NEON AND ARGON IN NATURAL GAS.

By HAMILTON P. CADY and DAVID F. MCFARLAND, University of Kansas, Lawrence. T THE last meeting of the Academy we reported the pres

ence of helium in natural gas. Since that time we have found argon in the Dexter gas and neon in that furnished the University by the Kansas Natural Gas Company.

It is not easy to detect neon in the presence of helium when working with small quantities of gas such as we used in the analyses previously reported, but during the preparation of considerable quantities of helium the neon showed its presence unmistakably.

The method for the isolation of these gases is briefly as follows: The natural gas is largely condensed in a bulb surrounded with liquid air. The uncondensed portion is then passed into bulbs filled with cocoanut charcoal and cooled with liquid air. Here all the gases except hydrogen, neon and helium are completely absorbed. Of these three gases hydrogen is rather freely absorbed, neon somewhat, and helium scarcely at all. In working with small quantities of gas the neon is very largely taken up by the charcoal, while with larger amounts the latter finally becomes saturated and then the lines of neon are plainly visible in the spectrum of the gas. The difference in the absorption of helium and of neon in cocoanut charcoal is great enough so that neon spectroscopically free from helium may be prepared as follows: A charcoal bulb is heated to about 400 degrees and exhausted as completely as possible. It is then cooled with liquid air and the unabsorbed gases from another charcoal bulb are passed in, allowed to stand for a time, and removed with a mercury pump. After a sufficient amount has been passed through the bulb, the helium is removed by pumping until a cathode-ray vacuum is obtained. The bulb is then warmed slightly and some gas removed; after this, by warming the bulb more strongly, neon spectroscopically free from helium is obtained.

The identity of the neon was established by measuring the wave-lengths of the lines in the spectrum and comparing them with those given by Baly (Phil. Trans. 1904, vol. 202, p. 183). Baly photographed something over a hundred lines in the visible portion of the spectrum. About half of these lines are

given as very faint. We have measured optically about sixtyfive of these lines, including all but the very faintest. The measurements, with one exception, agree with Baly's to within one angström. In this case it is quite evident that there is a misprint in Baly's table, for our measurement agrees well with that given by Living and Dewar (Proc. Roy. Soc. 67, 467, 1901) for this line, while it differs from Baly's by much more than the limit of error.

Besides the lines of neon, the tubes show about fifteen fairly strong lines which do not seem to belong to hydrogen, helium, or any of the familiar gases. These lines are given in a list of lines shown by the more volatile gases of the atmosphere and by the gases from Bath Springs as published by Living and Dewar (loc. cit.). This would seem to show that the lines are not accidental and that it is legitimate to look for a new very volatile gas in this residue.

The argon was identified from its spectrum by measurement of wave-lengths and by direct comparison of the spectrum with that given by a known specimen of argon.

SANITARY WATER ANALYSIS.

BY PROF. E. H. S. BAILEY, University of Kansas, Lawrence.

NOTWITHSTANDING all the advances that have been

made in analytical chemistry within the past twenty-five years, there seems to be little improvement in the methods employed for the sanitary analysis of potable waters. Referring back to the early methods of analysis that were worked out by Wanklyn and by Frankland, we find that there has been little change in these methods, except in the greater accuracy with which they can be carried out. The combustion method suggested by Frankland is so tedious that it is not often followed at present, and the value of the Wanklyn method is seriously questioned, especially by some of the members of the Geological Survey at Washington.

For the analysis of a water to determine whether it is suitable for domestic purposes, the most rational plan so far suggested seems to be to determine the nitrogen in the different forms of free ammonia, albuminoid ammonia, as nitrites and as nitrates. But there is considerable nitrogen in the various vegetable substances that may be dissolved in water, especially if the water flows over a rich or over a peaty soil. How shall we distinguish between the nitrogen from vegetable sources and that from sewage contamination, which might mean animal contamination? The organic bodies existing in the soil extract, or in the "humus," as this soil material is called, are very numerous and complex in structure. The free ammonia as obtained in an ordinary sanitary water analysis, if abundant, is supposed to be largely of animal origin. This is not necessarily true. Some waters are so loaded with organic matter, especially when associated with iron, that they yield a quantity of free ammonia which in other waters would characterize only sewage. This is often the case with artesian waters.

This fact can be best illustrated by referring to the analysis of the water used for supplying the city of Lawrence. These analyses were recently made by Dr. F. W. Bushong. The source of the supply is water obtained from "points" driven

in the bottom on the right bank of the Kansas river, a short

This water is pumped from the ground to the top of an inclined platform from which it falls over riffles into a basin. After being allowed to settle, it is run into another basin, from which the water is pumped into the city mains.

When first pumped the water is clear, but as soon as it is exposed to the air some of the gases escape, and some oxidation takes place, so that a large quantity of iron is precipitated, and the water becomes very turbid. The process of oxidation continues to some extent in the city mains, but even then the iron is not all precipitated, so that it frequently becomes turbid after being drawn from the faucet.

The organic matter of the soil has evidently yielded ammonia as a product of reduction. This may have been brought about by some vegetable organisms like the well-known Crenothrix, F, which contains iron as one of its necessary constituents. When the water has a chance to become aerated, most of the free ammonia escapes. The albuminoid ammonia retained is also small. There were no nitrates or nitrites in the freshly-drawn water, but it is evident that some of the ammonia has oxidized to these two bodies, hence we find them. in the city water. We have noticed that with this particular water the amount of nitrogen as nitrates found can be used an index of the thoroughness of the aeration at the plant.

There were no indications of bacterial contamination except in the case of the city water, and here during the past summer, and possibly at the present time, some river water has been pumped into the pipes, as the ground supply was not abundant enough. Of course it may be possible that the nitrates and nitrites came from the river water, but that is not probable, as the analysis has usually shown but a very small quantity of these in the river.

Water from the points does not show the presence of Colon bacillus, while the river water and the city supply do show it,

Later, the city water when obtained entirely from ground-water was free from bacterial contamination.

so we have concluded that when we get the pipes completely cleared of river water the city supply ought to show practically no bacterial contamination.

The analyses above quoted show the extreme importance of having a full and complete knowledge of the source of the water before passing any judgment on its quality. A decision that would apply to a river water would not apply to a well water, and a decision on the quality of a well water might not apply at all to an artesian-well water. The iron waters in the underflow of the valleys of the Middle West are a class by themselves.

The standards ordinarily applied to shallow-well waters do not apply at all here. Doctor Kinnicutt (Science, Vol. XXIII, p. 56) would exclude as suspicious a water that had more than the following quantities of nitrogen in 1,000,000 parts: In free ammonia, 0.05; albuminoid ammonia, 0.08; nitrates, 0.10.

Comparing the waters mentioned above with such a standard would be utterly futile. Then, afterwards the process of aeration entirely changes the character of the water, and it must be compared with surface-waters, in which more albuminoid ammonia would be allowable.

Facts of this kind only emphasize the necessity for more satisfactory chemical methods for water analysis, and a careful study of each water by itself, without too close reliance on standards that are fixed even for that particular class of waters.