Prof. Angel gives the following directions for testing water for Ammonia, which is well known to be the most sensitive test. He says: - "A minute and variable quantity of ammonia exists in the atmosphere. From this source rain-water receives it, which contains less than 0.5 part per million. The earth, in turn, absorbs it from rain-water, while some of it is destroyed by oxidation, so that rivers seldom contain more than 0.1 part per million, and perfectly pure spring or well-water contains only a mere trace.

"The ammonia process in water analysis is an indirect method of measuring the amount of organic matter which a water contains. Of course, all the ammonia, as such, that any natural water might ever contain, is perfectly harmless. The decay of organic matter produces ammonia, and importance is attached to the latter only as it indicates the existence of the former.

"In the laboratory two kinds of ammonia are recognized, 'free' and 'albuminoid' Free ammonia is that which has resulted naturally from the decay of organic matter contained in the water, and, other things being equal, shows how extensively such decomposition is going on. It is easily collected by distillation.

"Albuminoid ammonia is that which results from hastening decomposition artificially. It measures the amount of organic matter present which may decay, and is simply what would be produced naturally in the course of time.

"The ammonia process, when fully carried out, is the most reliable method known for determining the organic condition of water. To arrive at a correct conclusion in every case, it is necessary to estimate accurately both kinds of ammonia. The determination of albuminoid ammonia requires special apparatus, and is too complicated for general application; but the test for free ammonia is quite easily made, and, from a series of experiments and observations, it has been found that, generally, whenever a certain amount of free ammonia occurs in well-water, an excess of albuminoid ammonia is also sure to exist. So it is pretty safe to conclude that such water is polluted. Says an authority: 'When the free ammonia exceeds 0.08 part per million, it almost invariably proceeds from the fermentation of urea into carbonate of ammonia, and is a sign that the water in question consists of diluted urine in a very recent condition. In these instances the water will likewise be found to be loaded with chlorides.' Our experience places the amount a little higher than 0.08. We believe if a water contains 0.1 part per million of free ammonia, it should be regarded organically impure, especially if other indications point the same way. Of course there are exceptions. Some waters, organically pure, naturally contain much free ammonia, while others, that are badly polluted with vegetable matter, may contain sometimes much less than 0.1 part per million. In such cases the determination of albuminoid ammonia is indispensable to the detection of pollution. It is to be regretted that there is no simple and reliable method for doing this. But the cases are rare where water polluted with vegetable matter contains less than 0.1 part of free ammonia per million".

1. The following process for detecting and estimating free or carbonate of ammonia is sufficiently simple and accurate for general application:

Dissolve some mercuric chloride (corrosive sublimate, a poison) in a little water, making the solution quite strong. Also prepare a strong solution of carbonate of soda (common cooking soda will do, or caustic soda, or potash) by dissolving it in water. Place a tumbler, or clear glass, on a black surface in good light; fill it with the water to be tested, and then add a single drop of the solution of mercuric chloride, followed by a drop of the soda solution in the same place. Let the liquid stand without stirring. Look down through it, and if ammonia is present, even a minute quantity, a white cloud or opalescence, resembling white smoke, will be observed toward the bottom of the glass where the drops passed, which in the course of some hours will settle and cover the whole or part of the bottom of the glass with a white coating. If much ammonia is present, the reaction will be very marked, and almost instantaneous. Less ammonia requires more time, and the reaction is less marked.

The delicacy of the test is sufficient to give, within five minutes, a distinct reaction in water containing 1/1000000 Part of its weight of ammonia. Any one can satisfy himself of the delicacy of the test by the following: Add a spoonful of water, free from ammonia (water that has been boiled for some time), a single drop of ordinary ammonia; then add a drop of this to a tumbler of water that has been well boiled, and apply the test in the manner described above.

If water shows the reaction, it is far from the sanitary standard for purity, which, as has been said, is not more than 0.1 part per million, and this number is ten times less than 1/1000000, the limit of the test. Consequently, a water may contain too much ammonia and not show the reaction. To obviate this difficulty, a simple process of distillation must be employed.

2. Another test for detecting ammonia is to prepare what is called "Nessler's solution". Take 35 grains iodide of potassium, 13 grains of corrosive sublimate and about 800 grains of distilled water. Heat to boiling in a glass vessel, and stir until the salts dissolve. Add cautiously aqueous saturated solution of corrosive sublimate until the red iodide of mercury, which is produced as each drop of the solution falls into the liquid, just begins to be permanent. Then add 160 grains of caustic potash or 120 grains of caustic soda. When this is dissolved make the whole weigh 1000 grains by the addition of distilled water, add a little more cold saturated solution of corrosive sublimate and allow it to settle. When properly prepared it has a slightly yellow tint. If perfectly white it requires a little more corrosive sublimate. It should be tested before using by dropping a portion in a very weak solution of chloride of ammonium. If good it will at once strike a yellowish brown tint. Keep the solution in a well-stoppered bottle and carefully protect from the air. Then proceed to test whether there is free ammonia present, and for this purpose a test tube is filled with the water and held over a sheet of white paper, while a few drops of the Nessler solution are dropped into it. If then a yellowish brown color is produced ammonia is present, and, according to the depth of this tint, is the amount.

3. Ammonia may be detected also by slightly acidifying the water with muriatic acid, evaporating to dryness and adding to residue some caustic potash or soda solution. On holding the glass rod, previously dipped in acetic acid or diluted nitric acid, in close proximity to the residue, a thick white fog will be visible. (Hager.)