And the key to that improvement is more efficient roots?

It's not surprising. If you look at plants closely, it's the root systems that account for the variation in nutrient uptake.

People have known that roots are important for a long time. But they haven't been studied much. There was no need -- historically, the United States has had fertile soils, plenty of land, and the wealth to use inputs. Root efficiency wasn't a concern.

To just say that you want more roots -- that's the wrong answer. Roots use up carbon and other resources. By having more roots, you have less yield. What you want somehow is a root system that's doing just the right thing at the right time in the right place -- but what is that exactly?

How do you approach that problem?

At first the complexity was daunting. It was like spaghetti. You pull up roots out of the soil, and then you have thousands of these little things. How do you make sense of it? What aspects of what you're looking at are really important? But over time, by a combination of field trials and computer modeling, we have identified specific traits that are associated with soil-resource acquisition.

Can you give some examples?

We've been looking for a long time at the common bean, the most important food legume on Earth. The major constraint to higher bean yields is low phosphorus. And we have found root traits associated with performance in low-phosphorus environments. Shallow roots, for example, are better than deep ones for exploring topsoil. Longer root hairs are better for phosphorus uptake.

In the case of corn, which is the most important cereal, the key nutrient is nitrogen, which washes quickly through the soil. So deep roots are an advantage.

In retrospect, with many of these traits you think, "Why didn't I see this before?" This is basically stuff that is easy to spot, once you know what to look for. That's one reason why these traits are useful for breeding. Because in Africa, breeders may not have sophisticated tools to look at molecular markers. If they can use a shovel, dig up a root system and notice it has certain traits, that's something they can use. It's what we call “shovelomics.” And that may be more important than genomics in promoting food security in poor nations.

What needs to happen next?

Breeding for yield at low fertility has been successful where attempted. We're working with breeders around the world. The bottleneck right now is phenomics -- identifying those root traits that are important and then, what's more difficult, understanding how they interact with one another. There are very few people doing this kind of work. It's amazing to realize at this point how little we still know.

Jonathan P. Lynch, is professor of plant nutrition in the College of Agricultural Sciences, jpl4@psu.edu. In 2010, in collaboration with the Howard G. Buffett Foundation, he established the Ukulima Root Biology Center at the Foundation’s farm in the Republic of South Africa. To learn more, see http://roots.psu.edu/ukulima.