JlLY 3, 1903.] 



SCIENCE. 



the red cells become fused and lost their 

 independence, but the capillary walls have 

 been injui"ed in a definite and unmistakable 

 manner. The extravasations take place not 

 by diapedesis, as is now believed, but 

 through actual rents in the -walls. The 

 explanation of the rents is of much in- 

 terest. That they are not simple ruptures 

 seems proved by the disappearance, as if 

 through solution, of the parts of the walls 

 at the site of the escape of corpuscles. 

 The solution of continuity is one-sided, and 

 in some instances is attended by a displace- 

 ment of the adjacent endothelial cells, 

 which are pushed outward, away from the 

 vessel, by the force of the escaping blood. 

 The escape of corpuscles by dissolution of 

 the vascular walls is limited to capillaries 

 and small veins. When acted upon by 

 venom the vessels show irregular bulging 

 of the walls, by ricin a localized dilatation 

 or congestion of the vessels, which give rise 

 to a glomerular appearance ; but in a small 

 number of places only does the extravasa- 

 tion occur. It is probable, therefore, that 

 the points of injury to the vascular coat 

 are many, but in a part only of these does 

 the vessel give entirely away. 



The existence, then, of a substance hav- 

 ing an especial affinity for vascular endo- 

 thelium may be considered as proved for 

 snake venom and ricin. For this principle, 

 if principle it be, we have proposed the 

 name of hamorrhagin, and we look upon 

 it as a cytolysin for endothelial cells of 

 blood vessels, the injury and destruction 

 of which is the direct cause of the escape 

 of blood into the surrounding tissues. It 

 remains to be shown whether such ha-mor- 

 rhagins are of common occurrence in na- 

 ture, and to what extent they may play a 

 part in animal pathology-. I regard this 

 as a most promising field of future ex- 

 ploration, for I conceive that bacteria and 

 other pathogenic microorganisms probably 



produce endotheliolysins, and the action of 

 luvmorrhagins may be the immediate cau.se 

 of the extravasation of blood in purpuric 

 states and some other forms of so-called 

 parenchymatous haemorrhage. 



Should this view of the causation of 

 hemorrhage be supported by future stud- 

 ies we may well consider the possibility of 

 preparing experimentally an antidotal 

 agent for the principle involved in its 

 production. We already have at hand 

 certain observations bearing upon this ques- 

 tion. You are all familiar with antivenin 

 as prepared by Calmette and Fraser as an 

 antidote to venom. Although contrary to 

 the generally prevalent opinion, it may be 

 affirmed that the value of the commercial 

 antivenin is in inverse ratio to the hivmor- 

 rhage-producing power of the venom. This 

 fact arises from the consideration that in 

 the preparation of antivenin, cobra venom 

 chiefly is employed, and it is almost devoid 

 of the hemorrhagic principle. 



Venoms which cause much hemorrhage 

 exert a very destructive local effect upon 

 the tissues. For this reason very little suc- 

 cess has attended the efforts to produce an 

 anti-toxin for viper venom. On this ac- 

 count, Dr. Noguchi and I have sought a 

 means of modifying viper venom so as to 

 remove the locally destructive effect and 

 yet leave the value of the hemorrhagic 

 principle with which an immunity might 

 be established. Through the use of hj^dro- 

 chloric acid we have succeeded in getting 

 rid of the local effects, including the hem- 

 orrhage, and yet have preserved the com- 

 bining value of the hemorrhagic principle, 

 so that successive injections in animals of 

 modified venom could be carried out. The 

 .serum of these animals contains an appre- 

 ciable quantity of anti-hemorrhagin as well 

 as other antidotal principles, and is capable 

 of neutralizing the local eft'eets of rattle- 

 snake venom. 



