C e THE THEORY OF THE GAS ENGINE. By DTJGALD CLERK. REPRINTED FROM VAN NEW YOKE: D. VAN NOSTRAND, PUBLISHER, 23 MURRAY AND 27 WARREN STREET. 1882 PRE F AC E. The continued existence of the Gas Engine as a competitor with the Steam, and Hot Air Engines is no longer an open question. It has fairly passed that phase of the experimental stage which determines its fitness to aid in the solution of the ^problem of the eco- nomical conversion of heat into mechan- ical work.- The improvements of the last few years have brought this motor into a conspicu- ous position. We believe this essay of Mr. Clerk's to be the best presentation of the theory of tke Gas Engine that has yet appeared. EDITOR OF MAGAZINE. The Theory of the Gas Engine. THE practical problem of the conver- sion of heat into mechanical work has long occupied the minds of engineers and scientists ; the steam engine is a partial solution, but although perfect as a machine, its efficiency is so low that it can hardly be considered as satisfactory and final. As the result of the best modern practice it may be taken that the steam engine does not convert more than 10 per cent, of the heat used by it into work, and this in engines of considerable size and with boilers and furnaces fairly efficient. In small engines it is much less, indeed it is certain that few among the thousands of steam engines in daily use below 6 HP. give an efficiency greater than 4 per cent. The great cause of loss is the amount of heat necessary to change the water from the liquid to the gaseous state, most of this heat being rejected with the exhaust either into the conden- ser or the atmosphere. Many attempts have been made to use liquids of lower specific heat than water, and requiring less heat for evaporation, the principal being alcohol, ether and carbon bisul- phide, but for obvious reasons no success has been attained. To heated air as a means of obtaining power, the objection of loss by latent heat does not apply, the air is already in the gaseous statej and any heat added at constant volume increases the tempera- ture, and therefore the pressure 2 without the complication of change of physical state. A high efficiency would therefore be expected, and according to Professor Rankine the efficiency of the fluid in the engines of the "Ericsson" was about 0.26 ; the efficiency of the furnace was however low, and accordingly the actual efficiency oc the engine was no higher than that of the best steam engines now in use. In the " Stirling" hot-air engine, he found the efficiency of the fluid to be 0.3 with a higher efficiency of furnace than in Ericsson's. In the Ericsson engine the air was heated at constant pressure, the volume augmenting ai\d the power being given by the increased volume of the air as it entered the motor cylinder from the re- servoir into which it had been compress- ed. The mean effective pressure was only 2.12 Ibs. on the square inch ; the size and friction of the engine for a given power was enormous. In the Stirling engine the air was heated at constant volume with increase in pressure, the power being obtained by subsequent ex- pansion ; the mean available pressure was 37 Ibs. per square inch, and the fric- tion of the engine only amounted to one- tenth of the total indicated power. Both engines used the now well known contri- vance, the regenerator, which was the invention of Dr. Stirling, and which is the cause in both of the high efficiency. The failure of these engines was due to the rapid burning out of the cylinder bottoms by the direct action of the fire, it being found impossible to heat the air rapidly enough to the required tempera- ture without maintaining the temperature of the metal surfaces much higher than the maximum temperature to be attained by the air. To overcome 'this slow heat- ing of the air when in mass has been the object of many inventors, and a type has often been proposed with a closed fur- nace, and the air forced through this furnace keeping up the combustion, the hot products going to the motor cylinder and there doing work. This method of internal heating, however, introduces dif- ficulties as grave as exist in the external method. The hot gases having to pass through pipes and valves to the motor cylinder renders it impossible to main- tain a very high temperature without damage to the machine. Sir George Cay ley was the first to make and work experimentally an engine of this type. In view of these futile attempts, until very recently hot air was considered as among the failures of the past, and it was believed that, imperfect as the steam 9 engine is, nothing was likely to succeed in producing a better result. The great progress made in recent years with the gas engine, and its advance from the state of an interesting but troublesome toy to a practic-el powerful rival of the steam engine, has shown that air may after all be the chief motive power of the future. In the gas engine chemical considerations greatly modify the theory and prevent it from ranking as a simple hot-air engine; but to be thoroughly understood it is better first to consider the power to be obtained from air under certain theoretical condi- tions. Three well defined types of engines have been proposed (1.) An engine drawing into its cylin- der gas and air at atmospheric pressure for a portion of its stroke, cutting off communication with the outer atmos- phere, and immediately igniting the mix- ture, the piston being pushed forward by the pressure of the ignited gases during the remainder of its stroke. The 10 in-stroke then discharges the products of combustion. (2.) An engine in which a mixture of gas and air is drawn into a pump, and is discharged by the return stroke into a reservoir in a state of compression. From the reservoir the mixture enters into a cylinder, being ignited as it en- ters, without rise in pressure, but simply increased in volume, and following the piston as it moves forward, the return stroke discharges the products of com- bustion. (3.) An engine in which a mixture of gas and air is compressed or introduced under compression into a cylinder, or space at the end of a cylinder, and then ignited while the volume remains con- stant and the pressure rises. Under this pressure the piston moves forward, and the return stroke discharges the exhaust. Several minor types have been pro- posed and many modifications of these three methods are used. A thorough understanding of these, however, renders 11 it possible to judge the merits of any other. Types 1 and 3 are explosion engines, the volume of the mixture remaining constant while the pressure increases. Type 2 is a gradual combustion engine in which the pressure is constant but the volume increases. The author, in the course of his ex- periments on gas engines, has found that 1 3 537 Centigrade is the temperature usually attained by the ignited gases in his engine, and he has accordingly in- vestigated' the behaviour of air under different conditions at this temperature. Type 1. Suppose an engine to have a piston with an area of 144 square inches and a stroke of 2 feet. Let the piston move through the first half of its stroke drawing into the cylinder air ; let enough heat be immediately added to this air to cause it to rise instantly to 1,537 Centi- grade, and the piston continue moving forward under the pressure produced. If there be no loss of heat through the sides of the cold cylinder, but the temperature 12 of the air fall only through performing work, how much work would be done when the piston completes its out-stroke ? The air before the heat is added is supposed to be at a temperature of 17 Centigrade (about 60 Fahrenheit), and the ordinary atmospheric pressure. In Fig. 1 the line marked adiabatic No. 2 is the curve showing the work which would be obtained under the supposed condi- tions. Fig. 2. is the indicator diagram such an engine would furnish. It is not necessary here to detail the calculations. With this paper is given a table of the data used, so that the numbers may be verified. The following are the results : 1 cubic foot of air (at 170 Centigrade, " and 760 millimetres mercury) re- maining at constant volume re- quires to heat it to 1,537 Centi- grade, an amount of heat equiva- lent to.. 26,762 foot-lbs. Maximum pressure in Ibs . per square j inch above atmosphere J ' J Pressure at the end of stroke per ) I. -i r 19. b Ibs. square inch above atmosphere J OP THK UNIVERSITY l== 'ijouj ajenbs jsd'sqj u; ajnssajd aj.n| 14 Mean pressure during available part ) ^ g ^ g of stroke ) Temperature of air at the end of ) -< QQO n stroke ) Work done on piston 5,731 foot-lbs. Duty of engine ' =0.21. As the engine is supposed to draw in air for half of its stroke, the last half of the stroke only is utilized for power ; the mean available pressure calculated for 89 8 the whole stroke is only ~ = 19.9 Ibs. per square inch. There is a considerable pressure at the end of the stroke which could be made to give more work by ex- panding further ; but for the purpose of comparison it is better to consider the three types of engine as each having a cylinder capacity swept by the piston of 2 cubic feet, and in each case using in its operation 1 cubic foot of air at each stroke. Type 2. Suppose an engine to draw into a pump 1 cubic foot of air, on its return stroke forcing the air into a reser- voir at a pressure of 76.6 Ibs. per square inch above the atmosphere. The motor 15 If! d b> I a qoui ej^nbs jed-sq| ui 9j9C(dsouu;B uodn ejnsse 16 piston is now at the beginning of its out- stroke, and as it moves forward air from the reservoir enters the cylinder, but as it enters it is heated to 1,537 Centigrade, without rise in pressure ; the motor pis- ton sweeps through 2 cubic feet. Fig. 3 -shows the indicated card of this engine. abed is the pump diagram. Air at 17 Centigrade is taken in, com- pressed without loss* of heat, the temper- ature rising under the compression to 217.5 Centigrade. When it is equal to the pressure in the reservoir it is forced into the reservoir, as is shown on the line b c. In all the operations no loss or gain of heat is assumed, except in doing work or in work being done on the air. In the motor diagram from c to E the air is flow- ing from the reservoir following the pis- ton, and the temperature is 1,537 Centi- grade during the whole admission. At e the communication with the reservoir is cut off, and the temperature falls while the air is expanding doing work, until it reaches the end of the stroke, when tha 17 H -> u la { 9 10 1 2 3 4 5 6 7 1 cubic foot 2 f / I537?0 7 ^) , / / . r 1 / - f* g | g go g o - LJOUJ ejisnbs jed - sq| ui 18 exhaust is discharged by the return stroke of the piston. For convenience the pump diagram is shown on the motor one, and the shaded portion represents the work done by the air as the result of the cycle. As the heat is added while the air expands in volume,' it takes considerably more to raise a cubic foot of air to the required temperature than in the case of type 1. 1 cubic foot of air (17 Centigrade^) and 760 millimeters mercury) at I constant pressure requires to heat }- 32,723 it from the temperature of com- foot-lbs. pression217.5 to l,537Centigrade I heat equivalent to J Maximum pressure inlbs. per square ) ~Q L inch above atmosphere ) Pressure at end of stroke above at- ) * Q mosphere - ) Mean pressure during available ) 47.1 Ibs. per part of stroke f square inch. Temperature of air at the end of ) $SL 1,089^ stroke J Centigrade Work done on piston 11,759 foot-lbs. , Duty of engine ^- - =0.36. 19 Type 3. Suppose an engine to draw into a pump 1 cubic foot of air, on its return stroke forcing it into a reservoir at a pressure of 40 Ibs. above the atmos- phere. The motor piston is now at the beginning of its out-stroke, and as it moves forward air from the reservoir enters the cylinder while the piston sweeps through 0.39 cubic feet. At this point communication is cut off, and the temperature suddenly raised to 1,537 Centigrade. Hitherto the air has re- mained at the temperature of compres- sion 150. 5. The pressure goes straight up to 220 Ibs. above the atmosphere. This is shown at Fig. 1, and also at Fig. 4, which is the diagram of this type of engine., a b c d is the compression dia- gram ; a b e f the motor diagram. The piston continues to move forward, and the air expands doing work. At the end of the stroke the pressure has fallen to 8.4 Ibs. per square inch above the atmos- phere. 20 1 cubic foot of sir (17 Centigrade, and 760 millimeters mercury) at constant volume requires to heat it from the temperature of compres- 24,416 foot-lbs. sion 150. 5 Centigrade to 1,537 Centigrade heat, equivalent to ---- Maximum pressure in Ibs. per square inch above atmosphere ............ Pressure at end of stroke ............ 8.4 Ibs. Mean pressure during available ) 47.8 Ibs. per part of stroke ............... [ square inch. ) QRjO Temperature at middle of stroke (j e ntirade Temperature at end of stroke. . 648 Centigrade. Work done on the piston _______ 11,090 foot-lbs. 11,090 , Duty of engine Fig. 5 shows the most important modi- fication of this type ; in it, instead of a separate reservoir, a space is left at the end of the cylinder, into which the piston does not enter, and in this space is com- pressed the gases forming the inflamma- ble mixture. The rise in pressure there- fore commences at the beginning of the stroke instead of when the piston has traveled out. In this diagram the volume 21 bJO 1 3> D CN 1 10 << 1 I 1 / cubic foot. s / 1 / - O -! i / i X s 2 * ^ O O C? O CO O 1 g J3 2 S qouiajBnbsr Jd 'sq| ui 22 swept by the piston and the clearance space together are supposed to be equal to 2 cubic feet. Comparing the results obtained from these three modes under precisely similar conditions, the same weight of air heated to the same degree, and used in cylinders of identical capa- city, there is a considerable difference in the results possible even under the purely theoretic conditions stated. The relative work obtained from 1 cubic foot of air heated to the assumed temperature is shown below. RESULTS FROM ENGINES OF EQUAL VOLUME SWEPT BY MOTOR PISTON. Type 1. 5,731 foot-lbs. work obtained 0.21 duty. 2. 11,759 ' " " 0.36 " 3. 11,090 " " " 0.45 " That is, in an engine of type 1, if 100 heat-units be used, 21 units will be con verted into mechanical work. In type 2, with the same amount of heat, 36 units will be given as work, and in type 3 no less than 45 units would be converted into work. 23 / / / I o 7f ri - e ^ M i qouj ajenbs jed -sq| ui 24 The great advantage of compression over no compression is clearly seen, by the simple operation of compressing be- fore heating ; the last type of engine gives for, the same expenditure of heat 2.1 times as much work as the first. Compression, as used by the second type, does not afford so favorable a re- sult ; but even then the advantage is apparent, 1.6 times the effect being pro- duced. By a greater degree of compres- sion before heating even better results are possible. In an engine of type 3 expanding to the same volume after igni- tion as before compression, the possible duty D is determined by the atmospheric absolute temperature T', and the abso- lute temperature after compression T; it is T T' D= m~~ whatever may be the maxi- mum temperature after ignition. In- creasing the temperature of ignition in- creases the power of the engine, but does not cause the conversion of a greater proportion of heat into work. With any given maximum temperature 25 the smaller the difference between that temperature and the temperature of compression, the greater is the propor- tion of added heat converted into work with any given amount of expansion. The greater the compression before igni- tion, the more closely the two tempera- tures come together, and the higher is the duty of the engine ; neglecting in the meantime the. practical conditions of loss What compression does is to enable a great fall of temperature to be obtained due to work done with but a small move- ment of the piston. In type 1 when the piston has reached the end of its stroke, the increase from the moment of ignition is only from one volume to two volumes, while in type 3 with the same total volume swept by the piston, it increases from one volume to five volumes. In the one case the ratio of expansion is two, while in the other it is five. This will be readily seen in Figs. 2 and 4. Now this increased expansion is not ob- tained at the cost of loss average press- ure ; in type 1 the mean available press- 26 ure over the whole stroke is nearly 20 Ibs. per square inch, while in type 3 it is 38.5 Ibs. per square inch ; that is, the compression engine for equal size and piston speed has nearly twice the power of the other. In the compression engine with a maximum temperature of 1,537 Centi- grade, the final temperature is 648 Centigrade, while in the other, with the same maximum temperature, the final temperature is 1,089 C Centigrade. It is true that by expanding sufficiently the same final temperature can be obtained without compression, but the average pressure will be low, and consequently less available for the production of power. To produce anything like an expansion of five times without compression the pressure would fall below the atmos- phere, and it would be necessary to ex- pand into a partial vacuum, and use a condenser and vacuum pump, as is done in the steam engine. Compression makes it possible to obtain from heated air a great amount of work with but a small 27 movement of piston, the smaller volume giving greater pressures, and thus ren- dering the power developed more mechani- cally available. The higher the maximum temperature the greater the amount of compression which can be used advan- tageously. There is a degree of com- pression for every temperature, beyond which any increase causes a diminution of the power of the engine for a given size. The compression in the author's engine is 40 Ibs. per square inch above the at- mosphere, and he has accordingly con- fined himself to the comparison of engines employing this amount of com- pression with those using no compres- sion. Now, seeing that this difference is produced between engines of types 1 and 3 by the simple difference of cycle, when there is no loss of heat through the sides of the cylinder, the question arises which engine would give the great- est effect, which, engine in actual prac- tice, with a cylinder kept cold by water, -would come nearest to theory ? In which 28 of the engines would there be the smaller loss of heat ? The amount of heat lost by a gas in contact with its enclosing cold surfaces depends, first, on the difference in tem- perature between the gas and the cooling surfaces ; secondly, on the extent of sur- face exposed ; and, thirdly, on the time of exposure. It would be very difficult to make an accurate numercial compari- son between the engines, but all to be shown is, that in the one the loss of heat must be less than in the other. To compare the two engines, take equal movements of the pistons from a maximum temperature of 1,537 Centi- grade. In the engine working without compression this temperature is attained at the middle of its stroke, when the piston has moved through 1 cubic foot ; the average temperature, while it moves to the end of its stroke, is about 1,300 Centigrade. Now, in the compression engine the maximum temperature is attained at a point when the piston has moved through 29 0.39 cubic foot: suppose it to move to 1.39 cubic foot, it has moved through 1 foot in the same time as the first engine. Then, as the temperature at the middle of the stroke is 953 (Fig. 4) it follows that the average during this movement is not higher than 1,000 Centigrade, but the space containing the heated air has increased from 0.39 cubic foot to 1.39 cubic foot, and with it the cooling sur- face ; whereas the space containing heat- ed air in the first engine has, during the same amount of movement, increased from 1 cubic foot to 2 cubic feet. It follows that as the temperature in the compression engine is 1,000 Centigrade during the same time as the temperature in the first engine is 1,300 Centigrade? and as the surface in it for cooling is also less, the amount of heat lost by the air must be less in the portion of the stroke under consideration. During the portion of the stroke remaining, 0.61 cubic foot, the temperature of the heated air is low, falling to 648 Centigrade at the end of the stroke ; it follows that 30 very small comparative loss results. Al- together the loss of heat by the com- pression engine will be the least. It will be seen from Fig. 1 that there is a further cause of advantage. While the pressure and temperature are falling on adiabatic line 1, the work done by 1 cubic foot of air on expanding to the middle of the stroke at a temperature of 953 Centigrade is 7,888 foot-pounds, from 953 Centigrade to G48 is 3,202 foot-pounds, that is, 7,888 foot-pounds of work are performed by the engine during a movement of the piston equal to 0.61, while in the engine without com- pression a movement of 1.00 cubic foot only does 5,731 foot-pounds. The compression engine during this portion of its stroke has converted the heat entrusted to it into work at twice the rate of the other engine. This is a great point. Any method which con- verts the heat into work with the utmost possible rapidity, by reducing the time of contact between the hot gases and the cylinder, saves heat and enables the 31 theory of the engine to be more nearly realized. Taking all circumstances into con- sideration, it is certainly not over esti- mating the relative advantage of the com- pression engine to say that it will, under practical conditions give, for a certain amount of heat, three times the work it is possible to get from the engine using no compression. It will not be necessary to discuss the theory of type 2 in respect of loss of heat to the sides of the cylinder, as it is not much used, and has hitherto failed to yield results in any way equal to type 3. It will be seen, however, from Fig. 3, that the conditions are not so favorable for a minimum loss of heat as in tvpe 3. The temperature from the moment of admission at c, to the point of cut-off at e, is kept constant at 1,537 Centi- grade, so that the loss- of heat must be great, both the surface exposed and the mean temperature being high. It is the less necessary to discuss this point in the slow combustion engine, as the pos- 32 sibility of using a hot cylinder and piston reduces the loss by attaining a tempera- ture not far removed from the entering air. It will be interesting to calculate the amounts of gas required by these three types under the supposed conditions, and for this purpose an analysis of Man- chester gas, and also of London gas, has been used as the basis of calculation. ANALYSIS OF MANCHESTER COAL GAS. BY BUNSEN AND ROSCOE. Hydrogen 45.58 March gas 34.90 Carbonic oxide 6.64 Olefiant gas or ethylene 4 . 08 Tetrylene 2.38 Sulphuretted hydrogen . 29 Nitrogen 2.46 Carbonic acid 3 . 67 100. 00 volumes. Of this gas 1 Ib. at atmospheric pressure and 17 Centigrade measures 30 cubic feet, and evolves on complete combustion 10,900 heat-units Centigrade, 33 equivalent to 15,146,640 foot-lbs. 1 cubic foot of this gas will therefore evolve on complete combustion heat equivalent to - = 504,888 foot- oU Ibs. To obtain an idea of the difference in heating power of the different gases, there is given here a recent analysis of London gas. ANALYSIS OF LONDON COAL GAS. (A.) (B.) Hydrogen 50.05 51.24 March gas 32.87 35.28 Carbonic oxide 12.89 7.40 defines 3.87 3.56 Nitrogen 2.24 Carbonic acid 0.32 0.38 Taking the average of the two analyses, 1 Ib. weight of this gas at atmospheric pressure. and 17 Centigrade, measures 35.5 cubic feet, and evolves on complete combustion 12,500 heat-units Centigrade, equivalent to 17,370,000 foot-lbs., 1 cubic foot of this gas will therefore evolve, on 34 complete combustion, heat equivalent to foot-lb, The difference between the heat evolved by these- gases is but small. As Glasgow coal gas is of a high illuminating power, it will be richer in olennes, and the heat evolved per cubic foot will be some- what greater. Taking 505,000 foot-lbs. as the amount of heat evolved by 1 cubic foot of coal gas, the result is probably very near the average to be obtained from the coal gas of most towns. The number of foot-lbs. required for 1 HP. for one hour are 33,000x60=1,980,000. It therefore follows that if the whole heat to be obtained from gas were con- verted into mechanical work, 1 HP. for 1,980,000 one hour requires =3.92 cubic 505,000 feet. Now, taking the three types of en- gines, the amount of gas required by each to give 1 IHP. per hour would be as follows : 35 AMOUNT OF GAS REQUIRED BY THREE TYPES OF ENGINE. 3 92 Typel. Q 1 oi=18.3 cubic ft. per HP. perhr If these engines be worked without loss of heat through the sides of the cylinders, but the expanding g/ises fall- ing in temperature only through doing work, the above results would be ob- tained. It is interesting to compare the con- sumption of gas by the engines in actual practice, to see in what order it stands. Results have not been obtained from en- gines of equal volume swept through by the piston, but it is at once seen that the order is in accordance with what is required by theory. 36 AMOUNT OF GAS CONSUMED BY THE THREE TYPES OF ENGINE HITHERTO IN PRACTICE. 1. Lenoir. .95 cu. ft. per indicated HP. per hr. Hugon..85 " 2. Brayton.50 " " 3. Otto 21 " " " " For the Lenoir and Hugon engines the results of experiments by Mr. Tresca, of Paris, have been taken, as stated by Professor Thurston, corrected for an error into which he has fallen. He states the consumption of the engine to be 32 cubic feet per IHP. per hour, and then goes on to say that on the brake 4 HP. is obtained, while 8.6 is indicated. He has neglected to deduct from the gross indicated power in the cylinder, the pump resistance, and thus calculates the consumption on the gross indicated, in- stead of on the available indicated power. The available indicated power is not more than 5.2 HP., and the con- sumption is not less than 50 cubic feet per IHP. per hour. 37 For the "Otto" engine have been taken the figures given by Mr. F. W. Crossley. It is seen that the results are much what would be anticipated from the theory already developed. The difference between types 1 and 3 is greater than theory would indicate ; but at the time the Lenoir engine was in use, the imperfection of the igniting ar- rangements and the rapid heating of piston, and consequently of the entering gases, made its action diverge much more widely from theory than in the case with the " Otto." The latter engine not only has the advantage of a better theoretical cycle, but the arrangements are of a nature to secure a greater per- fection of action, and consequently a still closer approach to theory. An amount of about 18 per cent, of the heat used by it is converted into work, but only 3.9 per cent, by the Hugon engine. In types 1, 2 and 3, which have been discussed, it has been assumed that in each case the expansion doing work was 38 carried to twice the volume of the air before compressing. Fig. 6 is a diagram from one of the author's engines which belongs to type 3. It will be observed that in this en- gine the expansion is only continued un- til the volume of the hot gases becomes equal to the volume before compression. Taking the amount of work to be ob- tained from a cubic foot of air com- pressed to 40 Ibs. above the atmosphere, and then heated to 1,537 Centigrade, expanding as the piston moves to its volume before compression, and then ex- hausting, it will be found to give the following results : 1 cubic foot of air (17 Centi- ] #rade and 760 milimeters mercury) at constant vol- ume requires to heat it i from the temperature of ^4,416 foot-lbs. compression 150. 5 Centi- grade to 1,537 Centi- grade, heat equivalent to j Maximum pressure in Ibs. \ per square inch above ( 220 Ibs. atmosphere ; 39 qoui ejenbs jed - sq| ui sjai^dsoaiie OAoqe e . " S . 40 Pressure at end of stroke in ) A( ^ ^ r 89.8 above atmosphere ....... ) Temperature at the end of } 953= Centigrade . the stroke .............. ) Work done on the piston. . 7,888foot-lbs. Now the work actually given by 1 cubic foot of combustible mixture in the author's engine, as will be seen from Fig. 6, is 6,851 foot-lbs. The full lines are the diagram lines from the engine ; the dotted lines are the lines of compression and expansion without loss or gain of heat, except by work done on or by the air under similar conditions of temperature and compression. It will be observed that the compression line and the dotted line are very close to- gether ; no heat seems to be lost to the sides of the cylinder during compres- sion ; the loss of heat to the water-jacket is balanced by the gain of heat from 41 the piston, which must necessarily be much hotter than the cylinder sides, as it only loses heat by contact with the cylinder and by the circulation of air in the trunk. The temperature at- tained at the commencement of the stroke is in both esses identical, 1,537 Centi- grade ; the temperature at the end of the stroke without loss of heat is 953; the temperature in the cylinder at the end of the stroke is 656 Centigrade. The diameter of the cylinder from which this diagram was taken is 6 inches, and ahe length of stroke 12 inches. This flppears a very small loss of heat from a tame filling the cylinder, considering the surface exposed and the great difference of temperature between the ignited gases and the enclosing walls. Is it to be con- cluded, then, that the loss of heat to the cylinder during the time of the forward stroke is only 953 -656 = 297 Centi- grade ? On this assumption the duty of the engine would be 42 and the consumption of gas per indicated HP. per hour would be 3 92 OT286 =18 - 7 CUbi feet) but the consumption is 22 cubic feet per indicated HP. per hour, so that there has in some way been lost much more heat than is to be accounted for by the temperatures as determined by the dia- gram. The duty of the engine is The duty of the engine expanding to the same volume as the mixed gases be- fore compression is Gas required per IHP. perhr. Cub. ft. Duty without loss of heat to / sides of cylinder .......... Duty with loss of heat as shown by diagram Duty as determined by experi- ) Q ^ ^20 ment ..................... ) Now the number of cubic feet of com- bustible mixture required to produce 1 HP. for one hour in the author'* 1,980,000 6,851 The amount of gas in the engine per 22 cubic foot of mixture, ^ =0.0761 cubic foot, or =-Q of the total volume of gaseous JLO mixture passed into the engine. If only the amount of gas necessary to heat the air to the required temperature is pres- ent, 1 cubic foot requires 0.0482 cubic foot of coal gas, or about of its vol- iJL ume ; that is, although to heat up a cubic foot of inflammable mixture from 150 Centigrade to 1,537' only 0.0482 cubic foot of coal gas is required, yet although there is present 0.0761 cubic foot, or 1.58 time the amount necessary, the tempera- ture does not rise any higher. Why is this? Before going into the question, it is better to determine as nearly as possible what becomes of 100 heat units used bv 44 the engine. The exhaust being dis- charged at a temperature of 656, and the temperature of the air before com- pression being assumed at 17, it fol- lows that the exhaust from 1 cubic foot carries away with it (65617) X 17.61 = 11,253 foot-lbs. The work done by the cubic foot of mixture is 6,851 foot-lbs., and the equiva- lent in foot-lbs. of the gas present in 1 cubic foot of explosive mixture is 0.0761 X 505,000=38,430 foot-lbs. The heat is therefore disposed of as follows : Heat-units Foot-lbs. percent. Work done by 1 cubic foot ) Q^ ** Q o > O,oOJL l/.oo of mixture ) Mechanical equivalent of \ heat discharged with the ( 11,253 29 .8 exhaust ) Mechanical equivalent of j heat passing through sides > 20,326 52 . 89 of cylinder ) 38,430 100.00 This investigation is only approximate. The determination, with anything like 45 possible physical accuracy, would require an examination of many points involving months of continuous work. It is the author's intention to make an accurate research into the phenomenon attending the use of the gas engine, for the pur- pose of obtaining the physical constants necessary to calculate exactly the con- sumption of any power, size, and theory of gas engine, such as it may be possible to construct in the future. For the present, however, it is only necessary to discuss the principles in such a manner us to clearly show where original re- search is required. More than one-half of the total heat given to the engine passes through the sides of the cylinder and is lost. How is this enormous loss of heat sustained, while only a compara- tively small fall of temperature takes place below the adiabatic curve ? This leads back to the question of the gas present in excess of the amount necessary to raise the temperature to 1,537 which has already been noticed. At this point it is necessary to consider 46 the gas engine as something different from a hot-air engine. The chemical phenomena attending combustion now require consideration. If 2 volumes of hydrogen be mixed with 1 volume of oxygen (the proportions necessary for complete combination of both gases to form water), and be ignited in a closed vessel in such a manner that the maximum pressure may be measured, it will be found that the pressure is a much lower one than would be expected if the complete combination of the two gases took place at once, and the whole heat due to this combination were de- veloped. That this is not due to loss of heat to the sides of the vessel has been shown by Bunsen. He proved that the ratio of rise in pressure is exceedingly rapid compared to the rate of fall of pressure. The time taken for the in- flammation of the whole volume of mix- ture is the time of attainment of the maximum pressure. In his experiments he used only a very small tube, which contained a volume of gaseous mixture, 47 8.15 centimeters long, by 1.7 centimeter in diameter, and the entire length of this column was traversed by the electric spark, in order that the inflammation of the whole mass in the tube might be as nearly instantaneous as possible. In practice he succeeded in producing a maximum temperature in so short a time as j-yVo part of a second. By examining the light from the explosion through a revolving disc provided with radiating segments, the rate of revolution of the disc being known, he determined the duration of light within the tube, and therefore the duration of a temperature not far removed from the maximum. The duration of the illumination was found to be -^of a second. A maximum pressure, obtained in so short a time, with a duration so relatively long, makes it impossible that loss of heat through the sides of his tube could have affected his experiments. The cause, therefore, of the pressure falling so far short of what it would be if the combination took place Sompletely. is simply this, that the 48 temperature is so high that complete com- bustion is impossible. The temperature, and therefore the pressure produced by the combination of any gases, is limited by the dissociation or decomposition of their products of combustion. When any two gases combine, say (H) and (O) to produce water, what happens is this. The temperature rises till a point is reached, when any further rise would decompose the water which is already formed ; and if the gases are kept at this temperature, no further com- bination will take place. If the tempera- ture is lowered, further combination takes place until it is low enough to allow of the existence of steam without decomposition. The temperature at which steam can exist as steam without its partial resolu- tion into hydrogen and oxygen gases is not a high one. At 960 to 1,000 Centi- grade Deville has proved that it com- mences to decompose, and at 1,200 Centigrade, considerable decomposition takes place, the amount of decomposition 49 increasing as the temperature rises : for each temperature there is a proportion of steam to free gases, which is constant, and does not change till the temperature changes. The same law holds true for carbon dioxide ; at high temperatures it decomposes into carbonic oxide and free oxygen. Bunsen attempted to determine the temperature attained on the explosion of a mixture of hydrogen and oxygen, a pure electrolytic mixture. He found that the maximum pressure attained by such a mixture is 10 atmospheres, the temperature before ignition being 5 Centigrade. From this he calculated the temperature produced, but in doing so, as Berth elot afterwards pointed out, he neglected the fact that when these gases combine, 3 volumes of the gases form 2 volumes of steam gas, and con- sequently if complete combination is assumed, and it be supposed that the pressure is produced by steam only, the volume, before ignition, must be calcu- lated at two-thirds of that taken by the 50 mixed gases. But as it is known that combination is incomplete, at the lowest assignable temperature of the combus- tion, and it is not possible to tell the amount of combination at a given press- ure without knowing the temperature, this cannot be assumed. As in determining temperature by an air thermometer it is necessary that the amount of air in the thermometer should be constant at the different temperatures, it is evident that the temperature of an explosion cannot be known from the in- crease in pressure unless the chemical changes taking place do not alter the volume of gases under observation. In calculating the temperatures at tained in the author's engine, this fact has been kept in view. The capacity ol the space at the end of the cylinder was carefully taken by filling with water and weighing, the water. As the proportion of the combining gases to the excess of oxygen or free nitrogen is very small, only one-thirteenth of the whole volume used being combustible gas, the space 51 may be considered as simply filled with heated air, and the contraction caused by the formation of H^O and CO, neglected, especially as an increase in volume fol- lows the combination of the olefines with oxygen. 2 volumes of H combine with 1 volume of O, forming 2 volumes of steam. 2 volumes of marsh gas (CH 4 ) require for complete combustion 4 vol- umes of O, and form 4 volumes of H 2 and 2 volumes of CO.,. 2 volumes of carbonic oxide (CO) unite with 1 volume of O, forming 2 volumes of CO 2 . If the olefines in coal gas be taken as of an average composition of C 3 H 6 , then 2 volumes require for complete combustion 9 volumes of oxygen, forming 6 volumes of H 2 O and 6 volumes of CO 2 . Now taking the composition of coal gas as below the noted amounts of oxy- gen are required for combustion, and the given volumes of the products are formed 52 vols. vols. vols. H=50 requires 25 O=50 H 8 O produced. CH 4 =33 " 66 O=99 CO 4 &H 2 O " CO=13 " 6.5 O=13 CO 2 C 3 H 6 = 4 " IS O=24 CO 2 &H 2 O " 100 + 115.5=225.5 gives 186 vols. The amount of contraction due to com- plete combustion of this coal gas is small even when burning with pure oxygen, 225 volumes of the mixed gases becom- ing 186 volumes after combustion. When diluted with nitrogen the proportion of contraction is less and introduces no serious error. With a mixture of 1 volume of gas to 12 volumes of air, 125 volumes of the mixture before combina- tion become 122 volumes when complete- ly combined, at the original temperature, assuming the water to remain gaseous. If the curve of the dissociation of water and carbonic dioxide were known, it would be possible to show on the indica- tor diagram the reserve of heat available at each point of the fall. What the engineer requires of the 53 scientific chemist is a curve of the disso- ciation of water and carbonic acid, at temperatures ranging from the maximum produced by combustion down to the point at which it may be safely assumed that complete combination is possible. In Fig. 6 the dotted line shows a fall of temperature, by hot air doing work without loss of heat through the cylinder, and the black line shows the actual fall of temperature in the author's engine, with ]oss of heat through the sides of the cylinder. It is evident then that the cause of so near an apparent approach to theory is, that at the maximum tempera- ture, complete combination of gases with oxygen is impossible, and cannot take place until the temperature falls. As the temperature falls the gases further combine, until a temperature is reached at which combination is complete. The loss of heat through the sides of the cylinder is therefore much greater than would appear from the diagram. In calculating the efficiency of the gas en- gine, all previous observers have assumed 54 that the loss of heat to the cylinder is to be obtained from the comparison on the indicator diagram of the actual expan- sion-line with an adiabatic line from the same maximum temperature and press- ure. So far as the author is aware, Professor Riicker, of Leeds, was the first to point out the necessity of taking into account the phenomena of dissocia- tion in making such comparisons. Ac- cordingly, all previous estimates of effi- ciency, based on the indicator diagrani, are much too high. The gas engine, then, differs from the hot-air engine, using air heated in the manner assumed in the first part of this paper, in this, that the temperature is sustained, notwithstanding the enormous flow of heat through the sides of the cylinder, by the continuous combination of the dissociated gases. Figs. 7 and 8, have been taken from the "Journal of the Franklin Institute." They are Lenoir engine diagrams, and in them the same phenomena are apparent ; although running at a very slow speed, 55 the pressure is most perfectly sustained, the dotted lines showing the adiabatic, and the full lines the actual diagram. The author of the paper in which they occur, gives the probable maximum tem- Fig.7. LENOIR ENGINE. Diagram at 50 revolutions, cylinder 8^ inches diameter, 16^ inches stroke. LENOIR ENGINE. Diagram at 45 revolutions, 1 inch =32 Ibs. 56 perature attained at about 1,356 Centi- grade, and he says, "The dotted line represents the theoretical curve of ex- pansion, taking into account the loss of heat and consequent fall of pressure, due to the work done (which is the proper theoretical curve for an indicated dia- gram). The temperature at the end of the stroke, indicated by this line, would be 2,156 Fahrenheit (1,180 Centigrade). The final temperature shown by the dia- gram, supposing there be no leakage, is 1,438 Fahrenheit (781 Centigrade), and the difference 718 Fahrenheit (399 Centigrade), is the quantity of heat ab- sorbed by the water-jacket by which the cylinder is surrounded." " It will be observed that the explo- sion takes place so late in the stroke that there is a considerable available pressure at the end of the stroke, which of course is not utilized." Now if the Lenoir engine had only lost this amount of heat through the sides of the cylinder it would have been very economical, and would have ap- 57 preached the theoretic consumption mentioned in the earlier part of this paper; but the causes of loss are so great that it never did come anything near this figure, and an erro"r is intro- duced through neglecting the effects of dissociation. Interesting information, however, is to be obtained from these diagrams as to the proportion of gas and air in the mix- ture used by the Lenoir engine. When these diagrams were taken the maximum temperature after ignition was 1,356 Centigrade ; now in the author's present engine the maximum temperature is 1,537; it follows that Lenoir used a more diluted mixture as the temperature after ignition was lower. The engine giving this diagram could not have been using an ignitable mixture containing more gas than one-fourteenth of its vol- umea mixture which the author finds to be easily ignited at ordinary atmos- pheric pressure. The statement is often made that such a mixture will not ex- plode except it be first compressed ; this 58 is incorrect, it is possible to ignite even a weaker mixture without compression. Coquillon has determined the limits be- tween which a mixture of marsh gas (CH 4 ) and .air can be exploded. Mixtures of marsh gas and air in different propor- tions were introduced into a eudiometer and fired by the electric spark, with the following results : Marsh gas 1 volume, air 5 volumes. The spark is without effect. Marsh gas 1 volume, air 6 volumes. Explosion only occurs- in a succession of shocks. This is the first limit of possible explosion ; the marsh gas is in excess. Marsh gas 1 volume and 7, 8 and 9 volumes of air give a sharp explosion. With 12, 13, 14, 15 volumes of air for 1 volume of marsh gas the explosion occurs, but grows gr^dual- ly weaker. With 16 volumes of air the effect is reduced to a series of slight in- termittent commotions. This is the second limit ; the air is in excess. In Fig. 8, ignitions will be observed very late in the stroke ; these misses were caused by the points between which 59 the electric spark is discharged getting wet and thus preventing the passage of the spark at the proper time. From these diagrams, the time, from the begin- ning of rise in pressure to the attainment of maximum pressure, is found to be from one twenty-seventh to one-thirtieth of a second ; when the ignitions are late it takes longer, one-twentieth of a second being required ; that is, the flame has spread completely through the mass in one-twentieth part of a second. Now in the author's engine, calculating from the moment when the ignition port is opening to the flame, to the moment of maximum pressure as found from the diagrams, it has been ascertained that the time occupied is an average of one twenty-fifth of a second, a time nearly identical with that found for the Lenoir engine. If it be admitted that the flame has spread completely through the mass when the maximum pressure is attained in the Lenoir engine, it cannot be sup- posed that it has not spread in like man- 60 ner throughout the mass of ignitable mixture in the modern compression en- gine. Maximum pressure is the only outward indication of complete inflamma- tion ; by complete inflammation is not meant the thorough chemical combina- tion of the active gases present, but the spread of the flame through the entire mass. That when maximum pressure has been reached complete inflammation has also been attained has hitherto been considered self-evident. It is only lately that the theory has been advanced by Mr. Otto that in the modern compression engine attaining maximum pressure at the beginning of the stroke, the flame has not spread throughout the mass of the ignitable mixture in the cylinder; but that as the piston moves forward the pressure js sustained by the gradual spread of the flame. This supposed phenomenon has been erroneously called slow combustion ; if it has any existence it should be called slow inflammation. It has a real existence in the Otto engine only when it is working badly ; but even then maximum temper- 61 ature is attained, and very distinctly marks the point of completed inflamma- tion. The time taken to attain maximum pressure is longer in a large engine than in a small one, because the distance through which the flame has to travel is greater. During the investigation al- ready referred to, Professor Bunsen determined the celerity of the propaga- tion of ignition through a pure explosive mixture of hydrogen and oxygen in the following manner : the explosive mixture was allowed to burn from a fine orifice of known diameter, and the current of the rate of the gaseous mixture was carefully regulated by diminishing the pressure^ to the point at which the flame passed back through the orifice and ignited the gases below it. This passing back of the flame occurs when the velocity with which the gaseous mixture issues from the orifice is inappreciably less than the velocity with which the inflammation of the upper layers of burning gas is propa- gated to the lower and unignited layers. 62 The rate of the propagation of the ignition in pure hydrogen was found to be 34 meters per second. In a maximum explosive mixture of carbonic oxide and oxygen it was not quite 1 meter per second. Mr. Mallard has determined the rapid- ity of the propagation of inflammation through mixtures of coal gas and air by this method, and found that the maxi- mum rate of propagation was attained with a mixture of 1 volume of coal gas with 5 volumes of air, and it is 1.01 meter per second. One volume of coal gas with 6J volumes of air gave a rate of 0.285 meter, or 11 inches per second. This is the rate of ignition, it must be remembered, at constant pressure ; in a closed tube fired at one end it would ig- nite with much greater rapidity. In a closed space the conditions of inflamma- tion are quite different. The ignited portion instantly expands, compressing the portion still remaining, and thus car- ries the flame further into the mass, so that to the rate of ignition at constant 63 pressure is added the projection of the flame into the mass by its expansion. To determine from the rate of ignition at constant pressure the time necessary to completely inflame a given volume of mixture at constant volume is a very complicated problem, which it is proba- ble can only be solved experimentally. The author has found it possible to ignite a whole mass in any given time between the limits of one-tenth and one- hundredth part of a second, by so arrang- ing the plan of ignition that a small vol- ume of gaseous mixture is first ignited, expanding and projecting a flame through a passage into the mass of inflammable mixture, and thus adding to the rate of ignition the mechanical disturbance pro- duced by the entering flame. He has succeeded by this means in producing maximum pressure in one-hundredth part of a second in a space containing 200 cubic inches. This rate of ignition is too rapid, and would not give the en- gine time to take up the slack in bearings, connecting rods, &c. But by firing a 64 mixture with varying amounts of mechan- ical disturbance almost any time of igni- tion can be obtained between -^--^ and -^ of a second. It does not matter whether the mixture used is rich or weak in gas ; the rich mixture can be fired slowly and the weak one rapidly, just as may be re- quired. The rate of ignition of the strongest possible mixture is so slow that the time of attaining complete inflamma- tion depends on the amount of mechani- cal disturbance permitted. Fig. 9, a diagram from an Otto engine, shows what happens in a compression engine of type 3 when the ignition comes late and the movement of the piston overruns the rate of the spread of the flame. It is then seen that the maximum pressure is not attained until far on in the stroke, and as a consequence great loss of power results, the pressure at- taining its maximum when it is time for the exhaust valve to open. This may happen from several causes, a too diluted mixture, or too little mechanical disturb- ance by the entering flame; or the igni- 65 66 tion may be missed until the pressure begins to fall by the forward movement on the piston, when the rate of inflamma- tion begins to come more nearly to Mal- lard's number of 11 inches per second. This slow combustion, or rather slow in- flammation, is to be avoided in the gas engine. Every effort should be made to secure complete inflammation as soon after ignition as is practicable. The lines in the diagram show this very clearly ; the normal lines are those in which the rise is almost straight up from the point of the beginning of the ignition ; they are marked a and b ; the line c, although com- mencing from the beginning of the stroke, does not record the maximum pressure till the piston has moved forward one- third of its stroke, while the line d does not depart from the compression line until one-tenth of the forward movement, and does not attain its maximum till near the end of the stroke. In the last case the ignition has been missed until the pis- ton is in rapid motion, and consequently the flame is at first unable to overtake it. 67 The rate of inflammation at constant pressure has been determined only for atmospheric pressure ; were it known for higher pressures it would be possible to calculate exactly the piston speed which would prevent any rise in pressure at all. Fig/ 10 was taken by the author from the motor cylinder of an American Bray- ton engine of type 2. It shows how the pressure is sustained as the ignited gases enter the motor cylinder in flame. This is the true slow inflammation engine ; in it the pressure after ignition is not al- lowed to rise, but only increase of volume takes place ; at about the middle of the stroke the supply of flame is cut off and the piston moves on, and the heated gases expand doing work. Fig. 11 is the compression pump dia- gram, which must be deducted before getting the available indicated power. The motor-piston was of the same area as the pump, but had double the length of stroke. This type of engine is not a good one for a cold cylinder, the loss of heat through the cylinder being much 68 more than in type 3 ; but, as it has been before said, the possibility of using the w ft tf g O o g 2 -< theory in the future with a hot piston and cylinder renders reference to this engine 69 interesting. Slow inflammation is a mis- take if applied to engines of types 1 and 3 with cold cylinders ; in type 1, if the piston were moving rapidly enough, the inflammation could be so slow that with 70 a perfectly sustained temperature no power at all could be obtained. That is, the air would simply expand in volume without rising in pressure above the atmosphere, and even without loss of heat to the sides of the cylinder the whole heat would be uselessly discharged. In type 3 the perfection of slow com- bustion would be attained when the flame spread just as rapidly as the piston moves forward, and the pressure was never raised above that due to compression. The pressure diagram would then give the ideal results of " gradual expansion of gases" and a "perfectly sustained pressure." But this is just the condition of greatest loss of heat ; sustained press- ure means sustained, indeed increasing temperature, and the object to be attained in a good gas engine is to produce the most rapid possible fall of temperature due to work performed, to keep the mean temperature as low as possible, and it is only so far as this is successfully done that economy is possible. Slow inflam- mation causes loss of heat and power; 71 rapid inflammation reduces the loss to a minimum while attaining the maximum possible power. One more engine may be noticed ; its diagram is given at Fig. 12. In action it comes under type 1, but uses a very large amount of expansion, and is further complicated by cooling. It is the well- known Otto and Langen engine of the free piston type ; in it gas and air are taken in, for a portion of the stroke at atmospheric pressure and then ignited while the piston remains at rest until the pressure sets it in motion; the piston is free to move apart from the shaft altogether, and on the up- stroke it does no work. From f to a air and gas are taken into the cylinder. At a the mixture is ignited and the .piston moves to c with considerable velocity when the pressure has fallen to the atmosphere. From c to e it continues to move with continually diminishing velocity, until at e it comes to rest and then returns doing work, the work being equal to the diagram d g e added, to the weight of the piston and lOOr 60 72 Fig.12, 10 20 30 40 50 60 70 80 OTTO AND LANGEN ENGINE (FREE PISTON). Percentage of stroke. 73 rack through the stroke. It will at once be seen that as the gases only do work on the piston from a to c, and this work is absorbed in giving a certain velocity to the piston, and from c to e the velocity of the piston is being gradually checked by doing work on air, assuming the pis- ton to have no weight, the area of the portion of the diagram a c b must be equal to the part c e d. It is evident that the lines in the dia- gram are incorrect ; the explosion cannot fall nearly so rapidly as shown ; c should be much nearer e. The oscillations of the indicator have been so great that ac- curacy is impossible. The fall of the line d g below d e is caused by the cool- ing of the gases on the return strokte. In this engine the advantage consists more in the large amount of expansion than the velocity of the forward move- ment of the piston. The diagram has been taken from a paper by Mr. F. W. Crossley ; with ref- erence to it he says : " The very sudden and extreme rise in 74 pressure at the moment of explosion is due simply to the expansion of the gases under the temperature of the flame. If this temperature be taken at 5,000 Fahrenheit, and divided by 520 for the rate of expansion from an initial tempera- ture of about 60, it gives an expansion of about 10 times ; and as the gas com- pound occupied one-eleventh of the cylinder at the moment of ignition, if it expands ten times it gives very nearly the stroke actually takc?n by the piston. The 5,000 is an assumption only, but seems to be confirmed by the amount of expansion which follows it. After the explosion the temperature falls almost instantaneously, as shown by the sudden drop of pressure in the diagram." In the author's opinion Mr. Crossley has completely misinterpreted his dia- gram. Taking the temperature before ignition at 60 Fahrenheit, and the maxi- mum pressure shown on the diagram as 100 Ibs. absolute, it follows that the maximum temperature is not greater than 2,900 Fahrenheit (1,590 Centi- 75 grade). It is difficult to see how 5,000 Fahrenheit can be assumed. The ex- pansion of the gases by the extreme movement of the piston following igni- tion has no necessary relation to the temperature of the explosion ; but it is determined wholly by the work done on the piston by the explosion between the maximum and atmospheric pressures. Whenever the gases in the cylinder fall to the pressure of the atmosphere, which happens according to the diagram at about 0.35 of the stroke, the piston is doing work on air, and the mean press- ure below the atmosphere from c to e is the exact measure of the work previously done on the piston by the explosion, which has been expended in giving the piston velocity. This energy of motion is now being expended by compressing the atmosphere. Taking into consider- ation the weight of the piston and fric- tion of the rings, rack and clutch, it is certain that the area of the part of the diagram a b c must be considerably greater than c e d\ in the diagram it ap- 76 pears much less. It should be greater by the amount of work expended in giving the piston energy of position, and the amount lost by friction on the up-stroke. As a means of showing the nature of the explosion this diagram is mislead- ing ; it is certain that the maximum press- ure was less, and that the fall of press- ure is nothing like so rapid as it there appears. Comparing Fig. 12 with Figs. 7 and 8 the difference in appearance is so striking that it looks as if in one case the fall in pressure was instantaneous and in the other gradual ; this would be remarkable, considering that the maxi- mum temperatures are very similar. If the lines in Fig. 12 be corrected and drawn with the same relation of scale between pressures and strokes, it will be. found to be very similar in appearance to Figs. 7 and 8, so far as rate of fall is concerned. Indeed the advantage claimed for this engine is a movement of piston so rapid that its expansion is complete before much heat is lost to the sides of the cylinder, which is inconsistent with a 77 fall of pressure more rapid than in the Lenoir engine. To go completely into the points of originality in these engines would require a paper on the " History of the Gas En- gine ; " but it may be well to state the name of the first to propose each type : Year. Type 1. Explosion acting on piston con- nected to crank. . .W. L.Wright 1833 Explosion acting on free piston, Barsanti & Matteuci 1857 Type 2. Compression after ignition but at constant presssure. C. W. Siemen s 1 860 Compression with increase in vol- ume F. Millon 1861 Type 3. Compression with incrro.se in pressure - F. Millon 186L After ignition but at constant volume So far as the author has been able to ascertain, these are the names of the first to propose distinctly each of the three types of gas engine. From the considerations advanced in the course of this paper, it will be seen that the cause of the comparative ef- ficiency of the modern type of gas en- 78 gines over the old Lenoir and Hugon is to be summed up in one word, " com- pression." Without compression before ignition an engine cannot be produced giving power economically and with small bulk. The mixture used may be diluted, air may be introduced in front of gas and air, or an elaborate system of stratification may be adopted, but with- out compression no good effect will be produced. The proportion of gas to air is the same in the modern gas engine as was formerly used in the Lenoir, the time taken to ignite the mixture is the same, the only difference is compression. The combustion, or rather the rate of inflam- mation, is indeed quicker in the modern engine because the volume of mixture used at each stroke is greater, and yet the time taken to completely inflame the mixture is no more than in the old type. The cause of the sustained pressure shown by the diagrams is not slow in- flammation (or slow combustion as it has been called), but the dissociation of the products of combustion, and ual combination as the temperature falls, and combination becomes possible. This takes place in any gas engine, whether using a dilute mixture or not, whether using pressure before ignition or not, and indeed it takes place to a greater ex- sent in a strong explosive mixture than in a weak one. The modern gas engine does not use slow inflammation (or slow combustion if the term be preferred), but when work- ing as it is intended to do, completely in- flames its gaseous mixture under com- pression at the beginning of the stroke. By complete inflammation is meant com- plete spread of the flame throughout the mass, not complete burning or combus- tion. If by some fault in the engine or igniting arrangement the inflammation is a gradual one, then the maximum press- ure is attained at the wrong end of the cylinder, and great loss of power results. Compression is the great advance on the old system ; the greater the compres- sion before ignition the more rapid will 80 be the transformation of heat into work by a given movement of the piston after ignition, and consequently the less will be the proportional loss of heat through the sides of the cylinder. The amount of compression is of course limited by the practical consideration of strength of the engine and leakage of the piston, but it is certain that compression will be carried advantageously to a much greater extent than at present. The greatest loss in the gas engine is that of heat through the sides of the cylinder, and this is not astonishing when the high temperature of the flame in the cylinder is considered. In larger engines using greater compression and greater expan- sion it will be much reduced. As an en- gine increases in size the volume of gas- eous mixture used increases as the cube, while the surface exposed only increases as the square, so that the proportion of volume of gaseous mixture used to sur- face cooling is less the larger the engine becomes. Taking this into consideration, it may be accepted as probable that an 81 engine of about 50 indicated HP. could be made to work on 12 cubic feet of coal gas per indicated HP. per hour, or a duty of about 32 per cent. The gas engine is as yet in its infancy, and many long years of work are neces- sary before it can rank with the steam engine in capacity for all manner of uses ; but it can and will be made as managea- ble as the steam engine in by no means a remote future. The time will come when factories, railways amd ships will be driven by gas engines as efficient as any steam engine, and much more safe and economical of fuel. Grs generators will replace steam boilers, and power will not be stored up in enormous reser- voirs, but generated from coal direct as required by the engine. The steam engine converts so small an amount of the heat used by it into work that, although it was the glory and honor of the first half of the century, it should be a standing reproach to engineers and scientists of the present time having con- stantly before them the researches of Mayer and Joule. 82 APPENDIX. DATA USED IN THE PAPER ON ' ' THE THEORY OF THE GAS ENGINE." Specific heat of air at ) A 1 ft constant volume. } = ' 169 : water IM Specific heat of air at [ _ ~ 9 oo constant pressure [ ~~ Mechanical eqivalent ) of heat foot-lbs. V =1389.6 Centigrade ) Specific heat of air at") constant volume | in foot-lbs. for II 1* a f + n cubic foot at irf 17.6 foot-lbs. C. anrl 760 mm. | barometer J Specific heat ( -f air at ] constant pressure I in foot-lbs. for 1 I 9 , ft 0 MG r ^ v v ' .\~~~ ~ / [ ^\ f i / 1 \ / X ^ N J i \ / ^"""~"~ *^L" ' f 1 h /!\ / \ / I N I N v / I V .j x w I I \ I I I N -^. I - . *" Ai B C :D "V - ^ i ''"V/ Vi^>- i ^-^""i x x / \ x x \ '""^ ! / x / \ \ \ \ y"^" i/ / \ \ /^ i- / \ \ \ / i / \ \ / i / \ \ // i V V \ \ \ ^--~-i \ \ 113 ascended to a point past the point of maximum pressure, viz., till the point K, at the commencement of the part KY (which was supposed to be exactly adi- abatic) was reached. From the point S this curve became in the actual diagram a straight line parallel to AB. If, how- ever, the theoretical diagram, allowing for loss by conduction, were taken, the curve PQRS would ascend throughout the stroke. Hence the maximum point on the diagram was simply the point where the increase of pressure due to combustion was balanced by the decrease of pressure due to the forward motion of the piston, and there was no reason for saying that this maximum point cor- responded to complete ignition. He had had an opportunity of taking diagrams from the Otto gas engine, which Pro- fessor Ayrton had at the City Guilds Technical School, Cowper Street. The engine was designed for the electric light, and the cam, controlled by the governor, was made in a series of steps. He therefore had the governor taken 114 off, and the cam and the roller on which it acted so arranged that it should work independently of the velocity of the engine on a given step, so that the charge might be, as nearly as possible, the same at all speeds. And he varied the load by braking the fly-wheel. The two sets of diagrams were taken, one at a speed of one hundred revolutions, and the other at two hundred ; thus might be seen the effect which must be due to the phenomenon he had spoken of the ignition traveling gradually ; it could not be due to dissociation, for the reason which Mr. Imray had pointed out. In the diagrams the phenomena of dissoci- ation ought to be exaggerated at the higher temperature, but instead of that, it would be seen that the effects at- tributed to dissociation were less at the higher temperature where dissociation should be most active, and greatest at temperature below the point of dissoci- ation ; he therefore did not see why the results should be attributed to the phe- nomena of dissociation, when they could 115 be perfectly explained by the rate of progress of ignition through the cylinder. With the full charge at one hundred and at two hundred revolutions the effect of difference of speed was small, as shown by the two diagrams in Fig. 15. In that case, the rate at which the ignition went through the cylinder was so great that it only made a very little difference in the curve when the rate got up to two hundred revolutions. He then fixed the roller on the third step, when there was a less charge of gas. The diagram, Fig. 16, showed the hundred- revolution curve, in which the gas had time to explode, and to carry the pencil indicator up to the maximum point, and then down to the adiabetic line. Going to two hundred revolutions with the more dilute mixture, the rate of propaga- tion of ignition was slower ; therefore at that speed, although the temperature was less, dissociation would have much more to do. The effect was much more marked, simply from the dilution of the mixture ; there was therefore a less rate 116 lO bb of propagation of ignition, and the curve took the form shown in the diagram. Fig. 17 showed the same effects on the diagram when the curve roller was on 118 the second step, and consequently still less gas was admitted. The five super- posed diagrams were taken at speeds be- tween one hundred and two hundred revolutions per minute. It would be observed that the curve at the higher speed generally went outside the door. There was less work done at the begin- ning, and more gas to be combined at the end, and therefore a greater amount of work done at the end of the stroke. He did not wish to carry the comparison all the way through, but he would leave it to the author to show how he ex- plained the diagrams under the dissoci- ation theory. In Fig. 18 there was the least amount of gas with which the en- gine would work, and the speed was one hundred and thirty revoluions. The compression was 30 Ibs. ; the compres- sion line was the same as the others. The working line was a line nearly par- allel with the atmospheric line, but slightly rising, and at the end the ig- nition was not finished, indeed, in this case, if a light was applied to the ex- I haust the contents would explode. Ac- cording to the author's theory, that maximum point near the end of the stroke in the last diagram was a point 120 CO bJD 121 where the ignition was complete, and therefore all the gas should have combined at that low temperature where no dis- sociation could take place. Those were points which the author would have to meet in order to support his theory. Many of the facts mentioned by the author were incontestable, and his chief dispute with him was as to the interpre- tation he had put upon them. The au- thor had said nothing against the theory to which he had referred except that it was new, no argument whatever being advanced against it. The author stated, "From the considerations advanced in the course of this paper, it will be seen that the cause of the comparative effi- ciency of the modern type of gas engines over the old Lenoir and Hugon is to be summed up in one word, 'compression.'" He had not had time to go carefully through the diagrams ; but he did not think that they were fair comparisons, and he thought that other elements ought to have been taken into account. The author had given the old Lenoir, 122 and had stated that the temperature was the same, that the mixture of gas was the same, and that the great advantage over the Lenoir was compression. Mr. Bousfield might be permitted to point out that, in the Lenoir engine, the adi- abatic line was much above the actual line. It would be fairer to substitute the word " dilution " for " compression," so that the sentence would read: "The cause of the comparative efficiency of the modern type of gas engines over the old Lenoir and Hugon is to be summed in one word, ' dilution.* " The fact, how- ever, was that it could not be summed up in one word ; the two should be taken together, compression and dilution. The author further stated: "The pro- portion of gas to air is the same in the modern gas engine as was formerly used in the Lenoir." He did not think so. He believed that the Lenoir worked up to 13 to 1, and could not get further. He did not know what proportion Otto used, but it was considerably more than that. It was also stated that the time 123 taken to ignite the mixture was the same ; but that was a gratuitous assumption. The author said : " The cause of the sus- tained pressure shown by the diagrams is not slow inflammation (or slow combustion as it has been called), but the dissociation of the products of combustion, and their gradual combination as the temperature falls, and combination becomes possible. This takes place in any gas engine, whether using a dilute mixture or not, whether using pressure before ignition or not, and indeed it takes place to a greater ex- tent in a strong explosive mixture than in a weak one." Dissociation took place far more at high temperatures than at low; and if the author's application of the theory were correct the phenomena of dissociation ought to play a much greater part at high than at low tem- peratures. He had pointed out that this was not so in the diagrams, and that it was not so with Lenoir's explosive en- gines where the curve fell far below the adiabatic line. The paper contained other matters which he had not time to 124 dwell upon ; but he thought he had said enough to challenge the author to. show how he got rid of the old theory, and ex- plained the facts to which Mr. Bousfield had referred. Dr. JOHN HOPKINSON said a very inter- esting question had been discussed by Professor Kiicker and Mr. Bousfield, to which he desired to refer. The author maintained that the ignition of the mix- ture of gases had extended throughout the whole space at a time approximately represented by the point of maximum pressure. Others, on the contrary, maintained that the ignition had not ex- tended through that space by that time, but that it took a time lasting into the descending part of the indicator diagram before the disturbance had extended throughout the whole of that space. The author attributed the maintenance of the temperature during the latter part of the curve, and its approximation to an adiabatic curve, to the gradual combination of the gas through the mass, that combination not occurring com- 125 pletely in the first instance owing to the temperature being so high that a certain measure of dissociation occurred, or at all events so high that comptete com- bination could not occur. He thought that the question might be submitted to a crucial test. Suppose the opponents of the author were right, if a given mix- ture of air and gases were exploded in a gas engine revolving at a low rate of speed or in an entirely closed space, it would be expected that the maximum pressure would approximate to that calculated from the heat due to the com- bustion of the gas present and the tem- perature resulting therefrom. If the gine were running slowly, or if the ex- plosion were made in a completely confined space, the pressure would be expected to rise to a point very greatly in excess of that observed in the gas engine running at its normal speed. Whether that were so he did not know. The experiment might be objected to on the ground that when the engine was running slowly there was a great loss of 126 heat through the walls of the cylinder. That would give rise to a second crucial experiment. If the author was right the maximum pressure in large and small en- gines would be about the same ; if those who differed from him were right, in a large engine the maximum pressure would probably be greatly in excess of that in a small engine, there being less loss of heat through the walls ' of the cylinder. What the answer might be he did not know, but it appeared to him that there were there the elements of settling the ques- tion. The author divided gas engines into three classes, and had made a com- parison of their theoretical efficiency. In the second the mixtures were ad- mitted into the cylinder, and, without increase of pressure, the heat produced was devoted to increase of volume. In the third the mixtures were introduced into the cylinder, and then burned with an increase of pressure without immedi- ate increase of volume ; and in those two cases he took, for the purpose of com- parison, different maximum pressures. 127 In the second type he took a pressure of 7(> Ibs., and in the third over 200 Ibs. Prima facie it would seem natural, in order to make a fair comparison, that the same maximum pressure should be taken in the two cases. Probably the author had a good reason to justify his making a comparison on. that basis, and, per- haps, in his reply he would point it out. He agreed with those who had so often spoken on the subject of the gas engine that in that engine lay the future of the production of power from heat of com- bustion. It was quite in its infancy, and it had already beaten the best steam en- gines in economy of fuel, for the obvious reason that it was practicable to use with it much higher temperatures. The steam engine tolerably approximated to the theoretical efficiency that might be expected from it, having regard to the temperatures between which it was prac- ticable to work it. That was not the case with the gas engine, there being still a very large margin for practical improve- ment. Having regard to the very short 128 time during which gas engines had been used, he thought that practical improve- ments would take place, and that, when such difficulties as that of starting a large engine as conveniently as steam en- gines could be started had been over- come, the gas engine would supersede the steam engine. Mr. E. F. BAMBER wished the author had commenced his paper with that por- tion which treated of the analysis of the gas, and had given the mechanical equiv- alent of a unit of the same both in the pure and diluted state. If the explana- tion had then followed, that the mechan- ical equivalent of the latent heat of ex- pansion per unit of the gaseous mixture per degree of temperature was nearly the same as for atmospheric air, the reason why the gas engine might be considered in theory as an air engine would have been clearer, namely, that the adiabatic curve, or curve of no* transmission of heat, was nearly the same for both. The author commenced by an attack upon the steam engine. Much heat was required 129 in evaporating water whose specific heat was high, and hence the efficiency of the steam engine was low, and something better was needed ; whereas it was clearly proved by Eankine, a quarter of a century ago, that the maximum efficiency of a theoretically perfect heat engine, working between given limits of temperature, was equal to the ratio of the range of temperature to the higher absolute limit of temperature, and quite independent of the fluid employed. Raising the tem- perature entirely by compression or using regenerators were the two means by which the actual efficiency might be made to approach the maximum limit. The author believed in compression, but his method of defence of it and his illustra- tions of its advantages did not appear to be quite correct. He took three types of engine : the first and third were ex- plosive gas engines ; the second was worked at constant pressure, and these he treated as air engines. The first and second were worked between the same limits of temperature, but in the second 130 compression was employed. What the author wished to prove by the theoretical diagrams of these types was that the constant-pressure engine using com- press 'on was more theoretically perfect than an explosive engine using none, whilst Jin explosive engine using compres- sion was the best of the three. But he had shown by type No. 2, that by the use of compression ai- efficiency could be attained higher than the maximum ef- ficiency of a perfect heat engine, which seemed to require some explanation. The maximum was equal to - J ? in ab- r, solute degrees of temperature, and was for 1,537 Centigrade and 1,089 Centi- grade equal to 0.247 for both types ; whereas the author made it 0.21 for the first and 0.36 for the second. The author allowed that type No 1 would be im- proved by further expansion, but that that would require a vacuum pump and condenser ; yet surely it made no differ- ence, so long as they both consumed the same quantity of heat, whether a com- 131 pression pump was used at the beginning or a vacuum pump at the end of the stroke, whilst indeed there might be theo- retical reasons in favor of the latter. Types 1 and 3 were respectively worked without and with compression ; they were both explosive engines, and the efficiency of the latter was made double that of the former, but the latter was made to dis- charge at 648 Centigrade, and the former at 1,089 Centigrade. If these figures had been reversed, so would have been the efficiencies. Had the author explained that there was a certain maxi- mum efficiency for heat engines, and that by means of compression a larger per- centage of that maximum could be at- tained than without it, there would have been no reason for objection ; but that was a very different thing from trying to show that it was possible to obtain more than the maximum efficiency of a theo- retically perfect heat engine. The real value of the gas engine was, that it contained the furnace and engine in one ; thus the necessary heat lost in 132 the furnace to make a draught, and the unnecessary loss of heat by radiation from a large steam boiler were both avoided in the gas engine, and, finally, the gas engine could be used safety at a maximum limit of temperature, which could not be employed in the steam en- gine. There was no doubt a great future for this class of motor. Sir WILLIAM THOMSON said that he had recently seen a very interesting experi- ment made by the author with a gas engine at Glasgow, which he thought had a most important bearing on the mode of action of the gas in the cylinder. The experiment was made in the presence of his brother Professor James Thomson and Professors Jack and Ferguson (of Mathematics and Chemistry in the Uni- versity of Glasgow), who were all much interested in the inquiry. The object was to test the nature of the mixture in close proximity to the piston, so as to be able to form some idea as to whether or not the explosion took place through the whole space ; to be judged by finding 133 whether, right up to contact with the piston, gas and air were present in pro- portions suitable for combustion. He need not enter into details as to the way in which the experiment ,was made, but he might say, in a general way, that while the piston was being pressed in to con- dense the mixture at a definite point of the stroke, communication was made with the cylinder. The small experimental cylinder and piston were placed in proper position, in communication with an aper- ture bored for the purpose in the main cylinder. The author of the paper would be able to explain the details better than Sir William Thomson could. It was sufficient to say that by an automatic arrangement, worked mechanically from the cross-head, the communication was made exactly at one definite point of the stroke, and the experimental piston was pressed up in the cylinder so as to let it fill. At any time afterwards the stop- cock could be opened by hand, and the nature of the contents tested. In every case the contents were found to be ex- 134 plosive an explosive mixture of gas and .air proving that up to the very point, which he understood was within about an inch from the piston, coal gas was present in suitable proportions for producing an explosion. There was one other matter to which he wished to refer, which had been noticed in the discussion. There appeared to be some difference of opinion upon it, but to his mind it scarcely ap- peared open to doubt that the diagram, which showed an exceedingly sudden rise and a gradual fall, proved that combus- tion was practically complete at a point corresponding to the summit of the curve. Literally and precisely the instant of the maximum of the curve was that at which the rate of loss of pressure by expansion, the much smaller rate of loss of pressure by loss of heat carried by convection of the fluid to the solid boun. dary and out by conduction through the metal, were exactly counterbalanced by the rate of combustion still going on. It seemed certain that the rate of loss by the two causes he had indicated was ex- 135 ceedingly small in comparison with the rate of rise by the initial progress of the explosion ; therefore, practically speaking, the maximum of the curve indicated truly the instant when the combustion was as complete as dissociation at the highest tempearature attained allowed it to be. Mr. D. CLERK, in reply upon the dis- cussion, said that two of the speakers seemed to think that the question at issue was one of infringement of patent, but he desired to arrive at the truth, apart from mere questions of personal interest. The question of infringement was to him one of complete indifference. The question he was anxious about was the purely scientific one. Was his theory of the action of the gas engine the true one, or was it Mr. Otto's ? This mat- ter might appear to some persons a small one, but he considered it of vital interest, being convinced that not many years hence the gas engine would have a science of itb own, and scientific names connected with it as much honored as any ever linked with the steam engine. Dr. Sie- 136 mens bad fully corroborated his view of dissociation, and in the effect it had on the gas engine diagram, in preventing the more rapid fall, which must otherwise occur ; but he did not agree with him in the necessity for further research on dis- sociation, believing that St. Claire Deville's work was sufficient. Dr. Sie- mens would observe that St. Claire Deville's researches were referred to in the paper; but what he asked for had never to his knowledge been published, that was a complete curve of the dissoci- ation of water and carbonic acid. St. Claire Deville's results were more of a qualitative than of a quantitative nature. He feared that the method used was not capable of the necessary accuracy. He thoroughly believed that the engine for the very large powers to be construct- ed in future must be of one type 2, with hot chamber or cylinder, and regenerative contrivance in some form ; indeed, about two years ago he constructed and experi- mented with such an engine, and he was continuing his experiments. 137 The mechanical difficulties were much greater than in the cold cylinder, type 3. It must be remembered that the cold cylinder gas engine was the engine of the present, and it was most satisfactory that even with the small sizes so high a duty should be obtained. It proved that when larger engines were made a much higher duty might be expected. The theory of the cold cylinder engine did not allow of the application of any regenerative con- trivance, and consequently arrangements must be made to get the greatest possi- ble fall of temperature due to work done. A very interesting account had been given by Professor R acker of his view of the problem, and the necessity of correct- ing the calculations of previous observers in the light of present knowledge of the laws of combustion had been demonstrat- ed. It was satisfactory that Professor Eiicker so thoroughly agreed with him on the necessity for considering dissocia- tion in any theory of the gas engine, and had independently arrived at similar con- clusions. The experiments of Messrs. 138 Mallard and Le Chatelier corroborated those of Professor Bunsen in this, that at the high temperature of combustion, a large amount of heat was rendered latent. So striking a fact could hardly have escaped the notice of many other experi- menters who might not have published their results. He had noticed it about five years ago, while making experiments on the maximum pressure obtainable from a pure explosive mixture of gas and air. A cylinder 9 inches in diameter and 9 inches long, was filled with a mixture of gas and air in the proportions for maxi- mum explosive effect, and ignited the mixture by means of a hollow stop- cock, after Barnett's style of igniting arrange- ment. With the temperature of the mix- ture before ignition at 12 Centigrade,, the highest pressure attained was 97 Ibs. per square inch above the atmosphere. The pressure was measured by a loaded valve of known area, as in Bunsen' s -ex- periments. The absolute pressure attain- ed was only about 7-J- atmospheres ; if complete combination had taken place r 139 and no heat kept back by dissociation or absorbed by change in specific heat, then the pressure should have been at the low- est estimate, 11 atmospheres. He con- cluded that Professor Bunsen's explana- tion of this fact was a true one. The effect was equally visible in the large cylinder used by him and in the small tube used by Professor Bunsen. These experiments, and the recent experiments of Messrs. Mallard and Le Chatelier, make it certain that in a uniformly ignited gas- eous mixture the temperature was limited, and the apparent loss of heat was very- slow, and that this effect was due to dis- sociation, either complete or incipient. Such a mixture in expanding during work would give rise to all the phenomena de- scribed in the paper. He was pleased that his conclusions on the relation be- tween rate of inflammation at constant pressure and constant volume had been experimentally proved by these gentle- men. He had been challenged by Mr. Imray to controvert his statement on the history of the introduction of the gas 140 engine. This he did not do, because he considered Mr. Imray's account fairly correct. The only remark of Mr. Imray on his theory was : " He would only refer to Fig. 9. If the theory of dissociation were true, it would follow that the lower the temperature the more dissociation would take place, which was undoubtedly altogether wrong." It was difficult to understand this statement, it was so ex- ceedingly irrelevant. He could hardly believe the speaker had ever studied the pressure, volume, and temperature rela- tions of gases. On the indicated diagram low pressure had been mistaken for low temperature, neglecting the increased volume due to the travel of the piston. Mr, Imray had supposed that the maxi- mum pressure on line d (Fig. 9), being- lower than on line a, therefore the tem- perature was also lower. He failed to see the bearing on the theory under dis- cussion of Mr. Bousfield's statement : " He did not say that when the explosion took place, there might not be a certain 141 quantity of ammonia and a certain quan- tity of nitric acid formed." The question why, when maximum pressure was reach- ed at the beginning of the stroke, he as- sumed that the flame had spread through- out the mass in the cylinder was much more to the point. From the original of the diagram, Fig. 6, he had taken the two extreme lines shown at diagram Fig. 19, a and b were the points of maximum pressure. In the paper he had not detailed the method used to calculate the temperature attained at the point of maximum pressure ; it was necessary to do so before proceeding fur- ther. First, he determined the exact volume of the space at the end of the cy- linder into which the mixture was com- pressed, then on the diagram he had drawn the adiabatic line of compression, it was the dotted line shown at Fig. 6 ; the lower black line was the actual com- pression line drawn by the indicator. It would be seen that the two were as near- ly as possible coincident. The cause of this had been pointed out. The temper- ature at the point c was known to be 142 Fig.19. 3347 ^3311 0T^ 2110 l> 1537 Engine speed 150 revolutions per minute. One division of circle =one-fiftietli part of a second at above speed. 143 150. 5 Centigrade, and the pressure 41 Ibs. above atmosphere, and assuming the volume to remain constant, the tempera- ture at a was calculated from the press- ure 220 Ibs. above atmosphere. Let P pressure befofe ignition, and P' pressure after ignition, T = tempera- ture before ignition, and T' temperature after ignition, then T , P'T both pressures and temperature absolute. In diagram Fig. 1 it was shown that the temperature of compression, correspond- ing to 40 Ibs. above the atmosphere, was 150. 5 Centigrade, and from these figures the temperature 1,537 was obtained. This was the minimum possible tempera- ture, as would be observed from certain considerations developed at p. 21. Whether the flame had spread through- out the mass of the mixture or not, this was the average temperature. From a, Fig. 19, was drawn an isothermal line, a b c, dotted ; at the point a the tempera- ture had commenced to fall, up to that 144 point it had been rising at a very rapid rate. The semicircle drawn below the atmospheric line showed the path of the crank-pin, and each division represented in time one-fiftieth of a second ; the en- gine was running at one hundred and fifty revolutions per minute when the diagram was taken. Comparing the condition of the gaseous mixture in one-fiftieth of a second before maximum pressure, and one-fiftieth of a second after maximum pressure, in the first one-fiftieth of a second the average temperature had in- creased 905 Centigrade, while in the second hundredth it had diminished about 189 Centigrade. Within a limit of one twenty-fifth of a second there was a point where the increase of temperature ceased, and where a fall of temperature began. What did this mean ? Why did the in- crease of temperature cease in so sudden a manner and a fall of temperature set in? From the point d to a the temperature had been increasing, this increase being due to the progress of the flame ; at the 145 point a the increase ceased, and a fall set in. Take the point 6, then the average temperature was 632 Centigrade ; from e to a the time taken one-fiftieth of a second, and the temperature rose to 1,537 Centigrade ; in that time it had increased by 905 ; suppose the same rate of increase to continue for another one-fiftieth of a second, the pressure would rise to the point /*, and the tem- perature would be 2,442 Centigrade, the points g and h showed the effect of fur- ther increase. But the increase had abruptly ceased at the point a ; from a to f the volume had changed so slightly that the rate of cooling could not have in- creased appreciably. The amount of ' work done in that movement was also relatively insignificant, and yet from some cause the increase of temperature going on with such rapidity, 905 in one-fiftieth of a second, had not only diminished, but an opposite effect had set in. It could not be supposed for a moment that the progress of the flame had been abruptly stopped by any cause other than com- 146 pleted inflammation of the whole mass. The flame which in one instant of time had been flashing through the explosion mixture had reached the enclosing walls, it had uniformly heated the whole com- bustible mass, and in the next instant the temperature began to fall ; the law of cooling took effect. The very rapid rate of rise, and the abrupt change from rapid rise to slow fall of temperature, at a given point, showed that at that point completed inflammation had been attained. The cooling which was so slow as to be unable to put an appreciable check on the rate of rise up to the point of maximum temper- ature, could not be supposed to suddenly increase to such an enormous extent as to completely absorb and overpower at that instant the effect of continual spread of flame. There could be no doubt that, as Sir Thomson had pointed out, on diagram Fig. 6, the maximum of the curve indicat- ed truly the instant when the combus- tion was as complete as dissociation allowed it to be. It was certain that at this point of the diagram the flame had 147 spread completely through the whole volume of inflammable mixture, and that in whatever way the sustaining of the pressure to nearly the adiabatic line was to be explained, it could not be accounted for on the hypothesis of a continued s read of flame. A little consideration of the conditions of the indicated diagram would show that the slower the rate of inflammation, rel- atively to the movement of the piston, the less distinct would the point of maxi- mum pressure become, and the more rounded would the apex of the diagram appear. Nevertheless the point of com- pleted inflammation was easily, deter- mined from the point of maximum tem- perature, when near the end of the stroke this point might not be the point of maxi- mum pressure. He had been careful to make this distinction, and had said, with reference to slow inflammation, p. 25: " This supposed phenomena has been erroneously called slow combustion ; if it has any existence it should be called slow inflammation. It has a real existence in 148 the Otto engine only when it is working badly ; but even then maximum tempera- ture is attained, and very distinctly marks the point of completed inflammation." On diagram Fig. 19 was shown the effect of increasing the speed of the engine while preserving a constant rate of in- flammation. If the speed were increased from one hundred and fifty revolutions per minute three times, or to four hun- dred and fifty revolutions per minute, it would be found that the point a would be moved forward to k and b to I. In both cases the temperature attained would be nearly 1,537 Centigrade, a slight fall would be observed due to in- creased cooling surface and to a part of the work being done before maximum temperature was attained. But in all cases the maximum temperature marked the point of completed inflammation and the temperature began to fall so soon as it was attained. For ignitions attaining their maximum very late in the stroke, maximum pressure need not coincide with maximum temperatures ; but a reference 149 to the isothermal line showed the point of highest temperature. Using an in- flammable mixture of constant composi- tion, and varying the speed of the engine, it was always found that ignitions at- tained maximum temperature later and later in the stroke always came very near the isothermal line drawn from the point of highest pressure at the beginning of the stroke. The lines never ran over this isothermal. This meant that, whether inflammation was completed early or late in the stroke, nearly the same maximum temperature was attained. It followed from the relations between isothermal and adiabatic lines, that the lines drawn by the indicator from late ignitions always crossed those from early ignitions. This was shown by the diagrams taken from an Otto engine by Mr. Bousfield, for which he must thank that gentleman. In these diagrams, however, it was evident that the mixture used had not been of constant composition at all speeds. This would be evident by examining Fig. 15. When the speed had been changed from 150 one hundred revolutions per minute in the ]arger diagram to two hundred in the smaller, the increased speed of the engine had caused it to take in a smaller weight of gaseous mixture, as was shown by the compression line leaving the atmospheric line later, and that the pressure on com- pletion of the in stroke only rose to 22 Ibs. per square inch instead of 30 Ibs., as in the other. If the mixture had been the same the point of maximum pressure would have crossed in the first diagram at this point, and the pressure line would have run into the first lower down, as was shown in his diagram at #, Fig. 19. In the Otto engine the hot exhaust re- maining in the space when each cycle was completed still further complicated the comparison between different speeds. At the higher speeds the walls of the cylinder had less time to cool the exhaust, and consequently the average temperature of the mixture before compression must be greater at high speeds. In his own gas engine this complication had no existence, because the whole charge was replaced at 151 every stroke. In Mr. Bousefield's dia- gram, Fig. 16, the same change of mix- ture was evident, but here the change of speed of the engine was relatively greater, and consequently the lower diagram crossed the upper one somewhat earlier. In Fig. 17 this was more and more evi- dent ; still no two of the compression lines coincided, showing the proportion of exhaust to inflammable mixture to be continually increasing, and the maximum temperature attainable by the ignition consequently becoming less and less. Even in diagram, Fig. 18, maximum tem- perature was attained, and could easily be discovered by calculating the average temperature at each point along the line of increasing volume. Mr. Bousfield stated that a light applied to the exhaust of an engine, giving diagram, Fig. 16,. caused explosion, and from that inferred that combustion was not completed at the end of the stroke. He would find that when this happened the engine was missing ignition altogether and discharg- ing the unburned contents into the ex- 152 liaust. He might observe that the hor- izontal line in that diagram did not mean constant temperature, but indicated constantly increasing temperature. Mr. Bousfield has evidently fallen into the same error as Mr. Imray. and confound- ed low pressure with low temperature without considering the change of vol- ume. It was a characteristic of the in- flammation of a gaseous mixture in mass, that so long as inflammation continued to spread, so long did the average tem- perature increase. Dissociation did not begin to sustain temperature until the temperature fell. In the construction of the theoretical diagram Mr. Bousfield had fallen into error. He drew from the points F G H, Fig. 14, to A L produced, lines which he described as adiabatics, .and then said that the curve drawn through P Q R "represented the press- ure at any time in the contents of the cylinder, supposing these contents remain confined in the space at the end of the cylinder, and not allowed to expand." the lines F G H should not be 153 adiabatics but isothermals, as Mr. Bous- field's object in constructing the diagram was to get the time taken in a closed space to attain the temperature existing in the engine at the points F G H. The points L M N should show the pressure at constant volume at these temperatures. If Mr. Bousfield calcu- lated the temperature from an actual diagram, he would find that maximum temperature coincided with maximum pressure when at the beginning of the stroke. He thought from his remaining criticisms that Mr. Bousfield had not un- derstood the nature of the proof advanced in the paper, and that when he had studied the subject and appreciated the nature of the considerations advanced, he would admit the truth of the theory set forth in the paper. It had been asked by Dr. Hopkinson whether the pressure rose higher when an engine was running slowly than when it was running fast ? Whether the press- ure attained on exploding a gaseous mix- ture in a closed space and in an engine 154 was the same ? Given the same propor- tion of gas to air and the same tempera- ture and pressure of mixture before igni- tion, then the pressure attained after ig- nition was the same in all stages where the maximum pressure was attained at the beginning of the stroke ; it was the same whether in a closed space or in an engine. But the ignition must be rapid enough at the higher rate of speed to give maximum pressure at the beginning of the stroke. As he had already pointed out, if an engine was to run fast enough it might overrun the rate of inflammation, and the maximum temperature would not be attained till towards the end of the stroke. If an engine was run at two hundred revolutions per minute and max- imum pressure was attained at* the begin- ing of the stroke, then however slowly that engine ran using the same mixture, the maximum pressure would always be the same, it would not increase. Dr. Hopkinson then asked, Was the maxi- mum pressure the same in large and in small engines.? When using a similar 155 mixture, the same pressure and tempera- ture before ignition, it was the same. In small engines the temperature fell more rapidly than in large ones because of the greater proportion of cooling surface to volume of gases, but the maximum press- ure attained was nevertheless the same because of the rapid rate of ignition. The results obtained in the large cylinder to which he had alluded, and those obtained by Professor Bunsen in a small tube, each showing a limit to the rise of tem- perature which could not be referred to cooling, and each showing complete spread of flame, proved that the maximum pressure to be obtained from an explosive mixture was independent of the dimen- sions of the vessel used. Dr. Hopkinson had asked why, in comparing types 2 and 3 of engine, he used different maximum pressures ; why in the second type he used 76 Ibs. per square inch above the atmosphere, and in the third over 200 Ibs. per square inch. His reason was this : the three types were taken under condi- tions which have been found in practice 156 to be the most favorable for each. He had compared the theory of these types of engine as nearly as possible under con- ditions used in practice, It was quite true that type 2 should be compared with type 3 under similar conditions of press- ure from a purely theoretic standpoint ; but the object of the paper had been to inquire into the cause of the greater effi- ciency of the third type as in use against the two first also in use. It would be seen that to attain a pressure of 200 Ibs. per square inch in type 2 it was necessary to compress the mixture to that pressure before ignition, the temperature of com- pression being nearly o65 Centigrade. This involved considerable loss of heat in the reservoir, and increased the chances of leakage while compressing ; in type 3 a pressure of 40 Ibs. per square inch be- fore ignition was all that was required to attain 200 Ibs. after ignition. He believed that type 2 could work advantageously at a much higher pressure than 76 Ibs. per square inch, but he questions whether it could do so at so high a pressure as 200 157 Ibs. The advantage of type 3 in this respect was a comparatively low pressure before ignition. With careful workman- ship doubtless it would be possible to use an engine of type 2, the theoretical effi- ciency of which would be quite as much as type 3, as given in the paper. The description by Mr. F. H. Wenham of his work on hot-air engines was inter- esting, and his distinction of the cylinder itself as the heat generator or furnace was the essential one between gas and hot-aii* engines, and was indeed the great cause of success in these engines. Mr. H. Davey had objected to his com- parison of the efficiency of gas and steam engines, and considered the basis of com- parison of efficiency used by him as an unfair one. In comparing engines of the same system it was right, as Mr. Davey stated, to use as the standard the mechan- ical equivalent of the total available heat ; but in engines of totally different nature the only basis of comparison was the number of heat-units given to the engine, and the number of these heat-units con- 158 verted into mechanical work. If one sys- tem was necessarily limited in range of temperature, as the steam engine was, then the inquiry must not be how near it approached perfection within that range, but how much heat could another sys- tem convert into work as compared with it. In comparing steam engines with steam engines Mr. Davey is perfectly right ; in comparing with gas engines the .general basis must be taken. He agreed that the speedy downfall of the steam engine was not to be anticipated ; he only held that the gas engine was now in its infancy, that it contained greater possibilities than the steam engine, and that in the future it was c^-tain to be in every way a great advance on the steam engine, and likely to supersede it. The propriety of treating the gas en- gine as an air engine had been called in question, and he had been asked whether the specific heats of air and the gaseous mixture used were in any way comparable. The specific heat of air at constant vol- ume was 0.169, and the specific heat of a 159 mixture of 1 volume of coal gas and 12 volumes of air could not exceed 0.200, so that for the purpose of approximate comparison their adiabatic curves might be considered as nearly identical. So little was known of the specific heat of gases at high temperature that Mr. Clerk considered it simply an affectation of ac- curacy to endeavor to make the com- parison closer. He was aware that the efficiency of a heat engine was independ- ent of the nature of the fluid employed, provided the temperatures between wliich the engines worked were the same that was provided there was the same differ- ence between source and refrigerator. But this was just where the steam engine failed. Given equal amounts of heat from the same source, in the steam engine the high temperatures could not be utilized, because, first, a certain quantity of heat had to be expended to change the phys- ical state of the water ; and as the steam produced was rejected as steam all the heat so expended was lost for the pur- pose of procuring high temperature. 160 * With air, on the other hand, the same quantity of heat from the same source, a much higher temperature was attained, and consequently a greater range of tem- perature due to work performed. The use of steam necessitated a limited range of temperature, and the discharge of all the heat used in converting water from a liquid to a gas. It had been argued that in engine type 2 he had over-estimated the efficiency, and made it greater than was possible from a perfect heat engine working between the limits of tempera- ture used. Mr. Bamber had fallen into error by mistaking the limits, and in this he was not alone. This type of engine presented very interesting peculiarities in theory, which, so far as he was aware, had hitherto been missed by writers on thermo-dynamics. Although 1,537 Centi- grade was the maximum temperature, and 1,089 Centigrade the temperature of discharge with the exhaust, yet these temperatures were not the limits within which the engine was working ; the re- frigerator, which was at atmosphere 161 temperature 17 Centigrade, was being used to a certain extent without being apparent. The diagram was not a simple one ; the efficiency 0.36 was the result of the united action within two different limits. The diagram from 1,537 Centigrade to 1,089 Centigrade was the same both in types 1 and 2, and working between these limits the maximum possible- efficiency was 0.247 ; but in type 1 this efficiency was not attained, because at 1,089 Centigrade the air had not the same density as be- fore expansion, and some work had been expended in changing the volume to twice its original amount. If before heating the air had been compressed slightly, then heated to 1,537 and expanded to its original volume, and lowered in tem- perature due to work done to 1,089 Q , the duty would be 0.247. If in type 1 a con- denser were used, and the temperature reduced to 17 Centigrade, the additional work obtained would raise its duty to 0.247, without this it remained at 0.21. In both types the efficiency between the 162 Kmits 1,537 Centigrade and 1,089 Centi- grade was the same ; but in type 2 a con- siderable amount of work was obtained in the earlier part of the diagram, a cer- tain amount of work was done on in- creasing temperature from 217. 5 Centi- grade to 1,537, and a considerable pro- portion of heat could be converted into work on an increasing temperature, still T T conforming to the law *= - as the maximum possible between the limits. In type 2, to a certain extent, the re- frigerator at atmosphere temperature was made available in a portion of the action, and consequently a portion of work done on increasing temperature, while the latter half of the stroke was accomplished on falling temperature. This was the reason why a greater efficiency was got than the apparent limits would allow. Mr. Bamber then argued that it made no difference whether it was necessary to use an air pump or not, if only the same quantity of heat were consumed and the same theoretic efficiency obtained. In 163 practice it made all the difference ; the great cause of failure with hot-air engines was not imperfect theory but very low available pressures combined with high maximum pressures. Nearly all the power indicated was used up in friction ; in the earlier gas engines the average pressures were very low also. The advantages of compression were a high available pressure, small cooling surfaces, and small loss by friction. There the efficiencies depended on the range of source and refrigeration ; but compression allowed all this to be at- tained under practical conditions. It was hardly necessary to explain that there was a certain maximum efficiency for heat engines. What he had shown in this paper was that a greater proportion of this was possible under working con- ditions with compression than without. The parallel by Air. Cowper between slow inflammation and imperfect admis- sion of steam in a cylinder was very just, and illustrates the great loss of power and heat involved by imperfect mixing 1G4 of gas and air, or by failing to attain maximum pressure as soon after firing as practicable. It was only by a constant application of theory to practice, and a constant testing of results obtained by varying conditions, that he had been able to produce the diagram which Mr. Cowper approved. The amount of gas consumed by his G-HP. engine was 22 cubic feet per 1 HP. per hour. Of course in cost this did not stand comparison with the coal used by a large modern steam engine; the steam engine had greatly the advantage ; but compared with a small ' steam engine it was economical. When gas was manufactured expressly for gas engines it need cost but little more than the coal used to produce it, and as the gas need not be illuminat- ing all the carbon might be converted into gas. The gas might be in fact a mixture of carbonic oxide and hydro- gen. [WVEBSITY) *** Any book in this Catalogue sent free by mail, on receipt of price. VALUABLE SCIENTIFIC BOOKS, PUBLISHED BY D, VAN NOSTRAND, 23 Murray Street, and 27 Warren Street, NEW YORK. WEISBACH. A MANUAL OF THEORETICAL ME- CHANICS. By Julius Weisbach, Ph. D. Trans- lated by Eckley B. Coxe. A.M., M.E., 1,100 pages and 902 wood-cut illustrations. 8vo, cloth $10.00 FRANCIS. LOWELL HYDRAULIC EXPERIMENTS being a Selection from Experiments on Hydraulic Motors, on the Flow of Water over Weirs, and in open Canals of Uniform, Rectangular Section, made at Lowell, Mass. By J. B. Francis, Civil Engineer. Third edition, revised and enlarged, with 23 copper - plates, beautifully engraved, and about 100 new pages of text. 4to. cloth 15.00 KIRK WOOD. ON THE FILTRATION OF RIVER WATERS, for the Supply of Cities, as prac- tised in Europe. By James P. Kirkwood. Illustrated by 30 double-plate engravings. 4to. cloth 15.00 1 I>. VAN NOSTRAND'S PUBLICATIONS. BANNING. A PRACTICAL TREATISE OF WATER SUPPLY ENGINEERING. Relating to the Hydrology, Hydrodynamics, and Prac- tical Construction of Water- Works, in North America, With numerous Tables and 180 illustrations. By T. T. Fanning, C.E. 650 pages. 8vo, cloth extra, , $6 08 WHIPPLE. AN ELEMENTARY TREATISE ON BRIDGE BUILDING. By 8. Whipple, C. E. New Edition Illustrated. 8vo, cloth, 4 Oi MERRILL. IRON TRUSS BRIDGES FOR RAIL- ROADS. The Method of Calculating Strains in Trusses, with fi careful comparison if the most prominent Trusses, in reference to economy in cotnoinatiou, etc., etc. By Bvt. Col. William E. Merrill, U. 8. A., Corps of Engineers. Nine lithographed plates of illustrations. Third edition. 4to, cloth, a .... 5 Of SHREVE. A TREATISE ON THE STRENGTH OP BRIDGES AND ROOFS. Comprising the de- termination of Algebraic formulas ^or Strains ir Horizontal, Inclined or Rafter, Triangular, Bowstring Lenticular and other Trusses, from fixed and moving loads, with practical applications and ex- amples, for the use of Students and En- gineers By Samuel H. Slireve. A. M. f Civil Engineer. Second edition, 87 wood- cut illustrations. 8vo, cloth, 3 5C KANSAS CITY BRIDGE. WITH AN ACCOUNT OP THE REGIMEN OF THE MISSOURI RIVER, and a description of the Methods used for Founding in that River. By O. Chanute, Chief Engineer, and George Morisoii, As- sistant Engineer. Illustrated w\la five lithographic views and twelve plates of plans. 4to,clotb, . . D. VAN NOSTRAND'S PUBLICATIONS. CLARKE. DESCRIPTION OF THE IRON RAILWAY BRIDGE Across the Mississippi River at Quincy, Illinois. By Thomas Curtis Clarke, Chief Engineer, With twenty-one litho- graphed Plans. 4to, cloth, . . . $7 50 ROEBLING. LON.I AND SHORT SPAN RAILWAY BRIDGES. By John A. Roebling, C. E. With large copperplate engravings of plans and views. Imperial folio, cloth, . 25 00 DUBOIS. THE NEW METHOD OF GRAPHICAL STATICS. By A. J. Dubois, C. E., Ph. D. _ 60 illustrations. 8VO, cloth, . . . . 1 50 McELROY. PAPERS ON HYDRAULIC ENGINEER- ING. The Hernpstead Storage Reservoir of Brooklyn, its Engineering Theory and Results, By Samuel McElroy, C, E. 8vo, paper, . 5O BOW. A TREATISE ON BRACING with its ap- plication to Bridges and other Structures of Wood or Iron. By Robert Henry Bow, C. E. 156 illustrations on stone. 8vo, cloth, 1 50 STONEY. THE THEORY OF STRAINS IN GIRDERS and Similar Structures with Observa- tions on the Application of Theory to Practice, and Tables of Strength and other Properties of Materials. By Bindon B. Stoney, B. A. New and Revised Edition, with numerous illustrations. Royal 8vo, 664 pp., cloth, 12 50 HENRICI. SKELETON STRUCTURES, especially in their Application to the building of Steel and Iron Bridges. By Olaus Henrici. 8vo, cloth, 1 5O KING. LESSONS AND PRACTICAL NOTES ON STEAM. The Steam Engine, Propellers, &c., &c., for Young Engineers. By the late W. R. King, U. 8. N., revised by Chief- i Engineer J. W. King, U. S. Navy. 19th edition. 8vo, cloth, . . , .200 i>. VAN M>RTRAND'S PUBLICATIONS. AUCHINCLOSS. APPLICATION OF THK SLIDB VALVE and Link Motion to Stationary, Portable. Locomotive and Marine Engines. By William 8 Auchincloss. Designed as a band book for Mechanical Engineers. With 37 wood-cuts and 21 lithographic plates, with copper-plate engraving of the Travel Scale. Sixth edition. 8vo, cloth. $3 00 BURGH. MODERN MARINE ENGINEERING, ap- plied to Paddle and Screw Propulsion. Consisting of 3f> Colored Plates, 259 Practi- cal W T ood-cut Illustrations, and 403 pages of Descriptive Matter, the whole being an ex- position ot the present practice of the fol- lowing tinus: Messrs J. Penn & Sons; Messrs. Maudslay, Sons ^IANI- CAL THEORY OF HEAT, AND ITS APPLICA- TIONS TO THE STEAM-ENGINE. By Prof. R. S. McCulloch, of the Washington and Lee University, Lexington, Va. 8vo, cloth 3 50 VAN BUREN. INVESTIGATIONS OF FORMU- LAS for the Strength of the Iron parts of Steam Machinery By J. D. Van Buren, Jr.. C. E. Illustrated. 8vo, cloth, . 2 00 STUART. How TO BECOME A SUCCESSFUL EN- GINEER. Being Hints to Youths intending to adopt the Profession. By Bernard Stuart, Engineer Sixth edition 18rno, boards, ..... 50 SHIELDS. NOTES ON ENGINEERING CONSTRUC- TION. Embracing Discussions of the Prin- ciples involved, and Descriptions of the Material employed in Tunneling, Bridging, Canal and Road Building, etc., etc. By J. E. Shields, C. E. 12mo. cloth, . . . 1 5O D. VAN NOSTKAXD'S IMPLICATIONS. WEYRAUCH. STRENGTH AND CALCULATION OP DIMENSIONS OK IRON AND STEEL CON- STRUCTIONS. Translated from the German of J. J. Weyrauch, Ph. D., with four fold- ing Plates. 12ino, cloth, . . . $1 00 STUART. THE NAVAL DRY DOCKS OF THE UNITED STATES. By Charles B. Stuart, Engineer in Chief, U. S. Navy. Twenty- four engravings on steel. Fourth edition. 4to, cloth, . . . . . 6 00 COLLINS. THE PRIVATE BOOK OF USEFUL AL- LOYS, and Memoranda for Goldsmiths, Jewellers, etc. By James E. Collins. 18mo| flexible cloth, 50 TUNNER. A TREATISE ON ROLL-TURNING FOR THE MANUFACTURE OF IRON. By Peter Tun- ner * Translated by John B. Pearse. With numerous wood-cuts, 8vo, and a folio Atlas of 10 lithographed plates of Rolls, Measurements, &c. Cloth, . . 10 00 GRUNER. THE MANUFACTURE OF STEEL. By M. L. Gruner. Translated from the French, by Lenox Sm'ith, A.M., E.M. ; with an Appendix on the Bessemer Pro- cess in the United States, by the transla- tor. Illustrated by lithographed drawings and wood-cuts. 8vo, cloth, . . . . 3 50 BARBA. THE USE OF STEEL IN CONSTRUCTION. Methods of Working, A pplviug, and Test- ing Plates and Bars. By J. Barba. Trans- lated from the French, with a Preface by A. L. Holley, P.B. Illustrated. I2mo, cloth, 1 50 SHOCK. STEAM BOILERS; THEIR DESIGN, CON- STRUCTION, AND MANAGLMEMT. ^y William A. Shock, Engineer-in- Chief. U.S.N.; Chief of Bureau of Steam Engineering, U.S. N". 480 pages. Illustrated with 150 woodcuts and 36 f'lJl-paere plates (20 double). 4to, illustrated, half morocco 15 00 6 I). VAX NOSTRAND'S PUBLICATIONS. WARD. STEAM FOR THE MILLION. A Popular Treaties on Steam and its Application to the Useful Arts, especially to Navigation. By J. H. Ward, Commander U. S. Navy. 8vo, cloth. . . $1 00 CLARK. A MANUAL OF RULES, TABLES AND DATA FOR MECHANICAL ENGINEERS. Based on the most recent investigations. By Thin. Kinnear Clark. Illustrated with numerous diagrams. 1012 pages. 8vo. Cloth ,$7 50; half morocco, . . . . 10 00 JOYNSON. THE METALS USED IN CONSTRUC- TION : Iron, Steel, Bessemer Metals, etc, By F. H. Joynson. Illustrated. 12mo, cloth, 75 DODD. DICTIONARY OF MANUFACTURES, MIN- ING, MACHINERY, AND THE INDUSTRIAL ARTS. By George Dodd. I2mo, cloth, 1 50 PRESGOTT. FIRST BOOK IN QUALITATIVE CHE- MISTRY. By Albert B. Pi escotfc, Professor of Organic and Applied Chemistry in the University of Michigan. 12mo, cloth, . . 1 50 PLATTNER. MANUAL OF QUALITATIVE AND QUANTITATIVE ANALYSIS WITH THE BLOW- PIPE. From the last German edition. Re- v i sed and enlarged. By Prof. Th. Richter, o the Royal Saxon Mining Academy. Translated by Professor H. B.Cornwall. With eighty-seven wood-cuts and lithogra- phic plate. Third edition, revised. 568pp. HVO, cloth, 5 00 PLYMPTON. THE BLOW-PIPE : A Guide to its Use in the Determination of Salts and Minerals. Compiled from various sources, by George W. Plyrupton, C. E., A. M., Pro- fessor of Physical Science in the Polytech- nic Institute, Brooklyn, N. Y. 12mo, c7oth 1 50 7 D. VAN NOSTRAND'S PUBLICATIONS. JANNETTAZ. A GUIDE TO THE DETERMINATION OF BOCKS ; being ail Introduction to Lith- olqgy. By Edward Jaunettaz, Docteur des Sciences. Translated from the French by G. W. Plymptcm, Professor of Physical Science at Brooklyn Polytechnic Institute. 12mo, cloth, . . . . . . . $1 50 MOTT. A PRACTICAL TREATISE ON CHEMISTRY (Qualitative and Quantitative Analysis), Stoichioiuetry, Blowpipe Analysis, Min- eralogy, Assaying, Pharmaceutical Prepa- rations Human Secretions, Specific Gravi- ties, Weights and Measures, etc., etc., etc. By Henry A. Mott, Jr., E. M., Ph. D. 650 pp. 8vo, cloth ; '6 00 PYNCHON. INTRODUCTION TO CHEMICAL PHY- SICS ; Designed for the Use of Academies, Colleges, and High Schools. Illustrated with numerous engravings, and containing copious experiments, with directions for preparing them. By Thomas Buggies Pyn- chon, D. D., M. A., President of Trinity Col- lege, Hartford. New edition, revised and enlarged. Crown vo, cloth, . . 3 00 PRESCOTT. CHEMICAL EXAMINATION OF ALCO- HOLIC LIQUORS. A Manual of the Constit- uents of the Distilled Spirits and Ferment- ed Liquors ot Commerce, and their Quali- tative and Quantitative Determinations. By Alb. B. Prescott, Prof, of Chemistry, University of Michigan, I2rno, cloth, . 1 60 ELIOT AND STORER A COMPENDIOUS MANUAL OF QUALITATIVE CHEMICAL ANALYSIS. By Charles W. Eliot and Frank H. Storer. Be- vised, with the co-operation of the Authors* by William Ripley Nichols, Professor of Chemistry in the Massachusetts Institute of Technology. New edition, revised Il- lustrated. I2mo, cloth, . . . 1 50 D. VAX NOSTRAND'S PUBLICATIONS. NAG.UET. LEGAL CHEMISTRY. A Guide to the Detection of Poisons, Falsification of Writ- ings, Adulteration of Alimentary and Phar- maceutical Substances ; Analysis of Ashes, and Examination of Hair, Coins, Fire-arms and Stains, as Applied to Chemical Juris- prudence. For the Use of Chemists, Phy- sicians, Lawyers, Pharmacists, and Ex- perts. Translated, with additions, includ- ing a List of Books and Memoirs on Toxi- cology, etc., from the French of A. Naquet, by J. P. Battershall, Ph. D. ; with a Preface by C. F. Chandler, Ph. D., M. D., LL. D. Illustrated. 12nio, cloth, . . . . $2 00 PRESCOTT. OUTLINES OF PROXIMATE ORGANIC ANALYSIS for the Identification, Separa- tion, and Quantitative Determination of the more commonly occurring Organic Compounds. By Albert B. Prescott, Pro- fessor of Chemistry, University of Michi- gan. 12mo, cloth, . . . 1 75 30UGLAS AND PRESCOTT. QUALITATIVE CHEM- ICAL ANALYSIS. A Guide in the Practical Study of Chemistry, and in the work of Analysis. By S. H. Douglas and A. B. Prescott; Professors in the University of Michigan. Third edition, revised. 8vo, cloth, 3 50 RAMMELSBERG. GUIDE TO A COURSE OF QUANTITATIVE CHEMICAL ANALYSIS, ESPE- CIALLY OF MINERALS AND FURNACE PRO- DUCTS. Illustrated by Examples. By C. F. Rammefsberg. Translated by J. Tow ler, M. D. 8vo, cloth 225 BEILSTEIN. AN INTRODUCTION TO QUALITATIVE CHEMICAL ANALYSIS. By F. Beilstein. Third edition. Translated by I. J. Osbun. I2mo. cloth, 75 POPE. A Hand-book for Electricians and Oper- ators. By Frank L. Pope. Ninth edition. Revised and enlarged, and fully illustrat- ed. 8vo, cloth, 200 D. VAN NOSTRAND'S PUBLICATIONS. SABINE. HISTORY. AND PROGRESS OF THE ELEC- TRIC TELEGRAPH, with Descriptions of some of the Apparatus. By Robert Sabine, C. E. Second edition. 12nu), cloth, . . $1 25 DAVIS AND RAE. HAND BOOK OF ELECTRICAL DIAGRAMS AND CONNECTIONS. By Charles H. Davis and Frank B. Rae. Illustrated with 32 full-page illustrations. Second edi- tion. Oblong 8vo, cloth extra, . . . 2 00 HASKINS. THE GALVANOMETER, AND ITS USES. A Manual for Electricians and Students. By C. H. Haskins. Illustrated. Pocket form, morocco, . . . . . 150 LARRABEE. CIPHER AND SECRET LETTER AND TELEGRAPAIC CODE, with Hogg's Improve- ments. By C. S. Larrabee. 18mo, flexi- ble cloth, 1 00 GILLMORE PRACTICAL TREATISE ON LIMES, HYDRAULIC CEMENT, AND MORTARS. By Q. A. Gillmore, Lt.-Col. U. S. Engineers, Brevet Major-General U. S. Army. Fifth edition, revised and enlarged. 8vo, cloth, 4 00 GILLMORE. COIGNET BETON AND OTHER ARTIFI- CIAL STONE. By Q. A. Gillmore, Lt. CoL U. 8. Engineers, Brevet Major-General U. S. Army. Nine plates, views, etc. 8vo, cloth, 2 50 GILLMORE. A PRACTICAL TREATISE ON THE CONSTRUCTION OF ROADS; STREETS, AND PAVEMENTS. By Q. A. Gillmore, Lt.-Col. U. S. Engineers, Brevet Major-General U, S. Army. Seventy illustrations. 12mo, clo., 200 GILLMORE. REPORT ON STRENGTH OF THE BUILD- ING STONES IN THE UNITED STATES, etc. 8vo, cloth, 1 00 HOLLEY. AMERICAN AND EUROPEAN RAILWAY PRACTICE, in the Economical Generation C' of Steam. By Alexander L. Holley. B. P. With 77 lithographed plates. Folio, cloth, 12 0