« Carnival of evolution | Main | Happy Darwin Day! »

Tradeoff-free drought resistance?

I'm working on the last chapter of my book, Darwinian Agriculture: where does Nature's wisdom lie?, and will be sending it to Princeton University Press soon for their review process. So I'm alert for any information that might make me change my main conclusions.

One theme of the book, and also of my recent talk at the Applied Evolution Summit, is that past natural selection is unlikely to have missed simple, tradeoff-free improvements. So I'm always skeptical when someone speculates that we could double crop yield just by increasing the expression of some newly discovered "drought-resistance gene." My rationale is that mutants with greater expression of any given gene are simple enough to have arisen repeatedly over the course of evolution. (This contrasts with more-complex adaptations, like the ability to form a symbiotic relationship with nitrogen-fixing bacteria, which natural selection may have had fewer opportunities to test.) If past natural selection has repeatedly rejected these "drought-resistant" mutations, then they must have been some negative effects on fitness, at least in past environments.

But could some mutations repeatedly rejected by past natural selection still be beneficial in agriculture? Maybe. Tradeoffs that constrained past natural selection need not always constrain us, as discussed below. But, I suggested at the meeting, we can't just ignore them.

However, someone called my attention to Drysdale wheat, developed in Australia, and reported to have higher yield under drought than older varieties, without any apparent yield penalty under wetter conditions. Is this an example of a tradeoff-free improvement missed by past natural selection?

The answer can be found in a very interesting paper, Breeding for high water use efficiency, published in Journal of Experimental Botany in 2004 by A.G. Condon, R.A. Richards, G.J. Rebetzke and G.D. Farquhar. If you are at all interested in this topic, the entire paper is well worth reading. But here are some key points....

Yield of Drysdale wheat, relative to a previously recommended "check" variety, as a function of increasing water availability (as indicated by yield of the check variety). Condon et al. 2004 J.Exp.Bot.55:2447.

As shown above, Drysdale outyields a good older variety by up to 40% under the driest conditions (when the older variety yields only about 1000 kg/ha or about 1000 pounds/acre), while having similar yield to the older variety under wetter conditions, when the old variety yields about 5000 kg/ha. This is clearly a major practical advance. But does it contradict my hypothesis, that natural selection is unlikely to have missed simple, tradeoff-free improvements?

No. There are many different kinds of tradeoffs. Actually, the paper discusses several of these tradeoffs in lucid detail, as summarized below. But some tradeoffs that were key to past natural selection are relatively unimportant to us today. For example, how well would Drysdale do under wetter conditions, beyond the range shown in the graph above? Such conditions might be very rare on Australian wheat farms. From a practical standpoint, it wouldn't matter if Drysdale has slightly lower yield under conditions that hardly ever occur. But the wild ancestors of wheat must occasionally have experienced wet years, so that even slight tradeoffs could have affected past evolution.

What about other tradeoffs? Any plant that survives drought may be considered "drought-tolerant." But Drysdale was selected for water-use efficiency (WUE), the ability to actually grow and produce grain with a limited water supply. Water-use efficiency at the leaf level is the ratio of photosynthetic uptake of carbon dioxide, divided by transpirational water loss from leaves. Carbon dioxide diffuses into leaves through the same stomata that water vapor diffuses out, so

WUE = photosynthesis/transpiration = (Ca-Ci)/(Wi-Wa)

where Ca and Ci are CO2 concentrations in the atmosphere and the leaf interior, while Wa and Wi are corresponding water vapor concentrations. (I have left out a constant that doesn't affect any of my conclusions.)

So, one way to increase WUE is to increase Wa, the CO2 concentration in the atmosphere. You can help by burning more coal. Unfortunately, this approach can have various negative side-effects, including increased temperatures, which can raise Wi and thereby lower WUE.

A more-practical approach is to decrease Ci. Drysdale may do this by closing its stomata more than other varieties. With partial stomatal closure, CO2 can't diffuse into the leaf interior quite as fast, so photosynthetic uptake pulls leaf-interior CO2 concentration (Ci) lower, increasing WUE.

So here's our first tradeoff: lower CO2 diffusion into the leaf can mean higher WUE, but a lower photosynthesis rate. Wouldn't a lower photosynthesis rate result in lower yield? Not necessarily. Lower total seasonal photosynthesis would probably mean lower yield. But, by using water more efficiently, Drysdale may photosynthesize for more weeks before running out of water. Slightly lower photosynthesis per day, times more days, could give greater total seasonal photosynthesis.

How would past natural selection have responded to a tradeoff between photosynthesis rate and water-use efficiency? A key point is that an individual plant gets to keep the carbon it takes up via photosynthesis, but the soil water it conserves by using it more slowly may be used by a profligate competitor nearby. Given this tradeoff between individual-plant fitness and the collective efficiency of the plant community, it's not surprising that past natural selection has left some opportunities to improve whole-crop performance.

Here are some other tradeoffs discussed in the paper. Another way to decrease Ci and increase WUE is to crank up photosynthesis, pulling more CO2 through given stomata. The key photosynthetic enzyme, like other proteins, contains nitrogen, so one way to increase WUE is to be sure our crops have enough nitrogen. The paper notes that a crop with limited access to nitrogen could decrease Ci by making a few high-nitrogen leaves rather than many low-nitrogen leaves. At the individual-leaf level, this would increase WUE. But a crop with fewer leaves won't capture as much sunlight for photosynthesis -- another tradeoff.

And it gets worse. Sunlight not intercepted by smaller, high-N leaves will instead hit the soil, increases evaporation of water from the soil surface. The water-use efficiency of water evaporating from the soil is a big, fat zero. So making lots of leaves fast may increase season-long WUE, even if individual-leaf WUE is lower. Dr. Roberts, the plant breeder who developed Drysdale, is working to increase "early vigor", to completely shade the soil surface as soon as possible.

A related tradeoff is apparently a side-effect of the Green Revolution. Shorter wheat plants tend to have higher yield, because they waste fewer resources on stem growth, but they are also less competitive. (This is another example of the tradeoff between individual competitiveness and the collective performance of plant communities.) A side-effect of shorter height at maturity was reduced ability of a young seedling to reach the soil surface when seeds are planted deep. But deep planting can be very useful, particularly when the surface is dry, but there is moisture for growth deeper in the soil. Dr. Richards has found a way to get the best of both worlds: good emergence from deep-planted seeds, yet short stems in mature wheat. So this particular tradeoff wasn't as intractable as those discussed above.

Finally, if you want really big increases in water-use efficiency, the paper suggests a very different approach, namely, increasing Wa, or humidity. This isn't always possible, unfortunately. But they mention one example of farmers in Syria doubling the yield of chickpea by growing it in winter, rather than in spring. This became possible, however, only after plant breeders developed varieties resistant to diseases that previously destroyed winter-grown crops. This is a long way from the naive, tradeoff-ignoring approach of looking for "drought-resistance" genes and increasing their expression!


Many of those drought experiment papers are really poorly designed and executed. I trust the conclusions of very few of them...

You aren't referring, by chance, to the PNAS paper from Monsanto that used 90% confidence for their stats rather than the usual 95% and didn't show yields under well-watered conditions?

I don't have any details on the experimental data in Condon's paper, but their analysis is unusually insightful and clear.

Not to mention the fact that plants with abiotic stress tolerance are also more likely to be favored by natural selection outside of the agroecosystem, where nitrogen and water are also limiting.

I'm curious as to how you can affect Wa (humidity)? That's intriguing but I question how you could achieve it on a large enough scale.

If it is possible, are you aware of any method to achieve the opposite? Interested in removing large quantities of Wa from the environment.

Thanks for considering...

We can't do much to change humidity outdoors (except by evaporating water, which defeats our purpose of growing crops with less water), but sometimes we can grow crops when humidity is high (e.g., earlier in spring).

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Type the characters you see in the picture above.