INFECTIOUS DISEASES OF MICE 423 



extensive phagocytosis of red blood cells by the pulp cells. In splenec- 

 tomized mice (2) the course is generally more acute than in normal mice, but 

 varies with the age of the animal. In those younger than 3 months, infec- 

 tion is chronic for a period as long as 4 months, with intermissions and 

 relapses. The red cell count may fall to approximately 2,000,000 per 

 cu. mm., and mononucleosis up to 18 per cent may occur. Most of the 

 animals recover. On the other hand, infection in splenectomized mice 

 6 months or more of age is usually acute and terminates fatally from 3 to 

 12 days after B. miiris appears in the blood. 



The function of the spleen in relation to immunity in bartonellosis is of 

 considerable interest and has been the subject of much investigation in rats 

 ^75- 76. 77. 64, 166, 4, 332), but a satisfactory explanation has not yet 

 been obtained. Weinman (332) has shown that the anemia has no apparent 

 relationship to immune bodies in the serum, but is hemolytic in type and 

 due to direct action of the organism on the erythrocytes. In mice, as in 

 rats, the spleen seems to offer some barrier to infection, but here also its 

 role in immunity is not clear. By partial ablation, Adler (2) has demon- 

 strated that about 30 per cent of the spleen is sufficient to control infection. 

 Repeated injections of B. muris cause splenomegaly, but an immunity per- 

 sisting indefinitely after splenectomy does not result. It seems likely that 

 the immunological aspects of this disease will not be clarified until cultural 

 methods are employed and larger numbers of virulent organisms are used for 

 immunization. 



Tyzzer (296) has recently demonstrated that the natural bartonella 

 infection in white-footed deer mice {Peromyscus leucopus) can be trans- 

 mitted to splenectomized normal white mice. A severe and occasionally 

 fatal anemia results. If the splenectomized mice suffer from infection with 

 Eperythrozoon coccoides, however, the bartonella infection is prevented from 

 developing or suppressed if already established. On the other hand, the 

 mouse eperythrozoon may be transmitted to the splenectomized vole (Micro- 

 ius pennsylvanicus) , but does not interfere with natural bartonellosis in this 

 animal. Such phenomena of interference, first noted with viruses, have as 

 yet been observed in relatively few instances and the mechanism is not 

 known. The phenomenon must be kept in mind, however, since, as Tyzzer 

 points out, experimental results may be altered by the presence of unrec- 

 ognized infection. 



Etiology. — The bartonellae of rodent origin are small, pleomorphic, 

 gram-negative bacteria which vary in width from o.i to 0.5 ^i and in length 

 from 0.5 to 2.0 ^l. Variations from coccoid to bacillary form mav be seen 



