Juvenile arthritis implies an onset of disease under 16 years with arthritis persisting in one or more joints for at least six weeks, and with the active exclusion of well defined illnesses, such as systemic lupus erythematosus. Prognosis implies the ability to predict outcome. Its accuracy depends on many factors with early recognition and appropriate care being important. However, response to treatment may be variable. In general, those with involvement of a few joints do better than those with systemic disease or seropositive juvenile rheumatoid arthritis both with regard to persistence of disease activity and complications. These include not just joint deformities, but osteoporosis, amyloidosis, alterations in growth with overall failure and local anomalies, chronic iridocyclitis and psychosocial problems. More aggressive therapy was only introduced in the 1990's, so it is important that multicentre studies are properly assessed in the context of the suggested International diagnostic criteria. One hundred years ago, George Fredric Still drew attention to the systemic form of the disease as distinct from pure polyarthritis [1], but it was only in the 1970s, as follow-up proceeded, that the separate identity of variants became clinically evident [2]. At the Park City meeting [3] and at the EULAR meeting in 1977 [4] when three subgroups (notably systemic, polyarthritis and pauci-articular onset) were defined, that subclassification became regularly used. However, since there were no absolute diagnostic tests there had to be exclusions. At that time the most common medications were aspirin and corticosteroids, although a few patients received gold or penicillamine. In their large group Wallace and Levinson (1990) [5] found that at the 10 year follow-up between 31% and 55% still had active disease. Girls appeared to have a five-fold greater risk for persistent activity than boys; disease duration was probably the most important factor influencing disease activity at follow-up as suggested previously [6]. It was not until the 1990's that the more aggressive therapy in the form of methotrexate--which Giannini had shown to be effective when given in appropriate dosage [7]--and sulphasalazine [8] and the long acting local corticosteroid triamcinolone hexatonide became regularly employed [9, 10]. At the ILAR Meeting in 1993 an international task force was set up under the chairmanship of Dr. C. Fink [11] to develop a classification for the idiopathic arthritides in children, defining childhood as up to 16 years of age. Active exclusion of well-recognised disorders such as rheumatic fever or systemic lupus erythematosus, still had to be made. The first proposed types, which are mutually exclusive, are shown in Table 1. A more recent meeting in Durban under the chairmanship of Dr. R. Petty is yet to be published, but considerable advances have been made, particularly in the definition of subgroups.