Rescuing endangered primary data is important (Nature 468, 17; 2010). Even more crucial is to archive it at the time it is generated. This lesson has been learned by the radiobiology community, who took on that challenge 25 years ago.

The past 60 years have seen many large-scale studies of radiation effects on animals. The scale, cost (estimated at US$2 billion today) and ethical aspects of these experiments make them unlikely to be repeated. However, the data could usefully be reanalysed in the light of new paradigms in radiobiology.

These endangered data were deposited between 1985 and 1999 in the International Radiobiology Archives (IRA): they include results collected during 1960–98 on more than 460,000 animals in Europe, the United States and Japan. The IRA is now integrated into a legacy database, the European Radiobiology Archive (ERA), a project funded by the European Commission (see http://go.nature.com/evbdFv).

Development of the ERA taught us that integrating legacy data into a standard format is difficult. But the data must be searchable and usable to prevent the database becoming an information graveyard.

Sustainability is also a problem. It is hard to find agencies and institutions that are prepared to fund long-term archives. Fortunately, Germany's Federal Office for Radiation Protection has committed to maintaining the ERA database — an important asset for the radiobiology community.

The European Commission's STORE project will act as a data warehouse and radiobiological resource directory. Its importance is recognized by the Multidisciplinary European Low-Dose Initiative (http://www.melodi-online.eu), which coordinates policy on low-dose radiation research.