310 CAUSES OF VARIATION 



immunity that might be expressed as 400 when compared with 

 the normal as I. The increase was slow at first, but had reached 

 twenty times the normal within eight days, after which it rose 

 rapidly. 1 



Still more significant are the facts connected with the prepa- 

 ration of diphtheria antitoxin from the horse. In this process 

 Roux first mixes the filtrate of the bacillus with iodin to reduce 

 its virulency. Only one quarter of a cubic centimeter of this 

 " iodized toxin" is used for the first injection, but on the thir- 

 teenth day a full cubic centimeter is used. u On the seventeenth 

 day one fourth cubic centimeter of the pure toxin is injected, 

 and this is gradually increased in amount till on the forty-first 

 day 10 cc. is injected and on the eightieth day no less than 

 250 cc. The virulency of the last dose must have been some 

 five thousand to ten thousand times greater than that of the first 

 dose," showing the extent to which the constitution of the horse 

 had become acclimated to the poison. 2 



Davenport and Neal reared two lots of stentors, one in pure 

 water, the other in a solution of 0.00005 P er cent HgCl 2 . After 

 two days both lots were put into a killing solution (o.ooi per cent 

 HgCl 2 ). The resistance period of the lot reared in plain water was 

 on the average but 83 seconds, while that of the lot grown in the 

 weak solution was 304 seconds. Other experiments gave like 

 results, and from these investigations the principle was deduced 

 that the resistance period varies directly with the strength of the 

 solution in which the protoplasm has been ciiltivated? As would 

 be expected, if the culture solution should chance to be only just 

 outside the lethal point, it might permanently weaken without 

 destroying life, in which case successful acclimatization would 

 not follow ; the animal would succumb through weakness, and 

 the principle stated would not hold good. 



In these experiments no deaths occurred, 4 hence the results 

 must be attributed entirely to alteration of the protoplasm and not 

 at all to selection. All these considerations point conclusively 



1 C. B. Davenport, Experimental Morphology, Part I, p. 29. 



2 Quoted from Vernon (Variation in Animals and Plants, p. 387), who quotes 

 from Crookshank's Text-Book of Bacteriology, 1896, p. 58. 



8 C. B. Davenport, Experimental Morphology, Part I, p. 30. 

 4 Ibid. p. 31. 



