[Illustration: _Figure 10._--4-Cylinder vertical engine: a, cylinder assembly with valve mechanism parts; b, cylinder disassembled, and parts; c, pistons and connecting rods; d, bottom side of piston; e, crankshaft, flywheel and crankcase end closure; f, crankcase, with compression release parts. (Pratt & Whitney photos D-14996, 15001, 14998, 14994, 14999, 14989, respectively.)] The camshaft construction was considerably altered from the 1903 design. Although the reason is not entirely clear, one indication suggests that breakage or distortion of the shaft may have been encountered: whereas in the 1903 engine there had been no relationship between the location of the cams and the camshaft bearings, in this engine the exhaust valves were carefully positioned so that all cams were located very close to the supporting bearings in the crankcase. Also, the camshaft was solid, although it would seem that the original hollow shaft construction could have provided equal stiffness with less weight. The final decision was possibly determined by the practicality that there existed no standard tubing even approximating the size and wall thickness desired. There still was no carburetor, a gear pump metering the fuel in the same manner as on the 1904-1905 engine. Basically, the intake charge was fed to the cylinders by a round gallery manifold running alongside the engine. This was split internally by a baffle extending almost from end to end, so that the fuel mixture entering the manifold on one side of the baffle was compelled to travel to the two ends before it could return to the inside cylinder, this feature being a copy of their 1903 general intake arrangement. Apparently various shapes and positions of entrance pipes with which to spray the fuel into the manifold were used; and the injection arrangement seems also to have been varied at different times. The fuel pump was not necessarily always used, as the engine in some of the illustrations did not incorporate one, the fuel apparently being fed by gravity, as on the original engine. Chenoweth describes an arrangement in which exhaust heat was applied to the inlet manifold to assist the fuel vaporization process, but it is believed that this was one of the many changes made in the engine during its lifetime and not necessarily a standard feature. A water circulation pump was provided, driven directly by the crankshaft through a two-arm universal joint intended to care for any misalignment between the shaft and the pump. The water was piped to a horizontal manifold running along the cylinders just below the intake manifold, and a similar manifold on the other side of the engine collected it for delivery to the radiator. It is a little difficult to understand why it was not introduced at the bottom of the water jackets.