compression pressure will probably be substantially retained. The low mean effective pressure in reducing heat stresses and temperature conditions within the cylinder makes for reliability in operation.

The vexed question of "solid" injection of the fuel becomes a simpler issue when associated with surface ignition. All considerations in the design of the fuel pump operating and controlling gear, and the injection means for semi-Diesel engines, have in the past been subservient to that of simplicity. With a demand for the same degree of flexibility, and a capacity to burn as wide a range of fuels, without recourse to the water drip, as obtains with the Diesel engine, considerable improvements after the war can confidently be anticipated.

Appendix I-Mechanical Efficiency of Internal-Combustion Engines.The following notes have special reference to internal-combustion engines of the trunk piston type; but apply equally, with slight modifications, for crosshead engines.

Mechanical efficiency (the ratio between brake horsepower and indicated horsepower) is affected by the number of auxiliaries which are driven by the main engine.

1. Except in so far as auxiliaries are concerned, the difference between the indicated horsepower and the brake horsepower can be apportioned as follows:

(a) 50 per cent is due to piston and piston-ring friction.

(b) 28 per cent can be attributed to main cylinder pumping losses, suction, exhaust and scavenging.

(c) 22 per cent is allocated to valve gear and bearing friction, etc., in which are included windage losses and other factors of little importance. 2. Piston friction depends primarily on the following factors:

(a) The quality of the metal of the liner, the piston and the piston rings. (b) The quality of the lubrication. (Certain tests which have been carried out go to prove that a diminution in viscosity of oil increases the mechanical efficiency. In one case the mechanical efficiency was increased by water injection into the combustion chamber.)

(c) The clearance between the piston and the cylinder walls, has an influence on efficiency.

(d) The m. e. p. the compression pressure, and the pressure between the liner and the piston, and the liner and the piston rings, can probably have a most suitable value for the reduction of friction loss to a minimum. (e) The fit and the condition of the piston rings.

(f) The temperature at which the engine runs will have an effect on the lubrication and on the clearance; and it has been substantially proved that there is a temperature of maximum mechanical efficiency.

3. The suction loss, 28 per cent of the total, is primarily a function of design of ports, valve setting, piston speed and gas speeds.

4. The valve gear and the bearings, 22 per cent of the total loss, will depend on the design of the engine, the alignment, the efficiency of the lubrication.

In addition to the foregoing, there are records of mechanical efficiency being reduced by increased weight of flywheel.

Generally, mechanical efficiency is adversely affected by increased speed and reduced m. e. p.; and decreased by mal-alignment, etc.

The mechanical efficiency may be affected by the form of the combustion chamber which may produce undue distortion of the piston working conditions, although this is probably extremely slight; distortion of the piston being more due rather to the condition of the gudgeon pin bearing than to any other cause.

The mechanical efficiency, assuming a constant m. e. p., is practically unaffected by the size of the engine.

In connection with the above, a large number of records of tests of engines have been investigated from Guldner, Supino, D. Clerk, etc.

Appendix II.—Reversibility.-Reversibility need only be discussed in relation to two-cycle semi-Diesel engines, since those working on the four-cycle principle have not, as yet, been made directly reversible. The compression of the scavenging air in the crank chamber fitted with automatic valves, and the operations of scavenging and exhaust within the working cylinder are directly reversible (see Fig. 2).

The only problem remaining is that of the timing of the injection of the fuel for the reverse direction of rotation. If the end of the delivery stroke of the fuel pump coincides with the top dead center of the main piston, no alteration in timing for fuel injection is required for astern running. This condition is satisfied in most of the multi-cylinder reversible semi-Diesel engines. However, should the timing of the fuel pump be such that the end of the delivery stroke does not coincide with the top dead center, then operating gears are necessary for the fuel pumps-one for ahead and one for astern-in an exactly similar manner to the arrangements adopted for Diesel engine cylinder head fuel injection valves with reversible marine engines.

The method adopted for changing the engine from the ahead direction of rotation to astern, may be:

(a) By means of a pre-injection and ignition of fuel.

(b) Pre-injection of starting air.

(c) Stopping the engine, reversing the driving mechanism of the starting air valves and possibly of the fuel pumps as well, and starting up the engine in the astern direction.

With (a) and (b) the fuel pump is timed so that the point of delivery coincides approximately with the top dead center of the main piston, and arrangements are made, that when the fuel is cut off, the engine is declutched, the speed falls to a predetermined minimum and an injection of fuel or starting air is effected by the governor on the up stroke of the piston, so driving it in the reversed direction. Method (c) is only applicable to four-cylinder engines, since engines with a lesser number of cylinders cannot be started from any position of the cranks at which the engine may have stopped. Although a clutch is generally fitted, it is not necessary for method (c). For a full description of method (c), see Engineering for August 24, 1917.—Engineering, 25/10.

AERONAUTICS

THE LIBERTY MOTOR.-Its Checkered Career and Details of Its Construction. At last the restrictions of the censor have been lifted and we are able, without in any way giving aid or comfort to the enemy, to disclose to our readers pictures and drawings of that engine of mystery, the Liberty motor.

The career of this motor has been a checkered one. Announced first as a five-day creation, a masterpiece of ingenuity which could be turned out immediately in tens of thousands by typically American quantity production methods, it was loudly acclaimed as one of the greatest inventions of the war that would give America the mastery of the air. Then the pendulum swung to the opposite extreme. In the sharp reaction from this grossly exaggerated opinion of the Liberty motor, the machine was pronounced a pitiful failure, an immensely heavy brute with a voracious appetite for fuel, a mere automobile engine, absolutely unfitted to take on wings and soar among the clouds. It was now held up as a glaring example of inefficiency and incompetency. Ugly stories of graft were whispered about-of hundreds of millions of dollars absolutely thrown away. However, by the time the pendulum had reached that extreme the Liberty motor had passed through the long and tedious period of experimentation and preparation for quantity manufacture and it had undergone a thousand and one changes in minute details, all of which consumed a great deal of time, and when the public criticism had reached its highest tide this engine was already being turned out by at least one

of the large manufacturing concerns in fairly large quantities and the daily production was steadily increasing.

It was on Thanksgiving Day, a year ago, that the first Liberty motor, built on an organized production basis, was wrapped in an American flag and shipped from the Packard plant to the aircraft forces. It was not until the following March that the Packard Company had organized its factory, produced the necessary tools and jigs, and completed the preparation for the production of the motor in quantity. This was the first large plant to undertake the manufacture of the Liberty motor. Five other plants began operations later and it was not until well along in the summer that production was proceeding at full capacity.

On the 21st of November, the Packard plant shipped its five thousandth Liberty motor, and there is every prospect that its entire order for 6000 motors will be completed well before Christmas Day. The Lincoln Motor Company also has a contract for 6000 motors; while the Ford Motor Company's contract calls for the construction of 5000; the Nordyke & Marmon Company for 3000 the General Motors Corporation for 2000, and the Trego Motor Company for 500, a total of 22,500 Liberty motors. Despite all criticism of delay and inefficiency, our country had 15,000 Liberty engines when the armistice was signed, all of which had been produced within eight months.

In the Senate report on the aircraft situation, as well as in the report of the Hughes investigation, it was declared that the Liberty motor had proved a decided success. To be sure, this motor is not adapted to all types of airplanes. Somehow, the public gained the notion that the Liberty motor would serve every purpose as an airplane power plant. As we pointed out very clearly many months ago, the Liberty motor, although one of the lightest, if not the lightest motor per horsepower, is entirely too heavy a machine for a light battleplane. It will be clearly evident to anyone who stops to consider that power is not the only requisite in an airplane, and a machine which carries an engine weighing over 800 pounds cannot possibly make the quick turns that are accomplished by a machine equipped with a 200- to 300-pound engine. The momentum of the engine will carry it forward despite the action the airplane's rudder, and it takes time to swing it from one direction to another. The Liberty motor, therefore, was never intended to be put in a small airplane with a single operator who aims his gun by steering his machine toward his target, but it is adapted for the larger machines which are used for bombing purposes, for observation, for reconnaissance, and also for the larger battleplanes, in which there is a machine-gun operator, as well as a pilot.

As we have stated in previous issues of the Scientific American, the Liberty motor weighs about 825 pounds. Its horsepower was raised during the development period from 367 to 450, and in some tests it has run up as high as 480.

In our story of the Liberty motor published in the issue of June 1 last, we showed that this engine was not designed in five days and that it was not a radically new invention. It was the result of a year and a half of experiment on the part of the Packard Company prior to the war, and in the famous five-day conference it was this engine which was modified to meet requirements of the government and then built very hastily within a month so that it could be delivered in Washington on Independence Day as the "Liberty" motor.

The original Liberty motor was fitted with eight cylinders, but this design was not accepted because word came from France that a motor of much higher power was needed, consequently the 12-cylinder type was decided upon. One of our photographs shows an 8-cylinder Liberty motor being tested out at the summit of Pike's Peake, in order to determine its operation under the rarified atmospheric conditions prevailing at that elevated spot. The motor was also tested out at sea level.

The first 12-cylinder Liberty motor was tested by mounting it on a large truck, as shown in one of our photographs, and driving the truck by means of an airplane propeller. This truck was propelled at speeds of over 40 miles per hour through the snow by means of this form of drive. On one occasion the brakes were set to lock the wheels, and yet the truck was moved bodily forward, with its wheels sliding along the ground, by the powerful thrust of the propeller. In our illustration it will be noticed that there is a nick in one of the blades of the propeller. This was produced by a bolt which was accidentally jarred off the truck and fell against the propeller blade. Because of the enormous speed of the propeller, this light piece of metal caused it to tear out a large piece of wood.

Coming down to the more minute details of the Liberty motor, we may refer to the line drawing representing a sectional view of the machine.

It will be noticed that the engine is of the V-type, with cylinders disposed at an angle of 45 degrees to each other. In other airplane motors the angle is usually 60 degrees, but this sharper angle was adopted to reduce head resistance and also to strengthen the crank case and reduce vibration. The cylinders are of 5 by 7 bore and stroke and have a cubic capacity of 905 inches. The motor has individual cylinder barrels and stamped steel water jackets. This construction permits of machining the barrels all over so as to have uniformity of section and a maximum of strength with a minimum of weight. The stamped steel jackets insure uniform water space and absence of steam pockets. The jackets are welded to flanges formed on the cylinder, the cross sectional area of the flanges being the same as that of the jackets so that both members can be brought to an equal degree of temperature, thus facilitating the welding operation. The valves in the head are set at an angle to insure the best shape of combustion chamber and the maximum possible valve size. To shorten the

travel of the gases the inlet valves are on the inside, while the exhaust valves are on the outside and carry off the hot gases rapidly from the valves and motors. There is a single cam shaft for each set of cylinders which is mounted between the valve stems, providing a very simple and direct action. Double valve springs are used to minimize breakage from

CROSS-SECTIONAL VIEW OF THE LIBERTY MOTOR.

vibration. There are two duplex carburetors, each bore serving three cylinders so as to give the best distribution of gas and to permit of ready and accurate synchronization. Pressure lubrication is used on all plain bearings. The oil supplied to the connecting rod bearings and cylinders is controlled by metering conduits in the main and connecting rod bearings. The pistons are an aluminum alloy chosen for lightness and good heatconducting properties. There are two spark plugs for each cylinder so as to halve the possibility of losing a cylinder due to spark plug trouble and in order to increase the rapidity of the flame propagation in the cylinders.