The Alamogordo test, General Groves explains (in Now It Can Be Told: The Story of the Manhattan Project), had not set aside all doubts about the bomb:
It proved merely that one implosion-type, plutonium bomb had worked; it did not prove that another would or that a uranium bomb of the gun type would. We had made every possible component test we could think of. We were reasonably sure of each one. We knew we could bring the U-235 portions of the bomb together in such a way that, if the theories of atomic energy were correct and U-235 behaved as plutonium had, the bomb should go off. But still no test had been made of the complete bomb. Nevertheless, the indications for success were strong enough so that no one urged us to change our plans of dropping the first gun-type bomb in combat without prior test. In any case, we simply had to take the chance, because the production of U-235 was so slow, even compared to plutonium, that we could not afford to use it in a test.
[…]
Early in the morning of July 16, the bomb — minus the last necessary bit of U-235 — was put aboard the cruiser Indianapolis, which sailed almost immediately. The Indianapolis had a fast but uneventful trip through Hawaii to Tinian.
[…]
The Indianapolis arrived at Tinian on July 26 and discharged her cargo that same day. She then put out to sea, headed for the Philippines, but her voyage came to a sudden and tragic end on July 30 when she was attacked by a Japanese submarine and went down with some nine hundred of her crew.
[…]
I had arranged with Lieutenant General H. L. George, the head of Air Transport, for the final parts, including some U-235, of the Hiroshima bomb to be flown from Albuquerque to Tinian. Because I did not want to risk having the plane disappear in flight with an extremely valuable though small piece of U-235, I asked for two large cargo planes in perfect condition, and the best possible crews. I told George that the cargo would be almost infinitesimal and that the second plane was wanted just in case of need, to tell us where the first one crashed — if it did. George had had his orders from Arnold and he simply said, “Whatever you want, you get.”
[…]
The flight to San Francisco was uneventful but when they were about forty-five minutes out over the Pacific, de Silva’s plane had to return to the mainland because of engine trouble. Contrary to what I had counted on, Taylor’s plane, instead of returning too, continued directly to Hawaii and arrived there several hours before de Silva’s plane came in. Despite our desire to keep the flight inconspicuous, a radio tip — probably sent unofficially — alerted Hawaii to a very important flight, and Taylor was met at the airfield by a group of senior officers who expected at least one high-ranking general with entourage instead of a single second lieutenant. Taylor had no written orders with him, since they were in de Silva’s hands, and soon found the local personnel determined, in accordance with custom, to load the empty plane to capacity. Fortunately, the plane crew had been convinced before departure from California that the trip was most unusual and with their support he was able to keep the ship unloaded until de Silva arrived. From then on the trip was uneventful.
[…]
One of our big problems had been whether to use each bomb as it became available or to store up a sufficient number to permit their employment in quantity. Most professional soldiers will go to almost any length to avoid piecemealing away their resources. However, in this case, we felt that the considerations of time and expected power justified our using the weapons as they became available. If our assessment of these factors had been erroneous, our position might very well have been extremely embarrassing, for our projected production rates called for one Fat Man of the type we had tested at Alamogordo (but not by an air drop) to be delivered at Tinian on about August 6, with a second one to be ready about August 24, and additional ones arriving in increasing numbers from there on.
As I have said, we had no assurance that a uranium bomb, the Thin Man, would work at all. Yet, that was the weapon that we employed against Hiroshima, since we had used up our entire immediate supply of plutonium in the test bomb.
[…]
General Marshall was also furnished with the probable readiness dates of the implosion bomb, and I explained that there would have to be a gap of at least three days between successive bombs, no matter what type was used. I did not give him the reason for this: we needed the time to assemble the bomb, and nothing should be hurried. Judging by the Alamogordo test, I said, we now expected this bomb to have an explosive force exceeding the equivalent of ten thousand tons of TNT and possibly reaching as high as thirty thousand tons.
[…]
Many small problems arose as the date of the first bombing approached. Two years before, the War Department had issued orders prohibiting persons with knowledge of future military operations from flying over enemy-held territory. The reason for this order was sound — to prevent leakage of information through possible prisoners of war — but it created an impossible situation for us. We had to have Parsons and Ashworth and their two technical assistants aboard the bombing planes. We also needed Tibbets and his best pilots on these flights, and we wanted to have certain scientists in the observation planes. We simply could not operate under such a restriction.
When this problem was brought to my attention in Washington, I immediately secured War Department authority to disregard the prohibition for both civilian and military personnel concerned with Centerboard, as our overseas operation had been named, provided their flights were essential.