Sir

Your opinion article (Nature 415, 101; 2002) provides timely cautions about the errors of Science Citation Index (SCI) data compiled by the Institute for Scientific Information (ISI). As Chinese scientific publishers, editors and researchers, we wish to point out that these errors are far more serious for journals published in non-English-speaking countries. The ISI's coverage of scientific journals from these countries is far too limited, and the livelihood of many decent journals has been an unintended casualty.

The ISI misses the fact that non-English-language journals often have alternative names. The Chinese Journal of Geophysics — Chinese Edition is also called Diqiu Wuli Xuebao, Acta Geophysica Sinica and Chinese J. Geophys. for historical reasons. In 2000, this journal was cited more than 260 times, yet the ISI's Journal Citation Reports 2001 gives it a total citation score of 13.

The ISI fails to note that many English-language editions of journals published in China have Chinese-edition counterparts, with different contents and domestic and ISSN registration numbers, and should be considered as different journals. Among China's 12 English-language journals indexed by SCI, 8 also have Chinese editions. In 1998, for example, 91 of the 147 cited items for papers published during 1996 in the Chinese Science Bulletin, a prominent science journal, were incorrect: 52 of the 91 errors were attributions to the bulletin, although these citations were actually to the Chinese edition.

In addition to your reminder to people to use citation statistics prudently, we suggest the ISI should pay more attention to journals in non-English-speaking countries. Even though the ISI should not be held responsible for problems in doing science in developing countries, it can certainly be more accurate in its analysis of scientific achievements in places such as China, and hence help to promote international scientific communication.