I am, still, trying to get my head around batteries and proper use and maintenance of them.
Reading other threads, there is really only one way to estimate actual battery capacity, rather than original capacity, which is:-
get the batteries fully charged,
remove the surface charge (by running the tunnel light for a quarter of an hour or so OR by leaving the batteries with no load of any nature for 12 hours or so),
apply a known load to drain the batteries until they reach a specified voltage (12.3v on a 12v system), and note the time taken to deplete the batteries to that level,
do some arithmetic to calculate to total load applied and, therefore, actual capacity.
Surface charge.
My guess is that the more fully charged a battery is, the longer it takes to force new charge deep into the plates, that surface charge is electricity that hasn't yet started / completed that journey, and that leaving the batteries to rest (without applying any load) gives the surface charge an opportunity to be absorbed. That description will give you a pretty idea of how technical I am!!
Now, I understand this is all theoretical and of little practical value, but . . . . if a known load is applied to discharge the batteries to 12.3v, and the batteries were left for another 12 hours with no load of any nature, would the measured voltage rise by a small amount? The reasoning is that absorbing / providing current is a chemical reaction that, once started, takes a while to slow down and stop?