[old_tv_nut]

That article is missing the original blue phosphor in the 15GP22, the NTSC specified one, which was less violet-blue than the ZnS:Ag. Can't recall the formula, off the top of my head, but it was used because they had trouble with copper contamination turning the blue sulfide phosphor green. The alternate got baked into the NTSC specs, then was replaced in tubes by ZnS:Ag at some point.

I have wondered for a long time if the blue change occurred in the middle of the 21AXP22 run time, or at the start, or later. The RCA triniscope had the more violet blue, and it seems likely that would have been the proposed NTSC spec if they hadn't run into the copper contamination problem. There is also some indication of a possible typo in some documents specifying the x,y color coordinates of the NTSC blue. Whether the final spec is the correct one or the typo isn't clear.

By studying the RCA TV schematics from year to year, you can see the point at which they changed the color difference matrix to correct for the non-standard phosphors. Prior to that point, they used by-the-book color difference processing. Doing that with the more violet blue phosphor shifted yellows towards green. That, combined with the very cyan white point used in receivers and the sensitivity of the TK-41 dichroic filters to polarized light, caused a significant probability of blond hair being rendered greenish, while blues were rendered more violet.