Film color sensitivity is relative, by which I mean that the portion of the spectrum whihc the film is "insensitive" to isn't ignored completely by the chemistry, it just takes a lot more intensity for the "insensitive" color to register on the film. For example a "red-insensitive" film emulsion will show a red object as being dark, and a blue object as being light, if both have the same brightness or intensity. However, a very bright red object will still look light if it is bright enough. Signal flares are intensely bright, even though they are red.
The primary source of bright red color in a modern signal flare is usually Strontium Nitrate, although there are other. I don't know what was used back in the 1930's, but wouldn't be surprised if it were the same thing as today. The propellant industry (I work for one now, which used to be part of Olin) tends to use successfuly chemistry forever.