efficiency (code item 73). The tendency of the 72* line to fall is apparent.

A glance at curve No. 2 shows that ash has a worse effect on amount of steam generated, the decrease of which is considerable, than on efficiency.

Curve No. 3, combustion-chamber temperature, shows a steady drop of a few hundred degrees due to increasingly poorer fuel-bed conditions along with falling rates of steam generation.

Curve No. 4, flue-gas temperature, shows a slight drop on account of lower furnace temperatures.

Curve No. 5, the pounds of dry chimney gases per pound of “combustible," shows a marked and steady rise which is the chief reason for the decreasing-furnace and flue-gas temperatures just referred to. It is usually hard to keep the fuel bed in good condition when a coal is high in ash.

Curve No. 6 gives the averages for each group of the ratios of the carbon to the total hydrogen in the dry coal. This line drops, showing that the coals become poorer. The coals in the left-hand groups contain less oxygen in their molecules, so that their hydrogen, though decreasing in percentage, is available in larger amount. The fact that on going to the right the carbon-hydrogen ratio decreases partly accounts for the falling efficiencies shown by curve No. 1. (See p. 68.) Curve No. 7, per cent of completeness of combustion" (E), indicates approximately the percentage of total heat liberated from the coal which ascended from the grate. It rises decidedly toward the right in spite of the poorer fuel-bed conditions. The explanation is that inasmuch as the furnace temperature was almost always above the ignition temperatures of all the combustible gases and smoke particles present combustion was retarded more on the left than on the right, because less oxygen was present. For this reason more of the possible heat was generated with the higher ash coals.

It is worthy of note that the high and low efficiencies (72*) in each group (values not shown here) do not vary much between groups. Commercially this similarity of highest efficiencies means that buying coal on the basis of per cent of ash shown by proximate analysis of coal as fired is an uncertain matter so far as obtaining high efficiency thereby is concerned. Classification of the tests made on Illinois, Indiana, and western Kentucky coals on an ash basis shows no more than the minor importance of ash above mentioned.

In making tests on sized coal, as given on pages 45 to 48, there would be an opportunity for research on the question of per cent of ash as affecting the results of boiler trials. Coals of the same size could be classified on the basis of per cent of ash and the effect of the ash thus studied separately from the effect of size. It does not seem, how

a See glossary (p. 181), under "Efficiencies."

ever, that for the range of per cent of ash found in coal this investigation would be important, for it is not the amount of ash, but its composition and distribution, that affects the results of a steaming test. With a knowledge of ash composition it could be decided how the coal was to be burned. For instance, a coal which did not clinker badly, but which burned leaving free ash, would be burned under ideal conditions on a rocking grate no matter how high the per cent. of ash. A coal which clinkered badly at high temperatures should

FIG. 25.--Relations of per cent of sulphur in coal to boiler efficiency 72* (curve No. 1); per cent of completeness of combustion (E3) (curve No. 2); pounds of dry chimney gases per pound of "combustible” (curve No. 3); and combustion-chamber temperature (°F.) (curve No. 4) Tests 89-400.

be gasified at a low temperature, thereby avoiding an amount of clinkering that would stop up the air passages and also preventing clinkers from adhering to the grate. It is a study of ash composition along with the manner of distribution of the ash in the coal that is needed. The curves of fig. 28 present a classification based on per cent of sulphur in dry coal. For the low sulphur values there is a wide variation in efficiency 72*. As the per cent of sulphur increases, up

to 5 per cent, there is a gradual drop in efficiency 72*. There were 16 tests in which the per cent of sulphur was over 5.4. At the high values of sulphur the efficiency 72* curve rises, probably because the sulphur is in different form and also differently distributed. A curve through the average efficiency points shows a drop of 5 per cent, with an increase of 5 per cent in the sulphur, or, in other words, every increase in the per cent of sulphur in dry coal decreases the efficiency in equal amount.

Curve No. 2 shows the per cent of completeness of combustion (E), and is nearly parallel to the efficiency 72* curve.

Curve No. 3 shows the pounds of dry chimney gases per pound of "combustible," and rises slightly with the per cent of sulphur.

Curve No. 4 shows the combustion-chamber temperature, which, except for the high sulphur values, decreases as the per cent of sulphur increases, indicating a value at the high sulphur values about 300° F. higher than an average curve through the combustionchamber temperature points would indicate that it should be. The curves of fig. 17 (p. 25) show that this increase of temperature nearly accounts for the increase of efficiency 72* at the high sulphur values. The curves of fig. 21 (p. 30) show that as the per cent of clinker in the refuse increases the per cent of sulphur increases; also that all the highest values for per cent of clinker are grouped with the higher values for per cent of sulphur. The low per cent of values for sulphur are grouped with the high values for per cent of efficiency 72*, and vice versa. At the highest values for per cent of clinker the efficiency 72* is lowest, and with high efficiency 72* the per cent of clinker is the lowest. This relation was determined from two different classifications-one on per cent of clinker in refuse and one on efficiency 72*. The classification on per cent of clinker in refuse shows that efficiency 72* is more influenced by per cent of clinker than the per cent of clinker is influenced by this efficiency. The latter influence was determined by classifying on an efficiency 72* basis.

On referring to fig. 24 (p. 36), a' classification on the basis of per cent of clinker in refuse, it is to be noted that for all values of per cent of clinker the combustion-chamber temperature and pounds of dry chimney gases are about constant.

The general conclusion is that the per cent of sulphur does affect efficiency 72* indirectly by the formation of clinker. Since the amount of air used is a constant value, it is quite possible that the decreasing value of efficiency 72* with increasing per cent of sulphur is due to the effect of clinker on the distribution of the air supply. Moreover, with the high values for per cent of clinker probably more of the air enters the furnace over the fire.

AVERAGE DIAMETER OF COAL.

Preliminary to the discussion of the results produced by using coal of different sizes it may be of value to state how the coal is received at the fuel-testing plant and how the size is determined, and to suggest how further research work might be done on the influence of size.

Nearly all of the coal tested was passed through a crusher, and consequently coals of widely varying characteristics were reduced to the same size. Therefore by classifying coals on a size basis, as we have done, we have averaged a good eastern coal with a poor-grade western coal. While size shows, on the average, some few general relations, it would seem that the best comparable data could be obtained by taking one coal, say, run of mine, and after separating the various sizes, making a series of tests on each size. Even such tests would not be made on the same grade of coal, as the smaller sizes are sure to be higher in ash; but nevertheless this side of the problem is of commercial importance.

After the sample of the coal tested has been quartered and a part sent to the chemist for analysis, the remainder is passed through a revolving screen perforated its entire length with round holes ranging from one-eighth to 1 inch in diameter. These holes are arranged in strips or sets of rows, the width of which decreases as the size of hole increases. The coal passing through each set of holes is weighed separately and the data from the boiler room are reported, as follows:

Determination of average diameter of coal samples.

Column 2 shows that 98 pounds of coal was sized. No attempt is ever made to weigh out exactly 100 pounds, so the weights must be reduced to 100, as has been done in column 3, thereby obtaining the percentages of the various sizes of coal. From the construction of the revolving screen it follows that in an average diameter each of the percentages has a "weight of observation" proportional to the

diameter which it represents. In getting average diameter, then, column 3 is multiplied by column 1, obtaining column 4, and the sum of this column of products divided by 100, therefore, gives the weighted average size of the coal.

The classification plotted in the curves of fig. 29 was made on the coals tested in tests 89 to 401, and shows that as the average diameter increased from 0.35 to 1.26 inches (1) the capacity increased about 15

FIG. 29.-Relations of size of coal to combustion-chamber temperature (°F.) (curve Nc. 1); flue-gas temperature (°F.) (curve No. 2); per cent of rated capacity developed (curve No. 3); pounds of dry chimney gases per pound of "combustible" (curve No. 4); efficiency 72* (curve No. 5); per cent of completeness of combustion (E3) (curve No. 6); carbon-hydrogen ratio in dry coal (curve No. 7); and ratio of carbon to available hydrogen (curve No. 8). Tests 89-401.

per cent (curve No. 3); (2) the highest efficiency 72* values were obtained with the smallest sizes (curve No. 5); (3) there was practically no change in the per cent of completeness of combustion (curve No. 6); (4) there was a slight decrease in the pounds of dry chimney gases per pound of "combustible" (curve No. 4); and (5) the carbonhydrogen ratio figured from the dry coal decreased considerably (curve No. 7), showing that the good coals fall in the groups of small