It has been found that the lowest dewpoint temperature will be coincident with thawing weather, following severe frost, when the cold water soaks deeper in the ground.

We did not arrive at this seemingly simple method, without meeting many contradictory and disappointing results that cannot be touched on without using too much of your valuable time, but I hope that you will permit me to speak a little further on the subject of photometry and the precautions necessary in getting reliable comparative results in two photometers situated one mile apart.

We read candle power hourly, both at the works and testing stations, the candle power readers having no other duties. They work in three shifts of eight hours, and at both the works and testing stations report to a chief photometrist, who, with the Superintendents, report to the Engineer of Works. In the works' photometer room, the works' Superintendent has no authority and has no control over candle power readers' work, except to request more frequent readings if desired.

Each works makes coal gas and water gas, and for photometric samples a separate coil immersed in a 70° bath is provided for each gas; that is, one for each of the coal gas,.

water gas, mixed gas at holder inlet, and commercial gas. Also a separate photometer meter is provided for each kind, through which the sample is kept constantly flowing, to be switched on to the photometer burner as desired. Hourly readings are taken of the candle power of each gas, and more frequently if desired by the Superintendent. At each works two bar photometers are provided and are in constant use in adjoining rooms, thus giving a check upon each other if there is reason to doubt the accuracy of either from some undetected maladjustment.

At the testing station the temperature of the gas must not be allowed to fall below main temperature outside.

For such frequent readings and particularly in a climate like that along the 42nd parallel of latitude, we found it necessary to abandon candles promptly. Our experience with candles gives good ground for the statement that, under the conditions enumerated, standard sperm candles are impracticable as a source of light.

The representative of the municipality, Dr. N. W. Thomas, after carefully weighing the merits and demerits of the different lamps, decided upon adopting the Pentane lamp. After six years' experience with this lamp, having 12 in constant use, I am inclined to urge this Institute to take the necessary action to make such further investigation of its peculiarities as will be necessary to adopt it as a standard of light for the gas companies of this country, or reject it in favor of something better.

While the Carcel, Hefner and Harcourt to candle Pentane lamps are all in use, our investigation discloses the fact that but one of these, the Harcourt 10 candle Pentane lamp, has been adopted as a standard by legislation. This lamp is the standard for the Gas Referees of London, the Examiner for the London County Council, the National British Physical Laboratory, and is also used as a standard in the Electrical Testing Laboratory in New York. In this country it has been the subject of investigation, and some of the results have been presented to the Gas Associations. They indicate the need of making further investigation as to certain peculiarities.

We are dilatory in taking action on standardizing photometrical methods in this country. The urgency of action

arises from the tendency of municipalities and state authorities to frame laws regarding inspection of our product. In the case of the District of Columbia, Congress legislated that the gas must be equivalent to 25 candle power, measured on a Bunsen Photometer. This was the same photometer specified in the Gas Works Clauses, Act of Parliament of 1871, which goes with considerable detail into the method and apparatus required. In later legislation by Parliament, providing for the appointment of Referees for certain companies, much of this method of testing described in the Act of 1871, was repealed and left to the Referees for those companies with which they were officially connected, but Parliament in all these Acts specified that for common illuminating gas, the candle power shall be determined by burning through a Sugg's London Argand Burner. The Gas Referees of London, having to deal with gas of 12 to 16 candle power, have made seemingly wide departures in the method of photometry, and the gas companies of this country have not followed them. If, therefore, we do not get together and formulate specifications based on proven scientific methods, we will have less cause to complain if we are served with something in the way of legislation that proves difficult to digest. As an illustration, this year, in one of the large cities, a University Professor was engaged by a newspaper to make observations of the candle power furnished by the gas companies; he used a photometer that recommended itself largely because it could be carried in an ordinary suit case. I will say that no observations are entirely accurate when the gas saturated with hydrocarbon vapor, or nearly so, is passed through rubber connection.

DISCUSSION.

THE PRESIDENT: Gentlemen, you have heard the reading of this most valuable paper, and I hope it will be given a very thorough discussion. Mr. Gartley, will you please have a seat on the platform? Mr. Forstall, will you have anything to say on the subject?

MR. WALTON FORSTALL : Mr. President, I regret it very

much, but my line of work has been for many years in connection with the distribution of gas exclusively.

THE PRESIDENT : Mr. Walton Clark, will you not have something to say on this subject?

MR. WALTON CLARK: Mr. Gartley has said about all that I can say. I will say, however, that it has been a very important matter with us at home in Philadelphia to maintain our standard of candle power with the least use of oil. We are very rigidly held to 22 candle power at a point a mile from the works. We have not felt, however, that we were entirely meeting the expectations of the people of Philadelphia simply supplying 22 candle power gas a mile from the works; and we have made every effort by experiment and investigation in the hope of being able to find a method by which we could maintain that candle power to the consumer without having it much above 22 at the testing station. Mr. Gartley set up photometers in many parts of this city, and has given a great deal of study to the question. How well he has succeeded the paper shows. He and Mr. Bond, and the other men who worked with them, have worked out their method. There may be other methods of doing it that are equally successful, but there is no method that is more successful, I think, because as he has shown in his paper, that, carrying a candle power of approximately 2234 at the testing station, a mile from the works, he is able to deliver 22 candle power to his consumers generally. That shows, I think, that the method is eminently successful.

THE PRESIDENT:

Mr. McDonald, will you discuss this?

MR. D. McDONALD, Louisville : Mr. President, I have nothing to say. I think it is all right because Mr. Gartley

says so.

MR. NETTLETON: May I say a word before Mr. Gartley replies? I confess that I am astonished at the results claimed. I am free to say, however, that I do not see exactly the difference between his method and the usual practice. As I understand Mr. Gartley, he is using an is using an oil that has the reputation of producing a gas, the candle power, of which

falls off badly in extremely cold weather. If Mr. Gartley will explain what he does to make his candle power hold up, I will be greatly obliged.

MR. GARTLEY: I do not wish to be understood as saying that there will be no falling off in candle power when the gas is cooled. One of the purposes of my paper is to show how much the falling off in candle power will be. If you will refer to Plate 4, you will find there a curve that gives the loss in candle power that is found in a gas made from Texas oil, between the limits of 70 degrees and 35 degrees. For instance, if it has been found that the gas will drop in the mains to 40 degrees in temperature, this curve shows that we must put into our commercial holder a gas of 25.3 candle power, the gas being cooled at the inlet to the works' photometer to 70 degrees to yield 22.5 candle power at 40 degrees. I do not mean that the gas must show forty degrees on a thermometer inserted in the main at the testing station; I mean that the gas shall have been cooled as low as 40 degrees somewhere between the works and the testing station, and this determination is made not by a thermometer, but by the hygrometer described in the paper. If it has been cooled at some intermediate point and is then increased in temperature again before arriving at the testing station, it would be a gas unsaturated with hydrocarbon vapors, and as curve 4 deals only with gases saturated, and as the gas after being cooled to 40 degrees cannot increase in candle power between that point and the testing station, the hygrometer will show the temperature which must be applied on curve 4 to tell us what candle power we must put into the commercial holder. Do not understand me as saying that the gas to the consumer will fall 2.8 candles under these conditions, because such is not the case. The difference is due to the fact that on a bar photometer (which is the only accurate means of determining candle power) the gas is passed through a wet meter, while the consumers' gas is passed through a dry meter. If a gas that has been cooled to 40 degrees passes into a photometer meter, the temperature of which is 70 degrees, the gas will absorb water vapor in the meter until it becomes saturated with water vapor at 70 degrees. At 40 degrees the gas contained .83 of 1% of water vapor. At 70 degrees it will have