In the previous portion of this article it was stated that the quantity of water to be distributed In a town was usually calculated at the rate of 40 gallons per individual per day ; but evidently local circumstances may affect this rate of consumption to a very serious extent. It is mere than questionable whether even the most luxurious persons require absolutely a greater quantity than 6 gallouti per day for their own immediate consumption ; and if the latter be carried to 10 gallons per day, it would suffice with proper care for any real demands. In calculating for a town supply, it is necessary, however, to allow for the consumption by trades using large quantities of water, and for muni cipal purposes, such as street-watering, flushing sewers, extinction of fires, &c.; and, in order to cover these demands, it was considered, so long as the old-fashioned intermittent supply existed, that an allowance of 20 gallons per head per day would be sufficient. The introduction of the constant-delivery system has of late years tended singularly to increase the quantity of water distributed, and indeed the waste of water which thus takes place is at times so great that it almost assumes the importance of a public calamity. In London, where the inter mittent supply still prevails, the consumption of water is now not less than at the rate of 40 gallons per head per day on the average ; in Croydon, at Boston, New York, &c., the rate under the constant delivery has been 80, 100, and even sometimes 500 gallons per head per day. No doubt a great part of this waste could be prevented, and it would never be allowed to exist if the water supply of towns were left in the hands of private companies ; but the tendency of the age is certainly to withdraw municipal services from the control of speculative adventurers, and it therefore becomes important to direct attention to the eventual danger arising from the waste which seems to be inherently attached to the modern system of distribution.
Whatever may be assumed to be the normal rate of consumption by the inhabitants of a town, it most be observed that, in the summer months,—that is to say, in June, July, August, and September,—the average consumption per individual per day increases in about the ratio of 10 per cent. upon the ordinary average ; whilst in December, January, and February it is about as much below the same average. In designing gravitation-works, especially, this becomes an important consideration ; for the consumption actually increases when the supply decreases, and the dimensions of the reservoirs must be calculated of sufficient size to meet the demand of the season which is the most unfavourable. The compensations to be given to the ancient pose more of water privileges must also be taken into account in designing any new water-works, especially in the case of gravitation-works, which interfere more with the natural hydrographical conditions of a district than ordinary pumping-works do ; and the loss of water by evaporation from the reservoirs in hot weather requires also to be allowed for. In most countries of Southern Eurepe, it is to be observed, moreover, that the rate of personal consumption of water, so to speak, is much less than the rate prevailing in the north, but that the quantity of water poured down the channels, or used in the monumental fountains, or in other contrivances for cooling and refreshing the air, is, on the contrary, very much greater. Thus, in Paris, it is stated officially that
the supply of water takes place at the average rate of 13;} gallons per head per day ; but, in fact, the water is sold by the pail, and in no house in l'aris is there a supply (in 1861) to the second floor, and nearly all the water which is brought into the city is poured into the street kennels or is employed in the fountains. All these details of the application of a town supply must, therefore, be carefully examined, and allowed for, in selecting the source and mode of distribution ; for it must be evident that every particular climate, and every phase of civilisation, requires to be treated upon its own special principles.
The quantity and the quality of the water supply being determined, a very important question arises, as to whether it should be obtained, at a great original outlay, under such conditions as to allow the water• to flow to the place of distribution, by gravitation ; or whether it should be obtained by means of pumping-machinery, at a smaller original outlay, but at a greater annual expense for working. This question is one which it is impossible to decide d priori, because it must be so materially affected by local considerations as to render it dangerous to lay down any absolute rules in the matter. But it must be evident that, wherever it is possible, as it was in the case of ancient Rome, to lead into a town streams of water rising at a natural elevation above the proposed points of distribution, there must be an eventual economy in so doing over the cost of another system by which the water would be raised from even a nearer point by the aid of a compli cated piece of machinery. The Romans, who were able to concentrate upon any work they undertook the resources of the known civilised world, adopted the former system ; and the highly-centralised govern ments of modern Europe evidently prefer it, for the French arc about to incur an enormous expense in order to lead to Paris, by gravitation, some springs which rise at a great distance from that town. In England and in America, however, where individual enterprise is so much more active, the tendency is to resort to sources of supply which admit of the application of machinery at a small primary outlay ; and, in fact, the choice between the two systems must greatly depend upon the power of commanding capital, and upon the balance of motives which might lead to imposing a burden upon the present generation for the sake of its successors. In either case, the works required for the filtration, storeage, and distribution of the waters are the same ; the only difference consists in the mode of leading the water to the point from whence the distribution is to take place.