674 BACTERIVM 



antigen, which was different from the Friedlander somatic antigen, but that each had 

 a distinct capsular antigen. One of the capsular antigens was specific ; one was related 

 to pneumococcus Type II, and one to both pneumococcus Type II and Friedlander Type B. 

 GosMngs and Snijders (1936), Wielenga (1937), and Osterman and Rettger (1941) like- 

 wise found considerable antigenic complexity in the aerogenes group. According to Goslings 

 and Snijders, even the somatic antigen varies in different strains of Bad. aerogenes, being 

 sometimes similar to that of the rhinoscleroma bacillus, sometimes to the ozaena bacillus, 

 and sometimes to neither. The capsular antigen of one strain was similar to that of 

 Friedlander Type B, but the rest were different from any of the A to F capsular antigens. 



Studying capsulated strains of diverse origin, Edwards (1929) was able to divide 

 them into three groups by agglutination : (1) strains, chiefly from human pneumonia 

 and i^leurisy, identical with JulianeUe's Type A ; (2) strains from metritis of mares, 

 together with some strains of Bad. aerogenes isolated from soU, water, or milk, and an 

 occasional strain of hiuuan origin, identical with Type B ; (3) some soil strains of Bad. 

 aerogenes and an inguinal granuloma strain. JuUanelle (1930) obtained evidence that 

 most Type A strains were of human, and most Type B strains of animal origin. 



It is clear that the relationship of these various capsulated organisms to each other 

 will not be understood until a sufficient number of fully representative strains of each 

 group have been examined by serological and other methods. 



Pathogenicity. — Organisms of this group are frequently encountered in diseases 

 of the respiratory tract in man. As a rule they appear to act chiefly as secondary 

 invaders, but it is possible that they are sometimes responsible for the primary 

 disease (see Bhatnagar and Singh 1935). They are not uncommonly isolated from 

 suppurative processes of various kinds throughout the body ; occasionally they 

 invade the blood stream, and give rise to septicaemia. Jampolis and his colleagues 

 (1932) record a hospital outbreak of infectious diarrhoea in infants, accompanied 

 by severe constitutional disturbance and a high case mortality. Friedlander's 

 bacillus was isolated from the nasal secretions, stomach contents, and stools of 

 most of the patients, and appeared to be the primary cause of the condition. 

 Webster (1928, 1930) has described a spontaneous respiratory epidemic in mice 

 caused by Friedlander's bacillus ; Edwards (1928) has isolated this organism from 

 metritis of mares ; while Wallace, Cahn, and Thomas (1933) have found it in a 

 paralytic disease of moose. It has been held responsible for the disease of human 

 beings known as granuloma inguinale ; but according to Anderson, DeMonbreun 

 and Goodpasture (1945), this disease is due to a different organism, cultivable in 

 embryonic egg yolk, for which the name Donovania granulomatis is suggested. 



Experimentally the virulence of Friedlander bacilli is subject to considerable variation. 

 The smooth type is pathogenic, the rough type non-pathogenic for laboratory animals 

 (Toenniessen 1914, JuUanelle 19266, Webster 1928). Other variants have differing grades 

 of pathogenicity. Many observers have drawn attention to the differences in pathogenicity 

 between Friedlander's baciUus, the ozsena bacillus, the rhinoscleroma baciUus, and other 

 members of this group, and have endeavoured to make use of these differences in classifica- 

 tion. Unfortunately the protocols in the literature are not sufficiently definite to enable 

 one to assert that there is any constant difference between these organisms. It is therefore 

 possible that the actual virulence of any strain depends upon the proportion of smooth 

 bacilli in the culture. When this consists entirely of the smooth type, it is usually virulent ; 

 in proportion as the rough type appears, the virulence falls ; and when the smooth has 

 been replaced completely by the rough type, the culture proves avirulent. On the whole 

 it would appear that Friedlander Types A and B bacilli are highly pathogenic to mice 

 when injected intraperitoneally, while Friedlander Type C, ozsena, and rhinoscleroma 

 strains are usually non-virulent. 



