This is, I guess what puzzles me. Passing a VFR joiner from Approach to Tower when it has "half a chance of seeing the other traffic" seems like a sensible transition point. However, sighting a square mile of airfield is likely from a much greater range than seeing a 30 foot plank of a wing with a small fuselage attached.
And thats the point, bookworm. You get the airfield in sight first and are transferred to tower. You now have time to continue towards the airfield whilst, changing freq, receiving joining instructions and traffic information, looking for the traffic and setting the aircraft up for landing. hopefully allowing a seamless join to land.
From a tower controllers point of view, they will now probably have you in sight (eagle eyes you see!) and be able to exercise reduced seperation in the vicinity of an aerodrome.