Sir

As David King reveals in his Feature “The scientific impact of nations” (Nature 430, 311–316; 200410.1038/430311a), Brazil compares unfavourably with most other countries. It produces a meagre 1.2 % of the world's publications, despite having roughly 3% of the world's population and income. If Brazil is to have any chance of improving its status, the current trend of decreasing financial support for science graduates should be reversed, as a matter of urgency.

During the past few decades, many more people have taken part in Brazil's graduate programmes. Brazil has a formal system for evaluating these courses, and despite many flaws (as noted by L. de Meis, M. S. do Carmo & C. de Meis, Nature 424, 723; 200310.1038/424723b), this has allowed some progress in science and technology.

Such progress has become evident in the increasing number of contributions made to international science by Brazilian scientists — up from 48,800 authors appearing in publications indexed by the ISI in 1998 to 58,000 in 2002. A closer analysis reveals that this increase was due exclusively to the work of postgraduate authors (who increased from 11,300 to 21,800 during that period), while the participation of career investigators has decreased. Thus, the significance of postgraduate students in the production of scientific knowledge is unquestionable.

However, in a country where science funding is mainly public, the resources allocated by the Brazilian government to science and technology have decreased over this same time period. Between 1998 and 2002 the National Council for Science and Technology Development (CNPq), a federal agency, cut the amount of money allocated for postgraduate fellowships by 50%. This loss of funding is even more drastic when you consider that the number of postgraduate students increased by 36% during the same period. Sadly, this downward trend has also been followed by other Brazilian agencies, such as CAPES (which funds graduate programmes in Brazil and Brazilian postgraduates studying abroad) and FAPESP, which mainly supports innovation in science and technology.

Inflation has further eroded the value of the remaining fellowships; in 1994, a PhD fellowship provided the equivalent of US$1,000 a month, whereas today it is worth a mere US$350.

Given these numbers, it is not far-fetched to say that these students are being exploited. How long can even the most motivated student be expected to survive on just US$350 a month?

All of these factors are seriously jeopardizing Brazilian science. Surprisingly and contradictorily, the Brazilian government has historically emphasized the importance of science in the development of the country. It is time to convert those words into actions.