Method of classification: First, the unit of classification used is the article. Despite the very great variation between articles in length and quality this was the unit chosen because this is the form in which contributions to the Bulletin are made. Secondly, all articles are examined and where possible classified in terms of subject area, drug, geographical focus or setting, and the author's national background. For the purposes of the analysis a category within each of the four groupings is assigned a frequency of 1 if an entire article is devoted to it, a value of 0.5 if it shares an article with one other category, a value of 0.25 if it shares an article with three other categories, and so forth. There was no attempt to cover all articles by all classifications; more articles are classifiable by subject area and author's national background than by drug or geographical setting. So as to treat the range of contents in a set of categories large enough to be meaningful, we omitted infrequent categories in the classification of articles by drug and subject area, whereas the range of geographical setting and author background categories exhaust all the articles that are classifiable along these dimensions. However, articles of unknown authorship were left out of the table.