Wells may be either shallow or deep. The latter are always preferable, but sometimes the former must be relied on. The great and serious danger in connection with shallow wells is their liability to pollution from cesspools and drains, whose liquid contents (fully as poisonous as the solid) filter through the surrounding soil and go to swell the volume of water in the well, especially if, as nearly always happens, the cesspool is much shallower than the well.

In country villages, frequently the cesspools and wells are so intermixed that the entire bed of water is polluted, and hence all the wells are unsafe. But in isolated houses, if the well and cesspool are some distance apart, pollution of the well will depend chiefly on the direction of the movement of the underground water. If this movement is from the cesspool towards the well, the polluted water will flow towards the well; if the movement is in the contrary direction, the polluted water will flow away from the well. Hence Field’s caution, that before sinking a shallow well where sources of contamination are in the neighbourhood, the direction of the flow of the underground water must first be carefully ascertained, bearing in mind that it is not safe to assume that this flow is in the direction of the fall of the land, though it very frequently is so: if there is the slightest doubt, levels must be taken of the underground water in different places, and the source of contamination be accurately localised. Contamination from surface soakage can frequently be prevented by raising the top of the well above the adjoining ground, and paving the surface round the well with a slope so that the rain-water runs away from it. Norton Tube wells, which consist of an iron tube driven into the ground and surmounted by a pump, are useful for excluding surface pollution. If the pollution is sufficient to contaminate the subsoil and reach the underground water, no precautions that can be taken in constructing the well will keep the pollution out.

Generally, deep wells are safer from contamination than shallow wells, but may still, under certain circumstances, be polluted.

On the question whether a well which has been-polluted by a cesspool will become fit for use after the cesspool has been removed, no rule can be laid down. If the removal of the sources of pollution has been thorough, the well will frequently recover its purity; but under other circumstances the well may remain impure. As to the least distance between wells and cesspools compatible with safety, while the Local Government Board of London is content with 20-30 yards, Dr. Frankland insists on at least 200 yards. It would be more rational to forbid cesspools of all kinds; at the same time, possible leakages from drains, through injury or otherwise, must not be omitted from the estimate of risk of pollution. Again, the effect of increased demand upon the contents of the well at once extends the danger, because as the water in the well is lowered so is the area from which the well draws its supply increased, the ratio varying from 20 to 100 times the depression. Where a whole day’s supply is pumped at once into an elevated tank, the maximum figure will be reached.

Those who intend sinking wells are advised first to read a little book by Ernest Spon, on the ‘Present Practice of Sinking and Boring Wells,’ 2nd edition, 1885.

Rain-water collected from roofs forms a valuable auxiliary supply too often disregarded. In towns it is rarely pure enough for domestic use, but in country districts it is generally wholesome.

A country resident thus describes the manner in which he utilises rain-water, falling upon an ordinary tin roof, covered with some sort of metallic paint, said to contain no lead, and flowing into a large cemented brick cistern, whence it is pumped into the kitchen. The cistern differs from the usual construction in this manner: across the bottom, about 3 ft. nearer one side than the other, is excavated a trough or ditch about 2 ft. wide and 2 ft. deep; along the centre of this depression is built a brick wall from the bottom up to the top of the cistern, and having a few openings left through it at the very bottom. The whole cistern, bottom, sides, and canal included, is cemented as usual, excepting the division wall. Upon each side of the wall, at its base, 6-12 in. of charcoal is laid, and covered with well-washed stones to a further height of 6 in., merely to keep the charcoal from floating. The rain-water running from the roof into the larger division of the cistern, passes through the stone covering, the charcoal, the wall, the charcoal upon the other side, lastly, the stones, and is now ready for the pump placed in this smaller part. It is much better that the water at first pass into the larger division, as the filtration will be slower, and the cistern not so likely to overflow under a very heavy rainfall. He has used this cistern for many years, and was troubled only once, when some toads made their entrance at the top, which was just at the surface of the ground, soon making their presence known by a decided change in the flavour of the water.

If the house chances to be in a dusty situation, several plans will suggest themselves whereby a few gallons at the first of each rain may be prevented from entering the cistern. Should the house be small, and therefore the supply of water from its roof be limited, do not lessen the size of the cistern, but rather increase it, for with one of less capacity some of the supply must occasionally be allowed to go to waste during a wet time, and you will suffer in a drought, whereas a cistern that never overflows is the more to be relied upon in a long season without rain.

Rainfall varies exceedingly in different places, and even in the same situation it is impossible to foretell the amount to be expected during any short period of time, but the most careful observations show that about 4 ft. in depth descends at New York and vicinity every year, or nearly 1 in. a week. If this amount were to be furnished uniformly every week, the size of a cistern need only be sufficient to contain one week’s supply, but we often have periods of 4 weeks without receiving the average of one, and we must build accordingly.