigm. They will fund a project on enhanced nutrition in rice, another in sorghum, and yet another in cassava nutrition, yet want to enrich them all with the same nutrients. The same genes transformed into rice (Paine et al. 2005) and into maize (Naqvi et al. 2009) enhanced provitamin A content, and this will be the case for most other micro-nutrients in most crops. The same agencies have not supported consortia for multi-disciplinary, multi-crop projects for dealing with multi-crop constraints such as parasitic weeds, storage pests, or gene flow.

Cross fertilization is the best way to generate the diversity needed to overcome constraints, whether it be cross fertilization of genes to overcome constraints, or cross fertilization of ideas about how to overcome the constraints. The greater the distance between individuals, the less likely cross fertilization will occur, whether the individuals are plants with genes or individuals with ideas. Modern molecular biology allows people with ideas to put genes into species that do not normally cross-fertilize. That is why if we wish to rapidly and effectively deal with HIV or any other constraint in agriculture, we need to have the minds synergistically working together, not in the vacuum of dispersed institutions.

There is still a need for crop research. The best gene in an inappropriate genetic background may look great in a laboratory and greenhouse, but perform awfully in the field. Genes must be bred into locally adapted backgrounds, joined with other genes needed to cultivate the crop in that locale. Unfortunately, breeders are not being trained to do this, and far more crop breeders are retiring than are being produced.

Nothing is sustainable in agriculture

Weed management relying on a single tactic is not sustainable as weeds have the plasticity and genetically-based adaptability to adjust to changes in the agroecosystem. This is well illustrated with the increasing emerging cases of resistance to the herbicide glyphosate, mostly associated with no-till agriculture, production of transgenic crops resistant to this herbicide and overdependence on glyphosate for weed control in plantation crops.

Although we have discussed solutions that do or may work to solve the HIV problems of agriculture, we must be objective and realize that these solutions are ephemeral. We have discussed biological constraints to agricultural production. The weeds will eventually evolve ways around our solutions or they will go extinct. If they go extinct, they will quickly be replaced by other weeds filling the vacant agro-ecological niche. One cynical definition of pessimist is “realistic optimist”. We are realistic optimists; agriculture has dealt with each evolutionary hurdle in its way and has stayed ahead of Malthus’s dire predictions. It will continue to do so, but will do so best with new paradigms. The old paradigm was that we need taller crops to compete with weeds. The green revolution, coupling dwarfing with synthetic herbicides has saved millions from starvation in contradiction to the conventional wisdom of the time. Some weeds appeared that had never been controlled by herbicides, and some evolved resistance to herbicides, as we have seen with Echinochloa. We will not revert back to tall crops with low yields, we will find new solutions to the new problems as they arise, for as long as they last. We will try to develop strategies to make the solutions last longer. It is incumbent on us to do this in the most efficient manner. For this to occur, there must be new solutions on how agricultural research is done, or we will have less food security.

Copyright information

© Springer Science + Business Media B.V. & International Society for Plant Pathology 2009