Since Frank King was there literally on the day the first Yellowstone was delivered, and he described the boiler jacketing as "Russian Iron" rather than some shade of green - and later in his book references the engine's gray boiler color - would lead me to believe the boilers never were green.
You know, this isn't a case where one reference says one thing and one says another, and nobody knows which is correct. There are dozens - maybe hundreds? - of references in books, magazines, DVD narration, etc. of DMIR engines in general and the Yellowstones specifically having gray boilers.
On the other hand, there is one website which gives an undocumented comment that the DMIR Yellowstones were delivered with dark green boilers. Maybe that's true, but I'd like to see some actual documentation / proof before I just accept it as a previously-unknown fact.
I do not have the information handy but I once modeled the Colorado NG before the turn of the Century (1800's) and many engines were referred to as being delivered with Russian Iron boiler jackets. There was a huge discussion and at least one printout discussing this material. I believe it came from Bob Richardson of Colorado RR Museum fame.
Apparently it was something in the manufacturing process that could cause the metal to have a bluish or greenish cast to it. It depended on the process used and also the light it was viewed under. On a bright and clear day it might look more green than blue or vice-versa.
As has been pointed out so many times in the past here and other forums an individuals opinion on color is something of question. Perhaps the Yellowstones (and other engines) viewed by different people on different days and, I'm sure, at different ages could be any or all of these mentioned colors/shades?
I'm sure the maintenance on the original jackets was costly to keep up and the jackets may have been replaced by plain old sheet steel over the years and a shop painted them gray or green?
As a custom painter I have painted numerous DM&IR and T&P engines where the customer insisted on a gray or green boiler. Maybe they were or weren't?
oldline1
Here is an entry point to source material:
http://www.pacificng.com/template.php?page=/mr/russiairon/index.htm
If this does not give you the short page, click the small 'Model Building' tab to the side on this (or any) Pacific NG page and then 'Simulating Russia Iron'.
This has multiple references linked from one spot so is a better page to start; you can pull up the individual references and bookmark them from there.
The short link to the long (14-page) reference of accessible Internet sources is
http://www.pacificng.com/pdf/web/viewer.html?file=/ref/russiairon/pdf/SteveC_Research_Russia_Iron.pdf
and the companion file he references on planished iron.
It is not too difficult for modern metallurgists to understand why the early American processes would give a planished iron as thin as the Russian process, and visually similar, but lack the long-term high-temperature-stable weatherproof characteristics that made the material desirable for boiler clothing in an age before good high-temperature paint and good lagging.
Those with an interest in metallurgy will immediately note that 'oxidation' as we understand it now is not the sole operative mechanism (that is what 'bluing' is) and that the mottling in case-hardening is anomalous, but with nitriding added to carbiding.