A second basic determination which was made either concurrently or even possibly in advance of that of the general form and size was in the matter of the type of cylinder cooling to adopt. Based on current practice that had proven practical, there were three possibilities, all of which were in use in automobiles: air, water, or a combination of the two. It is an interesting commentary that Fernand Forest's[10] proposed 32-cylinder aircraft engine of 1888 was to be air-cooled, that Santos-Dumont utilized an air-cooled Clement engine in his dirigible flights of 1903, and that the Wrights had chosen air cooling for their shop engine. With the promise of simplicity and elimination of the radiator, water and piping, it would seem, offhand, that this would be the Wrights' choice for their airplane; but they were probably governed by the fact that not only was the water-cooled type predominant in automobile practice, but that the units giving the best and highest performance in general service were all water cooled. In their subsequent practice they never departed from this original decision, although Wilbur Wright's notebook of 1904-1907 contains an undated weight estimate by detailed parts for an 8-cylinder air-cooled engine. Unfortunately, the proposed power output is not recorded, so their conception of the relative weight of the air-cooled form is not disclosed. [Footnote 10: Fernand Forest, _Les Bateaux Automobiles_, 1906.] One of the most important decisions relating to the powerplant--one which was probably made long before they became committed to the design itself--was a determination of the method of transmission of power to the propeller, or propellers. A lingering impression exists that the utilization of a chain drive for this purpose was a natural inheritance from their bicycle background. No doubt this experience greatly simplified the task of adaptation but a merely cursory examination shows that even if they had never had any connection with bicycles, the chain drive was a logical solution, considering every important element of the problem. The vast majority of automobiles of the time were chain driven, and chains and sprockets capable of handling a wide range of power were completely developed and available. Further, at that time they had no accurate knowledge of desirable or limiting propeller and engine speeds. The chain drive offered a very simple and inexpensive method of providing for a completely flexible range of speed ratios. The other two possibilities were both undesirable: the first, a simple direct-driven single propeller connected to the crankshaft, provided essentially no flexibility whatsoever in experimentally varying engine or propeller speed ratios, it added an out-of-balance engine torque force to the problem of airplane control, and, finally, it dictated that the pilot would be in the propeller slipstream or the airflow to it; the second, drive shafts and gearing for dual propellers, would have been very