Step 1: Recall what happens physically inside a stereoscope.
A stereoscope works by presenting the left photograph to the left eye and the right photograph to the right eye, and the brain fuses the two slightly different views into a single 3D impression only if corresponding points differ in position mainly along one consistent direction, the direction of flight.
Step 2: Define the flight line on the photographs.
On each photo, the flight line is drawn by joining that photo's own principal point to the conjugate principal point, the principal point of the adjacent photo, transferred onto this one. It marks, on the photo itself, the direction the camera was travelling.
Step 3: State the orientation rule for correct fusion.
To view the pair stereoscopically without eye strain and without any vertical (y) mismatch between corresponding points, the two photos are physically arranged so that both flight lines fall on one single straight line, that is, the flight lines of the two photos are made collinear. This is the standard relative orientation setup used both for simple mirror stereoscope viewing and as the starting configuration before more rigorous analytical relative orientation.
Step 4: Rule out the wrong descriptions.
Arranging the photos on perpendicular planes or along perpendicular lines, as in options B and D, would introduce a large cross-track (y) parallax and destroy the stereo effect rather than enable it. A rule fixing the gap between the lines at "half the photo size," as in option C, has no basis in photogrammetric orientation theory, the correct separation is whatever reproduces the true air base at the chosen viewing scale, not a fixed fraction of photo format. Only collinearity of the flight lines is the genuine requirement.
\[ \boxed{\text{along a common straight line}} \]