In your News story “Doomsday food store takes pole position” (Nature 441, 912–913; 200610.1038/441912b) you report the establishment of a “doomsday seed bank from which the genetic riches of Earth's food crops could, if necessary, be reconstituted”. Unfortunately, this statement is wishful thinking.

As managers of one of the largest ex situ gene banks in the world, we have been surprised during the past few weeks by the increased number of comments by visitors about the possibility of seeds being kept viable for millennia if stored at sufficiently low temperatures. There are very few reported cases in which seeds were able to germinate after many decades or even more than a century of storage: most cannot survive long-term storage.

One prominent example is provided by a herbarium sample of the legume Albizzia julibrissin, which had been collected in China in 1793 and was deposited in the British Museum. The material was accidentally wetted in 1940 during firefighting, resulting in the germination of several seeds at least 147 years old (Nature 149, 658; 1942). There are a few other documented examples where seeds have been able to germinate after 100 years (details available on request from the authors), but these are rare exceptions. The suggestion that seeds can survive after thousands of years is wrong, and detrimental to conservation funding efforts.

Undoubtedly, it is important to keep safety duplicates for the estimated 6 million seed samples that are kept in ex situ gene banks worldwide. To this end, the food store featured in your News story is an important contribution. But the vast majority of plant seeds cannot be stored for more than 40 years without losing germination vigour, however low the temperature and however thick the concrete walls of the vault. Hence, gene banks have no other choice but to preserve their material by growing plants in greenhouses, field plots or in laboratories at regular intervals. This process requires manpower, knowledge and rigorous quality management. It is also the bottleneck in almost all the gene banks so far established: the seeds they contain urgently require regeneration, otherwise they will die.