438 
EXPERIMENTS WITH PAPER FILTERS. 
glass; therefore, those funnels having the greater portion of the paper free from the 
glass would be the best; that is, a broad-tlirocited funnel , other things being equal, will 
filter faster than a narrow-throated funnel. 
To test this point I selected two large funnels; No. 1 had three times as broad a 
throat as No. 2. With the first filters they ran:— 
117 : 100 123 : 100 133 : 100 118 : 100. 
The reason for this low difference was found in a thin spot near the point of No. 2. 
Other sets of filters gave:— 
2nd 
set, 
292 : 
100 
318 : 
100 
3rd 
set, 
288 : 
; 100 
335 : 
100 
4th 
set, 
300 : 
: 100 
burst 
5th 
set, 
384 : 
100 
407 : 
100 
6th 
set, 
242 : 
: 100 
482 : 100 
In the last set a porous filter, though off the same sheet as No. 1, was given to No. 2. 
Throughout the whole series of experiments every fair advantage was given to the 
weaker party, it being the first filled and the last emptied. 
To make assurance doubly sure, I tried filters in like funnels, stopping the pores of 
the paper at various points. Paraffin applied whilst liquid was the substance first used 
to prevent filtration. 
Two filters were chosen from the same sheet and of as uniform a texture as possible. 
No. 1 was stopped over one-third the radius from the point. No. 2, all but one-third 
the radius at the point. 
They filtered at nearly the same rate, No. 1 slightly the faster. The paraffin made 
the paper stiff, and as w r ater does not adhere to it, free passage was allowed between it 
and the funnel to the water of the upper part of No. 2. 
Here we see that one-ninth of the surface of the filter, when free, did as much work 
as eight-ninths adherent to the glass. The experiment was repeated with glycerin 
instead of paraffin. No. 1 ran 28 per cent, the faster. It might be objected that 
glycerin would wash out; so a pap of paraffin and spirits of turpentine was used to 
repeat the experiments. Each filter, after being painted, was wetted to prevent the 
spread of the mixture by capillary action. The trials were not sufficiently numerous to 
fiud a true mean, but the free point invariably ran the most—from 4 or 5 per cent, 
excess to 100 per cent. The point was assumed to be a circle one-third the radius of the 
filter. 
I understood the idea of the Fleitmann filter to be this, that, likening a plain filter to 
a peat-bed resting upon an impermeable subsoil, it might be compared to a porous sub¬ 
stratum interpolated between the swamp and the clay bottom. 
To test this idea, a Fleitmann filter was made and wetted, carefully patting down 
and smoothing out any irregularities. It was tried against a plain filter which was placed 
in a funnel with but two-thirds as wide a throat as that of the Fleitmann. It ran 
114 : 100; that is, the passages kept open by the elasticity of the paper, the creases and 
abutting edges liken this filter to tile drainage. 
To increase the size and number of passages I tried putting the inner filter into a 
plaited filter of coarse paper. Changing the filters after each trial, I found this form 
gave the following results as compared with the plain filter, calling the latter one 
hundred:— 
1st trial, 184 : 100 4th trial, 66 : 100 
2nd „ 201 : 100 5th „ 170 : 100 
3rd „ 250 : 100 
I afterwards found a thin spot in the plain filter of the fourth trial. 
Next a precipitate of sulphate of calcium was tried; the filtrates were as 131 : 100. 
On weighing the precipitates collected they were as 200 : 100. 
This form of filter w r as abandoned, since it does not filter as fast, is not as strong, 
takes more time to make and care to use than the form next to be described. It is, 
however, better than the plain filter as regards speed of filtration, equal to the plaited 
in this respect and stronger than it. 
To admit the use of very broad-throated funnels, the number of outside filters was 
