May 5. 1898] 



NATURE 



21 



There is also some relation between the activity of the mud vol- 

 canoes and the height of the river. The explanation is that 

 ;uring rains the ground-water presses on the petroliferous sands, 

 nd it is noteworthy that the main bed of oil-sand is found at 

 ;)out the level of high-water of the river. 

 Some signs of exhaustion in the oil-field are noticed by the 

 author, but it is possible that further productive beds may be 

 found by boring. H. B. W. 



EXPERIMENTS ON THE WORKING OF 

 GAS-ENGINES. 

 A GENERAL meeting of the Institution of Mechanical 

 Engineers was held last week, when the President, Mr. 

 Samuel W. Johnson, delivered an inaugural address dealing 

 chiefly with the progress of locomotive engineering on 

 the Midland Railway, of which he is chief mechanical 

 engineer. The most interesting feature in the ordinary 

 proceeding was the discussion of the first report to the j 

 Gas-Engine Research Committee of the Institution. The 

 author of the report was Prof. Frederick W. Burstall, under 

 whose superintendence the investigations had been carried out. 

 The object of the experiments was to determine the effect 

 produced upon the economy of gas-engines by altering one or 

 more of the conditions which governed their working. In in- 

 ternal combustion engines there are a much larger number of 

 factors to consider than in steam-engines, and it is difficult to 

 ascertain where to look for economy. The factors to be con- 

 sidered are the amount of compression, the speed, the ratio of 

 air to gas, and the amount of heat which is to be ejected through 

 the walls of the cylinder. An increase of compression, for i 

 example, is often regarded as conducive to more economical re- 

 sults ; but it is uncertain whether the attendant increase in 

 economy is really due to compression alone. To ascertain this, 

 the conditions of working should be altered successively one at 

 a time. This has been done for the steam-engine, but all pub- 

 lished results of tests made on gas-engines are based upon only 

 one fixed set of conditions. 



A small engine was used by the committee, but was one 

 specially constructed for experimental purposes. Small size 

 was an advantage, inasmuch as it allowed measurements, such as 

 those of volumes of air, to be made with accuracy. The work of 

 the committee appears to have been undertaken with commend- 

 able care, and a precision was aimed at more typical of the 

 physical laboratory than of ordinary engineering experiments. 

 This is particularly noticeable in the arrangement of the apparatus 

 and methods of calibration followed. It would take far too 

 much space to follow these in detail, interesting and instructive 

 as they are to engineers, and we can only hope to give a partial 

 idea of the methods followed. This report, it should be remem- 

 bered, is but introductory to the description of the actual work 

 of testing, most of which has yet to be undertaken. As the 

 author stated, experimental work is often compromised by being 

 carried out with instruments upon the accuracy of which no in- 

 formation is furnished. When a comparison is made of a number 

 of results, it is always difficult to discover how far differences 

 are due to working conditions or to inevitable experimental 

 error. In purely physical experiments, the report continued, 

 accuracy may be obtained to the degree of one part in 

 a thousand ; in a few special cases, even better results 

 may be reached. In an engineering experiment it is 

 hopeless to expect such accuracy, owing to the great diffi- 

 culty of keeping the working conditions sufficiently steady 

 from beginning to end of the experiment. With ordinary 

 care, and the use of appliances which are found in all works, 

 probably all that can be expected is to get results correct to 3 or 

 I 4 per cent. With special care, half of i per cent, may be reached ; 



but the author does not suggest that all the results of the 

 experiments made by the committee have this high degree of 

 accuracy, but in the principal measurements probably the 

 experimerital error involved does not in any case exceed i per 

 cent. 



The engine used was of 2-horse nominal power, capable of 

 developing a maximum of 5 I.H.P. ; it has a 6-inch cylinder and 

 12-inch stroke. The valves are worked in the ordinary manner ; 

 there is an ordinary Watt governor acting on a small roller, and 

 causing a charge of gas to be cut off when the speed is too 

 high. To effect changes in compression the connecting-rod is 

 made so that its length can be varied. Compressions employed 



NO. 1488, VOL. 58] 



in the experiment varied between 35 and 90 lbs. per square inch ; 

 variation in the amount of gas admitted was effected by throttl- 

 ing. For measuring the supply of gas a calibrated holder was- 

 used ; the vvet test meter being discarded, as it does not control 

 the fluctuations of pressure in the mains. By this instrument 

 accuracy to the extent of one-tenth of i per cent, was obtained ; 

 calibration was effected by means of a standard cubic foot 

 measure. To determine the air supply per stroke, a meter was- 

 used in place of trusting to the usual method of calculation. 

 The arrangement followed was practically that employed by 

 Dr. A. Slaby, of Berlin. The meter employed was a 400-light 

 standard wet meter made by Alexander Wright and Co., of 

 Westminster. Air is passed in by a blower, the pressure being 

 kept constant by a governor. After passing through the meter 

 the air is delivered into a safety-box, which is used to prevent 

 inflamniable gas from passing back into the meter, and also to 

 give relief in case of back ignition. A rubber-bag is used \.o 

 prevent fluctuations in the meter during the suction stroke. The 

 direct measurement of air supply is usually considered a diffi- 

 cult and dangerous undertaking ; but the author stated that no 

 trouble had been found with this portion of the apparatus. The 

 air meter was checked by blowing air through it into the gas- 

 holder, and was found to be correct to the half of i per cent. 



The amount of heat passed into the jacket was measured by 

 running all the cooling water for a single test into a tank, and 

 taking the temperature by means of thermometers. Samples of 

 exhaust gases were taken and analysed. In this detail the great 

 difficulty is not in making the analysis, but in obtaining a true 

 sample. A single bubble of gas was taken from just below the 

 exhaust valve after each explosion. The apparatus for doing 

 this was illustrated by means of wall diagrams, which showed 

 that the object aimed at was obtained by an electrical 

 relay which actuated a small needle valve that allowed 

 a single bubble of gas to be sucked into the gas receiver. 

 Power developed was ascertained by a Wayne indicator ; 

 an instrument found superior to others tried. Prof. Burstall 

 states that it is in careful hands, apparently the most accurate 

 indicator of the present time. It has a rotating piston in 

 place of the ordinary reciprocating piston. This piston does 

 not touch the containing cylinder at its outer extremities, 

 but is guided at the centre on circular bearings. In this 

 way friction is small and not liable to change, because the 

 bearings can be well lubricated. There are many interesting 

 points about its mechanism which were described in the report. 

 Thin sheets of smoked mica are used in place of the ordinary 

 metallic faced paper or "cards." This device is highly spoken 

 of by those who have had experience in its use. 



As the engine was not fitted with a timing valve — a device 

 which the author considers absolutely necessary to all sizes of 

 gas-engines— it was decided to attempt to ignite the charge by 

 means of an electric spark, and it was hoped that electric 

 ignition would prove more certain than any form of hot-tube 

 igniter. This, however, did not prove to be the case ; and not 

 the least interesting part of the report is contained in the dis- 

 cussion of the failure in this detail. The rope-break used was 

 of the ordinary kind, having dead weights on the lower end of 

 the rope and a spring balance at the upper end. A Harding 

 counter for ascertaining the number of revolutions was employed, 

 and analyses of the coal-gas were made by Mr. G. N. Huntly, 

 who also supervised the analyses of the exhaust gas. The 

 results of seventeen preliminary experiments made were given 

 in a table contained in the report, and on copies of indicator 

 diagrams attached. The mechanical efficiency of the engine 

 varied from 76 to 84 per cent., the mean value of the whole 

 seventeen tests being 81 per cent. It must be remembered, 

 however, that these experiments are, as stated, preliminary, and, 

 it may be added, they have been carried out under circumstances 

 of exceptional difficulty, which conditions, however, will not 

 recur. The report states that it would seem probable that the 

 influence of increased compression on economy is due to the fact 

 that weaker charges can be burnt completely during the stroke 

 when the compression is high. The tests seem to indicate, the 

 report continues, that economy depends on the choice of the 

 correct ratio of air to gas ; and this ratio increases with the com- 

 pression. The number of experiments, however, are, as the re- 

 port states, not yet sufficient to determine what this ratio is for 

 any given compression. It is intended to make a series of tests 

 sufficient for determining this important point. Further experi- 

 ments are to be made at a constant speeid ; the variables being 

 the load, the ratio of air to gas, and the compression. It is 



