The many benefits of growing a mixture of crop varieties together have now been demonstrated for many crops under many conditions. Latest entry is in a kind of specialised niche — organic tomatoes for processing — and the results are a little underwhelming. Three scientists at the University of California, Davis, grew one, three or five tomato varieties in soil that was either fallow or had a mustard cover crop the preceding winter. ((Barrios-Masias, F., Cantwell, M., & Jackson, L. (2010). Cultivar mixtures of processing tomato in an organic agroecosystem Organic Agriculture, 1 (1), 17-30 DOI: 10.1007/s13165-010-0002-z)) Although there were differences between the mixtures and the monocrop, they were not very pronounced: more shoots, and more fruits that were somewhat redder.
The 3-cv mixture thus had some minor advantages compared with the monoculture, but overall, there was little evidence of higher ecosystem functions from mixtures vs. monoculture.
Leaving aside the question of whether the growing conditions on the farm, as described in the paper, accord with organic ideals, given that they do abide by the rules, it is hard to know what to make of the poor showing of mixtures. The authors concede that there may simply be no benefit to be had because the conditions on California organic processing tomato farms don’t stress the crop to the point where a mixture might be a good thing. It is also possible that the varieties — AB-2, CXD-19, H-2601, H-8892 and Red Spring — do not actually encompass enough trait diversity to offer any benefits. But then, they were chosen as simply the best-performing processing varieties. If mixtures were assembled deliberately to deliver potential benefits, the results might well be different.
Still, good to know.
This seems to be typical; you need big differences to get significant benefits from complementarity. For example, diversity in maturity didn’t increase yield until harvest dates differed by at least 40 days (Baker et al. (1979) Exptl. Agric. 15:41 ). Disease resistance may be different, but think about spatial scale. A big field of a resistant cultivar is a better barrier between susceptible fields than a field with a mixture of resistant and susceptible cultivars.
Good point, although somehow I don’t think large maturity differences in a field of processing tomatoes would be a good idea. thanks for the link to those Baker papers from Nigeria.
A three cultivar mixture of one crop is, by definition, a monoculture, that is, one crop species (tomato) grown in a field – the number of varieties is not important. With the current anti-monoculture mania it has become fashionable to change the meaning of `monoculture’, notably in the oft-cited Zhu (2000) paper in `Nature’ on varietal mixture of rice which tried (and failed) to show that varietal mixtures of rice were more protected from blast than `monocultures’ (here used wrongly). In fact the experiments were all on monocultures – of a single crop, Asian rice, with several varieties. The paper then failed to demonstrate disease reduction because we now know the best explanation was the physical support of a tall rice variety by a short-strawed variety which prevented lodging: nothing to do with disease but a bonus for the use of varietal mixtures.
’It is a–MOST–PROVOKING–thing … when a person doesn’t know a cravat from a belt!’
Dave,
Are there published data supporting your “physical support” hypothesis? The different varieties were in different rows, so it’s interesting if one row could really prevent lodging of other rows. (Their diagram also shows that at least one of the single-variety treatments was at suboptimal density, raising doubts about claimed increase in overall yield.)
I agree that varietal mixtures don’t have the complementarity potential of species mixtures and that intercropping usually refers to the latter. But crops or varieties different enough to be complementary will often present practical problems, as Jeremy’s comment exemplifies.
Ford: Yes, there are.
Jill Lenne provided me with the following:
“A recent study in Yunnan has clearly shown that prevention of lodging of a tall, blast-susceptible glutinous rice variety was a measurable and important advantage of growing it in a mixture with a resistant hybrid (Revilla-Molina et al., 2009). Prevention of lodging has also been recorded as a positive character in mixtures of barley (Stutzel and Aufhammer, 1989) and wheat (Jackson and Wennig, 1997).
References:
Revilla-Molina, I. M., Bastiaans, L., Van Keulen, H, Kropff, M. J., Hui, F., Castilla, N. P., Mew, T. W., Zhu, Y. Y. and Leung, H. (2009) Does resource complementarity or prevention of lodging contribute to the increased productivity of rice varietal mixtures in Yunnan, China. Field Crops Research 111, 303-307.
Stutzel, H. and Aufhammer, W. (1989) Effects of winter barley cultivar mixtures on lodging. Journal of Agricultural Science 112, 47-55.
Jackson, L. J. and Wennig, R. W. (1997) Use of wheat cultivar blends to improve grain yield and quality and reduce disease and lodging. Field Crops Research 52, 261-269.”
Jill was a co-author of a paper on varietal mixtures: Smithson, J.B. and Lenné, J.M. Annals of Applied Biology 128, 127-158 (1996)
From my memory, in the original Zhu et al. study the rows were actually six of the modern, short-straw variety to one of the tall glutinous traditional variety.
At the time we challenged Mundt (one of Zhu’s co-authors) as follows: “In the Zhu et al. experiment, increased yields are ascribed to reduction in panicle blast severity.”
To which Mundt replied:
Yet, despite these caveats (and what I think was poor experimental design and confusing analysis), the Zhu paper continues to be cited as a prime example of the susceptibility of monocultures to disease.
And I’ll go along with Jeremy — there is now confusion as to just what monocultures are.
I agree with your second point about degree of complementarity between varieties versus species, especially for disease resistance and physical structure — as with the maize-beans-squash intercrop.
@Dave: I don’t understand the central point here (and we’ve strayed well away from tomatoes, but that doesn’t matter).
You seem to be saying that disease reduction was not the cause of yield benefits in rice mixtures. The prevention of lodging was the key. I’m happy to accept that prevention of lodging was an important factor, as is Zhu himself. It is also the case that rice blast was much less severe in mixtures than in pure stands. As far as I can tell from Revilla-Molina’s thesis, there was no comparison of blast severity with support against blast severity without support. So nothing much can be said about the relative importance of lower blast severity versus reduced lodging.
Secondly, I don’t see Zhu’s paper being used as evidence that monocultures are more susceptible to disease. I see it as supporting the idea that mixtures are a solution to the problem that monocultures are more susceptible to disease, as any fule know. All the Wolfe papers, and many others, point to this. Even breeders acknowledge it.
And as a final point, why is it important to partition the gains so accurately to lodging, or disease reduction, or sub-soil interactions, or light interception, or what have you? Isn’t the point that farmers can use diversity — at all levels — to gain benefits in yield and, more importantly, yield stability and incomes?
The important point is that monocultures are not always more susceptible to disease: it depends on the variety. The entire Zhu experiment used the fact that the modern hybrid – when grown in monoculture – was highly resistant to disease. By embedding the susceptible traditional variety within the modern variety, the expectation was that the traditional variety could be protected. But this did not happen as expected: in fact, in Zhu (2000) the site/year – Jianshui/99 – had the highest panicle blast severity and the highest overall yields. I spotted this as a botanist, not a pathologist, and it certainly should have rung alarms bells with the Zhu team. This shows that disease was not the major factor in yield reduction: there must be some other reason and Revilla-Molina et al., 2009 uncovered the lodging effect. By the time they did their experiment the modern variety, normally grown in monocultures, had become so effective at eliminating disease regionally that there was no disease to be measured in their experimental plots. The effect they measured was yield increase by reduction of lodging – a very likely reason for the results of the Zhu et al. study. But everyone associated with the Zhu experiment – and Wolfe’s puff in `Nature’ – wanted to believe in the `mixtures reduce disease effect’ and did not dig deep enough.
This leads to the highly important point for conservation: resistant modern varieties can so reduce disease regionally that, within the matrix of a modern variety, susceptible landraces can survive on their merits of a favoured grain type and not their disease resistance. This is certainly a plus for farmers wishing to grow landraces.
Apart from the fudged Zhu experiment there is a demonstrable but small mixture effect on yield. Jill Lenné has just written the following – not yet published but copied here with her permission.
“From a review of over 120 published studies, mostly in temperate regions under modern agriculture, Smithson and Lenné (1996) showed that improved stability and decreased disease severity were common features of mixtures relative to their components in pure stands. However, in the majority of cases, the yield advantage of mixtures was small being highest for wheat at 5.4%. A recent meta-analysis of 50 published studies on cereal mixtures confirmed the previous study finding an overall yield advantage of 2.7% only (Kiaer et al., 2009). At the same time, a number of studies on soyabean, groundnut, barley, maize, and wheat found yields of the mixtures to be significantly lower than the poorest component (Smithson and Lenné, 1996). It is therefore clear that mixtures per se do not guarantee yield improvements and, indeed, may produce considerably smaller yields, especially if inappropriate combinations of varieties are used.”
Kiaer, L. P., Skovgaard, I. M. and Ostergard, H. (2009) Grain yield increase in cereal variety mixtures: a meta-analysis of field trials. Field Crops Research 114, 361-373.
Smithson, J. B. and Lenné, J. M. (1996) Varietal mixtures: a viable strategy for sustainable productivity in subsistence agriculture. Annals of Applied Biology 128, 127-158.”
My (Dave’s) take on this is that if the average yield increases in varietal mixtures under modern agriculture are as low as 2.7-5.4%, then the possibility of recognizing or even obtaining any yield increase under developing country farming conditions with generally lower levels of management is very unlikely. But I could be wrong. And certainly for mixtures of different crops – notably cereal-legume intercrops – I would expect yields to be larger than the comparable monocultures. As to yield stability, the Green Revolution increased this over traditional cultivars under traditional management (and, of course, increased absolute yields substantially).
The Zhu et al paper has been systematically mis-interpreted & oversold. It shows that you can better grow a variety that is susceptible to blast if you plant it inbetween a resistant variety. This could be smart if you want to plant a susceptible variety because it fetches a higher price on the market. But for the high yield and low fungicide use you could simply leave out the susceptible variety.
This is entirely different from trying to maximize yield with mixtures of varieties with different levels of partial resistance, or with different crops.
There is no support for the notions that “rice blast was much less severe in mixtures than in pure stands.” (not in pure stands of the resistant variety!) and that “Zhu’s paper (…) supports the idea that mixtures are a solution to the problem that monocultures are more susceptible to disease, as any fule know”.
Mixed planting of these cultivars requires less fungicide per ha than pure stands of the susceptible variety; but more fungicide than pure stands of the resistant variety. If we believe the reported exponential expansion of this system, it should have led to an overall increase of the area with the susceptible variety and an overall increase of fungicide use in the region. Perhaps that’s worth the benefit of high value rice; but that is not the health or environmental benefit that is often mentioned. It would be nice to know how what the current area planted with this system is (I have heard is has declined). Another question would be how intercropping might affect the durability of the modern cultivars’ resistance.
Dave suggests that planting blast resistant rice varieties has changed regional disease pressure, such that traditional varieties in effect become more useful. The demise of stem-borers in the USA because of GMO corn is another example of such an effect. Organic growers benefit from GMO corn planted by their neighbors.
I wasn’t clear about monocultures and susceptibility; I was referring to the general case that a monoculture, especially of genetically more uniform individuals, will often be more affected by pests and diseases to which it is susceptible than a mixture of genetically less uniform individuals some of which are less susceptible to those pests and diseases.
Of course a monoculture is going to be better off if the threat is one it is resistant to, but one of the points that emerged from Wolfe’s barley work was that the unpredictability of the particular race of powdery mildew that would be dominant from year to year gave a mixture the edge over the combined yield of the pure stands. One pure stand might well do better; the difficulty is to know in advance which one.
Another point that I think is being lost here is the impct of mixtures on the whole farm system. Robert hints at this in his point about amount of fungicide per ha. Admittedly one doesn’t have a very good handle here on how and why nixed cropping is being adopted as widely as claimed, but there are some interesting figures in Revilla-Molina’s thesis.
Results show that farms under mixtures have a lower incidence of blast disease and farmer-adopters on average spent only US$10.50 ha–1 for pesticides, compared with nonadopters’ pesticide cost of US$42.92 ha–1 in 1999. Farmer-adopters had higher yields in 1999 than in 1996. In comparison with the yield on nonadopters’ farms, glutinous rice yield is 84% higher on mixture farms. The yield of hybrid rice was almost the same with 20% less land that was allocated to plant one row of glutinous rice between four rows of hybrid rice. Overall, the yield was 7% higher and gross return was 14% more since the price of glutinous rice, because of its higher quality, was twice that of hybrid rice. The net gain in farm-operator surplus was estimated at 25%.
To me, those whole-farm figures are compelling, although I agree it would be good to have new figures to see whether the impact is sustained.
It is quite possible, indeed likely, that total costs of production would be lower with monocultures of blast-resistant varieties, modern and, as Robert suggests, “traditional-improved”. But currently, we don’t have that, and it wouldn’t include the social costs of reduced labour requirements on the farms.
One more thing: what would keep (Chinese, IRRI) rice breeding programs from producing short, disease tolerant, high yielding varieties, that are glutinous and taste good? How hard can that be? A matter of (a lack of) focus?
Good question.