It is also noted by some people who have measured PAR for their light, that bulbs with different color temperatures, 6500K and 10000K for example, produce different PAR. For now I think it is hard enough to characterize what a given light fixture produces in PAR vs distance, that the finer points, like color temperature, will just have to wait a few years. The difference in PAR from a 10000K and a 6500K bulb is probably within the uncertainty range for most lights. And, all of this data is from lights operated out in the air, not in a water filled tank, where the difference is also probably within the uncertainty range for most lights.
Tonight I was measuring PAR from a screw-in CFL bulb in a dome type reflector. I noticed that on turning on the light the PAR slowly increased for a minute or two, then even more slowly dropped part of the way back. It was an old bulb, so this may be a problem with the bulb's built in ballast. Imagine the error this could introduce if you didn't allow for it.