In skilled hands, the small laboratory-apparatus abovementioned is able to furnish very trustworthy results; but it has the drawback of almost invariably attributing to the coal a higher value than is borne out on the working scale. And, paradoxical as it may seem, this failing is likely to be the more accentuated in the hands of the less experienced operator. The reason for this will be seen on a brief review of some of the particulars wherein the laboratory-apparatus necessarily differs from the actual plant employed in the gas-works:

(1) The retort is of iron, which is a better conductor of heat than clay, and can always be ascertained to be perfectly sound before making a test; whereas the clay-retort in the gas-works is of a porous nature, and has to be made tight by rendering with cement or by filling up its pores with deposited carbon, and is therefore more liable to spring accidental leaks. Then, (2) the temperature is under complete control and, if not high enough, a test can be delayed until the proper heat is attained; or experiments may be made to find the most suitable temperature for obtaining the best results from the coal. Again, the coal will usually be perfectly dry before it is introduced into the retort, and that introduction will be effected with such celerity that no gas is lost in the operation; whereas, in practical working, there is necessarily a considerable loss of gas ere the lid of the retort can be closed and sealed. And there is some slight further advantage in the fact that the gas is measured at a pressure little if at all above the pressure of the atmosphere; whereas, in the gas-works, there is an additional pressure, generally of 6 or 7 inches of water, due to the weight of the gasholder.

But there is another particular in which the laboratorymethod differs from the working-scale operation that, perhaps more than all those above enumerated, allows of unduly favourable results being obtained; and that is the different treatment in the matter of condensing, washing and scrubbing, which the gas receives. With the very best intention to obtain trustworthy results, it is impossible to subject the gas to the same rigorous treatment in the laboratory as it receives in the gas-works. Owing to the smallness of the scale on which the operation is conducted, the gas cannot be washed and scrubbed with ammoniacal liquor and clean water, as is done in the gas-works, without using a proportionately much greater quantity of the

liquids, and so overdoing the cleansing-to the great detriment of the gas. Therefore it is usual to dispense with the washing process, and the experimenter is content with cooling the gas to a sufficient degree, before purifying. Unfortunately mere cooling, except when carried to an extreme, does not suffice to rid the gas of the minute vesicles of tarry hydrocarbons (which are carried along with the stream in an exceedingly minute state of division, and require prolonged contact with wetted surfaces, combined with some stagnation of flow, to enable them to coalesce into drops of liquid, and so allow of their removal from the gas). If every trace of these condensable hydrocarbons be not removed before the gas is tested for its illuminating power, its quality will appear unduly exalted. For the illuminating power of coal-gas is due to its containing from 4 to 6 or 7 per cent. of heavy hydrocarbons, which, in their composition and nature, greatly resemble much of the liquid constituents of the tar; but differ from the latter in the important property of being, under ordinary conditions, uncondensable from the gas. A very slight increase in the percentage of hydrocarbons, which may be caused by inefficient condensation, is therefore calculated to make a considerable difference in the result on the photometer. And if such a result may obtain with every desire to be fair, what may not be done when the gas is purposely coddled with a view to high results?

Sufficient has been said to show how a much higher value can be attributed to the coal than it may really have for the practical gas-maker. By how much the estimate should exceed the reality has never been determined, though many opinions have been hazarded. It is, in fact, incapable of solution. However carefully and conscientiously the test may have been carried out, it is hopeless to attempt to predicate from the result the corresponding result that will be obtained in the gas-works. It must obviously depend upon the skill and intelligence brought to bear upon the working, as well as on the degree of perfection of the plant employed. While, therefore, it will be the aim of the analyst to arrive at the ultimate value of the coal, he must ever keep in view the actual conditions of gas-making, and not take advantage of his favouring circumstances to obtain results which cannot, even with the exercise of great care and skill, be realized in practice.

There is one other matter that requires to be taken into

VOL. XXIV.-1902-1908.

21

consideration in interpreting the results of a coal-test; and that is the question of the burner with which the illuminating power of the gas is determined, including the manner in which it is used. Obviously the burner used should be the standard one for the particular quality of gas, that is to say, the London argand for qualities up to 18 candlepower, and the batswing for qualities of 19 candlepower and upwards. But it is neither necessary nor desirable that it should always be used under standard conditions. Under the absurd regulations which, everywhere but in London, govern the testing of coal-gas, the illuminating power has to be determined with a fixed consumption of exactly 5 cubic feet per hour. When burning different qualities of gas in the argand burner, a uniform consumption must be prejudicial to the lower qualities; because more air is drawn upon the flame than is required, thus cooling and over-oxidizing it, with the result of depreciating the illuminating power. The analyst should therefore vary the consumption to suit the quality of the gas, and calculate the result to the 5 cubic-feet rate. But, if this be done, it is very important that the fact should be stated in the report, so that the gas-manager may know that he has to expect a less satisfactory result when the gas is consumed at the standard rate.

Mr. W. DOIG GIBB (Newcastle-upon-Tyne) wrote that the teaching of Mr. James Stewart's paper on "The Valuation of Gas-coals" was that what was known as a " laboratory test "would not give results comparable to those which could be obtained in actual working. This had been known for a long time. The paper was, however, of value in that it marshals the arguments for and against the "laboratory test" in a terse and lucid manner. In larger gas-works, where the capital expenditure required could be afforded, there was no doubt that an experimental plant on a working scale was of much greater advantage to gas engineers generally than a laboratory plant could be, but even then it was still a benefit to have a laboratory-plant in addition, since, with the latter, tests could be taken quickly and without great cost, and though the results might not be comparable to the actual workingresults they were trustworthy (if proper care were taken in testing) in comparing the different results got from the various coals tested. Mr. Stewart did not seem to believe in the future

possibility of the tests on a laboratory-plant being made in such a manner as would approximate the results got on a working scale. It might not be possible, but at all events it would be a step in the right direction if standard laboratory apparatus and standard methods of using the same were adopted. This would at all events result in the different published analyses of coal being comparable with each other. At present, owing to the different methods employed in sampling and testing, and also owing to the absence of any information on the printed analyses sheets as to the methods, etc., employed, an interested reader had great difficulty in comparing, in any accurate way, the value of one with another.

The PRESIDENT (Sir Lindsay Wood, Bart.) moved a vote of thanks to Mr. Stewart for his interesting paper.

Mr. G. MAY seconded the vote of thanks, which was cordially approved.

DISCUSSION OF MR. G. P. LISHMAN'S PAPER ON "THE ANALYTICAL VALUATION OF GAS-COALS."*

Mr. G. P. LISHMAN wrote that the main intention of his paper had been to emphasize the advantage derived from the regular introduction of a coal of known value into the testing of an unknown coal. This certainly introduced other difficulties, as Dr. Pattinson had pointed out, but in his (Mr. Lishman's) opinion they were considerably less than those which were overcome. A colliery-manager knew fairly well that certain of his seams varied in quality, and that others were very constant; and this, combined with his (Mr. Lishman's) own knowledge, derived from testing, led him to adopt the coal of the Maudlin seam at one of the Lambton collieries as his standard, whereby others were checked. The method had been in use, with great advantage, at the Lambton collieries for over two years; still it was not claimed that perfection had been reached, as the introduction of further refinements was desirable and might reasonably be hoped for. He (Mr. Lishman) did not entirely understand Mr. W. D. Gibb's alternative suggestion from the outline given, but possibly there might be some manner of combining it with the standard-coal method.

* Trans. Inst. M.E., 1902, vol. xxiii., page 567; and vol. xxiv, page 166.

The difficulty was that in winter the iron of the purifiers and pipes (apart from the condensers, which might if necessary be kept warm) was very cold, and owing to this the illuminating power of the gas made in winter-tests was usually of 14 to 16 candlepower as against 16 to 18 candlepower obtained in summer. The purifiers and all the other pipes, except the condensers, were not water-jacketted-it would hardly be practicable to have them so this perhaps best supplied the answer to Dr. Pattinson's query on the subject. No doubt, the effect of condensation was increased by the slower passage of the gas; and, if, as mentioned by Mr. Gibb, the entire room in which the gas was treated were kept at a constant temperature in summer and winter, there would be no necessity either for water-jacketted condensers or even a standard coal. Probably, a room could be arranged so that this might be effected, but in his (Mr. Lishman's) particular case the controlling of the room-temperature was found to be so troublesome that the idea was given up.

The time required for distillation was a question of retorttemperature, and when he (Mr. Lishman) stated 60 to 90 minutes, the latter was an outside limit seldom reached: one hour was regarded as the regular time for a test, but if any further gas was coming away at the end of that time the distillation was allowed to proceed until it ceased, or nearly so. The usual time actually taken varied from 60 to 70 minutes. It was frequently stated that the results from a laboratory coal-testing plant were higher than those likely to be obtained in a gasworks, but this by no means followed, and it depended entirely on the amount of condensation or scrubbing applied. The effect of the lower retort-temperature was to increase the illuminatingpower, the yield being reduced.

Mr. Stewart took the writer to task on the two statements that there is an almost total absence in scientific journals of papers on the testing of gas-coal" and that "although coaltesting plants are attached to most gas-works now, they are usually of but limited use to the engineer, who still has to rely mainly on his working-scale results." He (Mr. Lishman) did not wish to appear contentious on minor points such as these, more especially as Mr. Stewart agreed with him on the main issues of the paper, but considering the enormous number of papers published on all gas-matters in these days, the three or