Pondering polycultures: or, arguments within English permaculture…

…to affect a grandiloquent paraphrase.

So, first a happy new year to everyone. Looking at your editor’s 2015 workload on and off the farm I fear that my blogging is going to be quite infrequent this year, but let me start with good intentions and something meaty. I’m currently in the middle of a series of posts about eco-panglossianism, but I thought I’d take a short break from it to address the question of polycultures (ie the practice of growing 3 or more different crops together). Last November, Patrick Whitefield took me to task for ignoring or belittling the evidence that polycultures could outyield monocultures in one of my posts, so I want to pick up on this issue in a little more detail here.

To begin, I’d like to reprise what I actually said, which was this:

“biodiversity in the wild usually results from niche occupation by organisms with specialist skills in tapping often recalcitrant resources, whereas human cultivation usually relies on getting high returns from a small number of organisms that respond impressively to high resource availability when humans make conditions favorable for them. This explains why, at least at a given level of the system (a vegetable bed, for example), there is little compelling evidence that polycultures or companion planting are, in general, more productive than monocultures.”

Patrick was unimpressed with this, citing various studies from Altieri’s book Agroecology and others from his own excellent Earth Care Manual which showed positive land equivalent ratios for polycultures. He also accused me of trying to damn polycultures by associating them with the ‘wishful thinking’ of companion planting. And in the course of the debate he wrote, “The longer I practice permaculture…the more I’m convinced that dogma and off-the-peg solutions don’t help at all. Every situation is unique. Every piece of land is unique and so are the poeople who work it. It behoves us to choose the unique solution that will work best in each situation.”

I’m a great admirer of Patrick and his work, though if I might make so bold I don’t think my statement on the general productivity of monocultures vis-à-vis polycultures is negated by the existence of some studies reporting superior polycultural yields. Nor, I think, can my comments reasonably be interpreted as a wholesale dismissal of any kind of polyculture. But anyway, let me try to untease some of the underlying issues.

One way to begin would be to think about my own farming – do I practice polyculture? Well, each year on my 18 acre site I’d guess that there are well over 100 species that I’ve deliberately introduced co-existing, and many more species (in the permanent pasture, for example) that I haven’t introduced myself but am happy to make use of. Compare that with most 18 acre blocks of agricultural land in the vicinity where you’ll typically only find one crop growing at a given time, and I’m inclined to say that, yes, I’m a polyculturist! However, if you were to take any given square metre of cultivated land on my holding, you’d probably find only one or sometimes two crops growing there, so at that level – like a lot of commercial growers – perhaps I’m a monoculturist after all. My point is that scale may be important here. When does a monoculture become a polyculture? There are also scale effects which pose interesting problems for agricultural policy: there has been both tropical and temperate research that suggests increased crop diversity at the farm level may not have much effect on crop yields or wild biodiversity, but this finding is reversed when crop diversity is practiced at the whole farmed landscape level.

Let me make one more generic point before moving on to specifics. My natural sympathies incline to polycultures because it suits my politics: I like the idea of different entities (plants, people, nations) coexisting peaceably and strengthening each other through diversity. I’m also drawn to the idea that certain things (like plant polycultures) may work for mysterious reasons that are too complex for people to understand at present, and possibly ever. But at the same time, there are dangers here: nature works in all sorts of ways (like natural selection) that don’t really suit my politics at all, which isn’t a problem for my politics because human politics are completely different from inter-specific interactions, but it can be a problem if I try to read my politics into the script of the natural world (the same, of course, applies to right-wing ‘red in tooth and claw’ types). And likewise, though I’m drawn to the mystery of a functioning plant polyculture, I think it’s usually a good idea to try to understand as clearly as possible why it seems to work. I’ll come back to these points again at the end.

Hell, I’ll come back to the last one straight away. Let me suggest the usefulness of limiting factors as a way of thinking about polycultures – the most important ones, I think, being space and/or sunlight, fertility, water, pest pressure and labour. So let’s imagine some kind of generic patch of ground for growing crops, with a given soil and climate (seasons, rainfall etc).  I want to produce the optimum amount of crop biomass that I can eat, burn, weave or otherwise make use of from this patch by capturing sunlight, water and nutrients, hopefully in such a way that I don’t deplete the opportunities for doing the same again in the future. Let’s imagine how a few monoculture and polyculture scenarios might play out here. Maybe I can increase my returns by planting a crop mix which makes full use of the year’s solar radiation over time (early/late photosynthesis). Or maybe I could do the same with a crop mix that makes full use of solar radiation in space (canopy and ground layer crops, climbers etc). Again, perhaps I could increase returns by making fuller use of the space available in the rhizosphere – in this instance the ‘space’ available for my crops’ roots is probably a function of the availability of various key nutrients in the soil, which in turn will be associated with my fertility building strategy. Perhaps I can plant some crop mixes that will help with this strategy. Maybe crop productivity in my locality is also constrained by water availability, so I might improve returns with crop mixes that reduce water evaporation from the soil. Maybe it’s constrained by the depredations of a certain crop pest or pests, and again I might be able to reduce this with crop mixes that deter or confuse the pest, or promote the flourishing of one of its predators.

There are surely going to be tradeoffs involved in many of these possible polycultures. Maybe a crop mix will help deter one pest, but promote another. Or it will increase total yield, but also increase the labour needed for establishment or harvesting. And the nature of the limiting factors and the tradeoffs will likely depend on the type of grower and growing space. Someone planting a small home garden will probably be limited for space, but not for labour (it’s a hobby), variable returns from their polyculture experiments (it’s a hobby), or fertility (it’s easy to import external fertility). Therefore a space-stacked, labour-intensive polyculture may commend itself in this situation. A commercial organic grower in a wealthy country, on the other hand, will probably not be limited by space, but will be limited by labour (probably the costliest input for a commercial grower when you’re balancing hourly wages with financial returns to veg growing), fertility (it’s hard to generate enough organic fertility broadscale) and variable returns (you can’t afford crop failure).

To give an example, this year I undersowed my cabbages with a trefoil and white clover mix with the aim of boosting fertility and deterring caterpillars. It was probably successful in this, but it also promoted slugs and, I suspect, water competition with the cabbages during the hot summer, with the result that I lost a lot of crop. Maybe weather conditions will be more propitious for it next year, and my skill in timing the sowing will be better. But then again, maybe I’d be better off just adding some muck and netting the cabbages – I don’t want to lose as many cabbages next year, so I need to be pretty sure that the polyculture is a better solution than the monoculture plus muck and netting. If, on the other hand, I were a poor peasant farmer without access to costly nets or bought in fertility, but with a lot of available labour, then the polyculture solution would probably be best in this situation: many indigenous peasant agricultures have figured out such polycultures over the long term, and in my opinion it’s probably best for poor small-scale farmers to stick to them rather than be tempted by the blandishments of agricultural ‘improvers’ into growing cash-crop monocultures involving a lot of fancy inputs.

So the moral of the story so far, as I see it, is that Patrick is right in saying that it behoves us to choose the unique solution that will work best in each situation. And I accept that in some situations that solution will be a polyculture – though, in view of my preceding comments about politics, mysteries and explanations, I think that when that’s the case it may be better not to make too much of the fact that the solution is a polycultural one as if that’s somehow a good thing in itself, but to appreciate precisely why in that particular situation the polycultural solution deals best with the various limiting factors at play. This minimises the danger of people inferring that there’s something intrinsic to polycultures themselves that makes them the more optimal solution and then seek to apply polycultural solutions willy nilly in other situations. It’s similar to the notion that perennial crops are somehow intrinsically better than annual ones – but more on that in an upcoming post.

With all of the examples I mentioned above, the reason the polyculture works is additive – it’s not that there’s anything mysterious about the combination of the different crops, it’s just that together they make better use of total available resources than a single crop can. Now, with one very important caveat which I will discuss in a moment, I accept that there is evidence that polycultures can be more productive than monocultures – though as mentioned above I don’t see this as being inconsistent with my original comments. Perhaps my experience as a commercial grower is relevant here – organic commercial growers rarely grow polycultures (except in the sense I mentioned above of their whole farms or their rotations being polycultures), I suspect because the benefits of doing so are usually outweighed by the costs at their particular scale of operation. In his book, Patrick writes “Because more diverse systems are more complex they often also require more day-to-day management and more labour than monocultures….They substitute human input, mainly in the form of skill, for heavy inputs of chemicals and machinery” (Earth Care Manual p.266), and he goes on to point out the difficulties of adopting these systems commercially in an economy where the relative prices of human labour and fossil energy are stacked heavily in favour of the latter. I agree. My only slight misgiving, if I may make so bold, is I think I detect a certain sniffiness in these words about the superiority of labour-intensive skill over capital-intensive input. Well, I guess I share it myself, which is why I spent time last year buggering about with trefoil and clover mixes because it felt to me an intrinsically more elegant solution than muck and netting. More elegant yes, but more labour intensive…and not as effective. Now, I’ve long advocated on this blog the benefits of a more labour-intensive agriculture, but I’m inclined to reject the duality of skilled/labour-intensive vs unskilled/input-intensive as a little too simplistic. It’s true that commercial growers often have to adopt more simplified cropping systems than those that may commend themselves in a domestic garden – however, I don’t think it’s true that running a successful commercial growing operation involves less agronomic skill than running a successful domestic garden.

OK, nuff said on all that. I want to move on now from the notion of polycultures as additive in overcoming limiting factors to the possibility of them being interactive. In other words, it’s not just that I can tap a bit more total solar energy per unit area by training a squash plant up a maize stalk but that there is some specific beneficial interaction between these plants (or any other specific mix of plants that you care to mention). This is the essence of companion planting (plant x complements plant y). It really wasn’t my intention to damn polycultures by association with companion planting, but it interests me that Patrick dismisses companion planting as mythological wishful thinking, which suggests that he’s not persuaded that there are many beneficial interactive effects between specific plants – perhaps the mysteries of which I spoke earlier in fact are few.

Well, let me speak up for one such interactive effect, which stands out loud and clear from the article on ‘Polyculture cropping systems’ by Matt Liebman in Altieri’s Agroecology book to which Patrick  refers. This is the association between legumes and various other plants – for the vast majority of the examples of successful polycultures (really, bicultures) mentioned by Liebman in fact are associations between a legume and another crop. This is no mystery, of course – the association between legumes and rhizobia, and thus the ability of legumes to fertilise the soil with nitrogenous compounds is well understood. Liebman states on the basis of studies on sorghum/pigeon pea bicrops that (1) higher gross yields of the bicrop don’t result from the fertilising effect of the legume, and (2) that though association with sorghum reduced the size of pigeon pea plants, it increased allocation to seeds so that seed yields were still ‘quite high’. The first finding seems to relate to the additive effect of complementary root foraging in the rhizosphere in situations of high fertility, whereas the second one is indeed a bit mysterious but perhaps relates to interspecific competition forcing seed allocation – I’d like to read the original paper to get beyond the vagueness of the ‘quite high’ yield, but unfortunately I can’t breach Cambridge University’s paywall.

So, yes, there’s definitely scope for pursuing legume bicultures. But I think Patrick’s suspicion over the ubiquity of interactive effects that he expresses in relation to companion planting is probably well founded. Since atmospheric nitrogen is an effectively unlimited resource once it’s synthesised into soluble compounds, it’s not surprising that legumes splash it around in the rhizosphere, but we can’t necessarily expect such generosity to be widespread in relation to more recalcitrant resources. Maybe something similar can occur in relation to phosphates, as in the current rage for buckwheat – another companion crop I’ve been messing around with in my market garden. And there are doubtless some worthwhile polycultural solutions to pest problems (in fact, Perfecto et al have a very interesting discussion of this in relation to coffee, which I previously discussed) – though according to Ford Denison the evidence for increased yields resulting from polycultural solutions to pest problems isn’t that compelling, and I guess it was this comment that lay behind my original claim and that got me into trouble with Patrick. There do also seem to be some non-leguminous bicrop or polycrop mixtures that Patrick reports in his book and on my blog with increased gross yields – so maybe there is scope for a bit of interactive mystery here after all.  Then again, I’d want to look with a bit of care at the possible reasons: there can be artefactual effects of including a heavy yielder in the crop mix, or it may be something as simple as wind protection which may – may – be better provided by another means.

So by way of conclusion, I’d like to make the following eight propositions for debate:

1. Mainstream agriculture has become too dependent on monocultures of a small number of high-yielding and low labour/high energy input crops such as wheat, oilseed rape and perennial ryegrass.

2. It’s a good thing for numerous reasons to develop a more diverse farmed landscape, including lots of small farms growing many different crops. One of these reasons is that this may improve yields per unit area, but this isn’t always necessarily the case, and sometimes rotational monocropping within an overall diversely cropped farm may be appropriate.

3. Polycultures can nevertheless improve per unit area yields in some cases – usually because, additively, their various components can make better total use of resources.

4. In a given situation, a polyculture may or may not be the best solution – and this will probably depend on the scale of operation and the nature of the labour available. Generally speaking, polycultural solutions are more likely to commend themselves in space-constrained, labour-abundant non-market growing situations than in space-abundant, labour-constrained market growing ones.

5. At a micro level, polycultures are not intrinsically ‘better’ solutions than monocultures, though in a particular situation a polycultural solution may be better than a monocultural one.

6. There are likely to be complex tradeoffs with any type of agronomic solution, polycultural or monocultural: the solution will probably reduce some problems but compound others.

7. Evidence for interactive rather than just additive effects (or, if you will, for the efficacy of companion planting) is limited, but not wholly absent. The most important example by far is the well understood one of legumes…

8. But there are other examples too. Long live mysteries! So long as we don’t get too mystical about them…

Of course, I’d welcome further comments on this. And my thanks to Patrick for stimulating the debate.


16 thoughts on “Pondering polycultures: or, arguments within English permaculture…

  1. Do you have the specific citation for the sorghum/pigeon pea article cited by Matt? I’ll not even try to breach Cambridge U’s paywall… but I may have a mysterious way of divining said content. Once upon a time, in a land far, far away I was gainfully employed doing research on a grass known as Sorghum bicolor. And I still know a few folk who care.

  2. As usual a very interesting article. You’ve concentrated on theory but I certainly would have liked the odd post on such leguminous projects (including photos). I’ve really got to pop down on one of your open days and have a look, we’re practically neighbours.

  3. Thanks, gents. Sorry, I’m not very good at taking photos of things, but once I’ve got the eco-panglossians out of my system I’m certainly planning to write a bit more about practical projects on the farm. Looking around at the mess of it and all the unfinished projects here at the moment, I feel nervous about inviting anyone to visit but of course you’d be very welcome to come. Perhaps I should advertise open days here on the blog as well as on our Vallis Veg site – I tend to think of the blog as less for local consumption…

  4. Sorry, Clem, forgot to add that the reference is Natarajan, M. and Willey, R. (1980) ‘Sorghum pigeon pea intercropping and the effects of plant population density’ J. Agri. Sci. 95: 59-65.

    Any thoughts on Sorghum bicolor vs Sorghum halepense?

    • Sorghum bicolor v. S. halepense?? Someone put you up to that, right?

      So the first is an excellent grain crop and the second is a miserable weed. Wait, I guess its not that simple. For starters, they are cross compatible [the stomach turns just imagining the horrors of hybrid Johnson grass roaming the countryside willy nilly].

      Johnson grass (halepense) does have some value as a grass in pastures (though from my comments here you’d rightly suspect I’m not a fan). It is perennial, has rhizomes, and can be pretty tough to control. Sorghum bicolor is grain sorghum (or milo) and while it is perennial in a strict sense, it is grown as an annual (is not winter hardy) in most of the areas I’m familiar with.

      I’m curious though – your interest? I’d be surprised to learn that S. halepense will grow in your environment. I’d have guessed it too sensitive… (is typically a weed south of me… Brian likely knows it well – though I do occasionally see it Ohio my comments above come from a childhood in southern Illinois where it thrives).

      In a commercial veg garden, S. halepense will get you’re attention.

        • So the challenge becomes recognizing it as a seedling in the garden. On the pasture side of the fence… no big deal, on the garden side of the fence – controlling the seedlings important. Preventing plants on the pasture side going to seed is a BIG help.

          • Now that I think back we had a heck of a time eradicating JG from the garden space. Mr. Johnson was an improver in South Carolina in the 1800’s. He brought in the grass to help with erosion, about as effective as using kudzu.

          • Kudzu is worse than Johnson grass. Hands down.

            But Kudzu is a nitrogen fixing legume I hear Tom saying. So how can it be worse? It can host the fungus that causes soybean rust.

  5. I’ll start the engines on tracking down the article on sorghum and pigeon pea. But while I’m off on that errand, please allow me to make one challenge to your assertion that you’ve deliberately introduced “well over 100 species”…

    And while I’m not asking you to publish the entire list, I’ll offer that the home farm I grew up on had a pretty serious commercial veg operation in addition to some livestock (cattle) pasture and attendant hayfields. There was also a fair sized woodlot with occasional firewood harvest. Just out of curiosity I started a list of the veges and various other crops we planted. Added in a smattering of the more common timber species from the wood that we made use of. I’m still a bit shy of 100. Now it wasn’t our intent to pass 100 species, but I’m not the least embarrassed by the list that I’ve put together either. If I listed all the ‘varieties’ of vegetables (4-6 different green beans, 5 or more sweet corn hybrids, multiple tomato etc) then we’d easily pass the 100 mark (over a couple years effort). But considering broccoli, cauliflower, and brussels sprouts are all the same species, well… “well over 100 species” is a pretty impressive list!

    • I think the perennial grain breeders at the Land Institute have spent time playing around with bicolor x halepense. If you go to Kansas, be sure to wipe your boots!

      100 species? OK, I admit it was a bit of an unguarded off the cuff remark. The perils of blogging. And you’re right that I probably had varieties at the back of my mind rather than species as such. I guess I’ll try to compile a proper list one of these days – maybe I’ll post it up on here, depending on how embarrassed I am by the result.

      • OK, Clem – back of an envelope job this afternoon in an idle moment…I think I may be able to meet your ‘100 species challenge’, though truth to tell it’ll be more like ‘just about 100 species’ rather than ‘well over’. Anyway I’ll post on it in a while.

  6. Brian, thanks for the history lesson on the origins of ‘Johnson grass’…I had no idea. How terrifying to think that there could be a noxious weed in the future known as ‘Smaje grass’ as a result of my misguided enthusiasms. Fortunately, I suspect my botanically unadventurous tendencies make this unlikely.

    Also, some nodule pix courtesy of Clem for the edification of any legume enthusiasts who are reading:



Leave a Reply

Your email address will not be published. Required fields are marked *