bonic dioxide in 10,000 would be the standard of absolute purity. Authorities differ as to the greatest amount of carbon dioxide which might be permitted. It is quite certain that any unpleasant sensation is not experienced until the amount is increased to 10 or 12 parts in 10,000; yet authorities are generally agreed that the maximum amount should not exceed 10 parts in 10,000, at least for sleeping-rooms. The standard of good ventilation usually adopted at present would permit about 8 parts in 10,000 in the air. There has been a tendency to make the standard of ventilation higher and higher during the last few years, thus requiring the introduction of greater quantities of air.

Carbonic acid is continually increased by the processes of combustion and respiration, yet for the past thirty years the amount in the air has not sensibly changed.

Plant-growth and vegetable life assimilate carbonic acid and give off oxygen.* There exists in the air about 28 tons of carbonic acid to each acre of ground, yet an acre of beechforest annually absorbs about one ton, according to Chevandier; and no doubt the total vegetation growing is sufficient to absorb the excess due to combustion and respiration, so that the total does not experience much change.

Carbonic Oxide, CO.-This compound is not found in the air except under unusual circumstances. It is distinctly a poison, and has a characteristic reaction on the blood. Hempel, the German chemist, experimented on its poisonous

of organic matter with increase of carbon dioxide is a reasonable on The results of the experiment were as follows:

effects with a mouse. No symptoms of poisoning were detected until there were 6 parts CO in 10,000 of air, in which case after 3 hours' time respiration was difficult; in another case the mouse could scarcely breathe in 47 minutes. With 12 parts in 10,000 the mouse showed symptoms of poisoning in 7 minutes; with 29 parts in 10,000 the mouse died in convulsions in about two minutes.

25. Nitrogen-Argon.-The principal bulk of the earth's atmosphere is nitrogen, which exists uniformly diffused with oxygen and carbonic acid. This element is practically inert in all the processes of combustion or respiration. It is not affected in composition either by passing through a furnace during combustion or in passing through the lungs in process of respiration. Its action is to render the oxygen less active, and to absorb some part of the heat produced by the process of oxidation. It is an element very difficult to measure directly, as it can be made to enter into combination with only a few other elements, and then under peculiarly favorable circumstances.

A very small amount of ammonia, which is a compound of nitrogen and hydrogen, is found in the atmosphere.

Argon.-A constituent of the atmosphere recently discovered, which amounts to about one per cent of the total, was announced at the meeting of the Royal Society, January 31, 1895. This element is very soluble in water, and liquefies at a temperature 232° below zero F., under a pressure of 50.6 atmospheres. It is even more inert in action than nitrogen, and practically may be considered the same.

26. Analysis of Air.-The accurate analysis of air requires the determination of aqueous vapor, carbon dioxide, carbon monoxide, oxygen and ozone, but for sanitary purposes the determination of carbon dioxide and water is the most frequently called for. For a complete discussion of these various methods the reader is referred to Hempel's Gas Analysis, translated by Dennis and published by Macmillan & Co. The nitrogen of the atmosphere cannot be determined by any known method of analysis; it is obtained by deducting the sum of all the other elements from the total. The approximate determination of the oxygen is done very readily by

drawing a certain volume of the air into a measuring-vessel and then passing it over a mixture of pyrogallic acid and caustic potash; the oxygen is absorbed, reducing the volume in amount proportional to the quantity of oxygen. This process is, however, not of extreme accuracy, and for minute quantities very much more complicated methods must be resorted to.

Method of Finding Carbon Dioxide (CO).-The amount of this material present in the atmosphere is so small that the most delicate methods are required in order to measure it. The writer gives here the only simple method which can be rapidly applied, and which is said to be accurate to one part in one hundred thousand. This system of finding CO, was devised by Otto Pettersson and A. Palmqvist, two European chemists. The instrument used for this determination is shown in Fig. 8,

and can be had from any dealer in physical apparatus. It consists of a measuring-vessel, A, connected with a U-shaped burette B, from which communication can be made by a small stop-cock, b; a manometer, fg, containing a graduated scale nearly horizontal; and two stopcocks, f and g, by means of which communication can be made with the air. One side of this manometer, f, is in communication with the closed vessel C; the other side can be put in communication with the measuring-vessel A. The burette B contains a saturated solution of caustic potash (KOH). The flask E contains mercury, and by raising it, when the stop-cock c is open, the mercury will rise in the flask A, and the air will be driven out. If the flask E be lowered the mercury will flow from the measuring-tube, and the amount of air entering A can be measured by the gradua

tions. When the measuring-tube A is full of air, the stop-cocks c, b, f, and g being open, the position of the drop of liquid in the horizontal tube of the manometer is accurately read. The stop-cocks c, a, f, and g are then closed, that at b opened, and the vessel E raised, driving the air out of the measuring-tube A into the absorption burette B. This operation of raising and lowering the flask E is repeated several times; it is then lowered, and the air is drawn over into the measuring burette; the cock a is then opened and the vessel E manipulated until the reading of the manometer on the horizontal scale agrees with that in the beginning of the test. The reading of the graduated tube A gives directly the amount of CO,. The determinations are made with air of ordinary humidity, and there is a very slight correction due to this fact, which is not likely to equal, in any case, one part of CO, in one million parts of air.*

27. Determination of Humidity of the Air. The humidity of the air is determined by gradually cooling a body and observing at what temperature the vapor of the air condenses on the body as dew. When dew is deposited the air is saturated for the given temperature, and if the temperature of the air be known, at which dew will be deposited, and also the temperature of the air in its normal condition, we can compute the amount of moisture contained in the air. The instrument generally employed for this purpose consists of two thermometers, the bulb of one of which is exposed in its ordinary condition to the air; the bulb of the other is kept constantly wet by means of a bit of cloth extending to a vessel filled with water. If the air were saturated with moisture these two thermometers would give the same reading, but if the air is not saturated the readings will differ an amount depending upon the humidity. The table following, and a more complete one in the Appendix, give the amount of moisture expressed as percentage of saturation for different readings of the wet and dry bulb thermometer.

* For approximate methods of detern ining the purity of air see Appendix to book.

The first table gives the weight of moisture contained in one cubic foot of saturated air; the second shows the percentage of saturation for any difference in reading of the wet and the dry bulb thermometer. The weight of moisture is the product of the results. Thus, saturated air at 70° F. contains 7.98 grains per cubic foot, and if at the same time the differ. ence between the wet and dry bulb thermometers was 10, this air would be 66 per cent saturated, and would contain 66 per cent of 7.98 grains, or 5.26 grains. Since there are 7000 grains in one pound, this weight may, if desired, be reduced to pounds.

Moisture in air can also be determined approximately, but with sufficient accuracy for practical purposes, by the hair hygrometer. This instrument is illustrated in Fig. 9. It is constructed by fastening a hair, from which the oil has been removed, in the top part of a suitable frame, and winding the lower part on a cylinder which is free to revolve, and which carries a balanced pointer. The hair increases or diminishes in length, quite exactly, in proportion to the amount of moisture in the air, and this acquired property seems to be a permanent one. A scale graduated by comparison with determinations made with a wet and dry bulb thermometer serves to show the amount of moisture present, as a percentage of saturated air.

The degree of moisture in the air has an important in