1921] 
DUGGAR & KARRER—SIZES OF MOSAIC DISEASE PARTICLES 349 
particles which may pass through, but primarily the rate of 
passage. The writers feel that it may be necessary to determine 
carefully the influence of the time interval; but since in these 
experiments comparative rather than fundamental results were 
desired, the phase of the filtration problem just referred to has 
not been experimentally studied. 
With the celloidin membrane it was necessary to filter very 
cautiously so that a longer period of time at a pressure of 0.8 
atmosphere was given. In this case, too, the membrane was 
fastened over the bell of a thistle tube. The diseased juice was 
then added through the tube, and the thistle tube—with the 
stem of the latter inserted through a rubber cork—was placed 
in a wide-mouthed bottle and lowered almost to the bottom, 
sufficient water being added to the bottle to just cover the 
membrane. Aspiration was then applied to the bottle through a 
second tube entering to just below the surface of the cork. 
INOCULATION EXPERIMENTS WITH FILTERED JUICES 
Technique of 4noculation.—All inoculations were made by 
injuring the surface of the growing plant in 3 different areas, 
one from near the growing tip, one at the base of a young leaf, 
and another farther down the stem, or in the case of younger 
plants, just above the surface of the ground. These injuries 
were made with a needle or a fine pointed scalpel and in each 
case a drop of the infected juice was smeared over the injury 
and somewhat worked into it. This type of injury proved 
generally more effective than merely rubbing the stem or leaf 
as has been done in some cases. It was generally found advisable 
to make the inoculations in the late afternoon, the greenhouse 
being thoroughly watered afterward so as to prevent a too rapid 
drying of the injured surfaces. 
Since there was some danger that the operator handling the 
filtration apparatus might come more or less in contact with 
particles of the diseased juice it was arranged that all inoculation 
work should be carried out by a different operator. Moreover, 
in most cases the different inoculation experiments were made 
by different operators. Where this was not possible every 
