September 8, 1922] 



SCIENCE 



281 



further, the results became less regular, in fact 

 the dilution containing 3 X 10"-^ ee. of the 

 original toxin killed only 5 per cent, of the 

 animals tested. 



In the current issue of Science (August 4, 

 1922, page 143), Dr. Stehle questions this ob- 

 servation. He states that one cubic centimeter 

 of toxin does not contain enough molecules to 

 supply a molecule of toxin for each cubic 

 centimeter of the final dilution. Indeed, he 

 rightly concludes that "the average 3 X lO-^i 

 cc. quantity of solution would contain no 

 toxin." We agree with Dr. Stehle's conclu- 

 sion, but we think it offers an explanation of, 

 rather than a contradiction to our findings. 



Our findings do seem unbelievable. Moreover, 

 in view of the fad that mice are not vei'v 

 reliable as test animals for toxicity, together 

 with the possibility of some error in dilution 

 due to inaccuracy of pipettes, we do not insist 

 on the accuracy of the figures and offer them 

 merely as the best proof of the apparent ex- 

 treme increase of the potency of toxin — a con- 

 dition unknown with other bacterial toxins. 

 As a matter of fact, we doubted the figures 

 obtained originally and repeated the experi- 

 ment many times. Finally, we cheeked our 

 finding by a calculation similar to that used 

 by Dr. Stehle (in our calculation we followed 

 the reasoning found in Walker's "Introduction 

 to Physical Chemistry," Macmillan, 1907, 

 pages 214-219). This calculation, in fact, 

 gave us the audacity to offer our figures whicli 

 before seemed to us ridiculous. According to 

 our interpretation, this calculation showed that 

 each cubic centimeter of the dilution of the 

 toxin to the eighteenth power might still contain 

 enough of the specific substance to kill mice 

 regularly (over 80 per cent, of animals thus 

 treated died within 48 hours and about 10 per 

 cent, more died within the next 24 or 48 hours), 

 whereas in the dilution to the tAventy-flrst 

 power, many of the one cubic centimeter por- 

 tions of the solution might not have contained 

 even a single molecule of toxin, which appar- 

 ently explains why only 5 per cent, of the ani- 

 mals injected died. When identical results were 

 obtained on at least three different batches of 

 toxin, obtained on different lots of culture 



medium, and on repeated tests on each of these 

 batches, we decided to publish these unusual 

 findings. I wish to take this opportunity to 

 add that we still expect to find some source of 

 constant error in our procedure or in our cal- 

 culation. We know of no good proof indi- 

 cating that the toxin may act as a catalyst, 

 nor are we willing without further good reason 

 to believe in the existence in the body of some 

 vital center consisting of a small group of 

 cells (not more than 100 and possibly less), 

 the injury to which would lead to death in 80 

 per cent, of animals tested. Furthermore, con- 

 tinuing the above calculation, we were forced 

 to conclude that the molecular weight of such 

 an active toxin as we have seemingly obtained 

 when computed on the basis of total solids 

 could be about 380 and when calculated as 

 protein (on the basis of total nitrogen) could 

 be no greater than 260, which makes it ex- 

 tremely difficult to see how such a simple sub- 

 stance can possess the degree of specificity in 

 respect to antitoxin neutralization which our 

 solutions demonstrated. 



In order to clarify the question, I would like 

 to sum up the arguments which seem to favor 

 the validity of our observations. 



(1) When 3 X lO-^^ ce. of the acidified cul- 

 ture filtrate of Bacillus botulinus is injected 

 intraperitoneally into mice of 17-21 grams, the 

 animals die with all the symptoms of botulinus 

 poisoning in mice. 



(2) When receiving this small amount of 

 toxin the animals die within 24-48 hours, 

 which is a typical incubation period for botu- 

 linus toxin. 



(3) Animals receiving protective injection of 

 a homologous (type A) antitoxin invariably 

 survive the injection of the amount many mil- 

 lions of times greater than 3 X 10"^' cc. of this 

 active solution. 



(4) The animals receiving large amounts of 

 botulinus antitoxin type B (heterologous) are 

 not protected even against a single dose of 

 3 X 10-'* cc. of acidifiec} type A toxin. 



(5) The potency of acidified toxin is de- 

 stroyed by a very short exposure to heating at 

 80 degrees C. 



