It appears to me the problem with using any sort of float as shown in the manual from the L10 is the lack of a window which offers observation to the rear. It appears to me by eye without taking an actual measurement drift would need to be on the order of 30 degrees in order to be measured by the drift meter from either the right or left rear window. Again it appears to me the manual assumes a navigation bubble on the aircraft that allows a view of the horizon to the rear except directly behind the tail fin with some interference from the rear stabilizers. This said the lack of a view to the rear would seem to lower the probability such measurements were made. The only option I can think of, there certainly may be many others, is to observe from the front of the aircraft. This would entail dropping some sort of float, aluminium and lights were mentioned in the inventory and Noonan stated in a newspaper article that smoke had been tried and aluminium powder worked better, and turning the aircraft about and making the observations from the front. IF such a thing was assumed for amusement a line drawn from a reasonable place to mount the drift meter, the right front seat, over the nose should tell you how close you could be and see an object on the surface at a given altitude. Use of the drift meter by whatever method appears to me to be A reason for operation at 1000 feet. The drift angle is only valid at the altitude it is taken. The wish to be below cloud bottoms for search may be a simpler explanation. In any event floats, lights and the drift meter were present in the Luke Field inventory which leads me think some method to use them had been worked out, perhaps not logic would lead to believe someone one board would have known how to operate the DF loop. So much for logic.
Chris,
It appears this topic was discussed at great length some time ago. Please see:
http://tighar.org/Projects/Earhart/Archives/Forum/Highlights41_60/highlights52.htmlApparently someone, Noonan?

, did find a way to take drift bearings from the rear of the plane. There were brackets to hold the drift meter behind the doors and "steps" to hold the door open so Noonan could get his head out, so Ric stated. A person claiming experience using drift meters had found them inaccurate, due to aircraft motion, even without the distraction of a 150 mph wind. So was the provision made as eyewash, out of ignorance or for some reason
we've not yet contemplated. Fun discussion, probably not worth the memory to store it.