8 



Scientific Proceedings (30). 



clamp. Then both halves of the clamp are united and clamped. 

 I believe that when we deal with small blood vessels it is easier to 

 hook the walls on the pins than to turn them back like a cuff as 

 is done in Crile's cannula. When the clamp is closed, both blood 

 vessels are connected with the endothelial surfaces. 



I have performed several operations on dogs, uniting the 

 femoral or cervical vein of one dog to the femoral artery of an- 

 other. The transfusion was kept up for over half an hour, until 

 the donor was practically exsanguinated. There was no clotting, 

 leakage, or any other defect in the clamp. 



6 (344) 



The further separation of antitoxin from its associated proteins 



in horse serum. 



By EDWIN J. BANZHAF. 



[^From the Research Laboratory of the Department of Health, New 



York Cit)'.~\ 



The literature concerning means of purification of anti-bodies 

 and their chemical characteristics has been thoroughly reviewed 

 by Gibson,^ Ledingham,^ Banzhaf and Gibson,^ and Brieger and 

 Kraus.* 



Stark* was the first to report that by heating for one hour at 

 56° C, ovalbumin could be converted into a body, which, because 

 of its precipitation and solution reactions, and also its composi- 

 tion, was obviously a globulin. Later Noll® showed the same to 

 be true of albumin in rabbit, dog and horse serum. 



My experiments were to ascertain the resulting conditions after 

 heating antitoxic horse serum, citrated plasma, and Gibson's con- 

 centrated and partially purified antitoxic globulin solution. 



An antitoxic serum by the Gibson method ^ gave the following : 

 An elimination of 23 per cent, protein and an increase of antitoxic 

 units per gram protein of 30 per cent, over the native serum. A 



^ Journal of Biolog. Chem., I, p. i6i, 1906. 



^Journal of Hyg., vii, p. 65, 1907. 



^Journal of Biolog. Chem., iii, p. 253, 1907. 



^ Berl. klin. Woch., xliv, p. 946, 1907. 



^ Zeiischr. f Biol., xl, p. 494, 1900 (new series, vol. 22). 



* HofmeisUr's Beitrage,'iv, p. 563, 1 904. 



