This set up, without the glass, would produce a standard interference pattern, with a bright central maximum followed on both sides by progressively dimmer maxima.
Placing a glass plate over one of the slits would refract the light diffracting through the slit, resulting in the emergent rays being more "spread out" as they bend away from the normal.
I would assume that this has no impact on the interference pattern, and the central maximum would stay at O because the light from Q "spreads out" equally in both directions X and Y. However, apparently the central maximum would actually move towards Y, and the rest of the pattern would shift in accordance.
I cannot understand why this would happen. Where is the error in my understanding?