In trying to obviate this result I found, perhaps contrary to expectations, that the exudation liquid should be exposed to the air for a few hours before it is injected into a subsequent animal. This result was contradictory to the effect which Pasteur had found to be exercised by atmospheric oxygen on the virulence of microbes, and it requires at least some provisional explanation. The microbes of cholera differ from a certain number of other microbes in that they stand in need of a free and abundant access of air for growing and multiplying quite satisfactorily. They are deprived of this condition in the peritoneal cavity of an animal. It is possible, therefore, that a certain opposition between the maintenance or development of virulence on the one hand, and a lowering in vitality on the other, takes place while they are cultivated there, and a respite must be given them between each successive ‘passage’ through the Guinea pig by leaving them for a time in the free atmosphere. Be that explanation true or not, the result is that under such conditions the successive animals inoculated with the virus do succumb, and even in a shorter and shorter time, after the inoculation, the microbe apparently undergoing under such a treatment a progressive increase in virulence. A similar development up to a certain stage was observed by Pasteur when transferring the rabies virus from rabbit to rabbit. The last difficulty that presented itself was the following: The exudation liquid which is found in the peritoneal cavity post mortem varies in quantity; sometimes it is inconveniently large and diluted; sometimes, on the contrary, so scanty that it becomes difficult to collect and transfer it to another animal. I found that this variation stands in connection with the size of the animal, so that a diluted exudation fluid can be concentrated by injecting it into a small animal, while a too much concentrated exudate is rendered more dilute by transferring it to an animal of a larger size. Thus, by the initial use of more than a fatal dose, by alternating cultivation in an animal with exposure to air, and by attention to the size of the animal employed, a material was obtained which, as mentioned, increased in intensity from the first and proved fatal to animals in a shorter and shorter time after inoculation. Later the virus reached a stage when it killed a Guinea pig of three hundred and fifty grammes weight in eight hours. After that, in each further inoculation the time of eight hours remained stationary, showing that the virus has reached the condition of a ‘virus fixe.’ These experiments were conducted by me in the Pasteur Institute, in 1889 to 1893, simultaneously on the cholera microbe and on the bacillus of typhoid. The two exhibited a number of common features in their nature, and the results as above detailed for the cholera microbe were found valid for the typhoid bacillus also.