In particular, section a depicts a directional coupler working as incident wave separator, whereas section (b) shows the separation of the reflected wave.
The correct working is based on the assumption that the distance between the holes matches exactly one quarter of the wave length (l).
In fact, in the secondary waveguide, each hole will give rise to two waves going in opposite directions (one outside and the other inside the waveguide); consequently, in front of each hole, two waves are summed with their own phases.
The assumption made on the distance between the holes guarantees that, in front of one hole, (1) the two waves propagating outside the waveguide will be in phase, causing an enforcing effect in that direction; (2) while, in front of the other hole, the two waves (always propagating outside) will be in opposition, causing a canceling effect in that direction.