Sir

Electronic access to natural history collections data will enable biodiversity to be mapped with other parameters, improving science and environmental management. At least 10 major initiatives desire networked access to these data1. Studies of birds, important in environmental issues, may spearhead the development of networked databases2. But discussions at this year's meeting of the American Ornithologists' Union in August offered a sobering assessment. Widespread participation by curatorial staff and researchers is required for a networked system to succeed, and several issues prevent this.

Perhaps 50% of specimens in North America have yet to be computerized, and proofreading lags far behind. Quality is an issue when releasing error-ridden data sets to point-and-click users who have little familiarity with the biases and errors in collections. Electronic users are unlikely to check data against specimens, the primary data source. Invoking ‘downstream’ quality control is not comforting, because peer review is a poor safety net rarely applied to reports and management plans.

Systematics collections are underfunded3, and the community disapproves of mining collections data while this is the case. Most data-sharing initiatives1 overlook this. The value of the data is recognized, so the enterprise producing them must be supported. The resource is also dated, on average reflecting environmental conditions of the 1920s4. Recent specimens and their data comprise the ‘active zone’, representing continuing activities to fill knowledge gaps in a changing world. Intellectual-property rights and institutional investments are concentrated here, as are the data most relevant to environmental management efforts. These data will not be widely accessible until bought and paid for. The solution is to renew support (in funding and permits) for fieldwork.

A two-step process towards networked databases may be acceptable. First, old data would be made available. Second, increased support would purchase access to the critical ‘active zone’. Another viewpoint wants this support up front: commissioned biodiversity studies or subscriptions would generate new data and leverage access to the old.

The Freedom of Information Act causes concern in the United States5 because these databases are usually both public and private. But this act does uphold core values6. Data filters can protect vulnerable populations, intellectual-property rights, and rights to privacy.

The collections community is noted for sharing data, but leaving the tap on unminded may be an abrogation of responsibility. Progress on these issues, producing a model eliciting enthusiastic participation, is needed to fully access the wealth of data in natural history collections.