Thought this might be interesting...
On another forum a photographer was having problems with a coloration in random parts of his images. He was thinking the shutter was hosed in his Nikon. With some further information it was surmised that a fast shutter speed was capturing the flicker of the flourescent lighting he was using.
Some types of lighting actually flickers, but our eyes kind of blend everything together so it looks continuous and of the same color. Fluorescent lighting uses a phosphor that when it is excited by electrons will illuminate. The voltage used to excite the phosphor is usually alternating current which is turning on and off continually, causing the light to flicker, but fast enough our eyes don't sense it. During the dips and swell of the flickering it's concievable the color temperature can shift. It's possible when the shutter of your camera sweeps across the sensor it captures a part of the flicker that is off color and it would be at random times in the image.
So the home experiment to validate this....(sort of like MythBusters, but safer and less messy)
A box of TMax was taped to a white foam board and placed under a flourescent 2 tube 40" fixture. Various shots were taken around 1/1250 to 1/2000. The results show that the fast shutter speed was capturing the dip/swell flicker of the illumination and the color temp as it changed. I intentionally changed orientation of the shots to show the shutter path. I now wish I had tried a few shots at slower speeds, which I believe the AWB might have handled close to correct for the entire image.
Note: There were some images that the coloration showed in other areas of the image as well.
Canon 50D w/50mm f/1.4 lens
Full EXIF embedded in images. ISO1600 for all shots.