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.