Secondary sources in the sciences are often referred to as REVIEW articles. Review articles summarize, interpret, or analyze the findings of a scientist or group of scientists studying the same thing in an attempt to summarize the current state of knowledge on a topic. They may discuss the ways in which various researchers' work is related or consider the implications of the research, and pose broad or abstract questions about a field of study. They look more like traditional journal articles, not necessarily laid out in a linear way. These are the types of articles you might find in Discover magazine, although you will certainly find them in higher-level scholarly journals as well.
You should always consult the primary article under discussion when using secondary sources as support.
Here is an example of a secondary scientific article.
The title and abstract provide clues as to the nature of the article:
Williams VL. PERSPECTIVES IN BIOLOGY AND MEDICINE. Autumn 2004.Vol.47, Iss. 4; pg. 552.
Document type: Historical Article
Source type: Periodical
Abstract: The promise of molecular medicine is the prevention and treatment of illness. Understanding the mechanism of the disease should allow one to "fix" it. For sickle cell anemia, however, knowledge of the biochemical basis of the disease was only partly responsible for finding a means of treating the disease--of equal value were hypotheses and conclusions generated from clinical observations. This article describes the research path that led to the first effective treatment for sickle cell anemia, hydroxy-urea. Rather than exemplifying the "bench-to-bedside" model commonly used to describe the process of therapeutic innovation, this history of this research reveals that the critical advances for the development of treatment came not from basic research, but instead from clinical and patient-oriented research. Given that the linear approach is the prevailing paradigm of therapeutic innovation, this history is important because it indicates the inadequacy of this approach for a relatively straightforward single-gene mutation disease such as sickle cell anemia and suggests the need for multiple models of innovation for more complex diseases. Thus, this article questions the expectations of molecular medicine and the dominance of a linear model of therapeutic innovation, which often excludes or subordinates other models of developing treatments.