SANITARY NOTES ON POTABLE WATER. 563 
with. In this instance the meat had become putrid after 
three weeks. The explanation of this difference of behaviour 
is that the water* until the vessel has been filled* passes 
very rapidly through the spongy iron* carrying with it putre¬ 
factive elements which had originally been contained in the 
water or abstracted from the spongy iron. That the latter 
in the dry state does not prevent putrefaction of meat was 
proved by a separate experiment. Dry spongy iron may* there¬ 
fore, and probably does* contain bacteria of putrefaction* 
so that it was hardly to be expected that the meat could 
be disinfected when remaining in contact with the first water* 
which had passed through the filtering material* the more 
so as it filtered very rapidly without regulation. 
Thus far the meat was brought in contact with water con¬ 
taining the small quantity of iron—about 10 milligrammes 
per litre—which is always dissolved as ferrous hydrate or car¬ 
bonate* whenever ordinary water is passed through spongy 
iron. We know that some compounds of iron have antisep¬ 
tic properties. It therefore appeared of interest to ascertain 
what the effect would be if the iron were first separated 
before the water reaches the meat. To this end two filter 
cases were arranged as before* so as to have a continuous 
flow of water* but underneath the spongy iron I placed layers 
of pyrolusite and sand. Pyrolusite has the property of 
oxydising the proto-compounds of iron* which are mentioned 
above* converting them into per-compounds. They are thus 
rendered insoluble* and retained mechanically by the sand 
underneath. After the filters had thus been in operation 
for three weeks* the meat in both was again found to be 
putrid. Nevertheless* it did not appear to me likely that 
this result should be due to the abstraction of the iron* which 
had been in solution. 
The filtering vessels contained in the course of this experi¬ 
ment a comparatively large bulk of pyrolusite and sand, both 
of which cannot impart any antiseptic properties to the water; 
on the contrary* they contained unquestionably putrefactive 
matter* even to a larger extent, than the dry spongy iron in 
the previous experiment. Such putrefactive matter was 
carried down with the water for some time to the meat, 
causing it* in addition to the putrefactive matter originally 
adhering to the meat* to show incipient putrefaction before 
disinfection could take place by agency of the filtered 
water. 
To prove the correctness of this argument I had to demon¬ 
strate that if the filtering apparatus were first kept in action 
for some time before adding the meat* so as not to bring the 
