474 THE COLLIERY GUARDIAN. September 3, 1915. CURRENT SCIENCE Remarkable Origin of an Explosion. The report of Major A. Cooper-Key, H.M. Chief Inspector of Explosives, on the fatal explosion at a Stowmarket factory on May 10, states that on the morning of the accident cordite M.D. was in course of manufacture. At the moment of the accident two men in the building were on the point of handing out to two bogie runners the bags of cordite paste for removal to the hand mixing house. Before being placed on the bogie the bags are wiped down to get rid of the gun cotton dust, and also to remove to some extent the water from the bottom of the bags due to their having been standing on a wet floor. In order to avoid the use of a loose implement of any kind, this operation was done by hand, and not with a brush. It was a fine dry cold morning, and the two men in the building were wearing stout rubber overshoes, which would completely insulate them from electric contact with “ earth.” The bags were of rubber impregnated canvas, a marked dielectric, and in the case at any rate of a man with a naturally dry skin the operation of wiping the bags would undoubtedly have the effect of generating in his person a statical charge of electricity, which might well be sufficient to cause a spark to pass as soon as he should happen to “ earth ” himself by touching an uninsulated article. The Chief Inspector says that although this may perhaps appear at first sight to be a somewhat far- fetched explanation of the accident, he is convinced it is the true one. Mr. Duff Grant, the managing director of the New Explosives Company, records the fact that when on a visit to Canada, where, with the temperature below zero, the air is frequently practically free from moisture, he proved by personal trial that by rubbing the feet on a thick pile carpet, the gas could be lit by a spark from the end of the finger, and this not occa- sionally, but every time. It is stated that on cold dry days it is a favourite joke at five o’clock tea parties to shuffle the feet on the carpet before handing a plate of sandwiches, with the result that the person helping him- self would receive quite a considerable electric shock. A case has also been mentioned relating to a factory at Ardeer, where a workman wearing rubber shoes fired acetone vapour by a spark from his finger after merely allowing the thin strands of rifle cordite to pass over his hand. The above suggests many possibilities respecting the ignition of explosive mixtures of gas and air in coal mines. Pumping Water from Great Depths. Gilbert reports in Gliickauf on the results obtained in a Westphalian colliery with a plunger pump lifting 1,320 gals, per minute, and two centrifugal pumps each dealing with 1,075 gals., all working against a head of 2,650 ft. The water was highly charged with saline constituents. On installing the plunger pump, con- siderable difficulty was experienced in getting the valve tightly packed against the pump barrel with the usual rubber ring, the valve becoming so leaky after about 12 hours’ run that the pump had to be stopped. This was found, on examination, to be due to the ring having broken away in places, under the working of the valve on the seating surfaces, even when the screws were drawn up tight. The interposition of a protecting copper ring was tried, but without success. It was thereupon decided to dismantle the pump and obtain a tight joint by means of a feather and groove, with interposed Klingerite ring in the seating surface, cup leathers being aranged at the side, as shown in Pump barrel Packing ring • Cup leather Bearing ring Klingerite packing! Fig. 1.—Section of Valve with Improved Valve Packing. Packins' surface Corroded by water Packing surface Corroded by water. Figs 2 & 8.—Sections showing Corroded Parts of Packing Surface. fig. 1. In three cases this arrangement proved satis- factory, but the fourth barrel again began to leak at the end of the first week in use, because the screws did not press firmly enough against the seating surface, but allowed it to dance, so that the water wore away the edge of the groove over a width of about 6in., the material of the barrel being also damaged at the same spot (figs. 2 and 3). It was found possible to repair the groove by autogenous welding, the defective portion of the pump barrel being made good with manganese iron cement. Thus patched up, the pump ran for five weeks, by which time the capacity had receded by about 220 gals, per minute, the faulty barrel being then removed and replaced by a spare. All the other barrels have continued to work satisfactorily. In the case of the centrifugal pumps, difficulties of the above-mentioned kind are less serious, and can be more quickly remedied, for while four days are required to take out and replace four barrels of a plunger pump, AND TECHNOLOGY. with the assistance of a travelling crane, the work can be done in two days on a centrifugal pump of equivalent size; and though the efficiency of plunger pumps is greater, the other system is preferable for raising water from great depths. In deciding which system to use, the following points ought to be taken into consideration :—The efficiency of the plunger pump is quickly impaired by leakage in the valves, which may necessitate the pump being stopped altogether, to which must be added the greater frequency of repairs and higher running costs in comparison with the centrifugal type. The reliability of the latter type is increased by the fact that any damage can generally be repaired in a few hours, provided the several stages of the pump are detachable, so that the pump can be quickly taken apart and defective parts replaced. This cannot be done when the pumps are fully enclosed, and when the water contains much sediment, the work then often taking several days; and the vanes are frequently damaged by the force applied in jacking out the parts. Moreover, at least two centrifugal pumps of equal capa- city to a plunger pump can be installed at the same cost as the latter, thus doubling the security of the plant—an important consideration when the influx of water is large. The compressor required to fill the domes on the valve bodies and air chamber of the plunger pump is also very liable to get out of order, when mounted in the pump chamber, and in contact with highly saline water, the compressor valves soon wearing out, and having to be replaced, the non-return valves in the delivery pipe being corroded by the damp, salt- laden air, whilst, unless the domes are kept properly filled with air, the water in the upcast pipe runs back through these valves to the compressor, and corrodes the high-pressure valves. The only way to prevent this trouble is to instal the compressor at bank, and run an expensive air main down the shaft to the pumps. Finally, the upcast pipes often become leaky at the flanged joints when plunger pumps are used, owing to the vibration set up when the domes are not properly filled with air (which may easily happen through care- lessness on the part of the man in charge)—a source of expense altogether absent w7hen centrifugal pumps are employed. Photometric Researches at the National Physical Laboratory. In the report of the National Physical Laboratory for the year ending March 31, 1915, the following work on photometry is described :—The new realisation of the pentane unit of light in electric sub-standards is now7 complete, and the results are included in a paper which was read in March before the Physical Society. The chief points of interest in this are the agreement of the unit so obtained with the original N.P.L. unit of candle- power to within less than 0*1 per cent., and the informa- tion afforded as to the corrections to the lamp for atmo- spheric changes. The correction for barometric changes is identical with the previous determination, and that for humidity is only very little different (0-063 instead of 0-065). An analysis of the results has shown that the changes of candle-power w7hich are usually ascribed to the humidity of the atmosphere may in part be due to changes of temperature. As, however, in practical use changes of atmospheric humidity (measured volumetric- ally) are largely caused by changes of temperature, the usual method of analysing the results does not really permit the twro effects to be differentiated. The result, however, of combining the humidity and temperature connecting factors in one constant is that this constant may vary slightly in different localities. As a matter of fact, slightly different values for the constant have been found in different places, and the above is possibly an explanation of the discrepancies. The temperature- humidity correcting factor could not be the same for tw7o localities wdiich, on the average, had at one temperature the same humidity, and at another temperature different humidities. The differences in practice do not appear to be sufficient to affect ordinary photometric w7ork, but in measurements of highest accuracy they can be detected. The set of electric sub-standards in w7hich the newly- determined unit is held have tungsten filaments glowing at a very low efficiency, so that their light matches in colour that from the pentane lamp. The Fleming- Edisw^an lamps in w7hich the original unit was held operated at a slightly higher efficiency than this, and the tw7o groups of lamps now together form the fundamental sets from which the values of all higher efficiency lamps are derived. The work on the other sets of higher efficiency lamps, mentioned in previous reports, has now7 been completed up to the set running at 1’5 wratts per candle. The results of these cascade determinations have been analysed, and show a probable error of about 0-25 per cent, for the value of the unit held in lamps which work at 1*5 watts per candle. The direct comparison between the fundamental set and that running at 1-5 watts per candle gave the same result as that by the cascade method within the error of the experiment. All these determinations were made with the Lummer-Brodhun photometer head. Attempts to make the same deter- mination with a flicker photometer did not yield results of an accuracy comparable with that obtained with the Lummer-Brodhun photometer head, and w7ere not, therefore, proceeded with. All the sets of sub-standards (six in number) are now defined by the temperature at which their filaments operate. It is on the temperature of the filaments that the colour of the light emitted mainly depends, irrespective of whether the radiating substance is carbon or tungsten. The group of lamps having light of the same colour as that from the pentane lamp operates at 1,610 degs. Cent., whilst that running at 1-5 watts per candle (approximately) has an operating temperature of 2,300 degs. Cent. The other groups of lamps occupy intermediate positions, giving approximately equal steps. Experiments are now in hand to continue the series by means of gas-filled standard lamps, up to an efficiency of 0-6 watt per candle. For this purpose four addi- tional groups of lamps will be needed in order that the same colour steps may be maintained. In all of the work on these standards the thanks of the laboratory are due to the General Electric Company, who have pre- sented a large number of standard lamps specially for the purpose at the Osram Lamp Works. The determinations of the temperatures of these fila- ments has led to a preliminary investigation of the method of colour identity for the estimation of tempera- ture, the results of which have been embodied in a second paper to the Physical Society by Messrs. Paterson and Dudding. These preliminary experiments fully justify the continuance of the work on a more extensive scale, and a more complete investigation of the method of colour identity for practical use in the estimation of the temperatures of bodies emitting light whose spectral distribution is continuous. Experiments were made at the same time to determine the colour of light from molten platinum, as bearing on the usefulness of the Violle standard of light for practical photometry. The colour of light from molten platinum w7as found to be identical with that from a carbon filament glow7 lamp operating at 4-75 watts per mean spherical candle, corre- sponding with about 3’8 watts per mean horizontal candle. The investigation has been continued which has for its object the determination of the coefficient of reflection of the wdiite cards used in connection with illumination, photometers. These vary somewhat in the amount of light emitted in different directions, since their surfaces are not absolutely matt. The value for the ordinary white cards has now been determined, and the intention is to apply the same methods to other substances. Some high power gas-filled lamps, the gift of the British Thomson-Houston Company, have been of considerable assistance in these experiments. An investigation has been completed for a Government department on the possibility of distinguishing detail through materials of different translucencies. A test has been devised for examining materials of this character with a view to establishing a criterion. THE GERMAN AND AUSTRIAN COAL AND IRON TRADES. We give below7 further extracts from German periodicals that have reached us, showing the course of the coal and iron trades in Germany and Austria :— Coke versus Coal for Steam Raising. Attempts have been made in Germany for some time past to persuade coal consumers to employ coke as a substitute for the scarcer fuel, coal, travelling grates being recommended for this purpose. The experience of the Rhenish-Westphalian Electricity Works, how7- ever, indicates that the substitution is not altogether a success. Trials with an ordinary travelling grate having proved unsatisfactory, a vertical combustion chamber was constructed (of the type used for central heating boilers) for igniting the coke, with better results. Never- theless, the matter is not yet settled from the economic point of view7, and before the change from coal to coke can be successfully carried through, it will be necessary to ascertain whether coke can be obtained at a suitable price, and in sufficient quantity, at times when the steel industry is booming, because there will be considerable difficulty in reverting from coke to coal, and it can hardly be expected that owners of large boiler plants will agree to the prospect of having to make funda- mental alterations in their expensive firing appliances every time the market conditions change. German Steel Union Report for July. The total deliveries (in w*eight of raw steel) during July amounted to 258,092 tons, as compared with 318,952 tons in June, and 570,422 tons in July 1914. Of these, 61,678 tons (77,804 tons and 128,056 tons) were semi- manufactured goods, 118,727 tons (154,736 tons and 186,231 tons) railway superstructural material, and 77,587 tons (86,312 tons and 156,135 tons) sections. German Pig Iron Output in July. The Iron and Steel Makers’ Association reports the total output of pig iron for July in the German Customs Union to be 1,064,899 tons (993,496 tons in June), or 34,352 tons (33,116 tons) per working day. Of this total, 216,477 tons (203,849 tons) were foundry pig, 16,772 tons (18,887 tons) Bessemer pig, 654,479 tons (612,659 tons) basic pig, 158,029 tons (136,611 tons) steel iron and spiegeleisen, and 19,142 tons (21,490 tons) puddling pig. The production in the various districts w7as as follows :—Rhenish Westphalia, 456,515 tons (423,908 tons); Siegerland, Wetzlar, and Hesse-Nassau, 73,453 tons (67,202 tons); Silesia, 63,801 tons (63,291 tons); North Germany, 19,719 tons (18,504 tons); Mid- Germany, 33,429 tons (33,082 tons); South Germany and Thuringia, 20,132 tons (20,082 tons); the Saar dis- trict, 72,618 tons (68,734 tons); Lothringen, 165,195 tons (158,604 tons); and Luxemburg, 160,037 tons (140,089 tons). The Pig Iron Union report states that sales of high- grade pig for the third quarter of the year are now com- pleted, the w7hole of the possible output having been disposed of. The demand for haematite pig, steel iron, and spiegeleisen has been particularly active, the open- hearth steel makers being very busy. On the other hand, the enquiry for phosphoriferous pig has slackened. The foreign trade remains unchanged. During July the deliveries of high-grade pig amounted to 62-31 per cent, of the participation (57-25 per cent, in June). Selling prices for the fourth quarter remained unchanged.