Over much of the world, the growing season of 2050 will probably be warmer than the hottest of recent years, with more variable rainfall. If we continue to grow the same crops in the same way, climate change will contribute to yield declines in many places. With potentially less food to feed more people, we have no choice but to adapt agriculture to the new conditions.

To some extent, adaptation can be done by moving crops to more favourable areas and by agronomic tweaks. But that will almost certainly not be enough. We will have to give crops a genetic helping hand, infusing them with new genes to allow them to better cope with new climates, and the new pests and diseases they will bring. Where are these genes going to come from?

Some of them could come from completely unrelated organisms, to be spliced into their new genomic homes using advanced biotechnologies. However, there is significant public resistance to that strategy, and it is still unclear how effective genetically modified crops are at coping with heat and drought. We cannot risk putting all our eggs in that basket.

Another source of genes for crop improvement are traditional heirloom varieties, often called landraces, which are still grown by subsistence farmers in many parts of the world, although they are fast disappearing. Large collections of their seeds have been made over the years, creating genebanks that are scoured by plant breeders searching for crop diversity, and which helped spur the Green Revolution in agriculture from the late 1960s.

But there’s a limit to the diversity found in domesticated species, imposed by domestication itself. Cultivated species usually contain a fraction of the genetic diversity found in their closest wild relatives — a legacy of the ‘domestication bottleneck’. Ancient farmers selected relatively few plants from the progenitors of modern crops, in a limited number of places. Although there has been continuous gene flow between crops and their wild relatives where they coexist, a lot of genetic diversity has been lost as agriculture has developed.

We know that the ‘lost’ genetic diversity includes genes for resistance to high temperatures and drought, and to pests and diseases, as well as taste and nutritional composition, and even yield. If there was ever a time to go back and reclaim this diversity, that time is now. In fact, it is already being used more than many people realize. For instance, there is probably no widely grown rice cultivar that does not have some genes obtained by breeders from its wild relatives. But we could be making much more effective, and systematic, use of the reservoir of diversity our Neolithic ancestors left behind.

To read the full commentary, click here.