My bookcase holds many contemporary books on “style”—or to use the Aristotelian term, lexis—in written communication. They are largely concentrated on such matters as clarity, conciseness, and cons...

This article proposes a taxonomy of scientific titles: those staking claims; those setting problems; and those conveying themes. A close analysis of the deep structure of these titles suggests that their goal is the maximization of information content within a short compass, a compression that permits their easy retrieval in computerized searches. Placing these titles into the context provided by Gross, Harmon, and Reidy's Communicating Science suggests further that titles evolved to this point by adapting to changes in systems of information retrieval.

To gain a better sense of the metaphorical nature of the scientific research paper, the author reviewed 89 journal articles taken from the top 400 most‐cited documents in the Science Citation Index database for the period 1945–1988. Metaphorical constructions were found in a variety of forms: conceptual models, experimental designs, technical analogies, standard technical names, conventional figurative expressions, and even original figurative language normally associated with more‐literary writing. Examples are given for each mode of metaphor.

This article discusses the typical form and content of forty theoretical scientific papers. These papers were chosen from the 400 most-cited papers in the Science Citation Index for the period 1945–1988 (reported by Eugene Garfield in a series of recent essays appearing in Current Contents). It was found that the typical form for these papers is similar to that for experimental and methods papers, but the content differs substantially. In brief, the content follows the logical sequence: problem or need, assumptions made in attempting to solve problem or meet need, theorem derived from those assumptions and additional considerations, proof of theorem by logical reasoning or validation by comparison with what is established or establishable, conclusions from previous discussion, and recommendations on future experimental or theoretical work. Also, compared with experimental and methods papers, these theoretical papers have somewhat fewer figures and tables, but many more references and equations.

This article contains results from a literary analysis of fifty scientific papers selected from the top 100 most-cited papers appearing in the Science Citation Index for the period 1945–1988. Most papers are from the field of biochemistry and became citation superstars because their authors discovered a method or material that numerous others could use in their own research. The typical paper has two authors, two tables, six figures, and twenty-two references. It adheres to the conventional topical organization, with the topics distributed as follows: 2 percent abstract, 5 percent introduction, 25 percent methods and materials, 50 percent results, 10 percent discussion, 4 percent conclusion, and 4 percent reference list. Tables and figures occupy about 30 percent of the article. With respect to the writing style, the average sentence is somewhat long (24 words) but not unreasonably so, and the sentence structure is simple greater than half the time. Moreover, sentences tend to rely heavily on to be verbs (about 80% of sentences have at least one) and abstract nouns (0.66 per sentence). Explanations for the typical form and writing style in these papers are provided.

Technical periodicals and proceedings have been important instruments for transmitting news about scientific and technological discoveries for more than 300 years. The first such publications appeared amidst the birth of modem science when, for the first time, emphasis was placed on experiment as the basis for advancement of knowledge. Discussed in this article are the origin of the technical periodical and proceedings and their characteristics up to the end of the 18th century and the analysis of the writing style in an important technical paper written by Isaac Newton and published in 1672.

In contrast to the literary artist we expect the scientist-writer to transmit information to the intended audience as accurately and clearly as possible. Nevertheless, a few scientists have managed to slip into their prose such rhetorical devices as anagram, acrostic, pun, metaphor, litotes, and neologism.