152 MICROBES AND TOXINS 



suggested for the mosaic disease of the tobacco plant), of a 

 fluid, living contagium which is literally invisible? This 

 hypothesis of " soluble viruses " has been put forward, but 

 hitherto no positive proof has been given. 



The essential procedure in the definition of the ultra-micro- 

 scopic viruses is filtration ; it is the current method of isolation. 

 The liquid containing the virus is passed through a filter for 

 example, vaccinal pulp rubbed up in water; the filtrate 

 collected is virulent, and with it attempts at cultivation can be 

 made. The filters employed are of the well-known forms ; 

 the majority are hollow bougies, like those which are used for 

 filtering drinking water ; they are made of porcelain (Chamber- 

 land filter), of infusorian earth (Berkfeld filter), of asbestos, 

 charcoal, plaster, etc. 



These filters do not act towards microbes as does a sieve used 

 to sift seeds of unequal size, or as the metallic grids used for 

 separating sand of different coarseness. It would not be right 

 to conceive the large microbes as being kept back because they 

 are bigger than the meshes, whereas the little ones pass easily 

 through, just as the smaller fishes pass through the meshes of 

 a net. Even the bacteria of average size, such as the vibrio of 

 cholera, are smaller than the pores of our filters, and their size 

 would permit them, to use the simile of Duclaux, to pass 

 through, as a train passes through a tunnel, without rubbing 

 against the walls ; what keeps them back is that they are held 

 against the walls by the capillary pressure. 



Filtration is not a simple mechanical operation, various 

 factors act in it : the quantity of the virus, the motility of the 

 microbe, the pressure, the degree of dilution, the nature of the 

 liquid more or less albuminous, the temperature, the duration 

 of the filtration, and the texture of the filter. All these factors 

 have to be taken into account in these experiments. As a rule, 

 several nitrations are performed one after another, the first, 

 rougher and more rapid, prepare for the final one by freeing 

 the liquid from particles which block up the pores. One must 

 especially avoid having thick albuminoid substances present ; 

 they soon cover over the surface of the filters. By prolonging 



