1921] 
DUGGAR & KARRER—SIZES OF MOSAIC DISEASE PARTICLES 347 
juice through each of a series of porcelain or other filters of 
fairly well-determined porosity which might be subsequently 
standardized in a definite manner; but at the time no such 
series of filters could be found. Celloidin membranes did not 
seem to offer the range of porosity required. A little experi- 
mentation with rate of water flow, however, indicated that no 
inconsiderable range of possibilities was available in the form 
of the ordinary porcelain filters and atmometers of the laboratory. 
Accordingly, a series of filters was arranged consisting of a 
Mandler filter, à porous spherical atmometer cup, 2 cylindrical 
atmometer tubes, 2 cylindrical atmometer tubes infiltrated with 
precipitation films of AI(OH),, and 2 specially prepared celloidin 
membranes. Considerable preliminary work led to the selection 
of this series. It may be well also to indicate that the particular 
spherical atmometer cup used in this work proved to be the 
only one possessing pores noticeably finer than the average of 
these cups. This cup was one of the earlier ones distributed 
for work in atmometers. 
Filters employed.—The porcelain filters were, where necessary, 
thoroughly cleaned and all were boiled in distilled water prior 
to use. The Mandler filter employed was No. 5090 of the Arthur 
H. Thomas catalog, 244-514 inches, tested to 6-12 pounds air 
pressure without passing air bubbles. The cylindrical filters 
impregnated with Al(OH), were prepared in the following 
manner: The filter tubes were filled with 5 per cent AlCl, and 
after allowing time for this to penetrate the walls thoroughly 
the tubes were suspended in beakers of 1 per cent NH,OH until 
it appeared that the alkali had penetrated the cup, shown by a 
slight turbidity. The tubes were then carefully rinsed. 
The celloidin membranes were prepared according to the 
method of Brown (715) by which films of relatively great per- 
meability can be obtained. 'The membranes were formed on 
the inside of beakers. An 8 per cent solution of Schering’s 
celloidin in an equal volume of ether and absolute ethyl alcohol 
was poured into a beaker and allowed to drain over another 
beaker for 10 minutes. The beaker was then immediately 
immersed in distilled water. After about a minute the membrane 
was loosened from the sides of the beaker, washed in the water 
