various tests reported. As, for instance, indicating the substantially uniform rate of evolution of gas from a charge of coal irrespective of the weight of the charge, in view of the general knowledge that the capacity of a bench per day is substantially the same on varying lengths of charge, provided the weights charged correspond to the time.

The most interesting result to me is the general effect of the paper as confirming my own opinion that the finer points in the distillation of coal and differing results in different works, are elements which can never be wholly reconciled.

The title of the paper in describing the subject as Constants and Variables, is, I think, in itself somewhat misleading and would be more truly expressed as Known and Unknown Factors in the Destructive Distillation of Coal.

A "variable" in this problem exists only in the form of an unknown constant. In other words, like conditions inevitably and hence must invariably produce like results. We may accept this as a fundamental law of the universe and when we are looking for so-called variables in the process of making gas from coal, we must begin with the coal itself. We must remember that every foot of thickness of a vein of coal may comprehend as well 10,000 or 100,000 years. That a vein of slate may hold the story of a completely different climatic production of the vegetation comprising the coal below and above it. That it may represent an indefinite lapse of time and changed conditions. So we might say no two cubic feet of coal have the same organic composition.

Then we proceed to mine this coal, having all these possible differences, to ship it long distances exposed to varying weather conditions for varying periods, and to distil it in a retort in which it is not humanly possible that either the heat applied, the barometrical or static pressure,or the physical condition of the retort from day to day or even from hour to hour can be made uniform.

Any result, therefore, must be accepted at its very best as only a set of averages and when we apply this to the work done by the University students in the tests reported, it seems to me that it is a high tribute to the care, persistence and accuracy of Prof. White and his students. Not only have they all the

above inherent factors to deal with, but in addition they are operating from a single retort where every "variable" aggravates the extremes in the average.

The influence of pressure during the process of carbonization is perhaps one of the most important as well as one of the most neglected elements in the entire process, and it would be to me a matter of great interest if Prof. White should be able to carry out his suggestion of investigations along this line during the coming year in an effort to determine, if possible, not only the influence of differing degrees of pressure on the composition of the gas, but likewise the influence exerted by those sudden changes of pressure commonly termed pulsation.

I think that the Association is very much to be congratulated in having such hearty co-operation of not only the University officials, but as well of Mr. Douglas, the use of whose carbonizing plant enables the carrying on of these tests under the exigencies of practical operation, and to whose helpfulness very much more is due than appears on the surface.

V. F. Dewey, Detroit:

The points of particular interest to me are:

First, "Comparison of rate of evolution of gas from one retort with different charges." The conclusions as to the sizes of charges and the make of gas, the temperature remaining constant, are practically the same as has been found by most operators, as a consequence, general practice today is toward large charges and corresponding long periods of carbonization with no diminution in the make per retort per 24 hours.

Second, "The effect of Temperature changes on the make per retort." On this point we are certainly on debatable ground. The assumption that the increase in gas made will vary in terms of per cent increase in temperature demands much further data than already obtained to be supported.

I believe it can be granted that the rate of gas production is a function of the rate of heat transmission to the charge but I am not prepared to accept the statement that silica retorts give different results than ordinary fire-brick for the above reason,

because there has been nothing as yet brought out that states in definite terms that silica has any greater calorific conductivity than ordinary fire-brick. An interesting point was brought out in the paper in this connection, regarding the influence of carbon deposits on the rate of gas production. This may explain why certain kinds of material do give different and better results than others. It has been observed and noted without possibly appreciating its important bearing on the point in question that silica retorts do not carbon up as much as other retorts.

Third, "The influence of the variation of pressure on the retort." It is unfortunate that definite data were not obtained on this point. Experience has shown us that the maintaining of a very slight pressure with a variation of not to exceed 2-100ths of an inch water pressure does increase the yield without hurting the candle power and with very slight loss in B.T.U's. Average results in a large retort house operating under the above conditions show that the B.T.U. in gas from one pound of coal, m & a free, is from 3600 to 3700.

The paper is an extremely valuable contribution to the gas literature of the country and shows in no uncertain terms the value of the work being carried on by the University of Michigan and the Michigan Gas Association.

R. Schaddelee:

There is one variable that Professor White and Mr. Ferguson do not mention in their paper, which I think may have some bearing on the question. That is the question as to the introduction of the charges into the retort.

Were they introduced in exactly the same way in each case? My experience has been that, even though you put the same kind of coal in the retort, the quantity of gas you get out of the coal depends upon the way the coal is charged into the retort.

Mr. Tippy mentioned a very good thing in this connection, viz. putting a water gauge on the different standpipes to ascertain just what the pressure is in the retort.

I think that if we would try this oftener we would be convinced that we are far from getting a constant pressure on the

retorts.

This is due to two causes, viz.: fluctuations in pressure due to the pulsations of the exhauster and the intervention of either a liquor or tar seal between the exhauster and the retorts.

I was very interested to hear Mr. Batten state that in Detroit they are now unsealing the dip-pipes as soon as the charges are introduced. This is, in my opinion, an excellent idea.

W. S. Blauvelt, Detroit:

My attention was particularly attracted to the point Mr. Ball brought out, with regard to the Hellier coal, that weathering it 15 months seemed to make it a more valuable gas coal. The coal while exposed to the weather for a year apparently increased in value, which is very satisfactory though different from our almost universal experience. The explanation is, undoubtedly, that there were some variables that were not noted. Probably in the operation of the experimental plant, just as in the operation of a commercial plant, slight improvements were made during the year in a number of different places, which in the aggregate might more than counterbalance the deterioration of the coal due to weathering during one year.

There is one variable, however, that might affect the yield of gas and the total B. T. U. in the gas per lb. of coal, that is, the extent to which the volatile matter is expelled from the charge. We have no record of analysis of coke, but the report shows that in the later tests, the weathered coal with but one exception yielded less coke than was obtained from freshly mined coal. High volatile in coke might explain high yield of coke and low yield of gas.

You will notice that in these last two tests the yields of coke were very distinctly lower than in the others, which might possibly explain the difference in the yield of gas. In the one case a portion if the gas may have been left in the coke, in the other may have been more thoroughly expelled.

The best results were obtained at temperatures from 1800 to 1900 degrees. I think that possibly the reason for this is that with extremely high temperatures, a larger percentage of the gas is cracked, and as a result, the tar is high in free carbon, and there is a heavy deposit of carbon on the surface of the retort when the

temperature is low, much below 1800, on the contrary a very oily tar is obtained. This tar is low in free carbon but high in volatile oils which, if exposed before condensation to a slightly higher temperature, would have been cracked and would have appeared in the form of permanent mixed gases.

I am very glad to see the point brought out of the effect of variation in the pressure on the retort. In coke oven work, the variation in results due to variation in retort pressure is very marked. We find that our yields of gas, tar and ammonia are invariably affected over pressure variations. Both the calorific power and the illuminating power of the gas may be materially changed by very slight variations of the retort pressures.

I believe that in any method of manufacturing coal gas, it will be found eventually, that some form of retort governor which will control pressures within a very narrow range, say, three one-hundredths of an inch, will be found to be of very great value in increasing the yield and in insuring a uniformity in the quality of the gas obtained. I might say here too,that this problem is one that is being worked on at the university and that we hope to have installed as soon as possible an exhaust fan instead of the present exhauster, to take the gas from the retort. It is impractical, so far as we can see absolutely impossible, to secure a uniform pressure in the retort when the gas is drawn away with this small exhauster. The pulsations due to the exhauster induce a condition which makes it impossible to control the pressure. With a fan exhauter the pulsations will be eliminated, and we believe we can secure a satisfactory pressure control in this way. It will certainly be extremely interesting, should we succeed in this endeavor, then to go over some of the past year's work and find out whether we can get more satisfactory results than when pressure variations could not be properly controlled.