The first De Havilands arriving in France were immediately put together, such remediable imperfections as existed were corrected then and there, and the machines were flown to the training fields. The changing and increasing demands of the service indicated the advisability of certain changes of design. The foreign manufacturers had brought out a covering for the gasoline tanks, making them nearly leak-proof, even when perforated by a bullet. In the first De Havilands the location of the principal gas tanks between the pilot and the observer was not the best arrangement in that the men were too far apart from each other so that, if the machine went down, the pilot would be crushed by the gas tank. Also the radius of action was not considered to be great enough, even though the later machines of this type carried 88 gallons of gasoline. As a result the American aircraft designers brought out an improved De Haviland known as the 9-A. This carried a Liberty-12 engine; and the main differences between it and the De Haviland-4 were new locations for pilot and tanks, their positions being changed about, increased gasoline capacity, and increased wing surface. The machine was a cleaner, more finished design, showed slightly more speed, and had a greater radius of action than the De Haviland-4 which it was planned to succeed. We ordered 4,000 of these new machines from the Curtiss Co., but the armistice cut short this production. The difficulties in the way of producing new service planes on a great scale without previous experience in such construction is clearly shown in the attempts we made to duplicate other successful foreign planes. On September 12, 1917, we received from the aviation experts abroad a sample of the French Spad. We had previously been advised to go into a heavy production of this model and had made arrangements for the Curtiss Co. at Buffalo to undertake the work. This development was well under way when in December a cablegram was received from Gen. Pershing advising us to leave the production of all single-place fighters to Europe. As a result we canceled the Spad order, and after that we attempted to build no single-place pursuit planes. [Illustration: THE LEPERE CORPS OBSERVATION PLANE.] [Illustration: THE LEPERE (CAMOUFLAGED). THE ENGINE IS A LIBERTY 12-CYLINDER, 400-HORSEPOWER. This plane was developed in the United States.] At the time this course seemed to be justified. The day of the single seater seemed to be over. The lone occupant of the single seater can not keep his attention on all directions at once; and as the planes grew thicker in the air, the casualties among flyers increased. But the development of formation flying restored the single-place machine to favor. The formation had no blind spot, thus removing the principal objection to the single seater. The end of the war found the one-man airplane more useful than ever.