This project began when program managers in government agencies, recognizing that work on mapping and sequencing the genome of Arabidopsis was accelerating, convened an international series of workshops of leading scientists to devise a long-range plan. The continued commitment of these senior scientists to widespread sharing of information and materials, and the peer pressure and aggressive solicitation of stocks of mutant strains to be made available through distribution centers, have contributed to the almost universal sharing of materials in this community. Similarly the strong leadership of the 22 societies that provide oversight for the American Type Culture Collection (ATCC), and the strong scientific leadership and management of The Jackson Laboratories (TJL) are strengths of these successful repositories and distributors of resources. A most remarkable example is presented by the Human Genome Center of the Lawrence Livermore National Laboratory (LLNL), which, by default, has become a major supplier of material resources to the scientific community, without being supported for this function. The extent to which it has provided the leadership and the actual materials that have permitted widespread sharing of genetic materials and information and the forging of important collaborations is remarkable. LLNL has protected the use of this important resource for the research community.

Many of the important institutions in science have an ongoing responsibility to foster a culture of sharing and to continue to advocate for policies that assist the process. Professional societies and journal editors can support sharing of resources by developing appropriate policies guiding publications and responsibilities for making data available after publication. The Journal of Biological Chemistry, for example, has such a policy: ''Authors of papers published in the journal are obligated to honor any reasonable request by qualified investigators for unique propagative materials such as cell lines, hybridomas, and DNA clones that are described in the paper.'' Plans are under way to modify the phraseology to restrict the obligation to investigators who want to use the strain for noncommercial purposes and to include computer programs in the materials that have to be shared. In addition, after considerable debate, the policy was established that authors publishing crystallographic data must submit the details, coordinates, and related data to the Protein Data Bank at Brookhaven before publication. The appropriate accession number must be inserted into the manuscript; in a similar way, nucleotide sequences must be submitted to Genbank or a similar database, and the accession number must be inserted into the manuscript.