To determine the comparative usefulness of the C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) in the assessment of rheumatoid arthritis (RA) activity and to provide tables and nomograms of normative data in RA allowing the linking and interchange of test values. We studied 774 patients with RA seen in the clinic by obtaining complete rheumatologic examinations and laboratory studies. Clinical variables included visual analog scale pain and global severity, joint count, functional disability, depression, and a composite measure of disease activity. In addition, we measured ESR and hemoglobin, and rheumatoid factor (RF), CRP, IgG, IgA, IgM, haptoglobin, alpha 1-antitrypsin, albumin, pre-albumin, and C4 by nephelometry. Median values for CRP were 0.82 mg/dl and ESR 26 mm/h. The average correlation with 7 clinical variables was 0.248 for ESR compared to 0.259 for CRP. But partial correlation analysis showed that a substantial portion of the correlation with ESR is explained by the effect of immunoglobulins, RF, and hemoglobin rather than the acute phase response. Twenty-eight percent of results were discordant between ESR and CRP, and this discordance was explained by the above factors. When discordance occurred, CRP was a better measure of disease activity than ESR. Simple comparisons between ESR and CRP suggest that both tests are similar, but partial correlation analysis indicates that part of the correlation between ESR and clinical variables comes from non-acute phase factors. These factors, in turn, are responsible for most of the discordance between ESR and CRP results. Thus, CRP appears to be the better test regarding measurement of the acute phase. Because ESR is sensitive to immunoglobulins and RF, it may measure general severity better than CRP, even though it is a poorer measure of inflammation. This perhaps accounts for the relative equivalence of the tests. The combination of ESR and CRP yields useful information that is often not apparent when only a single test is used.