Silvia Matesanz and Ruben Milla have just published a very interesting paper comparing seven crops to their wild relatives using various measures of plasticity, specifically, their response to availability of nutrients or water.

Their figure 2a, showing treatment effects on vegetative (preflowering) plant weight, averaged across species, may stimulate the most discussion.  How would you expect the biomass of a crop to compare to its wild ancestor, given plenty of nutrients and water?  How about with less nutrients or less water?

I’m not sure what I would have predicted under the high-resource condition.  Because natural selection had millions of years to improve the efficiency of processes like photosynthesis in the wild ancestors, I have argued that further improvements will usually require accepting tradeoffs that were rejected by natural selection.  For example, genetic improvement has greatly increased production of seeds or fruit, but often at the expense of stems or leaves.  Sacrificing stem weight to increase grain weight was key to doubling the grain yield of rice and wheat during the Green Revolution.  So I would not necessarily have expected preflowering mass of high-resource plants to be greater for crops than for their ancestors.  But that’s what Matsanz and Milla found.

“When can humans find solutions beyond the reach of natural selection?”  That was the subtitle of our 1983 paper on Darwinian Agriculture.  One thing natural selection won’t do is to sacrifice individual-plant competitiveness for the collective benefit of groups of plants, even though it’s collective performance that determines crop yield.  Jacob Weiner and colleagues recently showed that highly-competitive wheat genotypes yield less (when grown together) than moderately-competitive ones, consistent with what Colin Donald predicted in 1968.  But tradeoffs between individual-plant competitiveness and whole-crop yield can’t explain Matesanz’s and Milla’s results, because they used individual plants in pots.

Of course, past natural selection adapted plants to past conditions.  We wouldn’t expect plants whose ancestors evolved with < 300 ppm CO2 to take full advantage of today’s 400 ppm CO2, for example.  And, until recently, nutrients and water were often more limiting than they are with modern agriculture.    Still, even before agriculture, plants sometimes got plenty of water from rain and plenty of nutrients from wild-animal manure.  Even if this only happened one year in five, I would have thought that would be enough for natural selection to retain the ability to respond to favorable conditions — unless the ability to respond to favorable conditions came with a serious tradeoff, such as the ability to grow well under low-resource conditions.

But Matesanz’s and Milla’s Figure 2a seems inconsistent with such tradeoffs.  Under low-water conditions, crops grew as much as their wild relatives, on average.  And under low-nutrient conditions (maybe not low enough?) the crops still outgrew their wild relatives.

If these results are confirmed, it appears that plant breeders have improved individual-plant growth, not just cooperation among plants, and not just allocation to fruit or grain, without obvious tradeoffs between low- and high-resource conditions.  And, as none of these crops have genetically-engineered traits for increased growth (replacing C3 photosynthesis with C4, for example), they did this through essentially the same process (nonrandom selection from among random variants) that shaped our crops’ wild ancestors.  The best hypothesis I can come up with to explain these results is that high-resource conditions were so rare, over most of the evolutionary history of our crops, that even minor tradeoffs between high- and low-resource conditions (not detected in this study) were enough to favor genotypes whose ability to respond to high-resource conditions was less than we see in modern crops.

I will be very interested in seeing follow-up work (commentary and experiments) related to this paper.