While useless waste should be prevented, it is most important that the legitimate use of water should be encouraged in every way. As Dr. Richardson has well pointed out, absolute cleanliness, properly understood, is the beginning and the end of sanitary design, and thorough cleanliness, of course, can never be obtained without an ample water supply. Not only should there be sufficient water for baths, lavatories, and washing of all kinds, but there should be a liberal allowance for flushing water-closets and all other sanitary appliances. Taking these sanitary considerations into account, as well as giving due weight to the observations which have been made by engineers and others on the quantity of water actually used in houses under different circumstances, it may be assumed that, if waste is efficiently prevented, a supply of 20-25 gallons per head per diem is sufficient in ordinary cases for houses with baths and water-closets. If horses are kept, a separate allowance should be made for them, and for stable purposes (a useful approximate rule being to reckon a horse as a man); and if water is used for watering gardens or ornamental purposes, this must also be reckoned separately. If earth-closets are adopted instead of water-closets, less water will be required, and 15-20 gallons per head per diem will be sufficient. In cottages with earth or other dry closets, the quantity of water required will be still less: 10 gallons per head will be an ample supply, and even 5-6 gallons may do in cases where it is absolutely necessary to limit the quantity used. _Sources of Supply._--Water for country houses is, in the vast majority of cases, derived from springs or wells. Rain-water collected from roofs is very frequently used as an auxiliary, and occasionally as the main supply. There are instances in which the supply is taken from streams or rivers, and even some in which water running off the surface of the ground is collected in “impounding reservoirs” (a mode often adopted for the water supply of towns); but these cases are exceptional, and attention will here be confined to springs, wells, and roof-water. The real source of all fresh water supply is rain. Springs and wells form no exception to this rule, though in their case the connection with the rainfall is not so clear at first sight as it is in the case of streams and open watercourses, because the passages by which the rain reaches springs or wells are not visible, and heavy rainfalls often have no apparent effect on their yield. In various parts of the country occur curious intermittent springs (locally called “bournes”), which burst out in some years and not in others, and the connection between which and the rainfall is still more obscure. Rain-water, before it issues from the ground as springs, accumulates in the porous strata beneath, and forms, as it were, large underground reservoirs; it is from these reservoirs that wells, sunk into the porous strata, derive their supply.