Some mentioned decay of ...

Some mentioned decay of chloride with the tower blow-down closed.  Most water has at least some chloride in it already, and if under some heat load, the evaporation during the test interval will also increase the chloride further than the spiked amount.  It is a valid means for the calculation as long as the test interval is short, but if you already have something like we do ~ 250 ppm chloride in the source water, and are running 3-4 cycles of concentration on a system that is north of 250,000 gallons, you are going to need a considerable amount of salt to be able to read that.  I would suggest using a tracer that is virtually not present in the source water, and hold the level in the tower basin very constant during test interval.  Molybdate ion is usually a good choice because it also will enhance anodic corrosion reaction of steel, although you might actually lose some of the molybdate ion to passivating reactions, depending on how active corrosion is in the system.