Gravimetric determinations were recorded for the total and several individual proteins (in the sodium oxalate plasma) fractioned with ammonium sulphate and sodium chloride. At precipitation, the plasma salt mixture had been diluted to a final volume of ten times the amount of plasma employed. Coagulations were on aliquot portions of filtrates, and the individual protein constituents (except serumalbumin) were calculated by difference. The eleven horses had been subjected to simultaneous immunization against diphtheria and tetanus toxins, each horse being subsequently continued on the toxin to which it responded best. Test bleedings of about 500 cubic centimeters only were made until maximum antitoxic potency (with almost coincident greatest variation See PDF for Structure in the contents of the several proteins) had been attained; routine bleedings of four to ten liters for antitoxin production were then instituted. The two refractory, one medium and the eight horses yielding a highly potent antitoxic plasma, all showed an increase of from 40 to 114 per cent. in the total serumglobulin. For the refractory animals, this increase was 59.3 and 87.9 per cents. In one refractory and in one high horse, the serumglobulin maximum preceded the highest concentration in antitoxin. In seven of the horses, the greatest increase in the total serumglobulin was coincident with the maximum antitoxic potency. The serumglobulin increase, however, tended relatively to precede that of the antitoxin. In the two other horses, both maxima also were observed together; no "intermediate" bleedings were made in these two instances. The greatest absolute increase in the serumglobulin was observed in the most potent plasma obtained in the series; the second place, however, went to a refractory horse. Subsequent to the maxima, the serumglobulin content was maintained at high concentration, in spite of repeated bleedings; it then only roughly paralleled the antitoxic variations in the plasma of the individual animals. At dilutions of the plasma in the precipitated mixtures of 1:15, 1:5 and 1:10, the ammonium sulphate "euglobulin" fraction amounted to about 60 to 70, 20 to 24 and 10 to 15 per cents., respectively, of the total serumglobulin in both the normal and the antitoxic plasma. In an 850 unit plasma, an increase in the "euglobulin" over the normal percentage was observed, but the high content of protein in this plasma probably influenced the precipitation limits. The influence of the protein concentration is indicated by the different percentages for the "euglobulin" obtained for the three dilutions of 1:15, 1:5 and 1:10. The "euglobulin" then was not increased to a greater extent than the "pseudoglobulin" as the result of immunization, as has at times been maintained. See PDF for Structure The "euglobulin" precipitated by saturating the plasma with sodium chloride (at ultimate dilution of the plasma 1:10) was much greater in normal plasma than the ammonium sulphate "euglobulin" at the same dilution. There was a tendency in early immunization for this sodium chloride "euglobulin" to increase along with the total serumglobulin; it rapidly diminished, however, until at the height of immunization and maximum serumglobulin concentration, it may have reached less than half the normal absolute amount. The serum albumin was diminished a third to a half the normal along with the serumglobulin increase. Subsequent to the anti- See PDF for Structure toxic and serumglobulin maxima, figures as low as a fifth of the original serumalbumin content have been noted. It is suggested that this diminution of the sodium chloride "euglobulin" and the serumalbumin is a physiological compensation for the greater viscosity of the plasma because of the increase in the more soluble serumglobulins. No characteristic alteration in the fibrinogen of the plasma was observed during immunization. Individual variations up to 0.5 gram per 100 cubic centimeters of plasma have been recorded. The influence of repeated bleedings does not essentially influence the protein changes induced by immunization. These remarkable regenerative processes are worthy of note. The results of our investigation indicate that in "forced" immunization, the same characteristic quantitative changes can occur in the blood proteins of both refractory horses and those yielding a highly potent antitoxic plasma. We cannot conclude, however, that the serumglobulin increase does not represent an accumulation of antitoxin, at least in part; it is possible that other antibodies may be formed either prior to or along with the specific antitoxin and that these may constitute a portion of the increase in the more soluble serumglobulin with which protein such substances are associated.