Salviati is performing a one photon at a time, Young's interference experiment.
After the first trial, Simplicius asks: Do they interfere? Salviati replies: I don't know.
After the 2nd trial: Sim: Do they interfere? Sal: I don't know.
Same question and same answer follow for the next few hundred trials.
However, at the nth trial: Sim: Do they interfere? Sal: Yes, they do.
Sim: Ah!, the nth trial has been the crucial one.
Sal: No, it was just like any other trial.
Sim: But what was special about it that made you decide that the photons beams do interfere?
Sal: It was the collection of trials that revealed the interference pattern.
Sim: Then, only a collection of photons interfere.
Sal: Not correct, in each trial, the photon must have interfered because the particular outcome of each trial is independent from the rest.
Sim: Why couldn't you see the interference in each trial then?
Sal: A dot obtained from each trial or a few scattered dots from several of them tell me very little. Yet, the collective result gives me information about each one of them.
Sim, (dubitative): Is the same true for the path of the two coloured photons?
Sal, (even more insecure): Well, er..., I guess so.
Sim: An interference pattern builds up between time $t$ and a short time afterwards, say $t+\delta t$.
Sal: No way to tell which path the photons took.
Sim: At a later time, a displaced interference pattern builds up.
Sal: If the displacement is upwards, the blue photons came through the lower slit B and the red ones through the upper slit A.
Sim: A measurement of the collection gives information about each photon then.
Sagredo (who has been listening to the discussion): Yes, although the photons at the detection plane are neither blue nor red but bear information of both.
After this phrase is said, the three of them stare at one another.
Simplicius seems quite satisfied but Salviati and Sagredo look somewhat troubled …