RICKETTSIA IN TISSUES 191 



from the blood vessels by migratory endothelial cells. The 

 analogous lesion in the central nervous system, particularly 

 in guinea-pigs, hardly permits of any other explanation. 



(w) Recapitulation of the vascular lesions: The earliest effect 

 is upon the endothelial cells, producing swelling and prolifera- 

 tion and a degree of injury resulting in thrombosis. Peri vascu- 

 lar collections of cells result from the further proliferation of 

 cells which we believe to be induced by the presence of rickett- 

 sia within cells of endothelial origin. Other cellular reactions 

 occur as in other disease processes; there is accumulation of 

 polymorphonuclear leucocytes, lymphoid and plasma cells, 

 and mast cells. Eosinophilic leucocytes are present in the 

 perivascular lesions in a few of our cases. 



The presence of bodies indistinguishable from rickettsia has 

 been demonstrated in twenty-one of the twenty-eight pieces of 

 skin excised ante-mortem. Our failures with this material is 

 directly attributable to faulty technic in embedding and 

 staining. 



Of the autopsy material, rickettsia have been found in the 

 skin of twenty-five cases which represents every case where the 

 post-mortem examination was made before the thirteenth day 

 of the disease and while the body was in a fresh condition. In 

 two cases rickettsia were found as late as the fifteenth day. 

 Other organs in which we have demonstrated rickettsia to 

 our own satisfaction are kidneys two cases, femoral vein one 

 case, testes and adnexa five cases, and the brain seven cases. 

 We have given a relatively short time to the search for rickett- 

 sia in the central nervous system as compared with that de- 

 voted to the skin in the search for organisms. Owing to the 

 small size of the involved vessels in the brain and the presence 

 of many granule bearing cells of neuroglia and endothelial 

 origin which make difficult the recognition of endothelial cells 

 in situ in capillaries, we have preferred to depend upon the 

 skin lesions for our numerical evidence in regard to the con- 

 stancy of rickettsia in typhus lesions. 



The possibility of mistaking granules of various sorts for 

 rickettsia impels us to state our criteria for identification. 



