TABLE 5.-Agglutination of the homologous antigen in the serum of a rabbit injected with strain 451, of the melitensis A variety, after absorption by various strains of human and bovine origin

1 See Table 1 for significance of the figures.

* Conditions for the absorptions were the same as in Table 4.

The data in the three tables show that the human strains 426 and 455 absorb the same quantity of agglutinins from the serums as the bovine strains 456 and 460, and they are distinctly differentiated from the strains of the melitensis A variety by the quantity of agglutinins absorbed. The same results were obtained irrespective of whether the serum used for the observations was homologous to the human strain or to the bovine strain. When a serum of the melitensis A variety was used for the absorptions, the human strains 426 and 455 were again shown to behave like the bovine strains, but unlike the human strains of the melitensis A variety.

The data leave no doubt that the human strains 426 and 455 are serologically identical with the group designated as the abortus variety, which includes the majority of bovine strains.

In Table 6 are presented the data by which the type of infection was determined with the serum from the Virginia case reported in this paper. The agglutinins from one sample of the serum were absorbed with an antigen of the abortus variety of Br. melitensis, and those from another sample were absorbed with an antigen of the melitensis A variety. The density of the absorbing antigens was equivalent to 40,000 parts per million of the silica standard, and the absorption was carried out in a 1:5 dilution of the serum. It was shown in the first paper of this bulletin that under these conditions the agglutinins are completely absorbed from a Br. melitensis serum of a titer of 1 to 640 or less if the antigen and antibody are homologous. The simple agglutination test was carried out with the absorbed serums, using both varieties of antigens.

The results of the absorption tests indicate that the infection was of the abortus (bovine) variety of Br. melitensis. The higher titer of the unabsorbed serum when tested with the melitensis A variety

as compared with the titer when tested with the abortus variety is of no significance, for it is shown in the first paper of this bulletin that the titer of a serum of either variety is frequently higher when tested with the heterologous antigen.

TABLE 6.-Agglutination tests with serum 16 A (see Table 2) before and after the absorptions show that it contained agglutinins specific to the abortus variety of Br. melitensis

When the serum was absorbed with the melitensis A variety of antigen, a part of the agglutinins specific to the abortus variety failed to be removed, which shows that the infection was not of the melitensis A variety. Absorption of the serum with the abortus variety of antigen removed all agglutinins from the serum, indicating that the antibodies and antigen were homologous and that the infection in this case was with the abortus variety.

In Table 7 are presented the data from which strain 528, from a human case of Malta fever in South Dakota, was identified with the abortus variety. A serum of the abortus variety and a serum of the melitensis A variety, both of a titer of 1:320, were absorbed with the unknown strain, and the simple agglutination in these absorbed serums was compared with the agglutination after absorption with the homologous antigens. The density of the absorbing antigens was equivalent to 60,000 parts per million of the silica standard, and the absorption was carried out, as usual, in a 1:5 dilution of the

serum.

TABLE 7.--Agglutinin absorptions with strain 528 indicate that it is of the abortus variety

1 See Table 1 for significance of the figures. The homologous antigens were used for the simple aggluti

nation tests.

The homologous antigen removed practically all agglutinins from serum 426. The slight agglutinins remaining in the lowest dilutions of the serum were due to the fact that heat-killed antigens were used in these tests. (See page 7 of the first paper of this bulletin for the behavior of antigens killed by heat.) Antigen 528 removed the agglutinins from serum 426 to a degree comparable with the homologous strain. The slight discrepancy falls within the limits of experimental error, whereas an absorbing antigen of the melitensis A variety would have left 12.5 per cent of the agglutinins, and the absorbed serum would have had a titer of 1:40. (See Chart 1 of the first paper in this bulletin.) Therefore the tests with serum 426 indicate that strain 528 is of the abortus variety. In spite of the fact that antigen 451 was killed by heat, it removed the agglutinins completely from serum 451. On the contrary, there remained agglutinins in dilutions up to 1:40 after absorption with strain 528, which shows that it does not belong to the melitensis A variety, and that it absorbs agglutinins from melitensis A serum to the same degree as strains of the abortus variety. (See Chart 1 of the first paper of this bulletin.) The results obtained in the tests with serum 426 were thus confirmed by the tests with serum 451.

Table 3 of the first paper of this bulletin shows that only two of purposes only the two varieties abortus and melitensis A need be country. Of 35 American strains 33 were of the abortus or melitensis A varieties. One bovine strain (457) was so closely related to the abortus variety that the distinction, which does not appear to be important, could not be made without a thorough study of the culture. The other American strain (481) which did not conform to the two common varieties is so different from them that the simple agglutination test alone would distinguish it. Hence, for practical purpose, only the two varieties abortus and melitensis A need be considered in identifying by the agglutinin absorption test a strain of Br. melitensis isolated in this country.

If a serum or strain belonging to one of the three Mediterranean varieties not yet found in this country (they are described in the first paper of this bulletin) should come to hand for identification, tests similar to those outlined above would reveal the irregularity, and further study would be necessary. It should be pointed out, however, that when the serum alone is available for identification the method given has certain limitations which are revealed by a study of Chart 1 of the first paper of this bulletin, where complete absorption by an antigen of a heterologous serologic group is shown in several instances. A study of the geographical distribution of strains reported in the first paper of this bulletin indicates, however, that the percentage of errors that might be made with the use of the given

method would be practically neligible for the identification of the type of infection with serums from patients in this country.

SUMMARY

A review is given of the literature on (a) cases of Malta fever which could not be traced to infection from goats; (b) the prevalence of Br. melitensis in cow's milk; (c) infectiousness of bovine strains of Br. melitensis for man; (d) titers considered indicative of Br. melitensis infections in man; (e) specificity of the serum reaction with Br. melitensis.

Five hundred human serums from patients suffering with a variety of diseases were tested for agglutinins specific to Br. melitensis. Fifty-nine, or 11.8 per cent, gave a definitely positive reaction in dilutions of 1 to 5 or higher.

One serum had a titer of 1:320, which would lead to an unquestioned diagnosis of Malta fever in regions where the disease is endemic. The patient was unaware of any possibility of having contracted an infection from goats. He was in the habit of drinking raw cow's milk. Absorption tests with the serum indicated that this patient was infected with the abortus (bovine) variety of Br. melitensis.

The titers of the remaining 58 serums which gave positive reactions varied from 1:5 to 1:40. The suggested explanation for these positive reactions is that the agglutinins were produced as a specific response to Br. melitensis ingested in cow's milk, although such an infection may not necessarily have caused a notable illness.

A description is given of the methods used for the identification of the variety of infecting organism.

NOTE. While this paper was in press, a serum from a human case of Brucella melitensis infection was received through the courtesy of Dr. Ernest C. Dickson, Stanford University Hospital, San Francisco, Calif. Agglutinin absorption tests showed definitely that the organism was not of the melitensis A variety, for agglutinins specific for the abortus variety remained in the serum after absorption with a saturating dose of the melitensis A variety. On the other hand, after absorption with the abortus variety no agglutinins remained for either the melitensis A or the abortus variety, indicating that the infection was with the abortus variety. No history of this case has yet been received.

STUDIES ON BRUCELLA MELITENSIS

V. HOGS AS A POSSIBLE SOURCE OF HUMAN INFECTION BY ALICE C. EVANS, Associate Bacteriologist, Hygienic Laboratory, United States Public Health Service

During the year 1924 two cultures of Brucella melitensis from human cases of Malta fever in the United States were sent to the Hygienic Laboratory for identification. Agglutinin absorption tests were carried out according to the method described in the first paper of this bulletin and both strains (Nos. 528 and 531, in Tables II and III of the first paper in this bulletin) were found to belong to the abortus variety.

Strain 528 was received from Dr. D. A. Gregory, of Sioux Falls, S. Dak., who supplied the following history: The culture had b en isolated from the blood of a patient who had had no contact with goats. He was a meat inspector who had been inspecting only hog carcasses for eight years. He also milked two cows daily. Doctor Gregory tested the serum of these cows and found no agglutinins for Br. melitensis. Infectious abortion was present in one of the dairy herds supplying the city, but the patient's milk supply was from a company which delivered only pasteurized milk.

Strain 531 was received from Dr. Kirby A. Martin, of the New Haven Hospital, New Haven, Conn. The history of the patient from whose blood the culture was isolated was obtained through the Courtesy of Dr. Millard Knowlton, director, bureau of preventable diseases, State of Connecticut Department of Health, Hartford, Conn. The patient had lived in New Haven for several years. He had not drunk goat's milk, nor was he in the habit of drinking cow's milk except in his coffee, and his boarding-house keeper stated that she boiled all the milk used in coffee. The patient was a carpenter by trade, but from the 1st of January until some time in June, 1924, he worked in a slaughterhouse, where his particular duty was "shaving ears and quartering" hogs. He became ill a few days after giving up this work.

Viviani has reported from Italy a human case of Br. melitensis infection which was traced to a sow which the patient had cared for after abortion.

The histories of the Sioux Falls, S. Dak., and the New Haven, Conn., patients suggest hog carcasses as the possible source of in

50587-255

57