If this thing really is on the camera negative and not on a copy it would be tricky but not impossible.AFAIK we know its on the original print, its not photoshop.
I'm not seeing how an in focus but tiny representation of a flash lamp ends up on the negative/final print, the shape is compelling though.
-It's a roll film, somewhere around 200 feet long.
-The image would have to be introduced before processing.
-I would use a positive image with a mask. This positive image transparency would basically be made the same way movie film is produced from a camera negative. (Maybe even a frame of 16mm film? Or a small section of a color slide?) The mask could be as simple as a sheet of construction paper with a small window cut out with an X-ACTO knife and the positive film taped to the front surface. Then give the whole thing a bit of light. (A filtered flash?)
The mask protects the rest of the film from light. The rest of the roll would have to be protected as well.
-This technique might leave a suspicious artifact on the negative, as the film in this small spot has been exposed to more light than has the surrounding film. A dark blotchiness, perhaps.
-It could be done before or after the film was in the camera.
- A complication is that the hoaxer would have no way of knowing where the frames on unprocessed film would be, let alone what exact bit of landscape was where. (The hoaxer might therefore expose the film in several places along the roll?)
This movie from 1940 gives a general idea of what this type of film looks like. Most film in the 1970's was machine processed but the old fashioned way was still used, I'll bet. Doesn't really matter in the end. The chicanery could be accomplished in either system.
The film processing up to the point that the film was fixed in the hypo would be done in total darkness, you understand.