Issued June 21, 1909. . S. DEPARTMENT OF AGRICULTURE, BUREAU OF ANIMAL INDUSTRY. BUI.I.HTIN 113. A. I). MELVIN, CHIEF OF BUREAU. TLTRATION EXPERIMENTS WITH BACILLUS CHOLERA SUIS. BRARY OF THI L A. 634 SOJJHffeTUKE AVE. LOS ANGJ BY C. N. McBRYDE, M. D., Senior Bacteriologist, Biochemic Division. WASHINGK )N : (InVhRNMliNT 1'KlN'l l\<. 1909. Issued June 21, 1909. U. S. DEPARTMENT OF AGRICULTURE, BUREAU OF ANIMAL INDUSTRY. BULLETIN 113. A. D. MELVIN, CHIEF OF BUREAU. FILTRATION EXPERIMENTS WITH BACILLUS CHOLERA SUIS. BY C. N. McBRYDE, M. D., Senior Bacteriologist, Biochemic Division. LIBRARY OF 1 L A. CO. MEDif 834 SOUTH WESTLAKfcAVE. LOS ANGELES BARLOW MEDICAL LIBRARY 742 NORTH BROADWAY LOS ANGELES. CAL WASHINGTON : GOVERNMENT PRINTING OFFICE. 1909. THE BUREAU OF ANIMAL INDUSTRY. Chief: A. D. MELVIN. Assistant Chief: A. M. FARRINGTON. Chief Clerk: CHARLES C. CARROLL. Biochemic Division: M. DORSET, chief; JAMES A. EMERY, assistant chief. Dairy Division: B. H. RAWL, chief. Inspection Division: RICE P. STEDDOM, chief; MORRIS WOODEN, R. A. RAMSAY, and ALBERT E. BEHNKE, associate chiefs. Pathological Division: JOHN R. MOHLER, chief; HENRY J. WASHBURN, assistant chief. Quarantine Division: RICHARD W. HICKMAN, chief. Zoological Division: B. H. RANSOM, chief. Experiment Station: E. C. SCHROEDER, superintendent; W. E. COTTON, assistant. Animal Husbandman: GEORGE M. ROMMEL. Editor: JAMES M. PICKENS. BIOCHEMIC DIVISION. Chief: M. Dorset. Assistant Chief: James A. Emery. Hog cholera investigations: Chief of Division in charge; W. B. Niles, veterinary inspector in charge of field experiments. Investigations of dips and disinfectants: R. M. Chapin, senior biochemist in charge. Bacteriological investigations of meat food products: C. N. McBryde, senior bacteri- ologist in charge. Tuberculin and mallein preparation and distribution: A. M. West, bacteriologist in charge. Meat inspection laboratories: Central laboratory, Washington, D. C., T. M. Price, senior biochemist in charge. Branch laboratories: Chicago, 111., Ralph Hoagland in charge; New York, X. Y., A. H. Roop in charge; Kansas City, Kans., C. H. Swanger in charge; South Omaha, Nebr., W. B. Smith in charge; East St. Louis, 111., E. A. Boyer in charge; San Francisco, Cal., A. E. Graham in charge. 2 LETTER OF TRANS/YUTTAL U. S. DEPARTMENT OF AGRICULTURE, BUREAU OF ANIMAL INDUSTRY, Washington, D. C., January 30, 1909. SIR: I have the honor to transmit herewith, and to recommend for publication as a bulletin of this Bureau, the accompanying manu- script entitled ''Filtration Experiments with Bacillus cholerse suis," by Dr. C. N. McBryde, of the Biochemic Division. The subject-matter of the paper, while technical in its character, has a direct bearing on the hog cholera problem. In a previous publication issued by this Bureau (Bulletin 72, The Etiology of Hog Cholera) it was first shown that the so-called hog cholera bacillus, B. cholerse. suis, could not be regarded as the definite and sole cause of hog cholera, since the disease could readily be produced by an ultra-microscopic virus in the blood of sick hogs, which virus had repeatedly passed through a filter that effectually prevented the passage of B. cholera suis. This discovery naturally caused widespread discussion among scientific investigators upon this subject throughout the world, some of whom have questioned the accuracy of the findings pub- lished in the bulletin referred to. The experiments described in the present paper have been carried out to meet these criticisms and have resulted in confirming the correctness of the work previously published. Respectfully, A. D. MELVIN, Chief of Bureau. lion. JAMES WILSON, Secretary of Agriculture. 3 CONTENTS. Page. Introductory 5 Filters used 7 Arrangement of filters .- 7 Technique employed in nitrations 8 Cultures used 9 Experiments with Berkefeld filters 11 Summary of experiments with Berkefeld filters .*. . . 13 Discussion of Berkefeld filtrations 13 Experiments with Pasteur-Chamberland filters 14 Summary of experiments with Pasteur-Chamberland filters 25 Discussion of Pasteur-Chamberland filtrations 26 Rate of filtration of Berkefeld and Pasteur-Chamberland filters 28 Granule formation in cultures of B. cholerse suis 29 Summary of results 30 Conclusions... 31 ILLUSTRATION. Page. FIG. 1. Apparatus for fractional filtration, designed for use with Pasteur- Chamberland or Berkefeld filters 8 4 FILTRATION EXPERIMENTS WITH BACILLUS CHOLERA SUIS. INTRODUCTORY. In the early part of 1905 Dorset, Bolton, and the writer* published a paper on the etiology of hog cholera, in which it was first definitely shown that hog cholera is due to a filterable virus and not to Bacillus cholerse suis as had been previously supposed. The conclusions reached in that paper were based on an extensive series of filtration experi- ments in which the blood serum of hogs sick of hog cholera was filtered through Berkef eld and Chamberland filters in order to exclude B. cholera suis. In these experiments a large number of nonimmune, healthy hogs were injected with the filtered serum, which was carefully tested in every instance prior to injection for B. cholerse, suis. The filtrates were proven to be free from that organism by either incubating the entire filtrate or else taking out large portions for subcultures, and also by the injection of large portions into guinea pigs and rabbits, animals which are extremely susceptible to B. cholerx suis. These experiments with the filtered serum of hogs sick of hog cholera and proven to be free from B. cholerse suis showed that if such serum was injected subcutaneously into well hogs it caused typical hog cholera with great regularity, whereas the same serum injected into small laboratory animals which are naturally very susceptible to B. cholerse suis was entirely without effect. In many of these experi- ments rabbits were injected with as much as 20 c. c. of the filtrates without any subsequent ill effects, whereas hogs injected with like amounts of the filtrates died of typical hog cholera. These results, coupled with others which developed in the course of this investigation, led to the following conclusions: (1) That in the disease known as hog cholera there is some other etiological factor present besides B. cholera suis; (2) that this other factor is an ultra- visible virus sufficiently small to pass through the pores of Berkefeld and Chamberland filters; (3) that this ultra-visible virus is the true cause of hog cholera, and that B. cholerse suis, when present in the blood and tissues of animals sick of hog cholera, is probably a sec- ondary invader and at most an accessory factor in the disease. These conclusions with regard to the etiology of hog cholera, being decidedly revolutionary in character, attracted widespread attention Dorset, M., Bolton, 15. M., and McBryde, C. N. The Etiology of Hog Cholera. U. S. Department of Agriculture, Bureau of Animal Industry, Bulletin 72. Wash- ington. 1905. 5 6 FILTRATION EXPERIMENTS WITH B. CHOLERA SUIS. * among scientific investigators in other countries, and it is gratifying to state that our results have been confirmed by many well-known workers in various parts of the world. Thus in Germany the same results were obtained by Ostertag and Stadie in the Hygienic Institute of the Veterinary High School at Berlin, and by Uhlenhuth and his coworkers, Hiibener, Xylander, and Botz, in the laboratories of the Imperial Board of Health at Berlin. Similar results were obtained by Hutyra in Austria, by Stockman and McFadyean in England, by Theiler in South Africa, and within the past year by Carre", Leclainche, and Vallee in France. In our own country our results have been confirmed by McClintock, Boxmeyer, and Siffer. It is thus seen that our conclusions with regard to the etiology of hog cholera are not lacking in confirmatory proof. Certain criticisms, however, have been made of our earlier work on the etiology of hog cholera, referred to above. In 1907 Lourens, subdirector of the State Serum Institute at Rotterdam, published an article on the filterability of Bacillus suipestifer in which he makes the statement that the so-called hog-cholera bacillus, known as B. cholerse suis or B. suipestifer, is capable of passing through filters of the Chamberland and Berkefeld types, composed, respectively, of unglazed porcelain and infusorial earth. He claims that the ability of B. cholerse, suis to pass through filters of the types mentioned is due to a property which this bacillus possesses of breaking up into granules sufficiently small to pass through the pores of the filter. He also asserts that B. cholerse, suis is the cause of hog cholera and that none of the investigators who have conducted filtration experi- ments to show that hog cholera is due to a filterable virus have afforded sufficient and convincing proof that the filtrates which they employed did not contain B. cholerse suis. While we had no doubts as to the accuracy of our earlier filtration work, nor any doubt whatever that hog cholera is due to a filterable virus, especially since our earlier work has been confirmed by such well-known investigators as those cited above, we nevertheless felt that such criticisms should not be allowed to pass unnoticed, and it was with a view to meeting these criticisms that the experiments described in the present paper were undertaken. In connection with our earlier work on the etiology of hog cholera the writer carried out a series of filtration experiments with cultures of B. cholerse suis. This was done as a further test of the efficiency of the filters used that is, to determine whether filters of the Berkefeld and Chamberland types could be relied upon absolutely to restrain or keep back B. cholerse suis. a Lourens, Louis F. D. E. Untersuchungen uber die Filtrierbarkeit der Schweine- pest-bacillen (Bac. suipestifer). Centralblatt fur Bakteriologie, abt. 1, orig., band 44, heft 5, pp. 420-127, Aug. 20; heft 6, pp. 504-512, Aug. 31; heft 7, pp. 630-648, Sept. 17. Jena, 1907. DESCRIPTION AND ARRANGEMENT OF FILTERS. 7 Since the publication of Lourens's paper a second series of filtra- tion experiments with bouillon cultures of B. cholerse suis has been carried out by the writer, and the results obtained, together with those obtained in the earlier experiments referred to above, none of which have hitherto been published, are now recorded in detail. FILTERS USED. The filters used were Berkefeld laboratory cylinders Nos. 5 and 7, and Pasteur-Chamberland bougies "F" and "B," and were of the same types as those used in the filtration experiments described in Bulletin 72. A circular which accompanied the Pasteur-Chamberland bougies stated that these niters were manufactured in France and imported into this country, and this was also stated on the cardboard boxes in which the bougies were packed. The individual bougies were labeled in French as follows: "F [or "B"l Filtre Chamberland, Systeme Pasteur, H. B. & Cie., Choisy-le-Roi." New filters were used for each experiment, except in the case of experiments 7 and 8 with the Pasteur-Chamberland filters, where filters were employed which had been used once and then cleansed according to the methods adopted by Lourens. Before being used the filters were first tested by submerging them in water and then forcing compressed air through their walls. If the air came through the walls of the filter in the form of small bubbles or beads the filter was regarded as being free from flaws, but if large air bubbles formed on the sides this was taken as an indication of possible defects or flaws, and such filters were discarded. This method of testing the filters is similar to that employed by Lourens. ARRANGEMENT OF FILTERS. In the first three experiments with Berkefekl filters and the first three experiments with the Pasteur-Chamberland type the filters were attached to ordinary side-arm or suction flasks and the filtrates collected in one portion. In the remainder of the experiments the fractional method of filtration was used, for which a special form of apparatus was devised. The apparatus, which is shown in figure 1, may be used with either the Berkefeld or the Chamberland type of filter. It consists essentially of a double side-arm suction flask with which the filters are connected above by means of rubber stoppers and a protected outlet below connected with the lower end of the suction flask by a short piece of heavy black rubber tubing. The outlet consists of a glass tube surrounded by a glass shield which is open at the lower end and is plugged with cotton to protect the out- let tube from dust. The outlet is controlled by a pinchcock above. 8 FILTRATION EXPERIMENTS WITH B. CHOLERJE SUIS. One arm of the suction flask is connected by means of heavy rubber tubing with a vacuum gauge and suction pump. The otherarmis connected by means of rubber tubing with a glass inlet tube plugged with cotton, the admission of air into the suction flask being con- trolled by means of a screw pinchcock. (See fig. 1 .) FIG. 1. Apparatus for fractional filtration, designed for use with Pasteur-Chamberland or Berkefeld filters, a, Glass mantle surrounding filter; 6, Chamberland filter; c, paraffin joint; d and e, rubber stoppers;/, double side-arm suction flask; g, pinchcock controlling outlet from suction Sask; h, outlet tube surrounded by glass shield and attached to lower end of suction flask by means of short rubber tubing; i, glass shield fused to and surrounding outlet tube as a protection against contamination when the filtrates are drawn off; j, glass Inlet tube plugged with cotton, for admitting air into suction flask; k, pinchcock governing the admission of air into flask; /, vacuum gauge; TO, stopcock connected with vacuum pump. TECHNIQUE EMPLOYED IN FILTRATIONS. In conducting a filtration with the apparatus described, the pinch- cocks g and fc are first closed off and the vacuum applied at m. The filtration is then allowed to proceed until the filtrate reaches the TECHNIQUE EMPLOYED AND CULTURES USED. 9 desired level in the suction flask /. As soon as the suction flask is exhausted of air the rubber tubing above the pinchcock g will be seen to collapse and will remain so as long as the vacuum is applied. When the filtrate reaches the desired level in the suction flask the vacuum is closed off and the air gradually admitted into the suction flask by means of the pinchcock k through the glass inlet tube plugged with cotton . The admission of air into the suction flask may be reg- ulated by watching the rubber tubing above the pinchcock g, which expands as soon as the vacuum within the suction flask has been neutralized. The filtrate which has collected in the suction flask is now drawn off into a sterile Erlenmeyer flask by means of the pinch- cock g through the glass outlet tube 7i. . In withdrawing the filtrate an assistant removes the cotton plug from the glass shield surround- ing the outlet tube, steadying the shield at the same time. The neck of the Erlenmeyer flask, which is to receive the filtrate, is then introduced within the shield beneath the outlet tube and the filtrate allowed to flow out by releasing the pinchcock above. The inlet, tube is now closed off again by means of the pinchcock k, the vacuum reapplied at m, and another portion of the filtrate collected and drawn off as before. In this way as many fractions may be collected as are desired. Dry heat is used for the sterilization of all glass portions of the apparatus except the glass mantle a surrounding the filter, which being outside of the filter does not require sterilization. Both arms of the suction flask should be plugged with cotton to prevent dust particles from being drawn into the flask. The filters and all rubber connections are sterilized by boiling in distilled water for from one- half hour to one hour. After the apparatus is set up all joints should be carefully paraffined. CULTURES USED. Five different strains of Bacillus cholerse, suis were used. These cultures were obtained from different sources, either directly from hogs which died of hog cholera or from guinea pigs inoculated with cultures obtained from hogs which had died of hog cholera. The cultures were grown on the various laboratory media and their biological and morphological characters carefully noted. All of the cultures used were virulent for guinea pigs and rabbits. A brief record of the different cultures is appended. 1. Culture Gp 3998 S. This culture was obtained from guinea pig 3998, which died from the inoculation of a culture obtained from hog 1086. Hog 1086 was a noriimmune animal which served as a check in the Iowa experiments. The animal was exposed, together with certain other treated hogs, to a natural outbreak of hog cholera. As a result of the exposure the animal contracted the disease and died with typical cholera lesions, B. cJiolene, suis being obtained in pure 72369 Bull. 113-09 2 10 FILTRATION EXPERIMENTS WITH B. CHOLERA SUIS. culture from the heart blood and spleen. A subculture obtained from agar plates made from the spleen of hog 1086 was used for the inoculation of guinea pig 3998. The guinea pig died within six days, and B. cholerse suis was recovered in pure culture from the spleen and heart blood. A subculture from agar plates made from the spleen of guinea pig 3998 was cultivated on the various laboratory media and found to correspond with B. cholerse, suis in all respects. This culture, which we have designated Gp 3998 S, was used in the first four experiments with Berkefeld filters, and the first two experiments with Pasteur-Chamberland filters. 2. Culture Gp4@92 S. Obtained from the spleen of guinea pig 4692, which died from the injection of 1 c. c. of unfiltered, undiluted blood serum of hog 1208, an experiment hog carrying the Scribner strain of disease in the experiments at Washington (see "Experiments with tail blood," Bulletin 72, page 38). This culture was used in a num- ber of the culture experiments described in Bulletin 72 and was found to correspond with B. cholerse suis in all of its cultural characters. 3. Culture H 2199 S. This culture was obtained directly from hog 2199, which served as a check in the Iowa immunity experiments. The animal died from the injection of 2 c. c. of unfiltered disease- producing blood from the Scribner and Syphax strains of disease. A description of the Scribner and Syphax outbreaks may be found in Bulletin 102 of this Bureau. The autopsy on hog 2199 revealed typical cholera lesions and B. cholerse suis was obtained in pure culture from the spleen. This culture, designated as H 2199 S,was cultivated on the different laboratory media and was found to correspond in all of its cultural characters with B. cholerse suis. 4. Culture H 2450 S. This was obtained directly from hog 2450, a western hog used in the Iowa experiments. Hog 2450 was an unin- oculated check and was placed in the exposure pen (for description of exposure pen see Bulletin 102, page 12) along with certain treated hogs in order to test the virulence of the disease prevailing in the expo- sure pen at the time. The animal contracted hog cholera in the usual time and died with characteristic lesions of the disease. Cultures from the spleen were grown on the Different laboratory media and found to correspond in all respects with B. cholerse suis. 5. Culture H 2228 S. This culture was obtained directly from hog 2228, which died as a result of an injection of 5 c. c. of defibrinated blood obtained from an outbreak of hog cholera in Iowa. Hog 2228 exhibited characteristic cholera lesions at autopsy and agar plates from the spleen revealed B. cholerse suis. This culture, like all of the others, was cultivated on the various laboratory media and was found to correspond with B. cholerse suis in all of its cultural characters. Dorset, M., McBryde, C. N., and Niles, W. B. Further Experiments Concerning the Production of Immunity from Hog Cholera. U. S. Department of Agriculture, Bureau of Animal Industry, Bulletin 102. Washington, 1907. BERKEFELD FILTER EXPERIMENTS. 11 EXPERIMENTS WITH BERKEFELD FILTERS. EXPERIMENT 1. A flask containing 400 c. c. of neutral beef broth was inoculated with culture Gp 3998 S and placed in the incubator. At the end of twenty- four hours the flask was well clouded and 200 c. c. of the culture was then passed through a Berkefeld laboratory cylinder No. 5. The vacuum used was 20 inches and the time occupied in filtration was one and one-half hours. The filtrate was collected in one portion in an ordinary side-arm flask and placed at once in the incubator. At the end of eighteen hours the filtrate, which came through the filter perfectly clear, was distinctly clouded. Agar plates from the filtrate showed B. cholerse suis in pure culture. A subculture made from the agar plates was grown on the different laboratory media and found to correspond in all cultural characteristics with the original culture. The results obtained in this experiment show that B. cholerse suis passed through the Berkefeld filter in the first 200 c. c. of filtrate. EXPERIMENT 2. A flask of neutral beef broth was inoculated with culture Gp 3998 S and incubated at 37 C. for twenty-four hours. The culture showed well-marked clouding and a microscopic examination showed the organisms to be actively motile. Filtration was made through a Berkefeld laboratory cylinder No. 7. The amount of culture filtered was 100 c. c. and the time occupied in filtration was one and one-half hours. The filtration was conducted under a vacuum of 22 inches. The filtrate was collected in one portion and was immediately placed in the incubator. After twenty-four hours in the incubator the filtrate showed distinct clouding and agar plates revealed a pure culture of B. cholerse suis. A subculture made from the agar plates was cultivated on the different laboratory media. This culture was found to correspond with the original culture in all of its cultural characteristics. In this experiment B. cholerse, suis passed through the Berkefeld filter in the first 100 c. c. of filtrate. KXPERIMKNT 3. This experiment was practically a repetition of experiment 2 and the results obtained were the same. The filter used was a Berkefeld laboratory cylinder No. 7. The amount of culture filtered was 100 c. c., which required one and one-half hours to pass through the filter. The vacuum employed was 23^ inches*. The entire filtrate was collected in one portion, as in the two preceding experiments. After forty-eight hours in the incubator the filtrate showed heavy clouding and agar plates revealed a pure culture of B. choleric, suis. A sub- culture was made from the agar plates and grown on the different laboratory media. This culture was found to correspond in all cultural characteristics with the original culture. 12 FILTRATION EXPERIMENTS WITH B. CHOLERA SUIS. EXPERIMENT 4. In this experiment the fractional method of nitration was used, the filtrates being collected in large sterile test tubes. In carrying out the experiment a twenty-four hour culture of Gp 3998 S, grown on neutral beef broth, was used. Filtration was made through a Berkefeld laboratory cylinder No. 7. The amount of culture filtered was 120 c. c. and the actual time consumed in the filtration was one hour and eighteen minutes. The suction varied from 20^ to 21^ inches. The details of the experiment are shown in the following table: Table showing details of experiment 4- Portion. Amount. Time consumed in passing filter. Suction. Result of incuba- tion. I c. c. 20 Minutes. 4 Inches. 20J Remained clear. II 20 7 20.J Do. III 20 10 20i Do. IV . 20 12 21 Do. V.. 10 7 21 Do. VI 10 10 21 Do. VII.. 10 11 21$ Clouded. VIII 10 17 21J Do. The test tubes containing the filtrates were placed in the incubator immediately after filtration. At the end of twenty-four hours por- tions VII and VIII showed very slight clouding; other tubes remained clear. At the end of forty-eight hours portions VII and VIII showed increased clouding, while the others remained clear. Agar plates were made from portions VII and VIII and revealed pure cultures of B. cholerse, suis. Subcultures from the agar plates were grown on the various laboratory media and found to correspond with the origi- nal culture. Portions I to VI were kept in the incubator for two weeks and remained perfectly clear. Hanging-drop preparations and agar plates from these portions failed to reveal any organisms. In this experiment the Berkefeld cylinder held back the organisms until 100 c. c. of the culture had been filtered, but the first portion of 10 c. c. collected after 100 c. c. of culture had passed the filter showed B. cholerse, suis. EXPERIMENT 5. In this experiment a twenty-six hour bouillon culture (II 2450 S) was used. The reaction of the beef broth in this instance was 1 per cent acid. The culture showed well-marked, uniform clouding, and a microscopic examination previous to filtration revealed active motility. Filtration was made through a Berkefeld cylinder No. 5 arranged for fractional filtration. The vacuum employed varied from 20 to 23 inches. It was the intention to filter several hundred cubic centimeters of the culture and collect the filtrate in separate SUMMARY AND DISCUSSION OF BERKEFELD FILTRATIONS. 13 fractions, as in the preceding experiment, but the filtration -proceeded very slowly and was discontinued at the end of forty-five minutes after 50 c. c. of culture had passed the filter. The filtrate was drawn off in a small sterile Erlenmeyer flask and placed in the incubator. At the end of eighteen hours the filtrate showed a distinct, uniform clouding and agar plates revealed a pure culture of B. cholerse suis. In this experiment B. cholerse suis passed through a Berkefeld cylinder in the first 50 c. c. of filtrate. SUMMARY OF EXPERIMENTS WITH BERKEFELD FILTERS. The results of the experiments with Berkefeld filters are summarized in the following table : Summary of Berkefeld filtration*. A . Culture used. 3 g i! Date begun. Grade of filter. Q GJ god o gl'-s Vacu- um. Result. No. Age. Medium. 6 5 O*i3 ^ ""< E-i 1904. Hrs. c. c. ft. m. Inches. i Feb. 28 No. 5.. Gp3998 S... 24 Neutral beef 200 1 30 20 B. cholerx suis in broth. filtrate. 2... Mar. 7 No. 7 . do... 24 ... do... 100 1 30 22 Do. 3 Mar. 9 No. 7 ...do... 24 ...do... 100 1 30 23J Do. 4 Mar. 15 No. 7 . do.. . 24 .. do 120 1 18 20t-2l| B. cholerx suis in filtrate after 100 c. c 1 . had passed 1908. filter. 5 June 12 No. 5.. H2450S 26 + 1 beef broth.. 50 45 20 -23 B. cholerx suis in filtrate. DISCUSSION OF BERKEFELU FILTRATIONS. In the foregoing experiments five Berkefeld cylinders, all of them new, were tested with bouillon cultures of B. cholerse. suis, and in each instance the organism passed through the walls of the filter. In the case of one cylinder (experiment 4) the organism did not appear in the filter until 100 c. c. of the culture had been filtered. In one instance (experiment 5) the organism appeared in the first 50 c. c. of filtrate. With two of the cylinders (experiments 2 and 3) the organism .passed through the walls of the filter in the first 100 c. c. of filtrate, and in the case of the remaining filter (experiment 1) the organism appeared in the first 200 c. c. of filtrate. These results would indicate that in the case of liquid cultures or relatively heavy suspensions of small micro-organisms like B. cholerse suis the small laboratory Berkefeld filter can not be relied upon to keej) back the organisms when any considerable amount of the material is filtered. That the results obtained in these experiments conform with those obtained by other investigators who have tested the Berkefeld filter is shown by the following brief review of the literature relating to Berkefeld filters. 14 FILTRATION EXPERIMENTS WITH B. CHOLERJE SUIS. Ill 1902 Wherry" found that the bacillus of pneumonia in guinea pigs, an organism somewhat shorter than, but very nearly as thick as, B. cliolerx &uis, would pass through the pores of a Berkefeld cylinder No. 5. Bouillon cultures one and three days old were used in these experiments, and the organism appeared in the filtrate after from 75 c. c. to 80 c. c. of the culture had been filtered. Pfuhl 6 in 1903 made a series of tests with four small Berkefeld cylinders and found that of these only one could be depended on with safety to deliver the first 100 c. c. of filtrate germ free. One of the filters gave 50 c. c. of germ-free filtrate, but no more; the other two filters failed to give even 50 c. c. of germ-free filtrate. The cul- tures used in these tests were B. coli communis and a vibrio of the same size as the cholera spirillum. In a parallel series of experiments with large Berkefeld cylinders, which have thicker walls, Pfuhl found that 50 per cent of these filters were unable to restrain organisms approximating the size of B. typhosus and B. dysenterix. Novy and MacNeal c in 1904 found that Trypanosoma leivisi could be passed through a Berkefeld filter. Novy and Knapp d in 1906 showed that even so large an organism as Spirochseta obermeieri, which is 7 to 19 microns or more in length, will pass through the small Berkefeld filters under a pressure of 50 pounds. Bulloch, Craw, and Atkin c in 1908 tested ten Berkefeld filters with tap water and found that only one gave a sterile filtrate on the first day under a maximum pressure of 32.5 pounds to the square inch. The remaining nine yielded contaminated filtrates within fifteen minutes, or practically as soon as the filters were started. EXPERIMENTS WITH PASTETJB-CHAMBEBLAND FILTERS. EXPERIMENT 1. A flask of neutral beef broth was inoculated with culture Gp 3998 S and incubated for twenty-four hours. At the end of this time the culture showed well-marked clouding and a microscopic Wherry, William B. Experiments on the permeability of the Berkefeld filter and the Pasteur-Chamberland bougie to bacteria of small size. Journal of Medical Re- search, vol. 8 (n. s., vol. 3), No. 2, pp. 322-328. Boston, Nov., 1902. b Pfuhl, E. Ergebnisse einer erneuten priifung einiger kieselgur- und porzellan- filter auf keimdichtigkeit. Festchrift zum 60th Geburtstag von Robert Koch, pp. 75-86. Jena, 1903. cNovy, F. G., and MacNeal, Ward J. On the nitration of trypanosomes. Michi- gan Academy of Science, Sixth Report, p. 180. Lansing, 1904. <*Novy, F. G., and Knapp, R. E. Studies on Spirillum obermeieri and related organisms. Journal of Infectious Diseases, vol. 3, No. 3, pp. 291-393. Chicago, May 18, 1906. Bulloch, William, Craw, J. Anderson, and Atkin, E. E. On the relative efficacy of the Doulton, Berkefeld, and Brownlow filters. Journal of Hygiene, vol. 8, No. 1, pp. 63-69. Cambridge, Jan., 1908. EXPERIMENTS WITH PASTEUB-CHAMBERLAND FILTEBS. 15 examination revealed active motility. The culture was then passed through an F bougie arranged to deliver into an ordinary side-arm suction flask. The amount of culture filtered was 200 c. c., the vacuum employed was 23 inches, and the time consumed in passing through the filter was thirty-five minutes. The entire filtrate was col- lected in one portion. When the filtration was complete the side-arm flask containing the filtrate was disconnected from the filter, plugged with a sterile cotton plug, and immediately placed in the incubator. At the end of eight days the filtrate was found to be perfectly clear, and a careful microscopic examination at this time failed to reveal any micro-organisms. A small portion of the filtrate was removed at this time by means of a sterile pipette and the remainder replaced in the incubator. A guinea pig was injected subcutaneously with 0.5 c. c. of the portion removed and showed no ill effects from the inoculation. The filtrate was kept in the incubator for six weeks and remained perfectly clear. The filtrate was again subjected to a careful microscopic examination before being discarded, and was apparently perfectly sterile. EXPERIMENT 2. A flask containing 400 c. c. of neutral beef broth was inoculated with culture Gp 3998 S and placed in the incubator for twenty-four hours. One-half of the culture (200 c. c.) was then passed through an F bougie and the other half through a B bougie, the two bougies being fitted to ordinary side-arm suction flasks. A vacuum of 21 to 22 inches was used, and the time required for the two filtrations, which were conducted simultaneously, was approximately forty-five minutes. The two side-arm flasks containing the filtrates were immediately placed in the incubator, where they remained for six weeks without showing any signs of bacterial growth. Before being finally discarded both filtrates were subjected to a careful micro- scopic examination, but no micro-organisms could be demonstrated in either hanging-drop or stained preparations. EXPERIMENT 3. A seven-day bouillon culture (Gp 4692 S) was used in this experi- ment. Filtration was made through a Pasteur-Ohamberland B bougie. The filtration was discontinued after 100 c. c. of filtrate had passed the filter. The vacuum gauge recorded 20 inches during the filtration, which occupied thirty-seven minutes. The filtrate was collected in one portion and immediately placed in the incubator. After two weeks in the incubator the filtrate showed a very slight opalescence, but a careful microscopic examination failed to reveal any micro-organisms and the opalescence was attributed to a slight precipitate from the beef broth. In order to test further the ster- 16 FILTRATION EXPERIMENTS WITH B. CHOLERA SUIS. ility of the filtrate, rabbits were injected as follows: Rabbit 1571 received 1 c. c. of filtrate intravenously; rabbit 1332, 3 c. c. intra- venously; rabbit 1570, 5 c. c. intravenously; rabbit 1574, 5 c. c. intraperitoneally. These animals were kept under observation for several months and showed no ill effects whatever from the injec- tions of filtrate. The possibility now suggested itself that' Bacillus cholerse suis might pass through the walls of the Pasteur-Chamberland filter in some form which would not develop in vitro, but which might develop within the animal body. In order to test this point collodion sacs were prepared according to the methods of Grubbs and Francis, of the United States Public Health and Marine-Hospital Service, and McCrae, 6 of the Johns Hopkins Hospital. Sacs prepared according to each of these methods were filled with the filtrate and inserted into the abdominal cavities of rabbits. After an interval of several weeks the animals were killed and the sacs removed. The filtrate within the sacs showed no apparent clouding, and a micro- scopic examination of the contents of the sacs and cultures from the same failed to reveal any micro-organisms. EXPERIMENT 4. A forty-eight-hour bouillon culture (H 2199 S) was used in this experiment. Filtration was made through a Pasteur-Chamberland F bougie. In this experiment and in those which follow the frac- tional method of filtration was used, the filtrates being collected in separate portions in small sterile Erlenmeyer flasks, which were placed in the incubator as soon as the filtrations were completed. Table showing details of experiment 4. Portion Amount. Time occupied in passing filter. Vacuum. Result of incubation. I. c. c. 75 5 minutes Inches. 21 Clouding at 48 hours. II 75 8 minutes 21 Do. III. 75 10 minutes 21 Clouding at 7 days. IV 75 35 minutes 21 Clouding at 48 hours. V... 35 3 hours 21 Remained clear. VI. 10 Over night. . No suction. Do. During the filtration of portions I, II, and III the bougie was entirely covered; during the filtration of portions IV, V, and VI it was only partly covered. Only about one- third of the filter was Grubbs, S. B., and Francis, Edward. Laboratory technique. Collodion sacs. U. S. Treasury Department, Public Health and Marine-Hospital Service, Hygienic Laboratory, Bulletin 7. Washington, 1902. b McCrae, John. Notes upon the agglutinations obtained by intraperitoneal inser- tion of celloidin capsules containing bacilli and upon a mode of preparing such cap- sules. Journal' of Experimental Medicine, vol. 5, No. 6, pp. 635-642. New York, Oct. 1, 1901. EXPERIMENTS WITH PASTETTR-CHAMBERLAND FILTERS. 17 covered with culture during the filtration of portion V, and this, together with the fact that the pores of the filter had become clogged, explains the length of time required for this portion to pass the filter. After drawing off portion V the suction was disconnected and the apparatus left in place overnight. On the following morning it was found that approximately 10 c. c. of the culture had passed the filter, and this portion was drawn off and incubated along with the other fractions. Portions I, II, and IV showed clouding at forty-eight hours. Microscopic examinations and agar plates showed in each case pure cultures of a micrococcus corresponding in morphology to StapTiy- lococcus pyoyenes albus, and probably due to outside contamination at the time these portions were drawn off. Guinea pigs weighing approximately 350 grams each were injected with 2 c. c. each of these portions and exhibited no ill effects from the injections. Portion III became cloudy after one week in the incubator. Micro- scopic examination and agar plates revealed a pure culture of a micrococcus similar to that found in portions I, II, and IV. The injection of 2 c. c. of this portion into a guinea pig was without effect. Portion V was kept in the incubator for ten days and remained perfectly clear. A portion of this filtrate was removed at the end of ten days by means of a sterile pipette, and a rabbit weighing 2,380 grams was given a subcutaneous injection of 10 c. c. without any subsequent ill effects. After drawing off the portion for the injection of the rabbit the remainder of the filtrate was inoculated with culture II 2199 S in order to see whether the filtrate would still furnish a suitable medium for the growth of B. cholerse suis. The inoculated portion was replaced in the incubator and showed well-marked clouding at twenty-four hours. Portion VI was kept in the incubator for ten days and remained perfectly clear. It was then inoculated witli culture II 2199 S, and showed well-marked clouding at twenty-four hours, thus demon- strating that B. cholerx, suis would have grown in this filtrate had it passed through the pores of the filter. The clouding of portions I, II, III, and IV in this experiment was due either to imperfect sterilization of the apparatus or to outside contamination at the time these portions were drawn off. That the clouding of these portions was not due to the passage of B. cholerse, suis through the filter is shown by the fact that guinea pigs inocu- lated with these portions showed no ill effects from the inoculations. It should be stated in this connection that the experiments described in this paper were conducted in a large general bacteriological work- room, where the conditions were more or less favorable for the outside contamination of the filtrates, owing to the constant passing of other workers and the impossibility of avoiding drafts and air currents. 18 FILTRATION EXPERIMENTS WITH B. CHOLERA SUIS. EXPERIMENT 5. In this experiment an eighteen-hour bouillon culture (H 2199 S) was used. The reaction of the bouillon in this experiment and those which follow was 1 per cent acid. Filtration was made through a Pasteur-Chamberland F bougie. The filtrate was collected in separate portions in small sterile Erlenmeyer flasks, which were placed in the incubator as soon as the filtration was completed. The bougie was kept entirely covered with culture during the filtra- tion. The amount of culture filtered was 300 c. c. Table showing details of experiment 5. Portion Amount. Time oc- cupied in filtration. Vacuum. Result of incuba- tion. I... c. c. 75 Minutes. 2 Inches. 20J Clouded. II 75 2 Remained clear. III... 75 3 Om Do. IV 75 4 20i Do. Portion I showed clouding after forty-eight hours in the incu- bator. Microscopic examination and agar plates revealed a pure culture of a large spore-bearing organism resembling Bacillus sub- tills. A guinea pig weighing 350 grams was injected with 2 c. c. of this portion but showed no ill effects from the injection. The organ- ism noted in this flask was evidently due to outside contamination at the time the portion was drawn off. Portions II, III, and IV were kept in the incubator for two weeks and remained perfectly clear. Two portions of 10 c. c. each were then removed from III and IV by means of sterile pipettes and injected into rabbits, as follows: Rabbit 2224 (weight 2,620 grams) received 10 c. c. of portion III; rabbit 2223 (weight 1,801 grams) received 10 c. c. of portion IV. The animals were kept under obser- vation for several months, and showed no ill effects whatever from the injections. After withdrawing portions from III and IV for the injection of rabbits, the remaining portions of these filtrates were inoculated with B. cholerse suis (culture H 2199 S) and returned to the incu- bator. After twenty-four hours both flasks showed well-marked clouding, and hanging-drop preparations and agar plates showed pure cultures of B. cholerse suis, proving that the absence of growth in these portions after filtration was not due to exhaustion of the bouillon. EXPERIMENT 6. Two flasks, each containing 400 c. c. of beef broth, were inoculated with B. cholerse suis (H 2199 S) and incubated for twenty-four hours. The contents of the two flasks were then poured together into a EXPERIMENTS WITH PASTEUB-CHAMBERLAND FILTERS. 19 large balloon flask and well mixed. The object of this was to give a sufficient quantity of culture to keep the bougie covered during the entire filtration. Filtration was made through a Pasteur-Chamber- land F bougie, which was kept entirely covered by the culture through- out th'e filtration. The filtrate was collected in separate portions of 75 c. c. each in seven flasks, the total amount of culture filtered being 525 c. c. Table showing details of experiment 6. Portion. Amount. Time of filtration. Vacuum. Result of incuba- tion. I... r. c. 75 Minute*. 2 Inches. 20 Clouded. II 75 3 20 Remained clear. III 75 7 23-24 Do. IV... 75 9 23^-24 Do. V 75 16 24-25 Do VI . 75 19 25 Do. VII . 75 27 25 Do. Portion I showed slight clouding at forty-eight hours, and a micro- scopic examination revealed a rather long, slender, nonmotile bacillus, apparently in pure culture. The flask was returned to the incubator and allowed to remain for six days, at the end of which time it showed well-marked, uniform clouding'. Hanging-drop preparations and agar plates revealed the same bacillus that had been noted at forty- eight hours. A guinea pig, weighing approximately 350 grams, was injected with 2 c. c. of this portion after incubation for six days, but showed no ill effects as a result of the injection. The clouding of this portion was thus proven to be due to a contaminating organism and not to B. cholerae, suis. Portions II, III, IV, V, VI, and VII were incubated for twelve days, and remained perfectly clear. Rabbits were injected with portions VI and VII as follows: Rabbit 2222 (weight 2,500 grams) received 10 c. c. of portion VI; rabbit 2227 (weight 2,088 grams) received 10 c. c. of portion VII. These animals were kept under observation for several months, and remained perfectly well. After removing portions for the injection of rabbits, the remaining portions of VI and VII were inoculated with B. cholerse, suis (II 2199 S) and replaced in the incubator. Both flasks showed well-marked clouding at twenty-four hours, and microscopic preparations and agar plates revealed pure cultures of B. cholcrse suis. This was done in order to prove that the bouillon had not been exhausted, but was still a suitable medium for the growth of B. cholenx, suis. KXPKRIMKNT 7. In this experiment and in the one which follows Pasteur-Cham- berland bougies were employed which had already been used in the preceding experiments. The bougies were cleansed according to 20 FILTRATION EXPERIMENTS WITH B. CHOLERA SUIS. methods recommended by Lourens before being used the second time. Lourens, it should be stated, did not use new filters in all of his experiments, but adopted certain methods for cleansing his filters, and the two experiments which follow were designed with the view to determining whether the methods which he adopted for cleansing his filters could have had any effect on the permeability of the filters and so have influenced his results. In carrying out experiment 7 an eighteen-hour bouillon culture of B. cholersR suis (H 2199 S) was used. The bougie used was a Pas- teur-Chamberland F which had been used once before in experiment 2. Before being used the second time it was treated by the method which Lourens first used for cleansing his filters, as follows : The filter was first washed with cold tap water, boiled for ten minutes in 1 per cent hydrochloric acid, then boiled for ten minutes in a 3 per cent sodium carbonate solution, after which a liter or more of cold dis- tilled water was passed through the filter until the filtrate showed only a faint alkaline reaction. The filter was sterilized by boiling for one-half hour in distilled water. The filtrate was collected in separate portions. The amount of culture filtered was 750 c. c. Table showing details of experiment 7. Portion. Amount. Time of filtration. Vacuum. Rasult of incuba- tion. I c. c. 75 1J minutes. Inches. 20J Remained clear. II 75 2 minutes. 20 -18 Do. III 75 3 minutes. 18 -21 Do. IV 75 6 minutes. 21 -22J Clouded. V... 75 9 minutes. 22J Remained clear. VI 75 12 minutes. 22J-21J Do. VII.. 75 16 minutes. 21J-22 Do. VIII 75 1J hours 22 Do. IX. - 150 Overnight... No suction. Do. During the filtration of the first seven portions the bougie was entirely covered with culture; during the filtration of portion VIII the bougie \vas partly exposed. After portion VIII was drawn off the vacuum tubing was disconnected from the side-arm flask and the apparatus left in place overnight. On the following morning, after standing for eighteen hours, 150 c. c. of filtrate had collected in the side-arm flask. This portion, which passed through the filter during the night, was then drawn off and placed in the incubator, together with the other portions. Portions I, II, III, V, VI, VII, VIII, and IX remained perfectly clear after eleven days in the incubator. Rabbits were then injected subcutaneously with portions VI and VII, as follows: Rabbit 2225 (weight 2,288 grams) received 10 c. c. of portion VI; rabbit 2226 (weight 2,185 grams) received 10 c. c. of portion VII. These animals were kept under observation for several months, and showed no ill effects from the inoculations. EXPERIMENTS WITH PASTEUR-CHAMBERLAND FILTERS. 21 Careful microscopic examinations were made of several of these filtrates, both in hanging-drop and in smear preparations stained with dilute carbol-fuchsin, and small, coccus-like bodies were noted, but cultures from the same portions gave negative results. Portion IV showed clouding after four days, and agar plates revealed a pure culture of a medium-sized micrococcus corresponding in mor- phology with S. pyogenes albus. A guinea pig was inoculated with 2 c. c. of this portion after clouding had developed, and showed no ill effects from the inoculation. The growth in this flask was evi- dently due to outside contamination when the portion was drawn off. This is shown by the fact that all of the portions drawn off before this one, as well as all of those drawn off after it, were sterile. Several of the filtrates after being incubated for eleven days with negative results were inoculated with B. cholerse suis and gave well- marked clouding at twenty-four hours, showing that the culture medium was still suitable for the growth of B. cholerse, suis. EXPERIMENT 8. The culture used in this experiment was a twenty-four-hour bouillon of B. cholerse, suis (H 2199 S). Filtration was made through a Pasteur- Chamberland F bougie which had been used once in experiment 6. The filter was cleansed according to the method finally adopted by Lourens as yielding the most satisfactory results. Following this method, the outside of the filter was first washed with cold tap water; a liter of cold distilled water was then passed through the filter; next a liter of potassium permanganate solution (1 gin. KMnO 4 , 6.5 gins. HC1, 1,000 c. c. water) was drawn through the filter; following the permanganate, a like amount of oxalic-acid solution (10 gms. oxalic acid to 1,000 c. c. water) was passed through the filter; hot water was next drawn through the filter until the filtrate was acid-free, and finally a liter of cold distilled water was passed through. Before being used the filter was sterilized in the usual manner by boiling for one-half hour in distilled water. The filtrate was collected in separate portions, the total amount of culture filtered being 525 c. c. During the filtration the entire filter was covered witli the culture. Table showing details of experiment