THE DISTRIBUTION OF THE IMMUNE BODIES OCCURRING IN ANTIPNEUMOCOCCUS SERUM.* BY OSWALD T. AVERY, M.D. (From the Hospital of The Rockefeller Institute for Medical Research.) By .a systematic study of the specific antigenic properties of the pneumococcus, Dochez and Gillespie1 have demonstrated four dis- tinct types of this organism occurring in -disease. Such a biologic classification of pneumococci affords the only rational basis for the clinical application of serum therapy in lobar pneumonia. The immunologic and clinical studies of Cole2 have shown that the pro- tective and curative value of antipneumococcus serum is dependent upon this group specificity, that a serum to be efficacious must be one produced in response to an organism of the same type as that causing the infection. The treatment by specific antisera of lobar pneumonia of pneumococcus.origin is at present possible only in infections due to organisms of types I and II. Immune serum pro- duced by the third group, Ptwmococcus miccosus, fails to confer passive immunity, and the heterogeneous nature of organisms of the fourth variety demands a specific serum for each individual strain,. making serum therapy in infections of this type impracticable. However, since approximately 70 per cent. of the cases of this dis- ease is due to organisms of the first two groups, the serum treatment of pneumonia is applicable in the majority of instances. Experi- mental evidence and clinical experience have `demonstrated Ithe necessity of administering relatively large doses of the appropriate serum in combating these infections. The present study was un- dertaken, therefore, with the hope of determining a method of con- centrating and purifying antipneumococcus serum, by which its anti- bacterial potency might be conserved with a minimum of foreign protein. It seems reasonable ta assume that such a process, by ' concentration of antibody content, might enhance the efficacy of ,.-, ._. *Received for publication, November 27, 1914. ' Dochez, A. R, and Gillespie, L. J., Jour. Am. Med., Amt., 19x3, lxi, 727. * Cole, R., Arch. Int. Med., IgIh'xiv, 56. .-- :' 133 134 immune Bo&s Occurring in Antipneumococcus kwum. serum treatment, and, by the elimination of certain irrelevant pro- tein constituents, lessen the incidence, or at least ameliorate the symptoms of serum disease. ,- .I Although much is known concerning the immunological reactions of various immune bodies, comparatively little is understood of their true chemical nature. Whether they are themselves protein in . character, or merely associated in some obscure chemical combina- tion with the globulin or albumin of immune sera is not definitely known. The fact that they may be thrown out of solution by cer- tain protein precipitants has been utilized in the concentration of antitoxic sera. Most investigators now agree that antibody pre- cipitation by chemical agents is not merely a mechanical or adsorp- tion phenomenon, but that these immune substances partake of the nature of the protein with which they are precipitated. In the present study it was first necessary, therefore, to determine with which fraction of the serum protein the pneumococcus immune bodies are associated. The serum used in these experiments was obtained by the intensive immunization of horses to virulent cultures of pneumococcus of types I and II. The various protein fractions isolated were tested for antibacterial action against virulent cul- tures of pneumococcus by protection experiments on white mice, and the results compared with the similar power of the whole serum. This method is less accurate than the more exact titration of anti- toxic sera. The potency of antipneumococcus serum can be meas- ured only by its abilijty to protect highly susceptible animals against actual infection, and the comparative antibody content of any given serum fraction can be interpreted only in terms of the death or survival of such animals. This reaction between antibody and bacterium is biological and not chemical, as in the neutral balance of antitoxin and toxin mixtures. Under these experimental con- ditions, however, some information has been gained concerning the distribution of the immune bodies occurring in antipneumococcus serum. Experiment I.-TO determine the protective value of the total globulins pre- cipitated by half saturation with ammonium sulphate. 300 C.C. of antipneumococcus serum I were diluted `with an equal volume of water, and 600 CC. of a saturated solution of ammonium sulphate were added, 0swa.M T. Avery. 138 The mixture was allowed to stand at room temperature over night, filtered, the precipitatte collected, pressed, and dialyzed for. eight days against running tap ,,.ater. To the filtrate crystals of ammonium sulphate were. added to full saturation. The precipitate consisting of albumin was filtered off, pressed, and dialyzed as above. The volume of globulin solution was 100 "cc., so that the globulins were three times as concentrated as in the whole serum. The volume of albumin solution was Ss cc., being 0.2 per cent. that of the original serum. Protective Power of Globulins, Albumin, and Serum I. Pneumococcus I. *mount of culture I. Serum I. Globulins. Albumin. Coatrols. Culture alone. 0.05 C.C. 0.1 C.C. 0.05 CL- 0.08 C.C. :: s. D. 18 . . . . . . . . . . . . S. D. 18' . . . . . . . . . . . . 0.05 C.C. 0.05 c.c! S. 0.05 C.C. 0.03 C.C. D. 72 :: D. 18 . . . . . . . . . . . . D. 18 . . . . . . . . . . . . 0.05 C.C. 0.01 C.C. D. 48 D. 72 D. 18 . . . . . .,. . . . . . 0.0001 C.C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. 36 0.00001 C.C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . _........... D. 36 0.000001 C.C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. 36 In the tables D. = died; S . = survived. The figures represent the number of hours before the death of the animal. Experiment z.-Antipneumococcus serum II was used, diluted with an equal volume of water, and the globulins were precipitated by half saturation with ammonium sulphate. The albumin was precipitated from the fitrate by complete saturation with crystals of ammonium sulphate. Both fractions were dialyzed for four days against running tap water, and made up to the original volume of serum with 0.85 per cent. salt solution. Comparative Protective Value of Globulins, Albumin, and Serum II. Pneumococcus II. Amount Amount of or culture II. seturn iractions. semm II. I Globulins. Albumin. Controls. culture alone. 0.1 C.C. 0.2 cc. D. 66 0.01 C.C. 0.2 cc. S. 0.001 C.C. 0.2 cc. 0.0001 C.C. 0.2 C.C. :: 0.00001 cc. 0.2 cc. S. 0.000001 C.C. 0.2 cc. S. D. 66 D.18 . . . . . . . . . . . . :: D. 19 . . . . . . . . . . . . D. 19 . . . . . . . . . . . . X: D. 24 D. 19 s. I :: :: Aggiutinution. 136 Immune Bo&ks Occurring in Antipneumococcus ikrum, . Experiment J.-The agglutination test was made by mixing equal volumes of an eighteen hour broth culture of pneumococcus `I and varying dilutions of serum fractions. The antigen used in the precipitin reaction was prepared from the washed bacterial residue of liter cultures of pneumococcus I frozen and desiccated in vacuum over sulphuric acid. The dry powder was dissolved in salt solution, with I mg. per cc., and a I : IO dilution of this was used in the test. Equal parts of antigen and serum dilutions were mixed and incubated in the water bath at 37" C. for two hours. Agglutination and Precipitation. . Agglutinuiml. Serum dilution. s- I. Globulins. Ahmh. Serum dilution. 1:1 I : IO z++ I : 100 - -+ Precipitation. ' Swum I. Globulins. $Z 1 I : I'd ++ I : 15 z ++ I : 20 ++ I : 25 : 1:30 z - From experiments I, II, and III it may be concluded that the immune bodies of antipneumoc&cus serum are not dialyzable, that they are precipitated with the globulins by half saturation with am- monium sulphate, and that the demonstrable antibodies, such as agglutinins, precipitins, and protective substances, are combined or associated with the globulins and not with the albumin of the im- mune serum. Experiment J.-TO determine the protective value of the euglobulin and pseudoglobulin. 3oo C.C. of antipneumococcus serum I were diluted with an equal volume of water and 6oo C.C. of a saturated solution of ammonium sulphate added, allowed to stand over night, and filtered, precipitate (a), filtrate (a). The precipitate (a) containing Iboth globulins was taken up in 6oo C.C. of water and saturated with crystals of sodium chloride (210 pm.) and filtered, precipitate (b), filtrate (b). The precipitate (b) was again taken up in water and resaturated with salt, the final precipitate pressed and dialyzed against running tap water until salt-free. .The dialysate was made slightly alkaline to effect solution, used undiluted, and called " euglobulin." To the filtrate (b) about 0.2 per cent. of acetic acid was added to precipitate the globulin remairting in solution. This precipitate was pressed and after dialyzing for twenty-four hours was neutrzl- ized to litmus by sodium carbonate, and dialysis continued for four days. The dialysate was used undiluted and called "pseudoglobulin." The original filtrate (a) was saturated with crystals of ammonium sulphate; the precipitate contain- ing the albumin was collected by filtration, .pressed, and dialyzed. Oswald T. Amy. , Comparative Protective Value. 0.1 c.c* 0.2 C.C. D. 7a D. 17 D. 20 D.17 . . . . . . . . . iJ.01 cc. 0.2 C.C. S. D. 84 D.17 . . . . . . . . . lJ.001 C.C. 0.2 C.C. :: s. D. 20 . . . . . . . . . O.lJOOI C.C. 0.2 C.C. E S. s. D. 24 D. 22 0.00001 C.C. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. 24 o.oooooIc.c........~... e.......... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. 36 0.01 C.C. 0.4 cc 0.01 C.C. 0.3 C.C. 2 :: D. 48 ......... D. 36 ......... 0.01 C.C. 0.2 C.C. ii: :: .. S. D.18 ......... 0.01 C.C. 0.1 C.C. S. D. 28 ......... 0.01 C.C. 0.01 cc. D. 40 D. 40 D. 40 D. 18 ......... 0.01 C.C. 0.005 C.C. . D. D. 18 D. 48 D.18 ......... 0.0001 C.C. ....................................................... D. 18 0.00001 C.C. ......................................................... D.28 fJ.000001c.c ........................................................ D. 36 Experiment 4 shows that by the technique employed the immune bodies of antipneumdcoccus serum are no; confined-to the so called euglobulin or pseudoglobulin, but occur in both these fractions. Their absence in the albumin of the serum confir& the previous experiment. Experikent S.-Fractional precipitation of heated serum by ammonium sulphate. 50 C.C. of antipneumococcus serum I were diluted with a half volume of distilled water tid I per cent. sodium chloride, and 38.6 C.C. of a saturated solu- tion of ammonium sulphate were added, making the total 34 per cent. satura- tion. The serum-sulphate mixture was heated' in a water bath at 56" C. for four hours, $3' C. for one hour, and 60" C. for five minutes. Filtered hot, the precipitate was pressed and dialyzed for four days against running water. To the filtrate, a saturated solution of ammonium sulphate was added up to 54 per cent. saturation. After standing at ropm temperature over night the precipitate was filtered off, pressed, and dialyzed for four days. To the final filtrate am- monium sulphate crystals were added to complete saturation, and the filtrate was pressed and dialyzed. The various fractions isolated were designated 34 Per cent. globulin, 54 per cent. glob&n, and albumin, respectively. After dialysis these fractions were each made up with 0.85 per cent. salt solution to. the volume of the original serum, and in the case of the 34 per cent. globulin suficient sodium car,bonate .was added to effect solution. 138 Immune Bodies Occurving in Alztipneumococcus Swum. Comparative Protective Value of Globulin Fractions Precigitated by Ammonium tEiiI" . * 0.1 CL 0.01 C.C. 0.001 C.C. 0.0001 C.C. 0.00001 cc. 0.000001 cc g&z&L gga. Albumin. wuy2 d$- pe!rim;ptr. suum I. CCWOIS. c,% -- D. 20 D. zo D. 20 S. . . . . . . . . . S. S. : D. 20 k . . . . . . , . . s: D. 20 2 . . . . . . . . . :: D. 20 :: S. D. 18 Z: D. 24 S. D. Ia S. D. 24 S. E: D. za Agglutination tests showed the presence of these antibodies in both fractions, and their absence in the albumin. Amount of frEEIs. 0.2 C.C. 0.2 cc. 0.2 cc. 0.2 C.C. 0.2 C.C. 0.2 C.C. Sulkhate from Heated Serum. - -..- The method followed in this experiment is essentially that devised by Banzhaf3 for the concentration of diphtheria antitoxin. In diphtheria immune serum the antitoxin is precipitated' with the pseudoglobulin. The addition of salt and the heating of serum- sulphate mixture converts some of the pseudoglobulin into the inac- tive euglobulin without material loss of its antitoxic potency. By the application of this method to antipneumococcus serum it was hoped that the residue of protective antibodies occurring in the pseudoglobulin might be changed over with the converted euglob ulin and the total potency of the serum be thus concentrated in the latter fraction. This, however, did not occur; both fractions still showed protection. Heating antipneumococcus serum to the tem- perature used in this experiment does not seem materially to affect the activity of the agglutinins and protective antibodies, although some loss of potency is suffered by .the latter. Experiment &-To determine the protective value of the globulin fraction precipitated `by passage of carbon dioxide through diluted serum. 50 C.C. of antipneumococcus serum II were diluted with ten volumes of dis- tilled water. Carboh dioxide was allowed to `bubble through slowly for t*`\`O hours, the diluted serum being kept at a temperature of 2" C. during the process; it was then placed in the ice box over night, and a sharp separation of precipitate and supernatant fluid was effected by centrifugalization. The carbon dioxide precipitate was dissolved in 50 C.C. of 0.85 per cent. salt solution. The supernatant fluid was half saturated with ammonium sulphate by the addi- tion of an equal volume of a saturated solution of this salt. This carried dotvn the residual globulin remaining after removal of the carbon dioxide-insoluble 8 Banzhaf, E. J., Studies from the Research Laboratory, Department Of Health, New York, rgra-rg, vii, 114 Oswald T. Avery. 139 f&on. After standing at room temperafure over night the precipitate was filtered, pressed, and dialyzed for four days and made up to a volume of 50 CC. ,,.ith 0.85 per cent. salt solution. L Protective Value of Carbon Diokde Globulins. -=.z- .hmmt of c"Iture II. 0.1 C.C. 0.01 C.C. 0.001 C.C. 0.0001 C.C. 0.00001 C.C. 0.000001 C.C. = s .- - 0.2 cc. D. 36 D. 36 0.2 cc. S. 0.2 C.C. S. :: 0.2 C.C. S. 0.2 C.C. : 0.2 C.C. :: s: = - ( _- - hrboa diaxidc- soluble COtltdS. globulin. Culture alone. D.18 s. S. S. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. zo D. 18 D. 18 Precipitation. . Antigen I: 1,000. 1:r ++ ++ +* 1:5 +`. I : 10 f z (7) From experiment 6 it appears that from diluted antipneumococ- cus serum there is precipitated by saturation with carbon dioxid.e a globulin possessing high protective value, and that by the method of titration employed a slightly less potent globulin fraction re- mains in solution, and can be separated from the `albumin by am- monium sulphate. "The somewhat anomalous fact, that each frac- tion of the carbon dioxide globulin apparently pokesses protective value equal to the whole serum, may be explained in part by the method of titration. The immunity unit, 0.b `of a cubic centimeter of serum, which has been found most suitable for protection tests, apparently contains an excess of antibody and the zone of carbon 140 Immune Bodies Occuwing in Antipneumoc&cus Swum. dioxide precipitation may so lie within the limits of this excess that, in the cleavage effected by acidification with carbon dioxide, suffi- cient antibody is carried over with each fraction of the globulin to afford an equivalent protection. The carbon dioxide-insoluble globulin, although representing approximately only one-fifteenth of the total protein, apparently contains more than half the antibodies. Ez-per&en: T.-TO determine the potency of the serum globulins soluble and insoluble in saturated sodium chloride. 50 C.C. of antipneumococcus serum II diluted with an equal volume of water were saturated with crystals of sodium chloride, and the salt-serum mixture was allowed to extract over night at room temperature. The portion of the globulin insoluble in saturated sodium chloride was filtered off, pressed, and dialyzed for four days against running water. To the filtrate containing the salt-soluble fraction was added an equal volume of a saturated solution of ammonium sulphate. The precipitate containing the salt-soluble globulin was collected, pressed, and dialyzed for four days. The fractions were all diluted with 0.85 per cent. salt solution to the volume of the original serum. Protective Value of Sodium Chloride Globulins. Amount Amount of Oligid Sodium chloride- Sodium chiorid~ COlltIOl5. of culture II. semm fractions. uxum II. insoluble globulin. soluble globulin. Culture alone. 0.01 C.C. 0.2 C.C. S. D. zo s. . . . . . . . . . . . . . 0.001 C.C. 0.2 C.C. S. S. s. . . . . . . . . . . . . . 0.0001 C.C. 0.2 C.C. S. S. 6. D. PO 0.00001 C.C. 0.2 C.C. S. S. I S. D. a4 0.000001 C.C. 0.2 C.C. S. S. s . . D. 36 * Agglutination.* Serum II ...................................... + + Sodium chloride-insoluble ...................... ++ Sodium chloride-soluble ........................ ++ Experiment 7 shows that the protective substances and the agglu- tinins of antipneumococcus serum are not confined to either of the globulin fractions separated by saturation with sodium chloride. The salt-insoluble globulin, according to Freund and Joachim,' represents the euglobulin, but includes only a part of the similar fraction precipitated by one-third saturation with ammonium sul- phate. Banzhaf and Gibson6 consider that this method yields as 4Freund, E., and Joachim, J., Ztschr. f. physiol. Chem., ~goz, xxxvi, 407. s Banzhaf, E. J., and Gibson, R. B., Studies from the Research LaboratorYI Department of Health, New York, 1go8-IO, iv-v, 202. Oswald T.. Avery. 141 sllarp a differentiation as possible betsveen the two globulin frac- tious, and that the sodium chloride separation probably represents ,,lore truly the common conception of euglobulin and pseudoglobu- liu than the more usual fractioning by ammonium sulphate. The re=ults obtained in this experiment are similar to and comparable l\ith those of experiment 4, in which the globulins were first pre- cipitated with ammonium sulphate, and then separated by precipita- riorl first with sodium chloride and then with acetic acid. &periment B.-To determine the potency of the water-soluble and -insoluble serum globulin obtained by dialysis. ICJO C.C. of antipneumococcus serum II, undiluted, were placed in a parchment bag and dialyzed for six days against running tap water. The dialyzing bag ,vas then washed out with distilled water, the precipitate collected by filtration, washed with water, taken up in 100 CL of physiological salt solution, and made slightly alkaline with sodium carbonate to effect complete solution. To the filtrate containing the water-soluble globulin, an equal volume of a saturated solution of ammonium sulphate was added, the filtrate collected, pressed, and dialyzed for four days, and then made up to volume ( IOO c.c.) with 0.85 per cent. salt solution. Protective Value of Globulin Fractions Separated by DialysiJ. Amount of Amount of Water-soluble Water-insoluble culture II. semm fractions. globulin. globulin. Oligld saum II. Controls. 0.01 C.C. 0.2 C.C. D. 24 s. . . . . . . . . . . . . . 0.001 C.C. 0.2 C.C. S. :: si . . . . . . . . ...*. 0.0001 C.C. 0.2 cc. s. S. S. D. 20 0.00001 C.C. 0.2 cc. S. D. 24 0.000001 C.C. 0.2 C.C. :: :: S. D. 40 Agglutination. Water-insoluble Water-soluble Pncttmococcnr 11. smm II. globulin. globulin. ++ ++ ++ Seng" showed that by dialysis of diphtheria antitoxin only a small part (1/23 td I/II) of the total globulin was insoluble and that this fraction possessed no antitoxic value. These results were later confirmed by Brieger and Krause.7 Freund and Joachim do not consider the globulins precipitated by fractionation with am- monium sulphate identical with the two groups obtained by dialysis. 6 Seng, W., Ztschr. f. Hyg. s. Infektionskmnkh., I&, xxxi, 513. ' Brieger, L., and Krause, M., Berl. klin. Wchnschr., 1907, xliv, g&. 142 Imniune Bo&es Occurring in Antipneumococcus Serum. Experiment 8 demonstrates the distribution of the antibacterial substances of antipneumococcus serum in both fractions of the glob- ulins separated by dialysis. These results indicate that the globulin precipitated by dialysis (water-insoluble), like that precipitated by carbon dioxide, contains a part, but not all of the immune bodies. Experiment P.-TO determine the relative potency of the globulin fractions obtained by the addition of progressive amounts of ammonium sulphate. SO C.C. of antipneumococcus serum II were diluted with a half volume of water and the whole was precipitated by 34 per cent. saturation with ammonium sulphate, 38.6 C.C. of saturated solution. From the filtrate the residual globulin was precipitated by adding sufficient saturated solution of ammonium sulphate to cause 54 per cent. saturation, allowing for the amount of this salt already present in the filtrate. The precipitates of both fractions were collected, pressed, and dialyzed for four days against running tap water. The dialysates were made up to the original volume of the serum with 0.85 per cent. salt solution, and were designated o to 34 per cent. globulin and 34 to 54 per cent. globulin, respectively. By the same technique a similar quantity of the same lot of serum was fractioned by salting out at 38 and 4.~ per cent. saturation, the residual glubulin in each case being separated from the albumin in the filtrate by precipitation at 54 per cent. saturation ammonium sulphate. Relative Potency of the Globulins Obtaiwd by Salting Out w'th Progt-estive Amounts of Ammonium Sulpkaf~. 0.01 C.C. 0.001 C.C. 0.0001 C.C. 0.00001 C.C. 0.000001 C.C. Globulins precipitated by ammonium sulphate Amount saturation within the limitll of culture II. fractions. AXE. I I Original c-34$ 34-M -3% 38-LiR -44 r-58 serum. __--- 0.2 C.C. . . . . . . . . . s. D. zo S. D. 22 0.2 C.C. . . . . . . . . . s. D. 72 S. D. 24 0.2 C.C. D. 36 S. D. 18 ! 0.2 C.C. D. 36 S. S. 0.2 cc. D. 36 S. S. Agglufinafion wifh Various Globulin Fractions. Serum II, lot 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ++ o-34 per cent. globulin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34-54 per cent. globulin 5+ a................................ _ o-38 per cent. globulin .,................................ ++ 38-54 per cent. globulin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - o-42 .per cent. globulin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ++ 42-54 per cent. globulin - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Experiments I to 8 show that the immune bodies occurring in antipneumococcus sera I and II are associated or combined with the globulins, but are not confined solely to any one of the globulin Oswal& T, Avery. 143 fractions obtained by the various methods of separation employed. Since certain of the experiments seemed to indicate that the major portion of the antibacterial substances. are crowded toward the true or euglobulin end of the protein spectrum, an attempt was made in experiment 9 to determine, by fractional salting out with progressive amounts of ammonium sulphate, a point where the zones of globulin and antibody precipitation might coincide and the total potency of the serum be conserved with a minimum of serum globulin. Experiment 9 indicates that this zone of joint precipita- tion of total antibody with minimal globulin is reached at about 38 per cent. saturation with ammonium sulphate. At this point a fraction of the serum globulin is precipitated which is as active in antibacterial potency as the original serum, while the residual glob- ulin is inactive and apparently unessential. DISCUSSION. The chemistry of the. proteins of immune serum and their rela- tion to various antibodies have been studied thoroughly in the anti- toxic sera. Numerous investigators have shown that diphtheria and tetanus antitoxin are precipitated with the serum globulins. More exact study of these globulins has demonstrated that they may be further subdivided according to their solubility and pre- cipitation by certain chemical reagents. Seng showed that the globulins of immune serum, and according to Marcus* those of nor- mal serum as well, are of two kinds: an insoluble globulin pre- cipitated by acetic acid, carbon dioxide, dilution with water, or dialysis ; and a soluble globulin unaffected by these reagents but precipitated by the neutral salts of ammo&m and magnesium sulphate. Seng showed, and many investigators have since con- firmed, the association ,of antitoxin with the soluble globulin. Fractional precipitation with the neutral salts of the heavy metals yields a similar, though not altogether comparable, separation of the globulins into a soluble and insoluble fraction, the so called euglobulin and pseudoglobulin. It is with tee latter of these, the pseudoglobulin, that antitoxin is combined. By the application of this principle Gibson and Banzhaf have devised an efficient and economic method for the concentration of'diphtheria antitoxin. s Marcus, E., Ztschr. f: physiol. Chem., r&g, xxviii, pig. 144 Immune Bodies Occurring in Antipneumococcus See-urn. It is now generally accepted that the antibodies occurring in immune serum are protein in character and are not readily disso- ciated from the protein of the serum in which they occur. Priischer'9' attempt to produce a non-protein antitoxin by digestion of antidiphtheria serum with trypsin has not been confirmed. Banz. haf" and Meilanby found that the ratio of protein digestion and antibody destruction are approximately the same. Seng, Joa- dhim If Atkinson,`* Ledingham,18 and Banzhaf found that certain quaniitative changes occur in the serum proteins of animals im- munized to bacterial toxins, and that the increase in the globulin content bears a marked relation to the increase in antitoxic potency of the serum. Landsteiner14 has shown that the antitryptic action of normal serum is associated with the albumin fraction, and Opie and BarkerL6 have found antileucoprotease in the same fraction. The distribution of immune bodies in the blood serum of dif- ferent animal species varies. Pickle found that in immune horse serum diphtheria antitoxin was associated with the pseudoglobulin, while in goat serum it is present in the euglobulin. The lysins and agglutinins of cholera occur in the euglobulin fraction of the serum of immunized goats and horses. Typhoid agglutinins in the immune serum of the horse are found in the pseudoglobulin, while in goats, rabbits, and guinea pigs they occur only in the euglobulin. Hartley17 has recently shown that the immune bodies of antirinderpest serum are present in the euglobulin fraction of bovine serum. The present work on the distribution of the immune bodies of antipneumococcus serum is confined to a study `of their occurrence in the blood serum of horses immunized to types I and II of the pneumococcus. The sera were produced by repeated intravenous injections of these organisms, beginning with dead bacteria and g Priischer, Miinchen. med. Wchnschr., 1go2, xlix, 1176. ' lOBanzhaf, E. J., Bull. Johns Hopkins Hosp., 1911, xxii, 106. 11 Joachim, J., Arch. f. d. ges. Physiol., 1go3, xciii, 558. 12 Atkinson, J. P., Jour. Ezper. Mea., ~gooa, v, 67. 1s Ledingham, J. C. G., Jour. Hyg., 1907, vii, 65. 14 Landsteiner, K., Centralbl. f. Bakteriol., zte Abt., Orig., rgoo, xxvii, 357. 15 Opie, E. L., and Barker, B. I., Jour. Exper. Mea., 1907, ix, 207. `6Pick, E. P., Be&r. Z. chew Phys.% Path., 1902, i, 351, 1' Hartley, P., Memoirs of #he Department of Agriculture, India, 1913-14, i, 177. Oswald T. Avery. 146 progressively increasing the doses,. until the animals were able to tolerate the bacterial residue of a liter of live virulent pneumococcus injected directly into the blood stream. The serum was obtained from horses which had been immunized over periods of one to two years. The method employed in obtaining antipneumococcus serum di[fers entirely from that used in the production of antitoxins. In the former, the immunizing response is provoked by a live virulent organism, while in the latter the stimulus is in the nature of a bac- teria-free toxin. The immunity mechanism involved in the elabo- ration of specific antipneumococcus substances may or may not _ differ physiologically from that concerned in the production of antitoxins, but the types of antibody evoked in each instance are distinct, and their distribution in the serum globulins is different. SUMMARY. The immune bodies of antipneumococcus serum are cdm- pletely precipitated by 38 to 42 per cent. saturation with ammonium sulphate. They are incompletely precipitated by (a) ammonium sul- phate in less than $3 per cent. saturation, (b) saturation with sod- ium chloride, (c) dilution and saturation with carbon dioxide, (d) ' removal of crystalloids by dialysis. The `immune bodies of antipneumococcus serum are, therefore, associated or combined with that fraction of. the globulins precipi- tated by 38 to 42 per cent. saturation with immonium sulphate. The immune body fraction doe; not correspond exactly with the ordinary euglobulin (one-third saturation with ammonium sulphate or complete saturation with sodium chloride). or with the insoluble globulins `precipitated by carbon dioxide or dialysis. These frac- tions carry with them only a part of the immune bodies. Neither the albumin nor that fraction of the globulin not precip- itated by 38 to 42 per cent. saturation of ammonium sulphate con- tain any of the demonstrable antibodies. The most promising method for the practical purification of the immune bodies occurring in antipneumococcus serum appears to be precipitation by 38 to 4 per cent. saturation with ammonium sulphate.