This is the final part of a two-part review. You can find part 1 here.
So is there anything I don’t like about Darwinian Agriculture? Actually, yes. Intercropping and polyculture, which is usually taken to mean mixing different crop species in space (and sometimes animals too), is, in Denison’s view, not proven to be more productive than growing the component species on their own in rotation. Fair enough; too often the long-term comparisons over a full cycle or two of rotations haven’t been done. But he doesn’t consider how polyculture might affect year-on-year stability; in marginal systems, especially, that may be a more important consideration than total productivity. The same goes for what I would argue is one of the clearest ways in which imitating Nature might be a good idea: mixtures of varieties of the same crop.
Denison explores the reasons for the 1971 epidemic of Southern Corn Leaf blight in maize, and uses it to make some telling points. But having said that “this disaster is often used, rightly, to show the need for more genetic diversity within crop species,” and promising that he would return in later chapters to the question “which is more useful, diversity within fields or diversity among fields?” he doesn’t, at least not for diversity within a species.
Of course it isn’t fair to criticise an author for not covering a topic close to my own heart, but I do wish he had given at least some space to the question of intra-specific diversity as a perfectly good Darwinian response to the threat posed by pests and diseases: lack of diversity in the response to a threat is risky. This idea comes into the management of resistance to GMO plants, with a good discussion of the need for refugia of non-engineered plants as part of a strategy to delay the emergence of resistance. But as a more general mechanism, it gets short shrift, which is a shame. And it isn’t as if the evidence is lacking, at least for damping down yearly yield variability.
You could get the impression from this review that Ford Denison is long on the problems with agriculture and short on solutions. You would be wrong. In fact, some of his proposed solutions are so exciting I honestly wished I were just starting on a long research career. Boost the cooperation between legumes and nitrogen-fixing bacteria? Count me in. Figure out how to use plant-animal signals more effectively than always-on alarm pheromone? You bet. Ask in detail how this year’s crop could benefit next year’s? Yes please. Just because trade-off blind biotechnology and unthinking mimicry have so little to offer is no reason to despair.
Perhaps the most important thing Denison has to ask, in what is either true humility or unbridled academic chutzpah, is “What if my proposed core principles turn out to be wrong?” This is the central argument for his rider, on the need for a bet-hedging approach that allows ideas to compete. This, too, appeals to my confirmation bias. Simplistic, whizz-bang approaches suck the air out of the room, leaving less shiny ideas gasping, and nutrition, breeding and sustainability are three particularly vulnerable areas. Right now a squillionaire philanthropist, with, say US$30m to invest, has a couple of choices. Increase the donor budget for simplistic whizz-bangery by somewhere between 1% and 10% – nice, but hardly a big deal. Or double the budget for “other” approaches, which could indeed make a real difference.
You’ll have gathered that I think Darwinian Agriculture is the best non-fiction I’ve read in a long, long time. Anecdotally, others in the field seem to agree. The big question is, how to get those who make the big decisions, and who clearly don’t understand either ecology or evolution as they apply to agriculture, to pay attention. Answers in the comments, please.
Jeremy,
Actually, Figure 7.1 compares within-species diversity within small fields versus between fields, but it implicitly assumes that we often don’t have enough locally adapted varieties to have diversity at both scales. That assumption interacts with my concerns that the high-cost of biotech approaches will leave us with fewer varieties of major crops, but it’s probably too pessimistic, at least for now. So, thanks for the correction.
Wasn’t it here that I read the suggestion that the yield benefits in the famous rice intercropping example from China are largely due to the resistant variety propping up the susceptible one, thereby reducing lodging? If so, could you provide a link, as an example of the complexity of benefits from fine-scale mixing?
Actually, it is Dave Wood who insists that Zhu’s results can be explained by support against lodging, although I have not seen a paper testing that directly.
Lodging would not explain Wolfe’s results with barley variety mixtures. There are a couple of other pretty good studies too, on cereals, e.g.:
Finckh, M. R., Gacek, E. S., Goyeau, H., Lannou, C., Merz, U., Mundt, C. C. et al. (2000). Cereal variety and species mixtures in practice, with emphasis on disease resistance. Agronomie, 20(7), 813–837.
And, of course, good theoretical considerations, from Mundt and Garrett. I have more refs if you want them.
There is also some suggestive evidence emerging slowly from Devra Jarvis’ multi-country project on pest & disease control. See Mulumba et al. for interesting data on bean diseases in Uganda. And little evidence for any protection in banana, which is what you might expect from theoretical considerations.
Jeremy: “insist” is a bit strong for me. You yourself noticed the lodging in the rice mixtures system here on 25 Sept. 2007, Zhu himself in a paper by Leung et al. commented on lodging in 2003, and paper by Revilla-Molina et al. showed it clearly in 2009 [http://dx.doi.org/10.1016/j.fcr.2009.01.003].
What was wrong with the Zhu et al. 2000 paper was that they set out to show that mixtures contributed to disease reduction and did so, ignoring other factors (like the anomaly that the highest overall yield was in the season with the highest disease pressure). Lodging is a major factor, but in claiming that the adjacent hybrid protects the traditional variety everyone ignores the obvious. It was in a region with already over 3000 ha of the resistant hybrid. This reduced disease regionally, not just in the next row. The traditional variety could probably have been grown in 1 ha fields and had the same protection.
The Malumba et al. paper you link too is a bit odd. I am trying to get my head round the sentence – repeated three times – that: “In sites with higher disease incidence, households with higher levels of diversity in their production systems had less damage to their standing crop in the field compared to sites with lower disease incidence.” Hmm.
Also, there was an absolute need for this paper to cite [DOI: 10.1111/j.1744-7348.1996.tb07096.x] the Smithson and Lenné 2008 paper on varietal mixtures: Barry Smithson was a bean breeder with extensive experience of Africa, and Lenné (a pathologist) had projects on beans mixtures in Tanzania. Lenné tells me that the mixture effect, after detailed rigorous field research on Phaseolus bean mixtures and three major diseases, is impossible to show under farming conditions (although they found, as here, some good resistances in particular lines).
My own collecting experience from Ethiopia, Yemen and Kenya was that farmers very often grow complex crop and varietal mixtures, and in Tanzania, for beans, separate seed from the mixed crop and then reconstitute ad hoc mixtures year by year with up to a dozen varieties (determined by seed-coat colour). There is very obviously some good reason(s) why farmers go to this trouble but it is too easy to make assumptions as to the reasons and conduct research to bolster these assumptions – as with the Zhu et al. 2000 paper.
Note that the bananas and beans of the Uganda study are both introduced crops. I am surprised that lists of beans varieties from four regions of Uganda did come up with Kabanima, bred by Colin Leakey in Kawanda in the 1960s and subsequently incorporated by farmers in what are probably now regarded as `traditional’ varieties (it could be that lots of the supposed traditional varieties listed by Mulumba et al. are rusticated improved varieties – as were a lot of the wheat varieties in India before the Green Revolution – from Albert Howard’s work in the 1920s at the original Pusa Institute).
This is an example of what I meant by “bloggers and commentators at Agricultural Biodiversity set the standard for insightful discussion.” Thanks for the information!
<blush />
Very kind!
Would it be useful if we tried to get Mulumba et al. to comment?
I wouldn’t hold your breath. I do think that the studies in that project — which includes similar work in China, Morocco and Ecuador — are very much more complex than a simple “do mixtures protect” might be. And as a result, the answers are not as clear as one might like them to be. Certainly having seen preliminary analyses, I’m not always fully convinced.
The sentence Dave Wood is trying to get his head around is intended to say, I think, that there is an interaction between the severity of disease pressure and the protection afforded by diversity. When there isn’t much disease, diversity offers less yield protection than when there is a lot of disease. In other words, it is a good bet-hedging strategy when one cannot predict how severe disease pressure will be.
Thank you, Jeremy, for letting me borrow your copy of the book, while waiting for my copy to be delivered. (and since Ford Denison is listening in it seems: thank you very much for writing this book!!) The book is fantastic and truly inspiring. And, well, I do have a (hopefully) long research career in plant genetics in front of me. I sincerely hope to translate the inspiration into something useful.