take it for granted that they utilize the exhaust. I would like to know what amount of back pressure it is customary to provide for in specifications?

Mr. Quay:-Well, five pounds is considered the maximum. With the use of any kind of a vacuum system to remove air from the radiators we heat from atmospheric pressure. We have two jobs working now where we have connected the Paul systern. We are heating both buildings at atmospheric pressure. We set the reducing pressure valve so that it works sometimes a little below, sometimes just at atmosphere. We found quite an interesting experience in this case where the engineer had not been used to heating at this pressure, and during a cold spell he thought he must have two or three pounds pressure. By a good deal of persuasion we induced him to try with atmospheric pressure and set the gauge at that point and it was only a short time when he came up in the building and said: "Our boilers are going to blow off." We told him to open his doors and let them cool down. He has spoken since about the economy of fuel that he has noticed, while not using exhaust steam entirely. But five pounds has been the maximum and very often we carry two pounds without any vacuum system.

Mr. Hoffman:-I live in the West and I know it is the custom almost altogether there, as Mr. Quay says, to use a single pipe system. I have an idea that one of the troubles that they sometimes have to contend with is getting water in the radiators on the descending pipe. If it is very cold the water will suck into the radiators and fill them, and the engineers have more or less trouble in getting rid of this difficulty. If I might step to the board I would like to give a little illustration of a T that I think would overcome the difficulty. (Mr. Hoffman made a sketch on the blackboard). Mr. Jellett: I have not seen that particular fitting.

Mr. Hoffman:-You have not seen that particular fitting, because I do not think it has ever been made.

Mr. Quay:-I do not think the water in the radiator is caused by what Mr. Hoffman says. We have heated buildings with temporary steam, I-pipe system, where the windows were not properly closed and where the buildings were wet and cold, the most unfavorable conditions you could find. In fact we sometimes found snow on the radiators before we turned the steam on-not afterwards (Laughter). But the difficulty is with a pipe either not having the proper pitch or not being the proper size, so that the water is held in the radiator, or not having the proper kind of a valve. We had a little experience a short time ago in regard to the kind of radiator valve that is used. We ordered corner valves, and through a clerical

error did not say "off-set" corner valves. It was a job out of the city. Our superintendent got the valves and put them on, and he first wrote me that the job was working very nicely. The next day he wrote us that the radiators were holding water and he had found that it was the fault of the valve. If you have noticed the ordinary corner radiator valve, the seat is raised so that it cuts off about half the outlet or the inlet of the valve. Now, this seat of the valve held the water in the radiator, cutting off half the opening, and there was a great deal of condensation and the water could not get back. I had the man take the valves all off and I wrote a letter to the manufacturer, who is one of the best in the country, rather condemning the manufacturer for putting such a valve on the market. He said that in the East they had been using that valve for years and he had never heard a complaint. I said to myself—I did not say it very loud-"I can understand why some people say that the 1-pipe system won't work; they have been using that valve with a 2-pipe system and if they had to put it on a 1-pipe system they would find right away that it would not work." There have been certain tees manufactured and put on the market with the idea that Mr. Hoffman suggests, but we never found it necessary to use them. I think if the piping is properly laid out, proper sizes and proper grades, and all those things considered carefully, that you will have no trouble about the water remaining in the radiator.

XXII. TOPIC No. 7.

Does the color of the finish on a radiator materially affect its radiating power?

Prof. Carpenter:-I would say that we made a few tests of the effect of paint on radiating surfaces and we did not find that the color made any particular difference. We did not have time to go through all the ranges of color in as thorough a manner as we should like. I believe we tested bronze, a coat of yellow paint, and of white paint. Regarding the results of these tests, see my paper in which as originally printed there is a mistake in the print. At the bottom of the table on page it is stated that a given pipe painted with asphaltum 113.5; pipe painted with light drab lead paint 126.7, as compared with unprotected pipe. The figures should be 103.5 and 106.7. The error was due to some mistake of my own, I suppose, in copying. The actual results of our tests showed that in both cases the painted surfaces were somewhat superior to the plain pipe unpainted. Beyond that we have not gone. We have not tested enameled surfaces and I have nothing to add in respect to it.

Mr. Fish-It is stated by some authorities that in pipes painted with black the efficiency of the radiator is increased.

Prof. Carpenter:-Yes, sir, we found about three per cent increase in efficiency.

Mr. Fish:-Walter Jones drew my attention to that fact and in his book he points that out, naming ordinary lamp black coating. In bronzing pipes and painting them red and green and having them coated over several times I never noticed any special differ

ence.

Mr. B. H. Carpenter:-I would like to ask if any one has had any experience or made any tests with second or third coatings, and what difference it makes?

Prof. Carpenter:-The bed pipe referred to in my paper was coated with two coats of lead paint. The results seemed to be a little better with lead than with asphaltum. It would seem from a theoretical standpoint that an enameled and smooth surface would have less radiating power than a rough surface, but personally I could not say what the results would be.

Mr. Fish:-Might I ask Prof. Carpenter to explain why, in his judgment, in pipes covered with lamp black or a black coating of lead paint the efficiency is increased? What is the theory of it?

Prof. Carpenter:-I think the same theory applies to all painted surfaces. If you will notice the table on page of my paper you will see that as the conductibility of coverings increase the amount of heat which is transmitted with certain coverings is increased rather than diminished. That, I think, applies in the same way to the metallic paints. The metallic paints unite very firmly with the metal. They increase the extent of surface and make, perhaps, a superior surface for delivery of the heat into the air, and in that way they increase rather than decrease the total amount of heat transmitted. You all know that the great resistance to the passage of heat from metal into the air is found on the exterior surface of the metal. Anything that we can do to lessen that resistance increases the efficiency of the surface; some of the paints seem to increase the efficiency of the surface very slightly. Whether all paints will can only be determined by trial.

Mr. Hoffman:-There is an old saying that nature abhors a vacuum, and isn't it possibly true that in the rough surfaces of the iron, as it first comes out, there is more or less space which would be very minute, but which would naturally be occupied by what we might term air globules? Now, we know that air will not conduct heat very rapidly. It is practically a non-conductor of heat; and is not that a possible explanation of the non-conductibility or superior

ity of the painted surface, because it smooths down the 'iron and fills up those little pores or openings so that the air cannot stay there?

Prof. Carpenter:-That might be. My idea is, however, a little different, although I am not positive. I had thought that a painted surface might be somewhat rougher than a plain surface; that is, under a very powerful microscope a painted surface might show more irregularities than the unpainted iron and in that way the surface would be somewhat increased by the paint. Perhaps I am wrong. It is a subject I have not investigated at all.

Mr. Hoffman:-I know it is an absolute fact, by test, that so far as a nickel-plated stove is concerned, there is very little radiating power in the nickel-plated part of the stove. I burned my hand one day in trying to demonstrate that. I got down to the stove before I knew I was anywhere near it.

Prof. Carpenter:-It is a well-known fact that the radiating power of polished surfaces is very small. But that is not referred to in this case at all.

XXII.-TOPIC No. 8.

In a fan system of warming, which is better-to draw the air through the heater or force it through?

Mr. B. H. Carpenter:--I have never had actual tests of drawing through a heater at all, but I have several times forced it through, and from plans that have been gotten out by the people who make fans and heaters, I notice in almost every case which I have seen that they have forced it through, and they generally recommend this method, but why, I do not know. I should like to hear from somebody who has had experience in both cases.

Mr. Jellett: I do not agree with Mr. Carpenter. I had experience with both blowing air and drawing air through heaters. While it is true that the fan makers' catalogues show more pictures of fans blowing through the heater, the same fan makers, in practice, put up more heaters the other way. The objection to the pulling through in many cases I can illustrate by an example which occurred some four years ago. There was not width enough in the room to draw through. There was a basement area window and the window was built in front of the blower. We found that in wet weather, in snowy weather, when it was particularly cold, that moisture condensed on the inlet of the fan and froze. I found on one occasion when there had been a heavy storm at night, the fan running at very slow speed during the night, that there was ice on the blades of the fan, and when the fan was speeded up in the morn

ing, the blades were torn out of the fan by reason of the ice. Now, this is not an isolated condition. I know of quite a number of cases where this same trouble has been experienced. I also find the trouble that at very low temperature of the air you cannot get an oil feeder that will work satisfactorily on slow moving fans. I have had trouble with hot boxes, with grinding out at times, and with the oil not feeding-simply that the temperature was so low that the oil or grease was held and the contact point not being on the shaft, but being fed automatically from above, we got no oil. This is overcome by drawing the air from the heater to the blower. We take the air from out-doors to the heater section, the air striking the first four to six pipes, meeting there the worst conditions. It has been our practice to feed this back section with a larger supply and return with a larger opening. Taking the air through this heater, it enters the blower at a high temperature. The blower is working then under the same normal condition of temperature at all times; the bearings are warm but not excessively warm; the temperature of the air from 130 to 160 degrees. It is not sufficient to affect the bearings in any way. It does keep the oil supply at all times free, and then, apart from that, you have a clear, straight delivery from the mouth of your fan to the ducts. I have found that a clear delivery from the mouth of the blower to the system of ducts gives a better distribution than blowing through the system of pipes first and afterwards to the ducts. My experience has been largely on the side of drawing through the heater rather than forcing it through.

XXII. TOPIC No. 9.

As the temperature of the air in heated rooms varies with the height above the floor at which it is measured, is it not better to specify the temperature of the air within a given distance of the floor rather than to say that the room must be kept at 70 degrees?

Mr. Payne: I always make it a rule to locate the thermometer as nearly as possible in the exact center of the room, and I have never found any party for whom I installed work to object to that, for I always reason that what I guarantee is of course the average temperature of the room. If you try to heat the whole of it to 70 degrees, some of it would have to be away up to 85 degrees or thereabouts.

Mr. B. H. Carpenter:-I think that would vary considerably with the style of heating you are putting in. If you are using one style of heating the difference of temperature may be quite considerable at different heights from the floor, while with another, where it is