Excellent question. I'm only going to address the Ca test in this response (it is late, and I am very tired). The Ca test uses a hydroxide to precipitate the Mg and then titrates the remaining Ca. Problem is, some Ca also gets precipitated as calcium hydroxide (kalk). When the test gets to the endpoint, the solution gets to this state where the equilibrium changes, and now because the titrant has bound up all the Ca in solution, the precipitated kalk wants to re-dissolve. But, this takes a little bit of time to happen. Because of this, the Ca test should not be considered complete until the change to the pure blue color persists for a few minutes.
While we're talking about this, there were a couple of clever oceanographic chemists (Pate and Robinson) back in the 1950s who published a paper about how to get better accuracy and precision out of this test with a simple change to the method. The improved method requires you to already have a reasonably accurate idea of what your Ca level is. The idea is this: instead of adding the hydroxide first, you add about 90-95% of the amount of titrant necessary to get to the endpoint first. Only then do you add the hydroxide to precipitate the Mg. What happens if you do this is that the titrant binds to the Ca more stongly than the hydroxide wants to, so much much less Ca gets precipitated as kalk. This gives a much more accurate and precise Ca reading.
If you don't know what your Ca reading is, then do a "pilot" test the normal way, adding the hydroxide first, and then do a second reading using the "95% titrant first" trick.
