AND ITS BEARINGS ON PATHOLOGY Sy jt 
therefore still in full activity, I introduced a small portion of the curd from 
one of the glasses into a glass of Pasteur’s solution, in order to study the 
behaviour of the organism in that fluid. In this experiment I made use of what 
I have called a ‘ separation-tube ’, a piece of glass tube bent at an angle of about 
45° (see Fig. 9), with one leg shorter than the other, the shorter leg closely 
packed with moist cotton-wool (e), which is purified together with the tube 
by boiling in water. The apparatus is then transferred to a pure liqueur-glass, 
in which it is arranged with the longer leg (/) placed vertically and kept in this 
position by a mounting of silver wire (g) applied to the tube before its purifica- 
tion. The liquid to be experimented on is now poured into the glass to a higher 
level than the mouth of the shorter leg of the separation tube, but not so high 
by a considerable interval as the top of the longer leg. 
The liquid finds its way slowly through the cotton and rises 
in the longer leg, and it is into the liquid in the longer leg 
of the separation tube that the material of inoculation is 
introduced. The cotton in the shorter leg is so closely 
rammed that it serves as an effectual filter to prevent the 
passage of any particles of ordinary solid matter; but if 
any organism capable of developing in the liquid is intro- 
duced into the longer leg, it will work its way through the 
cotton, and show itself in the general liquid in the glass. 
Even a motionless bacterium or a filamentous fungus will 
traverse the closely packed mass, at a rate corresponding 
with the rapidity of its development. Thus we have a 
simple means of separating things which are living and capable of developing in 
any particular liquid from particles which are not alive, or which, though 
living, are incapable of growing in that medium. It was into the longer leg 
of such a separation-tube in a glass of Pasteur’s solution that I introduced 
the morsel of curdled milk containing the actively growing Bacterium lactis. 
For a long time it appeared as if no growth whatever of the organism had 
taken place in its new habitat. But when nearly three weeks had elapsed, 
thinking that the little bit of curd looked somewhat swollen, I subjected 
it to microscopic examination, and found that the substance of the curd 
seemed to have disappeared, its place being occupied by small bacteria 
having the characters of Bacterium lactis, showing that the organism really 
had undergone some development. But it seemed to have grown only where 
it had the material of the curd to feed upon; for the outside liquid in the 
glass showed no trace of any organism either to the naked eye or under the 
microscope. Thus it appears that the Bacterium lactis, though not destroyed 
Bhi2 
