AVG Tomahawks - Again!

I see that the question of the undersurface colour of the AVG Tomahawks is riding the merry-go-round at Hyperscale yet again! In this case the presumption of a light gray and the request for a colour value for it initiated the discussion. That is one of the problems. The "light gray" is not so clearly identified when you peel back the onion skin, so how was it arrived at in the first place?

The origin lies in Terrill Clements' study of AVG colours, quoted elsewhere and re-affirmed by Tom Tullis in his 'Tigers Over China' book  (and recently in an article in Model/Military Aircraft Monthly (MAM) which appears to be derived from this). Unfortunately at the time these studies were written the appearance of 71-021 was unknown. Mr Tullis states in his book, and it is repeated in MAM, that "While the RAF used a greenish-blue color called "Sky Type S" for their fighters, there wasn't any DuPont color that closely matched it." This is incorrect. Du Pont 71-021 was designated as "Sky Type S Grey" and was shown in a contemporary Du Pont colour card to be a close match to AM Sky. The Du Pont 71-021 swatch was provided to British contractors responsible for supporting Curtiss P-40 aircraft in the UK via the DTD from Curtiss themselves and is an actual sample of paint. Why some researchers should accept and state that specific Du Pont colours were used on the upper surfaces of the Tomahawks but then vehemently refute that Du Pont's AM Sky equivalent was used on the under surfaces is beyond this writer.

Both authors speculate on the identity of a "light gray" colour from selected colour photographs, then choosing a USN colour 'Aircraft Gray' as the closest (later ANA 512 'Glossy Aircraft Gray' reported to be the same as or similar to FS 16473).  The colour photographs used for this identification were stated to be "the best" but it remains uncertain how that criteria was applied. The best at showing a gray rather than something more like Sky?  The tendency to "correct" colour photographs also throws up significant imponderables. In terms of established colour science attempting to identify precise colours from the visual assessment of colour photographs is fraught with unreliability. In Mr Tullis' book there are as many colour photographs of AVG Tomahawks showing colours that appear to resemble AM Sky as there are that are that appear to show grey - if not more. Both authors appear to have ignored the probability of the paint surface oxidising and "chalking" in the tropical environment to appear more greyish to the eye.

Again it is worth emphasising that light gray was the anomaly, the deviation; the AM wanted Sky, the factory drawings specified 71-021 (or "duck egg blue"). And apparently apart from those aircraft delivered to the AVG all the other Tomahawks arrived in under surface colours which were not remarked on at the time as being different or incorrect to specification, but rather were described by contemporary observers as "duck egg green" and in at least one RAF Squadron ORB as "grey green". So there should be pretty robust grounds for determining that a light gray was applied instead. Are there?

The White Ensign Models Colourcoats paint WEMCC US03 for the USN Navy colour 5L Light Grey has been cited in comments here as a good match for Du Pont 71-021 and a Floquil hobby paint of this colour is also recommended by Bruce Archer at Hyperscale. Bruce assets that "a Lt grey, similar to the FAA's 'Sky Grey'" was the colour applied to the undersides of the AVG Tomahawks but again provides no compelling evidence for this statement.

Steven Eisenmann asserts that the underside colour "definately (sic) was not Sky as we know it, probably anywhere from a light gray to a bluish gray. Who ever got them the paint the cheapest." That use of "definitely" makes it a bold statement but logic also suggests that, unless Steven was standing notebook in hand as each Tomahawk rolled out of the paint shop, that the "cheapest" paint may equally have been more like the required Sky than light gray. The proposition that anything went, that Curtiss were able to procure any old paint for the RAF order must, if accepted, raise as much doubt about the result being a light gray as anything else. Again the evidence for this statement is not provided but perhaps depends on the Curtiss paint contract records mentioned in posts by Bruce (which identify paint names from other suppliers including reference to "grays") but so far unpublished (and without any identified or matched colour values). AFAIK these records have not been cross-referenced to a specific chronology, dates or production batch serials. 

The unequivocacy of these statements and the determination to make very firm conclusions about the appearance of colour from digital images of colour photographs is intriguing. The use of Du Pont 71-021 has come under robust challenge - quite rightly, as peer review and critique is an integral stage of any serious research - but it is perplexing that the same rigourous challenge has not been applied to the "Sky Gray Theory"; some of the proponents of the theory appear to be unaware of this double standard or choose to ignore it. Why is that? Is it perhaps a case of the man being played rather than the ball?

I haven't checked out the Floquil or White Ensign Models Colourcoats paints for this colour but I was able to get some more information about the appearance of 5L from http://www.steelnavy.com/usnchips. There it is identified as being the light blue-grey of Munsell 10 B 7/1.5 whereas Du Pont 71-021 is closer to 10 GY 8/1. The DE2000 difference calculation between these colours is a significant 11.7 where a calculation of 2.0 or less is the equivalent to a close match. Viewed in juxtaposition to the 5L colour 71-021 looks even closer to the required and specified Sky. I'm not about to write a multi-page treatise here on metamerism but those sufficiently interested can explore the phenomenum via Google and there are also numerous demonstrations online to show that what we think we are seeing we are not!

  

The Steel Navy site also mentions that the approximate FS 595B equivalent to 5L is 36320, but the DE2000 calculation for the Munsell value to FS value difference is 8.25 (the author does note that the FS numbers are estimates based on his [visual] comparison of color chips to FS chips). The closest FS equivalent to 71-021 is actually 25622 @ 1.50 (so it is very close). Steven seems determined to assess 71-021 as a "bluish gray" even though he has never seen the actual paint. In colour value terms it is not a "bluish gray" but rather a paler, slightly more blue version of AM Sky (the clue is in the Munsell designator GY = Green Yellow!), although Steven's visual perception is quite understandable as one amongst many possible perceptions. Some people describe the colour as a pale green, others as a pale blue or blue-grey and some just see grey. But visual perception and assessment of colour are not the same as colour science or the identification of the colour values by measurement.

The closest FS 595B equivalent to the 5L Munsell value is actually FS 36375, but not very close @ 3.95 and darker and greyer than the bluish-grey of that value. FS 36473 is further away @ 5.67. In comparison to 71-021 the appearance of FS 36473 is significantly darker and greyer @ 10.4 and compared to FS 25622 even further away @ 11.5. Assuming Steel Navy's identification of 5L is correct those who may believe that 5L and FS 36473 are interchangeable as Tomahawk under surface colours are plain wrong, even before we consider 71-021, so what does that say about the theory?

Although this preliminary exploration of the colours suggests that 5L is not the same or even a similiar colour as 71-021 and/or FS 36473, the hobby paints for 5L may be closer. I shall check those out in due course.