200 ANNALS OF THE MISSOURI BOTANICAL GARDEN 



[Vol. 10 



others with pores or lacunae so fine as to prevent or greatly 

 inhibit the passage of such particles. A second phase of the work 

 involved a careful technique in the use of the ultrafilters. A 

 third phase required the inoculation of healthy plants with the 

 various filtrates obtained in order to determine the percentage 

 of dilution of the particles, if possible. Finally, some method of 

 standardization of the filters was necessary whereby their capacity 

 to permit or prevent the passage of particles might be related to 

 colloidal particles of known, or approximately known, sizes. 

 It will be unnecessary to go into the details of these experiments. 

 Two aspects of the results require emphasis. It was possible to 

 find a filter, in this case a cylindrical, porcelain atmometer cup, 

 which in a given interval of time, at a given pressure, and at the 

 reaction of the diseased tobacco juice, permitted only a relatively 

 small number of the infectious particles to pass through. Con- 

 siderable dilution of the juice from the standpoint of these 

 particles was therefore effected. This was shown by a reduction 

 in the incidence of infection from 90-100 per cent in the usual 

 controls to 5-20 per cent in the case of the porcelain filter only 

 partially permeable to the infective particles. 



Standardization of the filters was accomplished by the use of 

 hydrophilic colloids of biological origin. These were selected 

 in preference to sols of inorganic origin, such as gold sols, because 

 of possible greater complications (when employing the latter) 

 arising from electrical relations. The series of organic compounds 

 employed included casein, gelatin suspensions, lactalbumin, 

 hemoglobin, and dextrin. Fortunately, this series sufficed. 

 The results indicated that the hemoglobin content of a standard 

 hemoglobin solution prepared from fresh ox blood was diluted 

 to a very considerable degree in passing through the same filter 

 which obstructed to a large degree the passage of the infective 

 particles. In experimenting further with substances on either 

 side of hemoglobin, in reference to size range, it was clear that 

 from such filtration experiments the deduction must be made that 

 the infective particles of mosaic disease approximate in size those 



of a fresh 1 per cent hemoglobin solution. 



The best data on the size relations of hemoglobin particles 

 indicated a diameter of approximately 30 (X(x. It is presumable 



