were parts of piping schemes, bad burners and disgusted foremen and operators, and the net result of our looking around simmered down to the conviction that the majority of the failures was due to the fact that in most works tar was tar and that whatever effort they expended was always along the line of arranging piping and buying or devising burners with little

or no regard to preparing the tar so that it would flow smoothly. We started by building two strainers (see sketch A) as follows: Two 18" lengths of 3" W. I. pipe were cut off and threaded on each end. A 3" cap was bored in the center and tapped for a 16" length of 1" pipe, which was bored full of " holes and screwed through the cap from the inside, then the 3" cap was made upon one end of one of the 3" pipes which had been tapped to receive a 34" inlet about 4" from the capped end. Cylindrical strainers, one of perforated tin and a finer one of brass gauze (such as is used for dairy screens), were slipped concentrically over the I" outlet, but were held from making contact with it or with each other, by rings soldered at their tops and bottoms. Then the top of the 3" pipe was loosely capped to exclude dust but left so that the screens could readily be removed for cleaning.

One of these strainers was placed on the inlet of the tar reservoir and one between the tar reservoir and the burner. The tar reservoir was a steel tank holding 78 gallons and containing a small steam coil so that tar could be maintained at a temperature of from 100° to 110°F.

We had invested $15 in a well-known tar burner ("Parsons') in which the tar was brought to a central circular orifice by gravity and there it was supposed to ooze out into a circular jet of steam which surrounded the tar outlet; as an atomizer it was all that could be desired, the flame was smokeless and intensely white and within two hours the settings were dripping fused fire brick. Then the burner was tilted down and the flame baffled on a pile of broken brick thrown in onto the grate bars.

This gave a better distribution of heat, but, though we carefully experimented with the burner for nearly three weeks, we were not able to strike a combination that would heat the bench evenly and the yield per mouthpiece was correspondingly low.

The objections noted were bad enough, but added to this we found that in spite of the care with which we strained and maintained an even temperature on the tar, two men and a small boy could not keep that burner going and that in the face of all we could do it would stop up at intervals of from fifteen minutes to two hours.

In the burner mentioned, the flow of tar was regulated by a needle valve and, as we were cutting the consumption as low as possible, the valve was nearly closed and what tar did pass was forced into a very thin sheet surrounding the needle. Thinking this might account for its stopping, we carefully filed a slot of circular cross-section lengthwise of the needle, thus allowing the necessary amount of tar to flow in a solid body. This was a great help, but, as the burner did not operate on the syphon principle and there was no pressure on the tar except such as was afforded by about six feet of head, the trouble was not entirely eliminated, and as coupled with this we were unable to get anything like reliable operation, the results were far from satisfactory or economical.

As will be noted, the results obtained up to this point were entirely negative, but they have been given with the idea that it is frequently of value to know what not to do.

The first favorable results which we attained in burning tar were with burners which were built of pipe and fittings, and from

first to last, good, bad or indifferent, they have all operated on the ejector principle as this will largely eliminate stoppages due to slight changes in the consistency of the tar, and after many changes we have finally settled on the simple form (see sketch B) which we are using at present.

This burner can be built for 25 to 30 cents, including labor,

SKETCH

and is installed through the center of 42" opening in firing door, which we have bricked up with exception of opening mentioned.

The burner never stops up, requires practically no attention and running the tar through one screen having 16 meshes to the inch is all that is necessary, as the burner exerts considerable suction and has no constricted tar passages.

We are regularly carbonizing from 42 to 5 tons per day of 24 hours with an average consumption of 138 gallons, or about 37 gallons per ton; on tests this can readily be reduced to 134 gallons, but the average running will show up somewhat higher.

The figures given are hardly fair to the tar burner, as we find that the men pick lump coal for the coal-fired benches and give the tar-fired bench all the fine or wet stuff that remains. This can be done as the burner seems to have no trouble in holding up the heats.

It may be well to note that in using this burner we found it necessary to get rid of several preconceived ideas. In the first place, we used about 220 gallons of tar per day in the endeavor to get a white flame and still keep up the heats. By reducing the amount of steam admitted, we found that the tar could be reduced and no cooling of the bench was noticed. This was kept up until the burner was doing better work on 138 gallons than it did on 220 gallons, but the resulting flame from which the most economical results will be obtained is far from white. In fact, it is a rather slow, reddish yellow, but does not smoke or soot up the setting nor is any amount of smoke perceptible at the stack.

The setting with which the experiments were made is a Parker Russell Co. Standard half depth, semi-recuperating bench of sixes. We removed the arch and directed the flame upward directly among the retorts and against the saddles at a point about three feet back from the firing door, the flame first appearing on line with the lower retort, but in four months' run there has been no sign of melting them down or in any way working undue injury to the bench.

The tar supply is regulated with a 1⁄2" stove cock as shown in Sketch C, as this enables the operátor to observe the amount

of tar he is feeding. No breeze or other fuel is used and no fire brick is necessary to protect the grate bars, as the flame is

up among the retorts instead of being down in the furnace.

To burn tar in our electric plant a burner was constructed so nearly like the one which was later described by Mr. George Reed, of Binghamton, N. Y., in figure 15, page 570, American Gas Light Journal of April 10, 1905, and also on page 175, Progressive Age of April 15, 1905, that the description and drawing given will cover it, except that we used no fittings larger than 1⁄2" and did not draw down or contract the nozzle which was inserted through a 1⁄2" hole drilled in the boiler fronts above and in no way interfering with the firing doors.

The above arrangement made it possible to change from coal to tar at a moment's notice and in carrying a peak load we sometimes used both until we found that the tar burner alone was capable of making more steam than we could use.

The boilers are Class E. Sterling Water tube, rated at 202 H. P., having 2019 sq. ft. of heating surface and 42.66 sq. ft. of grate surface, stack draft about 6/10"

Tar used weighed 934 lbs. per gal. and the average evaporation was 12.82 lbs. of water per lb. of tar. This was largely due to an unfavorable load factor during several hours of the day, as full load tests indicated an evaporation of nearly 14 lbs. of water per lb. of tar.

The tar is pumped without any strainers whatever, direct from the tar well to the burners by a small Marsh steam pump working under a pressure of from 125 to 170 lbs. per inch. It will be noted that the tar pressure must be higher than the steam pressure used, and in our case this varies from 100 lbs. during light loads to 150 lbs. at the peak.

The flame is directed downward by bending the nozzle so that it will strike the grate bars about 32 feet back from the furnace doors, where it is received on a bed of about 8" of ash, clinker, coal or breeze. The ash pit doors are left open and ash pits are filled with water to prevent warping the grates, but as there is very little draft through the mass of material on the grates the principal air supply is an over draft admitted through shutters in the firing doors.

The burners rarely clog and require a minimum of attention. Flame is one whirl of intense white, there is no smoke and the entire installation can be made for $5 per boiler, exclusive of pump, which in our case cost about $31 and is large enough to supply two boilers to their greatest capacity.

In figuring the fuel value of tar against any fixed price per gallon, it must be remembered that the price received is never net, since an allowance must be made to cover pumping, barreling and carting the product. In burning tar, these costs, with the possible exception of the pumping, can be entirely eliminated and even this may be done away with in firing