This filter, however, at its best, is but a good strainer, and will only arrest the suspended particles. In a modern filter more perfect work is required, and another effect produced, in order that water containing objectionable matter in solution should be rendered fit for drinking purposes. Many persons when they see a water quite clear imagine that it must be in a good state for drinking. They should remember, however, that many substances which entirely dissolve in water do not diminish its clearness. Hence a clear, bright water may, despite its clearness, be charged with a poison or substances more or less injurious to health; such, for instance, as soluble animal matter. To make a perfect filter, which should have the double action of arresting the finest suspended matter and removing the matters held in solution, and the whole to cost but little and capable of being made by any housewife, has long been an object of much attention, and, after many experiments and testing various substances in many combinations, the following plan is suggested as giving very perfect results, and costing only about 8_s._ Purchase a common galvanised iron pail, which costs 2_s._ Take it to a tin-shop and have a hole cut in the centre of the bottom about ¼ in. diameter, and direct the workman to solder around it a piece of tin about ¾ in. deep, to form a spout to direct the flow of water downward in a uniform direction. Obtain about 2 qt. of small stones, and, after a good washing, place about 2 in. of these at bottom of pail to form a drain. On this lay a partition of horse-hair cloth or Canton flannel cut to size of pail. On this spread a layer of animal charcoal, sold by wholesale chemists as boneblack at about 5_d._ a lb. Select this about the size of gunpowder grains, and not in powder. This layer should be 3 or 4 in. A second partition having been placed, add 3 in. of sand, as clean and as fine as possible. Those within reach of glassmakers should purchase the sand there, as it is only with that quality of sand that the best results can be obtained. On this place another partition, and add more fine stones or shingle--say for 2 or 3 in. This serves as a weight to keep the upper partition in place, and completes the filter. By allowing the filtration to proceed in an upward instead of a downward direction much better results are obtained.