After a restful visit at home Frank Duryea returned to Springfield and finished the work on his burner. Now only the lack of a carburetor prevented a trial of the vehicle. He recalls that he studied several gasoline-engine catalogs and in one of them, a Fairbanks catalog he believes,[16] he saw a design that seemed to suit his needs. He decided to simplify the construction and operation of his carburetor and had a small bronze casting made to form the body of it. Inside was a gasoline chamber with two tapped openings, one to receive a pipe from the 2-gallon gasoline tank mounted above the engine, the other taking a pipe to the overflow tank underneath the engine, thus maintaining the gasoline level without the use of a float valve. This latter tank had a hand pump on one end so that the overflow gasoline could at times be pumped again into the main tank. Gasoline passed from the carburetor chamber through a needle valve, adjusted by a knob on top, then through a tiny tube that entered the pipe leading to the intake valve. It is not certain whether this intake pipe was at first fitted with the choke arrangement later used with the second engine. Frank, hoping at last to be rewarded for his efforts by the sound of explosions from the engine, was ready to give the carriage an indoor trial. Standing astraddle of the reach and facing to the rear, he spun the flywheel with both hands, taking care not to get his hands caught between the wheel and the frame. His efforts were in vain, as there was complete failure to obtain ignition. He then made a new ignition tube, nearly twice as long as the original 4-1/2-inch tube, and turned down its wall as thin as he thought safety allowed. The thinner wall did not conduct the heat off so rapidly and thus kept the tube hot enough to permit ignition. After this slight change, he was able to get a few occasional explosions but he does not now believe that the engine ever operated continuously. Each explosion was accompanied by a loud knock, due, undoubtedly, to the movement of the free piston. Had the engine operated continuously, it is likely that the action of the free piston would have shortly wrecked the engine. Further efforts appeared unwarranted until alterations could be made. [Illustration: ALL AGREEMENTS CONTINGENT UPON STRIKES, ACCIDENTS AND OTHER CAUSES BEYOND OUR CONTROL CABL ADDRESS "MOTODURYEA," PHILADELPHIA WESTINGHOUSE AND W. U. CODES