26 BULLETIN 879, IT. S. DEPARTMENT OE AGRICTJLTTJBE. 
types were supposedly of sufficient density to prevent the passage of 
visible organisms. 
It was hoped that a comparison of the filtrates from these different 
grades of filters would perhaps indicate the size of the particles 
responsible for infection, but the results did not warrant any con- 
elusion along this line, as no infection has occurred in a total of more 
than 100 plants inoculated with these filtrates. The filters of the 
Chamberland type, being of porcelain, are all denser than the Berke- 
feld, and filtration is therefore much slower. There is a possibility, 
therefore, that the infective principle may be held back by absorp- 
tion, as the suspended material in the plant juices rapidly clogs 
the filter and covers its surface with a gelatinous layer. 
Other filters. — Allard (4) reports that the juices of mosaic tobacco 
plants become noninfectious if passed through a Livingston atmom- 
eter. A filter made of a layer of powdered talc from seven-eighths 
to H inches thick also gave a noninfectious filtrate. Similar talc i 
filters have been used in experiments with cucumber mosaic, the 
layer of talc varying from three-fourths of an inch to 1 inch in thick- 
ness. Three trials have been made with these filters, and a total of 
37 plants have been inoculated from the filtrates. The results have 
been negative in all cases, however, and it is evident that the filtrate 
is rendered noninfectious, as in the case of tobacco. 
Iwanowski (16), Koning (21), and Beijerinck (6) have stated that 
the juices of mosaic tobacco plants do not lose their infectious nature 
when passed through Berkefeld and Chamberland filters, though 
Iwanowski found that only the first portion of the Chamberland 
filtrate was infectious. It is thus evident that the infective principle 
of both tobacco and cucumber mosaic is of such a nature that it can 
be removed by filters of the finer types. 
ENZYMS IN RELATION TO THE DISEASE. 
Since many writers have advanced the theory that enzyms, par- 
ticularly oxidases, are in some way connected with the mosaic of 
tobacco, a few attempts were made by the writer to isolate a pos- 
sible causal enzym from the juices of mosaic cucumber plants. The 
fact that the juices of such plants usually lose their power of infection 
within 24 to 48 hours has proved an almost insurmountable obstacle 
in such work, and so little has been done that it is not feasible to 
draw any definite conclusions. Tests of the juices of healthy and 
diseased plants which have been passed through filter paper have 
shown the presence of oxidases and peroxidases in both cases, the 
guaiacum reaction being used. Very little difference in the intensity 
was noted, but the juices of diseased plants seem to show a slightly 
stronger reaction for both enzyms. Similar tests of both healthy 
and mosaic plant juices after passage through a Berkefeld filter have 
shown a weaker test for both oxidase and peroxidase in all cases, 
