Juvenile chronic arthritis (JCA) is a heterogeneous disease, differing in most instances from adult rheumatoid arthritis. The diagnostic criteria generally used in Europe are those proposed in Oslo in 1977. There are three main types of onset: systemic, polyarticular and pauciartricular. However, when we observe patients during the course of the disease, this proposed classification seems too simple and does not reflect the reality of outcome. To be more efficient for therapeutic studies, we think that it should be more accurate to consider the type of course, rather than the type of onset. Thus, we will discuss in a first part of this paper subdivisions of the disease according to course, biology and genetics, and indicate in which circumstances a slow acting drug such as pencillamine should be a realistic therapeutic indication. In the second part of this paper, we will give the results of already published observations and of personal data on the effect and tolerance of D-penicillamine, and discuss a way of improving drug evaluation in JCA.