At work we have some MH lights that run 24/7, and others that get shut off every night. I shut the 24/7 lights off last night and when I started them again this morning, two of them during warmup started rapidly flashing pink.  What would cause this, just a settling of the salts?  The lamps that get shut of regularly never do this, so I wondered if this was something to do with them being on all the time. It has been almost a month since I have been able to shut them off.

I've observed both MH and MV through a welders glass while flashing pink during warm-up. It appears very small particles of debris comes loose and momentarily get caught in the arcstream. My theory on the lamps you mention is that since they are run constantly, there was a bunch of particles that were waiting to be knocked loose during the next start-up, which is the most traumatic part of the lamp's operating cycle. Only a theory of course.

It's called "flaring" and it happens frequently with metal halide lamps during warm up (or if they are bumped while in operation - mercury lamps might do it too). Some of my lamps do it...some of the time...others don't ever seem to. I've seen it occur on all types, from 35w to 1500w lamps.

Yeah I've come to determine that it is something to do with these 2 lamps.  I cycled them off again last night, after only being on constantly for one night and those 2 did it again this morning.  It is about a minute into warmup when it color shifts from blueish light to the warm white it is supposed to be. I believe they are Sylvania lamps.

I think it's normal, i sometimes see a color "disco" after power-up my HCI-TS70W/WDL. It's only sometimes, mostly it is in the order: incandescent white (only electrodes glow) -> bluish white (mercury) -> green (mercury + ??) -> final "830" white (all components), while each stage takes 5..10seconds.

It happens (my explanation), when halides change their layered order (moving with the lamp not yet cool, so they are still liquid), so then they evaporates irregularly (as the lower boiling point component is below the higher-boiling one, so it bubbles trough), so it creates "clouds" of different composition. And when the inner atmosphere circulates, the arc change it's color, depend on what "cloud" reached it. After a while, when it stabilizes and all fill gasses fully mix up, it settle the color and then burn normally.