The literature output over one year, 1990, of 22 International Agricultural Research Centers (IARCs), including 16 Consultative Group on International Agricultural Research (CGIAR) centers, was examined. Total output of the IARCs was 1,694 items, of which on average 42% were primary (refereed) literature; 24% were reports and monographs; 18% proceedings papers; 8% book chapters; and 8% semitechnical/popular literature. Total literature production from the IARCs is similar in magnitude to that of FAG, There were 1,230 internationally recruited scientists in the IARCs, with an average annual productivity of 1.38 items per scientist, including 0.58 primary literature articles. There was no correlation between scientific productivity and numbers of scientists in a center. However, there was a significant positive correlation between scientific productivity and center budget, indicating higher efficiency in the larger centers. In view of the nature of IARCs' literature output, we argue that IARCs should reject the trend for scientists to be assessed only by citations in ''core'' primary literature; and that IARCs should set up an international standard, perhaps based on the present proportionality of types of their literature output in order to assess IARC individual scientists and the ''health'' of their institutional output.
