[Vor. 8 
354 ANNALS OF THE MISSOURI BOTANICAL GARDEN 
DISCUSSION 
From the results presented it would seem clear that with 
approximately equal pressures and equal time intervals the in- 
fective particles of the juice of tobacco plants affected with the 
mosaic disease possess about the same capacity to pass through 
the pores of porcelain filters as do the colloidal particles of fresh 
hemoglobin prepared by standard methods. No determinable 
dilution or loss of infectivity of the tobacco juice was occasioned 
by filtration through the spherical atmometer cup used in these 
experiments. On the other hand, a dilution of approximately 
50 per cent resulted when a 1 per cent gelatin solution was 
filtered in the same cup. The sizes of the infective particles 
would therefore appear to be considerably less than those of 
gelatin particles, and since the particles of gelatin are not ap- 
parently very much larger than those of hemoglobin the con- 
clusion is further strengthened that the infective particles here 
in question have about the size relations of fresh hemoglobin. 
In considering the estimated size of hemoglobin particles referred 
to previously in connection with the work of Bechhold it should 
be pointed out that Bechhold seems to have worked with dried 
preparations of hemoglobin, and it is perhaps to be expected 
that these would be larger rather than smaller than those of the 
fresh product. All indications are that, in general, a relatively 
freshly made colloidal solution possesses particles more uniform 
in size, and this idea is tentatively accepted. Assuming that at 
most the hemoglobin particles worked with may have possessed 
a diameter of 30au, more or less, and that the average small 
diameter of bacterial plant pathogens is around 1000, (some 
being as low as 500 and others as large as 150044) we have 30: 
1000 to express roughly the diameter relations of mosaic disease 
particles in comparison with bacterial plant pathogens. On the 
basis of this average relation it is interesting to note that the 
volume relation would be about as follows: 1: 37,000, or about 
26: 1,000,000, assuming that in each case we may treat the 
bodies as spherical structures. 
The results of the filtration experiments have directed the 
attention of the writers to the possibility of the existence of 
