Yesterday I expressed my irritation at the ambiguities in the highlighting. I completely agree with durban that highlighting is unnecessary, so the ambiguity could have been avoided.
On the other issues I'll try to defend the setter's position. He has clearly tried to represent the experiment as completely as possible but has encountered dilemmas. The choice of L and S is technically wrong, but L and T would be implausible, L and K would be a mixture of units, and a small o and a large O in the grid would be too revealing. Every pair is open to criticism.
It's not clear exactly where the pair of letters should be, but it's fairly safe to assume that solvers won't be marked wrong if their relative positions conform to the conclusion of the experiment. In that respect I'd say that placing them at the top would be wrong, but anywhere from about a third of the way down to near the bottom should be correct. Bananabean, please note, I think my previous suggestion is wrong - if they are on the ground that doesn't show what needs to be shown.
Lastly, the annotation. I'm sure the marker will accept anything appropriate. (I'm tempted to write "A........ disproved" ). That requirement might have been in the preamble more for the initial, which reassures solvers that they have the right experiment.