Sir

The biodiversity community must learn from its counterparts in the physical and biomedical sciences and move towards the provision of unhindered access to its baseline data: taxonomic descriptions, imagery, geographical and temporal distribution, and characters — molecular, morphological and behavioural (see H. C. J. Godfray's Commentary “Challenges for taxonomy”, Nature 417, 17–20; 2002).

International codes of nomenclature require taxonomic actions to be published, and the data thus made available. Yet much of the underlying information is accessible only by examination of the specimens involved, so access is in effect limited to all but a few potential users. The assertion of copyright by publishers further limits the distribution of published information.

Very few libraries around the world have the financial capacity to carry the full range of literature in which systematic results are published. To take the ants as one example, the 11,000 species were first formally described in approximately 3,800 publications — roughly 100,000 printed pages — in more than 800 serials and monographs.

As F.-T. Krell noted in Correspondence (Nature 415, 957; 2002), the relevance of taxonomic publications remains high for many years. Although funding has been secured to make 80% of the ant pages accessible online within the next two years at http://www.antbase.org (see Nature 416, 115; 2002), many recent papers are not in the public domain because of publishers' copyright restrictions.

To chart even the 1–1.5 million “known” species of the world (E. O. Wilson, Science 289, 2279; 2000) is a daunting task. International initiatives such as the Global Biodiversity Information Facility are needed to help people work effectively towards that goal, as are the development of tools from information technologies and a new cultural approach to ownership and sharing of data.

In the genomics community, authors place all sequence data in a publicly accessible depository. As a result, the data themselves can be peer-reviewed, and new areas of investigation have developed through comparison and collation of data sets.

The biodiversity and conservation communities would greatly benefit from similar provision of open access to character and distributional data. Because of space and cost constraints, many of these data are unpublished. We sorely need a mechanism to provide access to these data, along the lines of GenBank, as well as the cultural imperative to deposit data (see Godfray's Commentary for a proposal to make taxonomy a web-based unitary discipline).

For now, it would be a tremendous benefit if publishers would make published taxonomic papers open-access, so that an equivalent to PubMed can make this important scientific information available to the broadest possible community.