The studies of Pumpelly and Smyth on soils as filters of different compositions and structures are very important and were made to ascertain:

1. Their action as filters for air or gases, generally.

2. Their action as filters for water and other liquids.

When we consider that every foul and defective cesspool is a center of pollution to the ground water which supplies our wells, and to a certain extent also the reservoirs for cities, we can understand the importance of determining to what extent average soils can filter them. Our houses become the ventilating chimneys for the surrounding ground, sucking in the air from its pores, and when this air is polluted by sewage it is important to know if these impurities can be removed from it by filtration. The experiments showed that the filters, especially where wet, were able, as might be expected from what we have learned about bacteria adhering to damp surfaces, to entangle and retain the germs which had been floating in the air, whereas when the germs were in water none of the filters tried could, for the same reason, extricate them from it. The germs could not escape from the water, but had to follow it between the grains of the filters.

Our next two pictures show the apparatus used for making these investigations (Figs. 47 and 48). Fig. 47 shows a simple test tube slightly contracted at the center, having an infusion of beef or some other easily putrescible substance at the bottom. A small piece of copper gauze rested on the shoulder formed by the contraction in the tube above the infusion, and upon this copper gauze as a support rested the substance to be tested in each case as a filter. The height of the filter column varied from about a sixteenth of an inch to about six inches. Nearly ninety filters were tried of asbestos, sand, loess, charcoal, animal charcoal and coal ashes. This apparatus and the others to be described showed that dry soils, even in comparatively coarse grains form good filters for germ-laden air, and the wet filters are still more efficacious than dry; and that putrid soils not only retain the germs which they already possess, but also extract others from air passing through them. The infusions under the filters remained sound for months, though separated from the bacteria floating in the surrounding air only by a very small layer of the filter.

In the second series of experiments the apparatus shown (Fig. 48) was used to show the effect of the filters upon a current of air passing through them. A flask provided with an air-tight rubber stopper was used. The filter was placed in a tube passing through the stopper and rested on a copper gauze within the tube. The aspirator tube also passed through the stopper and its lower end was protected by a filter in order to prevent microbes from entering the flask through it while it was disconnected with the aspirating mechanism which we have already described. The beef infusion was put in the bottom of the flask. The filters were subjected in this, and in another apparatus shown in the test tube Fig. 48a, at intervals during five months to rapid currents of air beginning at the rate of 1 quart in from 4 to 7 hours and increasing to 1 quart in 1 minutes. The two filter columns in the bent tube apparatus were about 4 inches (10 centimeters) high each, and consisted of sand screened through 30 to 50 meshes per inch; in others fine sand, 100 grains to an inch, was used, and in still others asbestos packed rather tightly. All the filters stood these extremely trying tests, the infusions remaining intact.

Figs. 47 and 48. Apparatus for testing air filters.

Figs. 47 and 48. Apparatus for testing air filters.

Fig. 48a. Apparatus for testing air filters.

Fig. 48a. Apparatus for testing air filters.

The experiments show that as long as the grains are small, all the substances tested, filter perfectly all organisms from the air passing through them.

Very different were the results of the attempts to filter germs from liquids.

In these experiments there were tried in the first apparatus consisting of the simple test tube, three sets of filtrations, the liquids being respectively (a) fresh infusions, (b) putrid infusions, and (c) water.

In the first set the filters employed were sand, charcoal, and animal charcoal, the finest of each being 100 grains to the inch, and asbestos in columns of about 6 inches high. All failed to filter out the germs except two filters out of twelve of asbestos, and of the finest animal charcoal; the remaining asbestos and the coarser animal charcoal 15 to an inch and the vegetable charcoal in all grades failing equally with the sand.

Similar filters were used in the second set of experiments (with putrid infusions). In this set only tightly packed asbestos stood the test; in all others the infusions passed through turbid.

In the third set (filtration of water) the substances tested were in columns 6 inches high, of tightly packed asbestos, charcoal, some 25, others 50, and others 100 grains to the inch; animal charcoal, some 15 and others 100 grains to an inch; loess, kaolin, coal ashes, sand of 25, 30, 50 and 100 grains to the inch, respectively.

The experiments were conducted so that only a drop or two of water should reach the infusion at first; more water being passed after intervals of several days, where the infusion remained uninfected.

Of these the tightly packed asbestos stood repeated filtrations for ten days, after which the infusion putrefied. Of the finest animal charcoal filters, one out of four filtered out the germs. The other three stood repeated filterings during 17 to 20 days, after which the infusion became infected.

The coal ashes were subjected to repeated filtrations during 10 to 19 days, after which the infusions broke down. With the loess the infusions were affected after six to eight days.

With kaolin and sand they broke down in from one to eight days, except two, which lasted 19 and 20 days respectively. In all the others the first drop that passed through infected the infusion.

The experimentors now made a bold jump from columns of a half a foot to columns 22 and 100 feet in height, and in them the filter used was sand mixed 18 and 100 grains to the inch. The apparatus used is shown in the coiled pipes in Fig. 49. To obtain sterilized columns of such a length the sand was intensely heated and poured into a lead pipe, which was then coiled. It was then placed, in a furnace and heated to between 250 and 300 degrees Centr., and then the lower end attached to the flask containing the infusion, as shown on the floor by the table. The flask was ventilated through a tube protected by sterilized asbestos, as shown.

The whole was then allowed to rest several weeks to be sure that the infusion had been properly sterilized. The pipe was then very slowly filled with water from the faucet. The first water that passed through each of the 22 and 100 foot columns carried infection with it.

The following conclusions were drawn from these experiments :

"I. All the substances operated on are excellent filters in eliminating germs from infected air passed through them, except when they are of a coarse grain - 10 to 20 grains to an inch - when the interstitial cavities become probably much less labyrinthine. And all these filters withstood the tests of currents having many thousand times the maximum velocity attained in the soil.

"II. All natural substances tried thus far, except the finest animal charcoal, and perhaps tightly packed asbestos, failed to eliminate wholly the germs from liquids."

Of the natural soils tried we find sands entirely without power to filter germs from water, and probably in columns 10,000 feet long as well as 100 feet. The loess and kaolin have more power, and in greater heights of columns it is possible they would be effective. The filtering capacity seems to be proportional to the smallness and intricacy of the interstitial cavities; and in dry air filters there is a critical limiting point beyond which there is no filtering.

With liquids, far greater fineness and compactness of grain and intricacy of passage is needed than with air. While sand of even 20 grains to the inch is an excellent air filter, it is worthless for water, even as fine as 100 grains to an inch and in very long columns, and the critical limiting point below which soils begin to exercise any filtering action on water probably verges on the size of the particles of an impalpable powder.

"From these results it appears very clearly," says the report, "that sand* interposes absolutely no barrier between wells and the bacterial infection from cesspools, cemeteries, etc., lying even at greater distances, in the lower wet stratum of sand. And it appears probable that a dry gravel or possibly a dry very coarse sand interposes no barrier to the "And from the results obtained from the two series of experiments, viz., in filtering air and in filtering water, we can now draw one very important practical conclusion which cannot be too strongly emphasized. That a house may be built on a thoroughly dry body of sand or gravel, and its cellar may be far above the level of the ground-water at all times, and it may yet be in danger of having the air of its rooms contaminated by germs from leaching cesspools and vaults; for, if the drift of the leaching be toward the cellar, very wet seasons may extend the polluted moisture to the cellar walls, whence, after evaporation, the germs will pass into the atmospheric circulation of the house."

*Successful water filtration through ordinary sand requires a surface coating of some much finer substance than the sand itself. This coating, in slow sand filtration, is automatically deposited partly by bacterial action, and in rapid mechanical filtration it is chemically precipitated. In both systems the amount of the coating must be periodically regulated in order to maintain the proper speed of filtration. This regulation of the amount of coating, which is sometimes called the "dirt covering" or "schmutzdecke," forms the most costly part of the operation of a filter. A number of plans for reducing this expense are now being developed, of which one by the writer is outlined in a later chapter on "Filtration." free entrance into houses built upon them of these organisms which smarm in the ground-air around leaching cesspools, leaky drains, etc., or in the filthy made-ground of cities.

Fig. 49. Long coil of pipe filled with sand.

Fig. 49. Long coil of pipe filled with sand.