LETTERSpoints require special attention when evaluating the role of these cells in RA: whether they accumulate in rheumatoid lesions and whether they express a specific family of y/S receptor. Considering the first point, available data again show conflicting results. Kjeldsen-Kragh et a1 (5). Wood et al (6), and Chaouni et a1 (7) found similar proportions of y/S T cells in SF and PB of RA patients. In contrast, our results (1) and those of Rkme et al (3) and Smith et a1 (4) show higher numbers of y/S T cells in the synovial compartment than in PB of RA patients. Reasons for these disparities remain obscure, but it seems obvious that further investigations of larger numbers of RA patients in the early stages of disease, as well as of comparison groups of patients with other chronic inflammatory arthropathies, are needed.As far as the expression of y/S T cell receptor subpopulations is concerned, most studies show that ylS T cells from the rheumatoid synovial compartment do not express V,9 gene products, which contrasts with $8 T cells from the PB (1,4,6,8). Furthermore, results of both phenotype and messenger RNA analysis suggest a preferential expression of V,1 gene products (1,4,6,8). We agree with Kjeldsen-Kragh et a1 that the phenotype sTCS1-, BB3-,TiyA-could be explained by several combinations of rearrangements between S gene segments and y gene seg-