this bulletin. Reference to that paper will be made frequently in the following pages.

Strains of Br. melitensis of human origin were used for the antigen. Some of the serums were tested with strain 451, which is of the melitensis A variety, and some were tested with strain 455, which is of the abortus variety. It was shown in the earlier publication that the simple agglutination test does not differentiate between these two varieties of Br. melitensis, and control tests with the antigens used showed that duplicate tests with the two antigens gave results as nearly alike as duplicate tests with either one.

RESULTS

Of the 500 serums tested, 59, or 11.8 per cent, were positive in dilutions of 1 to 5 or higher. (Sedimentation of 75 per cent or more of the antigen was considered a positive reaction.) The titers for those serums which gave a positive reaction are given in Table 2, together with the diagnosis of those cases for which a report could be obtained.

The data presented in Table 2 may be summarized as follows:
43 serums were positive in the 1:5 dilution.
11 serums were positive in the 1:10 dilution.
2 serums were positive in the 1:20 dilution.
2 serums were positive in the 1:40 dilution.
1 serum was positive in the 1:320 dilution.

DISCUSSION OF THE POSITIVE RESULTS

One of the serums tested gave a positive reaction in dilutions up to 1:320. A second sample of serum was obtained, and the positive results were confirmed with the inactivated serum as well as with the fresh serum. In our search through the literature the highest. titer considered the minimum necessary for conclusive evidence of Malta fever which was found was the 1:200 requirement of Fici. Hence, in regions where it is endemic, a diagnosis of Malta fever in this case would be unquestioned. The case was ambulatory, and the medical officers who made the examination found no clinical symptoms which would lead to the diagnosis of any other disease. Following the positive finding in the serologic test, they believed it to be a case of true Malta fever. According to the patient's statement he was in the habit of drinking raw cow's milk. The patient lived in Virginia, and so far as he knew he had never drunk goat's milk nor eaten cheese made from goat's milk, and he had never come in contact with goats. The serologic type of the infecting organism was determined by the agglutinin absorption test and found to agree with the abortus variety. The data will be given further on.

TABLE 2.-Titers of the positive serums and diagnoses of the cases

928.

SERUMS WHICH WERE HEATED TO 56° C. FOR 30 MINUTES

Dementia præcox.

Chancroid; syphilis (?)

Iritis; hypertension, nephritis.

Constitutional psychopathic inferiority; chronic bronchitis; mental deficiency.
Mental deficiency; effort syndrome; neurosyphilis (?)

1:5 Psychosis, undiagnosed; pulmonary tuberculosis, apparently arrested.

935.

1:10

Old fracture of left leg; hysteria.

Chronic amygdalitis; valvular heart disease; suppurative otitis media.

Chancroid.

Dementia præcox.

Mental deficiency; acute pharyngitis.

Chronic rhinitis; pulmonary fibrosis.

Lues.

Under the conditions of the test as applied to the 500 serums here considered, inactivation of the serum did not reduce the percentage of positive results. In fact, it happened that the inactivated serums gave a higher percentage of positive reactions than those which had not been inactivated. Of the 396 heated serums, 49, or 12.37 per

cent, gave positive reactions, whereas 10 of 104, or 9.61 per cent, of the unheated serums gave a positive reaction.

The significance of the positive reactions in low dilutions of the serums is problematical. There are four possible explanations for the reactions: (1) The reactions may not be specific; (2) the agglutinins may have been acquired in a secondary manner by absorption in the intestines from the agglutinins present in milk that had been ingested; (3) the agglutinins may have been produced as the result of an infection some time in the past; (4) they may indicate a present infection. It may be that one explanation would apply in some cases and another explanation in other cases.

It is possible that some of the reactions may not be specific. The tests were made before the cross reactions between Br. melitensis and Bact. tularense had been observed, hence the possibility that some of the patients giving a slightly positive reaction had at some time had tularaemia was not eliminated. Possibly there are other diseases which produce agglutinins that will react with Br. melitensis. But it is to be borne in mind that when nonspecific agglutinins reacting with Br. melitensis have been reported in the literature, in no case has the possibility been considered that they may have been the result of infection with the abortus variety of the organism.

In view of the fact that the abortus variety of Br. melitensis is common in cow's milk, the most probable explanation of the positive reactions in low dilutions of the 58 serums is that they are the result of ingestion of the specific organism. Nicolle and Conseil showed that the ingestion of Br. melitensis may give immunity. Human subjects vaccinated by feeding killed cultures were protected against subsequent injection with living organisms, although there were no agglutinins in the blood. Of the seven human subjects who drank raw cow's milk infected with the abortus variety of Br. melitensis in Cooledge's experiments, there was an increased antibody content in the serums of five, although all remained in good health. In one case the titer of agglutinins reached 1:40. The obvious conclusion was that the increase in antibody content of the serum was due to the ingestion of infected milk. The patients whose serums showed agglutinins for Br. melitensis in the tests under discussion could have acquired these antibodies in the same way as the subjects in Cooledge's experiments.

It does not appear reasonable to assume that any considerable titer of agglutinins could be acquired by absorption in the intestines from agglutinins in ingested milk; for, according to Smith, Orcutt, and Little, the titer of agglutinins in cow's milk is rarely higher than 1:40. Two of the serums (Nos. 845 and 1385) gave positive reactions in the 1:40 dilutions. It is difficult to conceive that a man

could passively accumulate in his serum as high a titer of agglutinins as are commonly present in milk.

A more reasonable explanation for the presence of agglutinins specific for Br. melitensis in human serums in cases of disease other that Malta fever is that they are the result of a mild infection, past or present.

That a Br. melitensis infection could occur without any notable illness was shown by the works of Shaw, Vaccaro, and San Roman. By agglutination reactions and by cultivation of the organism from blood or from the urine, these authors (previously quoted) showed that apparently healthy persons living in regions where Malta fever is endemic may be carriers of Br. melitensis infection. The fact that Malta fever is very little known in this country leads to the belief that the bovine type of Br. melitensis is less virulent for man than the caprine type. If strains of the variety known to be highly infectious for man can produce a very mild type of disease, it seems quite probable that strains of the presumably less virulent variety also may produce infections so mild that they are commonly ignored or undiagnosed.

In regions where Malta fever is endemic most authorities would consider the titer of 1 to 40 found in the case of the serums from two of the patients as indicative of a melitensis infection, and the lower titers of 1 to 5, 1 to 10, and 1 to 20 would be regarded by most workers as sufficient to suggest such an infection.

The positive results obtained with serums Nos. 1385 and 1387, having titers of 1 to 20 and 1 to 40, respectively, were reported to the medical officer in charge of the subdistrict office of the United States Veterans' Bureau which submitted the samples. A letter in reply stated that the positive Malta fever agglutination reactions had been confirmed.

The results from a group of serums from Alexandria, Va., may be considered together and compared with those from the remainder of the serums. At the time when the serums were tested, the milk supply of Alexandria was not under municipal control. Contagious abortion was known to exist on the farms in the surrounding country. Very likely a considerable proportion of the Alexandria patients were using raw cow's milk containing the abortus variety of Br. melitensis. Out of 51 serums from Alexandria, 12, or 23.5 per cent, were positive in a dilution of 1 to 5 or higher; out of the remaining 449 serums, 47, or 10.46 per cent, were positive. This comparison suggests again that Br. melitensis in raw cow's milk may be responsible for the positive results obtained when human serums are tested for agglutinins specific to that organism.

METHODS USED FOR THE IDENTIFICATION OF THE INFECTING ORGANISM

Since the simple agglutination test can not differentiate the abortus and melitensis A varieties, it is necessary to resort to the agglutinin absorption test to identify an infection. If a culture is not available, a presumptive identification can be made with the serum alone, providing it has a titer of 1 to 80 or higher. The data by which the infections were identified as abortus will be given for three cases, one in which both serum and culture were available, one in which only the serum was available, and one in which only the culture was available. The technic employed in carrying out the agglutinin absorption tests is described in the first paper of this bulletin.

The complete data identifying strain 455, from the Baltimore case reported by Keefer, are presented in Tables 3, 4, and 5. The history of the strains used for comparison and their serologic classification are given in the first paper of this bulletin. (Strain 426 is Feusier and Meyer's human strain of the abortus variety; strains 456 and 460 are from cases of bovine abortion; strain 428 is a European human strain of the melitensis A variety; and strain 451 is a human strain of the melitensis A variety from Phoenix, Ariz.)

TABLE 3.-Agglutination of the homologous antigen in the serum of the Balti more patient after absorption by various strains of bovine and human origin

1 See Table 1 for significance of figures.

Absorptions were accomplished by adding 0.2 c. c. of serum to 4.8 c. c. of antigen of a density of 5,000 parts. per million.

TABLE 4.-Agglutination of the homologous antigen in the serum of a rabbit injected with the bovine strain 456, after absorption by various strains of human and bovine origin

1 See Table 1 for significance of the figures.

2 Absorptions were accomplished by adding 0 2 c. c. of serum to 3.8 c. c. of antigen of a density of 40,000

parts per million.