294 DISTURBANCES OF CIRCULATION 



most important practically, but many poisons, such as phosphorus, 

 formaldehyde, yhytotoxins (ricin, abrin, and crotin), and zootoxms 

 (snake venoms) cause numerous and abundant hemorrhages. For- 

 merly, the tendency was to ascribe hemorrhages from the above causes 

 to mechanical injury of the vessels by thrombi, or by emboli of ag- 

 glutinated corpuscles, but the work of Flexner^^ has shown that 

 venoms cause hemorrhages by injuring the capillary walls, so that 

 actual rents are produced by the intravascular pressure, and it seems 

 highly probable that hemorrhages are produced by other chemical 

 substances in a similar way. We may, therefore, refer such hemor- 

 rhages to an endotheliotoxic action of the poison, or to a solvent effect 

 upon the intercellular cement substance. In the case of ordinary 

 chemical poisons the endotheliotoxic action is not specific, but with 

 some of the toxins it seems to be quite so; for example, rattlesnake 

 venom contains an endotheliotoxic substance ijiemorrhagin) , which 

 seems to be a specific poison for endothelium, and which is the most 

 dangerous constituent of the venom. If we immunize animals against 

 tissues containing much endothelium (e. g., lymph-glands), their serum 

 will be found to contain endotheliotoxins, so that when tliis serum 

 is injected subcutaneously into a susceptible animal, large local hemor- 

 rhages result; if injected into the peritoneal cavity, there results 

 marked desquamation of the endothelial cells, which soon undergo de- 

 generative changes (Ricketts).^^ It is quite probable that the bac- 

 terial poisons that cause marked hemorrhagic manifestations likewise 

 contain endotheliotoxins, although this matter does not seem to have 

 been investigated. 



Even hemorrhage by diapedesis seems to be due to, or at least 

 associated with, chemical changes in the capillary walls, for Arnold-*' 

 found that when capillaries from which diapedesis had occurred 

 were stained by silver nitrate, dark areas were found between the 

 endothelial cells. As silver nitrate is a stain for chlorides, and dark- 

 ens intercellular substance because it is rich in sodium chloride 

 (Macallum), it is probable that there is an increase in the amount 

 or a difference in the method of combination of the chlorides of the 

 cement substance between the endothelial cells at the places where 

 red corpuscles escape. M. H. Fischer-^ suggests that diapedesis 

 results from a change in the endothelial cells, which under the inthi- 

 ence of acids or other agents of metabolic origin become excessively 

 hydrophihc, swell up, and become so softened that corpuscles may 

 pass directly through the cell, just as a drop of mercury can pass 

 through a sufficiently soft jelly without leaving a hole in the jelly. 



Ilemori'hage in cachetic conditions is often ascribed to changes 

 in the vessel-walls due to malnutrition, but it is diffi.cult to imagine 



" Univ. of Penn. Med. Bull., 1902 (15), 355. 

 " Trans. Chicago Path. Soc, 1902 (5), 181. 

 20 Viichow's Arch., 1875 (()2), 157. 

 2' "Nephritis," New York, 1912, p. 78. 



