The selection of aluminum for the material was an integral part of the basic design decision. Despite the excellence and accuracy of the castings that could be obtained, there was nevertheless a minimum dimension beyond which wall thickness could not be reduced; and the use of either one of the two other proven materials, cast iron or bronze, would have made the body, as they called it, prohibitively heavy. The use of aluminum was not entirely novel at this time, as it had been utilized in many automobile engine parts, particularly crankcases; but its incorporation in this rather uncommon combination represented a bold step. There was no choice in the matter of the alloy to be used, the only proven one available was an 8 percent copper 92 percent aluminum combination.

By means of the proper webs, brackets and bosses, the crankcase would also carry the crankshaft, the rocker arms and bearings, and the intake manifold. The open section of the case at the top was covered with a screw-fastened thin sheet of cold-rolled steel. The main bearing bosses were split at a 45° angle for ease of assembly. The engine support and fastening were provided by four feet, or lugs, cast integral on the bottom corners of the case, and by accompanying bolts (Figure 2). Although the crankcase continued to be pretty much the "body" of the internal combustion aircraft engine throughout its life, the Wrights managed to incorporate in this original part a major portion of the overall engine, and certainly far more than had ever previously been included.

The design of the cylinder barrel presented fairly simple problems involving not much more than those of keeping the sections as thin as possible and devising means of fastening it and of keeping the water jacket tight. They saved considerable weight by making the barrel quite short, so that in operation a large part of the piston extended below the bottom of it; but this could be accepted, as there were no rings below the piston pin (Figure 6). The barrel material, a good grade of cast iron, was an almost automatic choice. In connection with these seemingly predetermined decisions, however, it should be remembered that their goal was an engine which would work without long-time development, and that, with no previous experience in lightweight construction to guide them they were nevertheless compelled to meet a weight limit, so that the thickness of every wall and flange and the length of every thread was important.

With the separate cylinder barrel they were now almost committed to a three-piece cylinder. It would have been possible to combine the barrel and head in a one-piece casting and then devise a method of attachment, but this would have been more complex and certainly heavier. For housing the valves, what was in effect a separate cylindrical, or tubular, box was decided upon. This would lie across the top of the cylinder proper at right angles to the cylinder axis, and the two valves would be carried in the two ends of this box. The cylinder barrel would be brought in at its head end to form a portion of the cylinder head and then extended along its axis in the form of a fairly large boss, a mating boss being provided on one side of the valve box. The cylinder barrel would then be threaded into the valve box and the whole tightened or fastened to the crankcase by means of two sets of threads, one at each end of the barrel proper. This meant that three joints had to be made tight with only two sets of threads. This was accomplished by accurate machining and possibly even hand fitting in combination with a rather thick gasket at the head end, one flat of which bore against two different surfaces. This can be seen in Figure 6, where the circular flange on the valve box contacts both the crankcase and the cylinder barrel. Altogether it was a simple, light, and ingenious solution to a rather complex problem.

At this point the question arises: Why was the engine layout such that the exhaust took place close to the operator's ears? It would have been possible, starting with the original design, to turn the engine around so that the exhaust was on the other side. This would have little effect on the location of the center of gravity, and the two main drive chains would then have been of more equal length. However, of the many factors involved, probably one of the principal considerations in arriving at their final decision was the location of the spark-advance control, which was in effect the only control they had of engine output, except for complete shutoff. In their design this was immediately adjacent to the operator; with a turned-around engine, an extension control mechanism of some sort would have been required. The noise of the exhaust apparently became of some concern to them, as Orville's diary in early 1904 contains an entry with a sketch labeled "Design for Muffler for Engine," but there is no further comment.

The problem of keeping joints tight, and for that matter the entire construction itself, were both greatly simplified by their decision to water-jacket only a part of the cylinder head proper, and the valve box not at all. This was undoubtedly the correct decision for their immediate purpose, as again they were effecting savings in time, cost, complexity, and weight. There is nothing in the record, however, to show why they continued this practice long after they had advanced to much greater power outputs and longer flight times. Their own statements show that they were well aware of the effect of the very hot cylinder head on power output and they must also have realized its influence on exhaust-valve temperature.

The cylinder assembly was made somewhat more complicated by their desire to oil the piston and cylinder by means of holes near the crankshaft end in what was, with the engine in the horizontal position, the upper side of the cylinder barrel. This complication was no doubt taken care of by not drilling the holes until a tight assembly had been made by screwing the barrel into place, and by marking the desired location on the barrel. Since this position was determined by a metal-to-metal jam fit of the crankcase and cylinder barrel flange, the barrel would reassemble with the holes in very nearly the same relative position after disassembly.

With the valve box, or housing, cylindrical, the task of locking and fastening the intake and exhaust valve guides and seats in place was easy. The guide was made integral with and in the center of one end of a circular cage, the other end of which contained the valve seat (see Figure 5). Four sections were cut out of the circular wall of the cage so that in effect the seat and guide were joined by four narrow legs, the spaces between which provided passages for the flow of the cylinder gases. These cages were then dropped into the ends of the valve boxes until they came up against machined shoulders and were held in place by internal ring nuts screwed into the valve box. The intake manifold or passage was placed over the intake valves so that the intake charge flowed directly into and through the valve cage around the open valve and into the cylinder. The exhaust gas, after flowing through the passages in the valve cage, was discharged directly to the atmosphere through a series of holes machined in one side of the valve box.

The intake and exhaust valves were identical and of two-piece construction, with the stems screwed tightly into and through the heads and the protruding ends then peened over. This construction was not novel, having had much usage behind it, and it continued for a long time in both automobile and aircraft practice. One-piece cast and forged valves were available but here again it was a choice of the quick, cheap, and proven answer.

The entire valve system, including guides and seats, was of cast iron, a favorite material of the Wrights, except for the valve stems, which were, at different times, of various carbon steels. Ordinary cold-rolled apparently was used in those of the original engine, but in later engines this was changed to a high-carbon steel.

The piston design presented no difficulty. In some measure this was due to the remarkable similarity that seems to have existed among all the different engines of the time in the construction of this particular part, for, although there were some major variations, it was, in fact, almost as if some standard had been adopted. Pistons all were of cast iron and comparatively quite long (it was a number of years before they evolved into the short ones of modern practice); they were almost invariably equipped with three wide piston rings between the piston pin and the head; and, although there were in existence a few pistons with four rings, no oil wiper or other ring seems to have been placed below the piston pin. The Wrights' piston was typical of the time, with the rings pinned in the grooves to prevent turning and the piston pin locked in the piston with a setscrew. In designing this first engine they were, however, apparently somewhat unsure about this latter feature, as they provided the rod with a split little end and a clamping bolt (see Figure 6), so that the pin could be held in the rod if desired; but no examples of this use have been encountered.

The Wrights' selection of an "automatic" or suction-operated inlet valve was entirely logical. Mechanically operated inlet valves were in use and their history went back many years, but the great majority of the engines of that time still had the automatic type, and with this construction one complete set of valve-operating mechanisms was eliminated. They were well aware of the loss of volumetric efficiency inherent in this valve, and apparently went to some pains to obtain from it the best performance possible. Speaking of the first engine, Orville Wright wrote, "Since putting in heavier springs to actuate the valves on our engine we have increased its power to nearly 16 hp and at the same time reduced the amount of gasoline consumed per hour to about one-half of what it was."[12]

Why they continued with this form on their later engines is a question a little more difficult to answer, as they were then seeking more and more power and were building larger engines. The advantages of simplicity and a reduced number of parts still existed, but there also was a sizable power increase to be had which possibly would have more than balanced off the increased cost and weight. They did not utilize mechanical operation until after a major redesign of their last engine model. Very possibly the answer lies in the phenomenon of fuel detonation. This was only beginning to be understood in the late 1920s, and it is quite evident from their writings that they had little knowledge of what made a good fuel in this respect. It is fairly certain, however, that they did know of the existence of cylinder "knock," or detonation, and particularly that the compression ratio had a major effect on it. The ratios they utilized on their different engines varied considerably, ranging from what, for that time, was medium to what was relatively high. The original flight engine had a compression ratio of 4.4:1. The last of their service engines had a compression ratio about twenty percent under that of the previous series—a clear indication that they considered that they had previously gone too high. Quite possibly they concluded that increasing the amount of the cylinder charge seemed to bring on detonation, and that the complication of the mechanical inlet valve was therefore not warranted.

The camshaft for the exhaust valves (101, Figure 6), was chain driven from the crankshaft and was carried along the bottom of the crankcase in three babbit-lined bearings in bearing boxes or lugs cast integral with the case. Both the driving chain and the sprockets were standard bicycle parts, and a number of bicycle thread standards and other items of bicycle practice were incorporated in several places in the engine, easing their construction task. The shaft itself, of mild carbon steel, was hollow and on each side of an end bearing sweated-on washers provided shoulders to locate it longitudinally. Its location adjacent to the valves, with the cam operating directly on the rocker arm, eliminated push rods and attendant parts, a major economy. The cams were machined as separate parts and then sweated onto the shaft. Their shape shows the principal concern in the design to have been obtaining maximum valve capacity—that is, a quite rapid opening with a long dwell. This apparent desire to get rid of the exhaust gas quickly is manifested again in the alacrity with which they adopted a piston-controlled exhaust port immediately they had really mastered flight and were contemplating more powerful and more durable engines. This maximum-capacity theory of valve operation, with its neglect of acceleration forces and seating velocities, may well have been at least partially if not largely the cause of their exhaust-valve troubles and the seemingly disproportionate amount of development they devoted to this part, as reported by Chenoweth, although it is also true that the exhaust valve continued to present a problem in the aircraft piston engine for a great many years after, even with the most scientific of cam designs.

The rocker arm (102, Figure 6) is probably the best example of a small part which met all of their many specific requirements with an extreme of simplicity. It consisted of two identical side pieces, or walls, of sheet steel shaped to the desired side contour of the assembly, in which were drilled three holes, one in each end, to carry the roller axles, and the third in the approximate middle for the rocker axle shaft proper. This consisted of a piece of solid rod positioned by cotter pins in each end outside the side walls (see Figure 5). The assembly was made by riveting over the ends of the roller axles so that the walls were held tightly against the shoulders on the axles, thus providing the correct clearance for the rollers. The construction was so light and serviceable that it was essentially carried over to the last engine the Wrights ever built.

The basic intake manifold (see Figure 5) consisted of a very low flat box of sheet steel which ran across the tops of the valve boxes and was directly connected to the top of each of them so that the cages, and thus the valves, were open to the interior of the manifold. Through an opening in the side toward the engine the manifold was connected to a flat induction chamber (21, Figure 5) which served to vaporize the fuel and mix it with the incoming air. This chamber was formed by screw-fastening a piece of sheet steel to vertical ribs cast integral with the crankcase, the crankcase wall itself thus forming the bottom of the chamber. A beaded sheet-steel cylinder resembling a can (73, Figure 6) but open at both ends was fastened upright to the top of this chamber. In the absence of anything else, this can could be called the carburetor, as a fuel supply line entered the cylinder near the top and discharged the fuel into the incoming air stream, both the fuel and air then going directly into the mixing chamber. The can was attached near one corner of the chamber, and vertical baffles, also cast integral with the case, were so located that the incoming mixture was forced to circulate over the entire area of exposed crankcase inside the chamber before it reached the outlet to the manifold proper, the hot surface vaporizing that part of the fuel still liquid.

The reasons behind selection of the type of ignition used, and the considerations entering into the decision, are open to speculation, as are those concerning many other elements that eventually made up the engine. Both the high-tension spark plug and low-tension make-and-break systems had been in wide use for many years, with the latter constituting the majority in 1902. Both were serviceable and therefore acceptable, and both required a "magneto". The art of the spark plug was in a sense esoteric (to a certain extent it so remains to this day), but the spark-plug system did involve a much simpler combination of parts: in addition to the plug and magneto there would be needed only a timer, or distributor, together with coils and points, or some substitute arrangement. The make-and-break system, on the other hand, required for each cylinder what was physically the equivalent of a spark plug, that is, a moving arm and contact point inside the cylinder, a spring-loaded snap mechanism to break the contact outside the cylinder, and a camshaft and cams to actuate the breaker mechanism at the proper time. Furthermore, as the Wrights applied it, the system required dry cells and a coil for starting, although these did not accompany the engine in flight. And finally there was the problem of keeping tight the joint where the oscillating shaft required to operate the moving point in the spark plug entered the cylinder.

This is one of the few occasions, if not the only one, when the Wrights chose the more complex solution in connection with a major part—in this particular case, one with far more bits and pieces. However, it did carry with it some quite major advantages. The common spark plug, always subject to fouling or failure to function because of a decreased gap, was not very reliable over a lengthy period, and was undoubtedly much more so in those days when control of the amount of oil inside the cylinder was not at all exact. Make-and-break points, on the other hand, were unaffected by excess oil in the cylinder. Because of this resistance to fouling, the system was particularly suitable for use with the compression-release method of power control which they later utilized, although they probably could not have been looking that far ahead at the time they chose it. High-tension current has always, and rightfully so, been thought of as a troublemaker in service; in Beaumont's 1900 edition of Motor Vehicles and Motors, which seems to have been technically the best volume of its time, the editor predicted that low-tension make-and-break ignition would ultimately supersede all other methods. And finally, the large number of small parts required for the make-and-break system could all be made in the Wright Brothers' shop or easily procured, and in the end this was probably the factor, plus reliability, that determined the decision which, all things considered, was the correct one.

There was nothing exceptional about the exact form the Wrights devised. It displayed the usual refined simplicity (the cams were made of a single small piece of strip steel bent to shape and clamped to the ignition camshaft with a simple self-locking screw), and lightness. The ignition camshaft (38, Figure 5), a piece of small-diameter bar stock, was located on the same side as the exhaust valve camshaft, approximately midway between it and the valve boxes, and was operated by the exhaust camshaft through spur gearing. That the Wrights were thinking of something beyond mere hops or short flights is shown by the fact that the ignition points were platinum-faced, whereas even soft iron would have been satisfactory for the duration of all their flying for many years.

The control of the spark timing was effected by advancing or retarding the ignition camshaft in relation to the exhaust valve camshaft. The spur gear (37, Figure 5) driving the ignition camshaft had its hub on one side extended out to provide what was in effect a sleeve around the camshaft integral with the gear. The gear and integral sleeve were slidable on the shaft and the sleeve at one place (39, Figure 5) was completely slotted through to the shaft at an angle of 45° to the longitudinal axis of the shaft. The shaft was driven by a pin tightly fitted in it and extending into the slot. The fore-and-aft position of the sleeve on the shaft was determined by a lever-operated cam (40, Figure 5) on one side and a spring on the other. The movement of the sleeve along the shaft would cause the shaft to rotate in relation to it because of the angle of the slot, thus providing the desired variation in timing of the spark. The "magneto" was a purchased item driven by means of a friction wheel contacting the flywheel, and several different makes were used later, but the original is indicated to have been a Miller-Knoblock (see Figure 5).

The connecting rod is another example of how, seemingly without trouble, they were able to meet the basic requirements they had set for themselves. It consisted of a piece of seamless steel tubing with each end fastened into a phosphor-bronze casting, these castings comprising the big and little ends, drilled through to make the bearings (See Figures 5 and 6). It was strong, stiff and light.[13] Forged rods were in rather wide use at the time and at least one existing engine even had a forged I-beam section design that was tapered down from big to little end. The Wrights' rod was obtained in little more time than it took to make the simple patterns for the two ends. The weight was easily controlled, no bearing liners were necessary, and a very minimum of machining was required. Concerning the big-end material, there exists a contradiction in the records: Baker, whose data are generally most accurate, states that these were babbited, but this must be in error, as the existing engine has straight bronze castings without babbiting, and there is no record, or drawing, or other indication of the bearings having been otherwise.

Different methods of assembling the rod were used. At one time the tube ends were screwed into the bronze castings and pinned, and at another the ends were pinned and soldered. There is an indication that at one time soldering and threads were used in combination. One of the many conflicts between the two primary sets of drawings exists at this point. The Smithsonian drawings show the use at each end of adapters between the rod and end castings, the adapters being first screwed into the castings and pinned and then brazed to the inside of the tube. The Science Museum drawings show the tube section threaded and screwed into the castings. The direct screw assembly method called for accurate machining and hand fitting in order to make the ends of the tubing jam against the bottom of the threaded holes in the castings, and at the same time have the end bearings properly lined up. The weakness of the basic design patently lies in the joints. It is an attempt to utilize what was probably in the beginning a combination five-piece assembly and later three, in a very highly stressed part where the load was reversing. It gave them considerable trouble from time to time, particularly in the 4-cylinder vertical engines, and was abandoned for a forged I-beam section type in their last engine model; but it was nevertheless the ideal solution for their first engine.

The crankshaft was made from a solid block of relatively high carbon steel which, aside from its bulk and the major amount of machining required, presented no special problems. It was heat-treated to a machinable hardness before being worked on, but was not further tempered. The design was an orthodox straight pin and cheek combination and, as previously noted, there were no counterweights to complicate the machining or assembly. A sizable bearing was provided on each side of each crank of the shaft, which helped reduce the stiffness requirement.

Their only serious design consideration was to maintain the desired strength and still keep within weight limitations. A fundamental that every professional designer knows is that it is with this particular sort of part that weight gets out of control; even an additional 1/16 in., if added in a few places, can balloon the weight. With their usual foresight and planning, the Wrights carefully checked and recorded the weight of each part as it was finished, but even this does not quite explain how these two individuals, inexperienced in multicylinder engines—much less in extra-light construction—could, in two months, bring through an engine which was both operable and somewhat lighter than their specification.

In one matter it would seem that they were quite fortunate. The records are not complete, but with one exception there is no indication of any chronic or even occasional crankshaft failure. This would seem to show that it apparently never happened that any of their designs came out such that the frequency of a vibrating force of any magnitude occurred at the natural frequency of the shaft. Much later, when this type of vibration became understood, it was found virtually impossible, with power outputs of any magnitude, to design an undampened shaft, within the space and weight limitations existing in an ordinary engine, strong enough to withstand the stress generated when the frequency of the imposed vibration approximated the natural frequency of the shaft. The vibratory forces were mostly relatively small in their engines, so that forced vibration probably was not encountered, and the operating speed range of the engines was so limited that the natural frequency always fell outside this range.

The flywheel was about the least complex of any of their engine parts and required little studied consideration, although they did have to balance its weight against the magnitude of the explosion forces which would reach the power transmission chains, with their complete lack of rigidity, a problem about which they were particularly concerned. The flywheel was made of cast iron and was both keyed to and shrunk on the shaft.

Some doubt still exists about the exact method of lubricating the first engine. The unit presently in the airplane has a gear-type oil pump driven by the crankshaft through a worm gear and cross shaft, and the Appendix to the Papers states that it was lubricated by a small pump; nevertheless Baker says, after careful research, that despite this evidence, it was not. Also, the drawings prepared by Christman (they were commenced under the supervision of Orville Wright) do not show the oil pump. In March 1905 Wilbur Wright wrote to Chanute, "However we have added oiling and feeding devices to the engine ..."; but this could possibly have referred to something other than an oil pump. But even if a pump was not included originally, its presence in the present engine is easily explained. Breakage of the crankcase casting caused the retirement of this engine, which was not rebuilt until much later, and the pattern for this part had no doubt long since been altered to incorporate a pump. It was therefore easier in rebuilding to include than to omit the pump, even though this required the addition of a cross shaft and worm gear combination. On later engines, when the pump was used, oil was carried to a small pipe, running along the inside of the case, which had four small drill holes so located as to throw the oil in a jet on the higher, thrust-loaded side of each cylinder. The rods had a sharp scupper on the outside of the big end so placed as also to throw the oil on this same thrust face. Some scuppers were drilled through to carry oil to the rod bearing and some were not.

The first engine was finished and assembled in February 1903 and given its first operating test on 22 February. The Wrights were quite pleased with its operation, and particularly with its smoothness. Their father, Bishop Wright, was the recorder of their satisfaction over its initial performance, but what he noted was probably the afterglow of the ineffable feeling of deep satisfaction that is the reward that comes to every maker of a new engine when it first comes to life and then throbs. They obtained 13 hp originally: later figures went as high as almost 16, but as different engine speeds were utilized it is rather difficult to settle on any single power figure. The most realistic is probably that given in the Papers as having been attained later, after an accurate check had been made of the power required to turn a set of propellers at a given rpm. This came out at approximately 12 hp, the design goal having been 8. Following exactly the procedure that exists to this day, the engine went through an extended development period, and it was the end of September 1903 before it was taken, with the airplane, to Kitty Hawk where the historic flights, which have had such a profound effect on the lives of all men, were made on 17 December 1903.

The Engines With Which They Mastered The Art of Flying

Two more engines of this first general design were built but they differed somewhat from each other as well as from the original. Together with a third 8-cylinder engine these were begun right after the first of the year in 1904, shortly after the Wrights' return from Kitty Hawk. In planning the 8-cylinder engine they were again only being forehanded, but considerably so, in providing more power for increased airplane performance beyond that which might possibly be obtained from the 4-cylinder units. Progress with the 4-cylinder engines was such that they fairly quickly concluded that the 8-cylinder size would not be necessary, and it was abandoned before completion. Exactly how far it was carried is not known. The record contains only a single note covering the final scrapping of the parts that had been completed; and apparently there were no drawings, so that even its intended appearance is not known with any exactness. It was probably a 90° V-type using their original basic cylinder construction.

The changes carried through in the two 4-cylinder engines were not major. The water-cooled area of the cylinder barrel was increased by nearly ten percent but the head remained only partially cooled. In hindsight, this consistent avoidance of complete cylinder-head cooling presents the one most inexplicable of the more important design decisions they made, as it does not appear logical. In the original engine, where the factors of time and simplicity were of paramount importance, this made sense, but now they were contemplating considerably increased power requirements, knowing the effect of temperature on both the cylinder and the weight of cylinder charge, and knowing that valve failure was one of their most troublesome service problems. Nor does it seem that they could have been avoiding complete cylinder cooling through fear of the slightly increased complexity or the difficulty of keeping the water connections and joints tight, for they had faced a much more severe problem in their first engine, where their basic design required that three joints be kept tight with only two sets of threads, and had rather easily mastered it; so there must have been some much more major but not easily discernible factor which governed, for they still continued to use the poorly cooled head, even carrying it over to their next engine series. Very probably they did not know the effect on detonation of a high-temperature fuel-charge.

One of the new engines was intended for use in their future experimental flying and has become known as No. 2. It had a bore of 4-1/8 in., incorporated an oil pump, and at some time shortly after its construction a fuel pump was added. The fuel pump was undoubtedly intended to provide a metering system responsive to engine speed and possibly also to eliminate the small inherent variation in flow of the original gravity system.

This engine incorporated a cylinder compression release device not on the original. The exact reason or reasons for the application of the compression release have not been determined, although the record shows it to have been utilized for several different purposes under different operating conditions. Whatever the motivation for its initial application, it was apparently useful, as it was retained in one form or another in subsequent engine models up to the last 6-cylinder design. Essentially it was a manually controlled mechanism whereby all the exhaust valves could be held open as long as desired, thus preventing any normal charge intake or compression in the cylinder. Its one certain and common use was to facilitate starting, the open exhaust valves easing the task of turning the engine over by hand and making priming easy. In flight, its operation had the effect of completely shutting off the power. The propellers would then "windmill" and keep the engine revolving. One advantage stated for this method of operation was that when power was required and the control released, the engine would be at fairly high speed, so that full power was delivered immediately fuel reached the engine. It is also reported to have been used both in making normal landings and in emergencies, when an instant power shutdown was desired. Although it is not clear whether the fuel shutoff cock was intended to be manipulated when the compression release was used for any of these reasons, over the many years of its availability, undoubtedly at one time or another every conceivable combination of operating conditions of the various elements was tried. Because of the pumping power required with at least one valve open during every stroke, the windmilling speed of the engine was probably less than with any other method of completely stopping power output, but whether this difference was large enough to be noticeable, or was even considered, is doubtful.

Since a simple ignition switch was all that was required to stop the power output, regardless of whether a fuel-control valve or a spark-advance control was used, it must be concluded that the primary function of the compression release was to facilitate starting, and any other useful result was something obtained at no cost. The compression release was later generally abandoned, and until the advent of the mechanical starter during the 1920s, starting an engine by "pulling the propeller through" could be a difficult task. With the Wrights' demonstrated belief that frugality was a first principle of design, it is hardly conceivable that they would have accepted for any other reason the complication of the compression-release mechanism if a simple ignition switch would have sufficed.

The compression-release mechanism was kept relatively simple, considering what it was required to accomplish. A small non-revolving shaft was located directly under the rocker arm rollers that actuated the exhaust valves. Four slidable stops were placed on this shaft, each in the proper location, so that at one extreme of their travel they would be directly underneath the rocker roller and at the other extreme completely in the clear. They were positioned along the shaft by a spring forcing them in one direction against a shoulder integral with the shaft, and the shaft was slidable in its bearings, its position being determined by a manually controlled lever. When the lever was moved in one direction the spring pressure then imposed on the stops would cause each of them to move under the corresponding rocker roller as the exhaust valve opened, thus holding the exhaust valve in the open position. When the shaft was moved in the other direction the collar on the shaft would mechanically move the stop from underneath the roller, allowing the valve to return to normal operation.

If the 1903 engine is the most significant of all that the Wrights built and flew, then certainly the No. 2 unit was the most useful, for it was their sole power source during all their flying of 1904 and 1905 and, as they affirmed, it was during this period that they perfected the art, progressing from a short straightaway flight of 59 seconds to a flight controllable in all directions with the duration limited only by the fuel supply. It is to be greatly regretted that no complete log or record was kept of this engine.

The Wrights again exhibited their engineering mastery of a novel basic situation when, starting out to make flight a practical thing, they provided engine No. 3 to be used for experimental purposes. In so doing they initiated a system which continues to be fundamental in the art of providing serviceable aircraft engines to this day—one that is expensive and time consuming, but for which no substitute has yet been found. Their two objectives were: improvement in performance and improvement in reliability, and the engine was operated rather continuously from early 1904 until well into 1906. Unfortunately, again, no complete record exists of the many changes made and the ideas tested, although occasional notes are scattered through the diaries and notebooks.

In its present form—it is on exhibition at the Engineers Club in Dayton, Ohio—the No. 3 engine embodies one feature which became standard construction on all the Wright 4-cylinder models. This was the addition of a number of holes in a line part way around the circumference of the cylinder barrel so that they were uncovered by the piston at the end of its stroke toward the shaft, thus becoming exhaust ports (see Figure 9). This arrangement, although not entirely novel, was just beginning to come into use, and in its original form the ports exhausted into a separate chamber, which in turn was evacuated by means of a mechanically operated valve, so that two exhaust valves were needed per cylinder. Elimination of this chamber and the valve arrangement is typical of the Wrights' simplifying procedure, and it would seem that they were among the very first to use this form.[14]

The primary purpose of the scheme was to reduce, by this early release and consequent pressure and temperature drop, the temperature of the exhaust gases passing the exhaust valve, this valve being one of their main sources of mechanical trouble. It is probable that with the automatic intake valves being used there was also a slight effect in the direction of increasing the inlet charge, although with the small area of the ports and the short time of opening, the amount of this was certainly minor. With the original one-piece crankcase and cylinder jacket construction, the incorporation of this auxiliary porting was not easy, but this difficulty was overcome in the development engine by making different castings for the crankcase itself and for the cylinder jacket and separating them by several inches, so that room was provided between the two for the ports.

This engine demonstrated the most power of any of the flat 4s, eventually reaching an output of approximately 25 hp, which was even somewhat more than that developed by the slightly larger 4-1/8-in.-bore flight engine, with which 21 hp was not exceeded. Indicative of the development that had taken place, the performance of the No. 3 engine was twice the utilized output of the original engine of the same size, an increase that was accomplished in a period of less than three years.

The Wrights were only twice charged with having plagiarized others' work, a somewhat unusual record in view of their successes, and both times apparently entirely without foundation. A statement was published that the 1903 flight engine was a reworked Pope Toledo automobile unit, and it was repeated in an English lecture on the Wright brothers. This was adequately refuted by McFarland but additionally, it must be noted, there was no Pope Toledo company or car when the Wright engine was built. This company, an outgrowth of another which had previously manufactured one-and two-cylinder automobiles, was formed, or reformed, and a Pope license arrangement entered into during the year 1903.

The other incident was connected with Whitehead's activities and designs. Whitehead was an early experimenter in flying, about the time of the Wrights, whose rather extraordinary claims of successful flight were published in the 1901-1903 period but received little attention until very much later. His first engines were designed by a clever engineer, Anton Pruckner, who left at the end of 1901, after which Whitehead himself became solely responsible for them. It was stated that the Wrights visited the Whitehead plant in Bridgeport, Connecticut, and that Wilbur remained for several days, spending his time in their machine shop. This was not only categorically denied by Orville Wright when he heard of it but it is quite obvious that the 1903 or any other of the Wright engine designs bears little resemblance to Pruckner's work. In fact, its principal design features are just the opposite of Pruckner's, who utilized vertical cylinders, the 2-stroke cycle, and air-cooling, which Whitehead at some point changed to water-cooling.[15]

The Four-Cylinder Vertical Demonstration Engine and the First Production Engine

In 1906, while still doing general development work on the flat experimental engine, the Wrights started two new engines, and for the first time the brothers engaged in separate efforts. One was "a modification of the old ones" by Wilbur and the other, "an entirely new pattern" by Orville. There is no record of any of the features of Wilbur's project or what was done in connection with it. Two months after the experimental operation of the two designs began, an entry in Wilbur's diary gives some weight and performance figures for the "4" x 4" rebuilt horizontal," and since Orville's design was vertical the data clearly refer to Wilbur's; but since the output is given only in test-fan rpm it does not serve to indicate what had been accomplished and there is no further mention of it.

Orville's design became the most used of any model they produced. It saw them through the years from 1906 to 1911 or 1912, which included the crucial European and United States Army demonstrations, and more engines of this model were manufactured than any of their others including their later 6-cylinder. Although its ancestry is traceable to the original 1903 engine, the design form, particularly the external configuration, was considerably altered. Along with many individual parts it retained the basic conception of four medium-size cylinders positioned in line and driving the propellers through two sprocket wheels. From the general tenor of the record it would seem, despite there being no specific indication, that from this time on Orville served as the leader in engine design, although this occurred with no effect whatsoever on their finely balanced, exactly equal partnership which endured until Wilbur's death in 1912.

The first major change from the 1903 design, putting the engine in an upright instead of flat position, was probably done primarily to provide for a minimum variation in the location of the center of gravity with and without a passenger. Whether or not it had any influence, the vertical cylinder arrangement was becoming predominant in automobile powerplants by this time, and the Wright engines now began to resemble this prevailing form of the internal combustion engine—a basic form that, in a wide variety of uses, was to endure for a long time.

Over the years, the Wrights seem to have made many changes in the engine: the bore was varied at different times, rod assembly methods were altered, and rod ends were changed from bronze to steel. Chenoweth states that on later engines an oil-control ring was added on the bottom of the piston, necessitating a considerable increase in the length of the cylinder barrel. This arrangement could not have been considered successful, as it apparently was applied to only a limited number of units and was not carried over to the later 6-cylinder engine model. There was much experimentation with cam shapes and most probably variations of these got into production.

With the crankcase, they did not go all the way to the modern two-piece form but instead retained the one-piece construction. Assembly was effected through the ends and a detachable plate was provided on one side for access to the interior. It is clear that they regarded this ability to get at the interior of the case without major disassembly as a valuable characteristic, and later featured it in their sales literature. They were apparently willing to accept the resultant weakening of the case and continued the construction through their last engine model. The integrally cast cylinder water jackets were abandoned and the top of the crankcase was machined flat to provide a mounting deck for individual cylinders. The use of aluminum alloy was continued, and the interior of the case was provided with strengthening webs of considerable thickness, together with supporting ribs. The cam shaft was supported directly in the case.

The individual cylinder design was of extreme simplicity, a single iron casting embodying everything except the water jacket. The valves seated directly on the cast-iron cylinder head and the guides and ports were all contained in an integral boss on top of the head. The exhaust valve location on the side of the engine opposite the pilot was a decided advantage over that of the 1903 design, where the exhaust was toward the pilot. A four-cornered flange near the bottom of the cylinder provided for fastening it to the crankcase, and a threaded hole in the top of the head received a vertical eyebolt which served as the rocker-arm support. The cylinder was machined all over; two flanges, one at the bottom and the other about two-thirds of the way down provided the surfaces against which the water jacket was shrunk. The jacket was an aluminum casting incorporating the necessary bosses and double shrunk on the barrel; that is, the jacket itself was shrunk on the cylinder-barrel flanges and then steel rings were shrunk on the ends of the jacket over the flanges. The jacket thickness was reduced by machining at the ends, making a semigroove into which the steel shrink rings fitted. These rings insured the maintenance of a tight joint despite the tendency of the aluminum jacket to expand away from the cast-iron barrel.