1921] 
DUGGAR & KARRER—SIZES OF MOSAIC DISEASE PARTICLES 351 
of next lower porosity, which proved to be the cylindrical at- 
mometer A. After the standardization was carried out, as 
discussed later, bearing out the importance of the work with 
these 2 cups a second series of inoculations was made with new 
filtrates of diseased juice through these 2 cups. Twenty tobacco 
plants were inoculated with each filtrate and numerous unin- 
oculated controls were kept in adjacent plats. Between 10 and 
18 days after the inoculations 19 plants developed the disease 
among those inoculated with the spherical cup filtrate and 5 
plants became diseased from the filtrate of the cylindrical cup 
A. Thus the previous test was admirably confirmed and even 
better indications were afforded that a small number of infected 
particles pass the cylindrical atmometer cup. 
STANDARDIZATION OF THE FILTERS 
In attempting to standardize the filters which had been em- 
ployed in this work there was the possibility of using the same 
filter after a thorough cleansing, or the possibility of employing a 
similar filter assumed to be of equal porosity. It became evident 
that direct standardization of the original filter employed was 
essential where this could be done without fear of change or 
injury. Consequently the first step in the standardization 
involved a thorough cleansing of the filters employed. The 
standardization process was delayed until after the results of 
the inoculation in order to limit the amount of unnecessary work. 
From the inoculation experiments it was clear that the sizes of 
the infective particles must lie between the pore sizes of the 
spherical atmometer cup used and that of the most porous 
cylindrical tube A, and probably close to the pore sizes of the 
latter. At the time we had no idea of the relation of these 
sizes, and had not the subsequent standardization experiments 
indicated that these two pore sizes were sufficiently close to- 
gether, it would have been necessary to seek further for a porous 
filter of intermediate pore dimensions. 
To avoid difficulties arising from adsorption or from the 
possible action of electrolytes derived from the filters, it was 
determined to use organic sols rather than metallic sols for 
