Sixty eight cases of Sydenham's chorea (SC) were studied with the purpose of characterizing biologically the choreic individual by means of some laboratorial data. Based on antecedents, on the presence of recent infectious disease, on clinical examination, on electrocardiographs and x-rays of the heart, and according to a modified Jones criteria the patients were initially divided in three groups: a)Group 1 — 30 patients (case 1 to 30) which presented SC associated with active rheumatic fever (RF); b) Group 2 — 20 patients (cases 31 to 50) which presented SC associated with a previou or present infectious state without active RF; c) Group 3 — 18 patients (cases 51 to 68) which presented "pure" SC, not having anything in their antecedents, or in present history, nor in their physical examinations that could justify calling them "rheumatic" or "infectious". However the analysis of the clinical data, by means of the homogenization tests (qui square or the exact Fisher test) and Goodman's contrast test showed the artificiality of this grouping, which could not be longer sustained. From the 68 cases studied the average age group was 9.9 years, with the maximum age being 17 years, and the minimum age being 4.5 years. 47 of the cases were females as compared to 21 males (2.2 to 1); 60 patients were white, 7 dark-skinned and one negro. The average evolution time of the choreic syndrome, at the time of the first consultation, was 6 months and 7 days, with a minimum of 13 days and a maximum of 60 months. The incidence of the outbreak as far as the season of the year is concerned, was as follows: 31 cases between autumn and winter; 14 cases in spring and 22 in summer. The following laboratory examinations have been made: a)"classical acute phase serum reagents" (APSR): sedimentation rate, differential blood count, Weltmann reaction, mucoproteins, C reactive protein, antistreptolysin-O titter, electrophoresis of serum proteins; b) copper and ceruloplasmin; c)magnesium; d)sulphur; e)electrophoresis of cerebrospinal fluid (CSF) proteins; f) electroencephalograms (EEG). For each of the three initial groups, considering the laboratory examinations carried out, the correlations between the APSRs were studied and compared with the electroencephalograms, in accordance with partial correlation tests. The analysis of association tables as made in accordance with the exact Fisher method, the McNemar method and also by using the Goodman method. The real average values were estimated from the APSR results and the electroencephalograms, by points and intervals. The significance levels for these tests were 5% and the confidence intervals (CI) referred to for the percentages obtained for each APSR and for the electroncephalograms were estimated using tables with a confidence of 95%. The APSRs that showed the highest positive indexes in the confidence intervale, independently of any correlation, were: a)antistreptolysin-O (n = 43) was high in 60.46% (confidence interval (CI) = 45.86 to 75.06%)), when the normal limit was considered up to 250 units; considering the normal limit of up to 333 units, therefore considering high titers starting at 400 units, the percentage of positivity fell to about half of the previous value (37.2%); b) copper (n = 39) appeared as the most altered APSR in our casuistics, with 94.97% (CI = 81.0 to 98.0%) of high values; the ceruloplasmin (n = 40) kept pace with copper and was increased in 90.0% (CI = 80.7 to 99.3%; c) the serum proteinogram (n = 24) showed patterns similar to those of the RF, that is, decrease in albumin in 66.66% and increase in alpha-2-globulin fraction in 91.60% (CI = 30.00 to 70.00%) of the cases; a simultaneous alteration of these two fractions observed in 11 cases (45.83%, CI = 25.9 to 65.76%); gamma-globulin, opposing the initial acute rheumatic patterns showed itself to be quite high 83.33% (CI = 68.43 to 98.23%); d)magnesium (n = 38) decreased in 86.84% (CI = 76.09 to 97.59%), increased in 4 cases (10.52%, CI = 2.5 to 25.0%) and was normal in one case (2.64%,); e)sulphur (like SO4 ion) was increased in 71.79% of the cases (CI = 57.68 to 85.90%); f) the cerebrospinal fluid's proteinogram (n = 13) showed an increase of prealbumin in 6 cases (46.15%); g)the EEG, during the acute phase (n = 42) showed anormality in 45.24% of the cases (CI = 30.19 to 60.29%); part of these anormalities persisted up to the phase of remission, when from 34 EEG only 29.41% (CI = 14.11 to 44.71%) showed anormalities (10 cases); among these, two were "disorganization of the alpha rhythm", one of which normalized rapidly, another showed only asymmetry and may be considered normal; there remained then only 7 cases in the remission phase, with persistent anormalities, which however, are discussible as to their pathologic significance in the ones labeled as "slow for the age" and "paroxystic dysrhytmia by slow waves" in the occipital areas, in view of the age group to which they belong (4 to 11 years). Only two cases, in the remission phase, had anormalities in the temporal and occipital areas: case 29, with suggestion of "paroxystic dysrhythmia in the left occipital region, by acute atypical waves", and case 33, with "paroxystic dysrhythmia in the left-temporal region, by hypersynchronic waves"; this one was the only case that required anticonvulsive treatment. In concluding, the attempt to characterize Sydenham's chorea by means of some laboratory serum examinations showed that: 1) The classical "acute phase serum reagents" (sedimentation rate, blood count, Weltmann reaction, C reactive protein and mucoproteins) were not conclusive for this purpose; 2) the "acute phase serum reagents" that presented the greatest alterations in Sydenham's chorea were copper and ceruloplasmin, electrophoresis of serum proteins (alpha-2 and gamma globulin fractions) and antistreptolysin-O; 3) the "acute phase serum reagents" when evaluated separately were of little value, but it is quite possible that two or more taken altogether could, in certain instances of the evolution of the disease, characterize "an infectious and/or inflammatory stage"; 4) eosinophilia was observed in a high percentage of cases (81.66%) requiring further studies for confirmation purposes; 5) the correlation of the various "acute phase serum reagents" studied among them and in comparison with other determined parameter (magnesium, sulphur, and EEGs) did not show statistical significance, not only when comparing the three initial groups, but also in its totality, facts that may suggest a unitarian pathogeny for Sydenham's chorea; 6) a high percentage of hypomagnesemia was found (86.84%) although at not too low levels; 7) sulphur was high in 71.79% of the cases. The EEG is not an element allowing distinction between a "pure" and a "rheumatic" chorea, and its valorization may not be based on a unique record, but on evolutive records, because only with a prolonged follow-up is it possible to be certain of the persistence or not of the encephalographic abnormalities.