(approximately) 1,100 square feet, and if the annual rainfall fluctuate between 24 inches and 30 inches, then the amount of rain falling upon the roof will vary from 2,200 cubic feet to 2,750 cubic feet. If we take a cubic foot as the equivalent of 6 gallons, then we may say that the amount of rain annually falling on the roof will fluctuate between 13,750 gallons and 17,187 gallons. If we put the average water-supply of the roof at 15,000 gallons a year, or rather more than forty gallons a day, we shall not be far wrong.

Water experts say that in towns we want a supply of forty gallons per head per diem. The dweller in the clean country is content with much less than this, and I feel convinced that ten gallons a day is an extravagant estimate for the daily supply of a perfectly clean peasant who does clothes-washing at home, but has not the power of wasting water.

The storage capacity of the tank is about 1,600 gallons, or forty gallons a day for a drought of six weeks. The water is excellent, odourless and colourless, and altogether very unlike ordinary rain-water.

The water of this cistern was analysed both chemically and bacterioscopically for the Royal Commission on Sewage Disposal on November 14, 1901, and with the following results:

once

0.23

After incubation at 80° F. for 6 days after 4 hours 0-79

'Sample clear but yellow, no sediment, peculiar faint smell rather like soot.'

To

The above is a typical analysis of rain-water. the eye and palate it is the best sample of rain-water I have ever seen, and it has been used for all domestic. purposes. It should be said that the yellow colour is very slight. Personally I cannot detect any smell, but there is a faint taste of terra-cotta.

Dr. Houston found 25 bacteria per c.c. on gelatine at 20° C., and 7 per c.c. on agar at 37° C. The tests for Bacillus coli and Bacillus enteritidis sporogenes gave negative results.

These analyses are full of instruction, and show how chemistry and bacteriology are needed to check each other, and how both need to be checked by a knowledge of source and circumstances.

The disposal of slop-water is always an important consideration in cottage management. Usually this means slop-water plus roof-water, but in this cottage the roof water has been provided for. The amount of slops, allowance being made for evaporation in cooking, and washing and drinking, must always be considerably less than the water consumed. Economy in the use of water lessens the slop difficulty.

In this instance the slops are strained and filtered, and allowed to flow away in a 'filtration gutter,' to be presently described. The arrangements are on the south side of the cottage, well exposed to the sun, so as to favour evaporation.

The sink is just beneath the window of the scullery, and the waste-pipe, without trap of any kind, passes through the wall, and terminates in a free end about 18 inches from the wall and 2 feet 6 inches above the

level of the ground. The waste-pipe empties itself into a strainer and filter, which are placed about 15 inches from the cottage wall, so as to avoid the risk of splash or back soakings or accumulations of dirt' and insects between the wall and the filter. The strainer is placed on the top of the filter, and the filter discharges its water on to a filtration gutter. This filter is shown in fig. 1 at the extreme right, and is marked with a cross. A longitudinal section of the arrangement is shown in fig. 3.

The strainer consists of a basket with a wisp of straw in it (B). This arrests all but the finest particles, and is the best fat-trap I know-the only one, in fact, which does its work efficiently and without offence. The straw may be changed as often as necessary—every day, once a week, once a month, according to the amount of accumulations, which will largely depend upon the thriftiness and knowledge of the cook. The contents of the strainer may be given to the chickens, put on the manure heap, or burnt. A new handful of straw is then put in and the strainer replaced. The changing of the straw has the advantage of giving a new direction to the water. Any old basket of suitable size which will hold the straw answers the purpose of a strainer. After months of use it will get greasy and rotten, and may then be burnt and be replaced by a new one. From the strainer the slops flow into the filter, which is simply a galvanised iron vessel, with an outlet at the bottom and filled with broken clinker varying in size from peas at the bottom to walnuts at the top. This filter effects a further purification of the slops, and acts partly mechanically and partly by virtue of the growth of bacteria, on the surface of the broken clinker. The filter shown has been specially constructed, and is duplicated (see fig. 4), and the waste

S, sink. W, wall of cottage. P, waste-pipe. B, basket containing straw. F, filter. G, cast-iron filtration gutter, supported in trench by (X) columns of bricks on edge.

pipe of the sink is provided with a reversible nozzle so that either half of the filter can be used. For a cottage, however, this is not necessary, and an old galvanised iron bucket with a hole in the bottom will be found to answer every purpose.

The filtration gutter consists of strong cast-iron guttering, perforated with conical holes, having the small ends upwards so that they cannot get jammed (see fig. 5). This guttering, which is 9 inches wide and in lengths of

FIG. 4.-DUPLICATED TANK FILTER.

6 feet, is laid upon loose porous rubble or gravel placed in a trench.

A trench 18 inches wide and 18 inches deep was first dug from the filter due south, care being taken that the bottom of the trench should slope away from the cottage, in order that water should not flow back towards the foundations of the building. The lengths of guttering are then laid on a level with the top of the trench, the level being maintained by means of bricks on edge, built