Today's science too often reduces cause and effect to single factor relationships, with little or no consideration for their context. So single genes, or in agriculture, single bugs cause disease, and we believe we can fix it by just addressing the cause.
Understanding interactions, even just between pairs of factors, is often a daunting challenge, and it's often unworldly when it comes to the real complexity of multiple interacting factors. We write all the time here on MT about the overly simplistic single-gene view of the world. It ignores genes in context, interacting with other genes, but also within a body living in and responding to the environment.
So, let's consider the asthma epidemic that began in the 1980s. Hundreds of millions of dollars, if not more, have been spent on searching for the gene/s for asthma. But why? Clearly something environmental caused the epidemic, not genes. But the high-tech search was on for genes for disease, a huge part of how this problem was framed. With, we might add, little success. (The epistemological problems with the environmental epidemiology of the problem are another story -- results have been all over the map, with the epidemic blamed on breast feeding and bottle feeding, hyper cleanliness and cockroaches and air pollution, and many more.)
The same drill-down approach was taken to agricultural pests and weeds. It seems to be at the root, for better or worse, of how science is done in our time, often with a technology-first approach. Build pesticides and herbicide-resistance right into the plant, problem solved. The consequences of this approach include, as was predicted by many, herbicide-resistant "superweeds" and pesticide-resistant insects, leading now to the return to the more toxic poisons that preceded high-tech, GM approaches to these problems.
Liebman reminds us that the results of perturbing an ecosystem can be unpredictable. Indeed, they can often only be understood after the fact. But, there are examples of integrated pest management that take advantage of relationships within existing ecosystems, and that have fewer destructive effects on the system than do toxic chemicals or genetically modified organisms.
An example that Liebman uses is integrated management of the weed, Striga hermonthica, stem borers and maize. Stem borers attack maize and sorghum in many regions of Africa, substantially reducing crop yields. The parasitic weed, Striga hermonthica, also causes major crop losses throughout Africa. It's particularly a problem in poor soils, and is hard to control with herbicides.
|Khan Z R et al. J. Exp. Bot. 2010;61:4185-4196|
Push-pull technologies exploit interrelationships between plants and pests, potentially controlling weeds and insects both. In the case of Striga and stem borers, Zhan et al. report the successful use of push-pull techniques to control both. Stem borer moths are repelled by Desmodium, which is planted between the crop rows, and emits unattractive, push volatiles, and they are attracted by Napier grass, which emits appealing, pull chemicals, and that also happen to be sticky. The moths stick to the plants, are immobilized and die. Further, chemicals emitted by the roots of the Desmodium, the push plant, prevent attachment of Striga to the maize roots, so the weed can't grow.
So, targeted weed and pest control, without additional disturbances to the system. The Desmodium increase the nutritive value of the soil for the maize, which helps to discourage Striga growth, and they also repel stem borers. The Napier grass attracts them, and unless you're a stem borer or Striga, this works for everyone. As Matt has said, some have suggested developing transgenic herbicide-resistant maize. But it wouldn't be nearly as beneficial to the ecosystem. And would surely have unintended consequences.
Biodiversity, Liebman points out, also can reduce the intensity of zoonotic diseases, diseases that are communicated from animals to people. (Just to illustrate how important this can be, David Quammen had a chilling piece on coming zoonotic diseases in the New York Times the other day.) Lyme disease is caused by a pathogen, Borrelia burgdorferi, that lives in the field mouse, or the white-tailed deer. When a deer tick, Ixodes scapularis, bites an infected host and then bites a person, the person too can become infected.
If biodiversity is high, however, this can dilute transmission of the pathogen, lessening the incidence of vector-borne illnesses like Lyme disease. And, decreased biodiversity can lead to increased incidence. So, if ticks have multiple food sources, such as opossums or other animals that don't harbor B. burgdorferi, fewer mice will become infected.
There are implications of these lessons on how farming is done today as well. Factory farming concentrates disease. Liebman notes that there are 60 million laying hens in the state of Iowa alone, producing 15 billion eggs per year. Each egg-producing farm houses about a million animals. That's a million. This concentrated, centralized egg production makes it very easy for salmonella to spread rapidly and wide. More farms with fewer hens distributing eggs shorter distances could mean that the next outbreak of salmonella would be smaller and more contained.
We write a lot about complexity here, largely in the context of genetics and human disease. Lots of people use the terms but often mean that what we need is just bigger samples and more data. That's a technology first, ideas (perhaps) later. Even at this late date we and others who see and write about this can feel like lone voices in the wilderness. For one thing, any challenge in any culture to the way things are done will be resisted, for good reasons some of the time, but also for selfish ones.
Perhaps it's fair to characterize twentieth century science, a legacy of the Enlightenment period 400 years ago, as the effort to reduce systems to their smallest components -- quarks, bosons, genes. The work of Matt Liebman and others in agronomy is a heartening sign that at least some science in the twenty-first century may well focus on the importance of considering whole systems. The challenge of doing so beyond some model systems that work, such as we have described, is a huge one. But the consequences of not doing so can be dire.