AIRCRAFT PROPULSION — TAYLOR 269 



num (fig. 12, left) . This cylinder design suffered from the same trou- 

 ble as the early "Hispano-Suiza" engines, namely, poor exhaust-valve 

 cooling because of poor contact between the head and the separate 

 cooling element. However, the first radial engine using this cylinder 

 type, the A.B.C. "Wasp" of 4I/2'' bore, was successful enough to gain 

 the support of the British government for its development in a larger 

 version, the "Dragonfly" of 5^2^ bore. Evidently the fact that cooling 

 problems increase with increased cylinder size was not realized at the 

 time. 



This development finally became, through several changes in owner- 

 ship, the Bristol "Jupiter" engine (pi. 13, fig. 1), which was built and 

 used in considerable quantities in England and in Europe, chiefly for 

 military purposes. It was never a really satisfactory aircraft engine, 

 because of poor exhaust- valve cooling (S.D. Heron said that its con- 

 sumption should be given in terms of pounds of exhaust-valves, rather 

 than in pounds of fuel, per horsepower-hour ! ) . Finally realizing this 

 fact, Bristol changed from steel heads to aluminum heads with the 

 Jupiter "F," about 1930. 



Meanwhile the Gibson-Heron type cylinders had been further 

 developed by Armstrong Siddeley, and were used on the "Jaguar" 

 two-row radial (pi. 13, fig. 2) which passed its type test in 1922, 10 

 years after Gibson started his work. 



A parallel development of air-cooled engines with aluminum cylin- 

 ders was begun about 1916 by C. L. Lawrance. Starting in 1916 with 

 a 2-cylinder opposed engine, he built a 3-cylinder engine in 1919, and 

 finally a 9-cylinder 200-hp. radial in 1921, the J-1 (pi. 14, fig. 1), sup- 

 ported by an order for 200 engines from the U.S. Navy. In 1922 

 this company was absorbed by the Wright Aeronautical Corpora- 

 tion" and, with Navy support, the 9-cylinder engine was built in 

 improved models known as the Wright J-3, J-4, and J^b, all with 

 essentially the Lawrance cylinder design. During the same period, 

 1918-1926, S. D. Heron had left England and had been employed by 

 the U.S. Army Air Service at McCook Field, Dayton, Ohio, to assist 

 in the development of large radial engines.^* Heron was a devoted 

 worker and an able engineer, and by 1921 had developed successful 

 air-cooled cylinders of nearly 6 inches bore, based on his work with 

 Gibson plus his own improvements worked out at McCook Field. 



Against considerable resistance from their chief engineers, who at 

 the time were thoroughly committed to watercooling, the Army con- 

 tracted for sample radial engines from both Curtiss and Wright 

 Aeronautical. Such engines were built, but in very small numbers. 



The Lawrance and Heron developments were brought together 

 when Heron joined Wright Aeronautical Corporation in 1926, the 



" See previous remarks regarding the Hlspano-Sulza engine. 



" The author was In charge of the engine laboratory at McCook Field, and was closely 

 associated with Heron and his work during this period. 



