The Ship Model Forum

The reason why I discovered your forum this evening was because I started wondering what color the Pearl Harbor ships were, which of course led me here & elsewhere, and to Tracy White's & others excellent research. I did some reading & would like to offer the following points to consider.

There seems to be a debate as to whether the Pearl Harbor 5-D was a neutral or blueish gray, and much of debate centers on Munsell data. I'd like to take a deeper look at the Munsell data here:http://www.shipcamouflage.com/pearl_harbor_experiments.htm

Specifically the
Dark Gray 5-D, one coat 5PB 2.7/0.8 6.0
&
Dark Gray 5-D, two coats 5PB 2.8/0.8 6.6 from a 1941 report

I thought I had seen everything until I had seen this. What makes it so unusual is that they are Munsell notations apparently derived from spectral data prior to 1943, which is when the Munsell spacing was first mathematically revised to allow this to be done. What makes it more interesting is that the reflectance values don't correspond to either the old or revised Munsell Value tables. But somehow, the NBS physicists came up with these Munsell notations which would have corresponded to a pre 1953 Munsell Book of Color. OK. Of course, if someone can find the original data, either the xyY, XYZ, the curve plot, or a table of reflectance vs. wavelength, we would have a definitive answer and would need not worry about old Munsell notations. But till then,

In the Journal of the Optical Society of America, Volume 43, Number 3, March 1953, the old Munsell chip 5PB3/1 was measured & converted to the new system as 5.1PB2.98/1.3, which means if you want to see the closest chip to the old Munsell measurement in todays Book of Color, take a look at the 5PB3/1 chip. This is a lucky case where the old & new chips are roughly the same. Although the old chip measured out at a chroma of /1.3, remember we were starting with a chroma of /0.8, so subtracting 0.2 from the 1.3 leaves us with 1.1 which is close enough to a Munsell /1.0 chip.

But really, I think both camps are going to be right. It is a nearly neutral. It still has some blue in it, & I bet it behaves quite differently in true sun as opposed to the daylight lamp I'm using now. But as it is still dark, will need to wait a few hours for that comparison.

OK it was late last night. Now the report makes more sense. The NBS guys, if I had to guess, Dean Judd & Kenneth Kelley, used instrumentation for the reflectance only, which is why it is unique. That means the the Munsell notation was derived from the conventional interpolation/extrapolation method used at the time. That means they had the 5-D paint applied to a swatch, found that it matched closest to the 1929 (5)PB3/1 chip, and used their own method (which could have been a visual estimate or a more precise mathematical approach) to declare the Value lower (2.7-2.8 ) and the Chroma lower (0.8 ). Looking at the current Munsell 5PB3/1 chip would be replicating their experiment. Luckily the chroma works out so you wouldn't have to use imagination to make it more lower, but for the greatest accuracy, keep in mind that the chip's Value 3/ is slightly too high & you need to imagine it lower by .25ish. Luckily, Munsell sells a separate fan deck for the Value scale where the grey chips are broken down into .25 scale increments.

To put that all together, look at the current 5PB3/1 chip. This will correctly tell you the amount of chroma (or blue) in the paint. However, the chip will be slightly lighter than the 5-D paint swatch the NBS scientists were looking at. To get an appreciation for the difference, get a Munsell Neutrals fan deck & look at the N3.0 vs. N2.75 chips. Not much of a difference, but for the truly curious.

Next question, what was the gloss level of the paint?

regards,

Mark

Oh one more thing, when looking at dark chips like these, the proper technique is a mask. To analyze the dark chips, it's best to use black as opposed to grey masks, and a three dimensional mask is preferred. That is, instead of using black paper & cutting a hole it in, go to a crafts store & get a sheet of black velvet craft paper. Make a mask out of that instead. The 3D nature of the black velvet (the hairiness) creates a totally dark surrounding for the chip, thus the features hue, value & chroma can be more finely studied

No, don't look at the current chip--look at the chip in the 1929 Munsell Book of Color. The two are completely different, and 5-D was based on the 1929 chip.

_________________
Cheers,
John Snyder
Shady Grove Farm
Shady Grove Farm on Facebook: https://www.facebook.com/ShadyGroveDuckEggs

Well surprise to us both. They never published the /1 chroma chips in either the 1929 BoC or the 1942 reprint. The left most column of chips on every page isn't a /1 column, it's the /0 neutral column. It's labeled as such in the 1929 edition. They forgot to label the column in the 1942 reprint. So it's left to mathematics on this problem.

Based on the 5PB3/1 target chip, I took some measurements today of some chips that people may have in their homes that can help bracket in the color.

595B 35042 fandeck 0.98PB2.74/0.93
595B 36076 fandeck 4.44PB3.37/1.16
595C 35042 fandeck 0.56PB2.76/0.89
595C 36076 fandeck 4.92PB3.30/1.17
Snyder & Short 5-D Dark Gray 2.58PB2.69/0.54

The FS595 chips give a good representation of the chroma suggested by the 1941 NBS report


Here's a graphic some might be familiar with. Goes back to when the NBS was trying to use Munsell for naming colors.

Thanks for your posts... my head hurts!

_________________
Tracy White -Researcher@Large

"Let the evidence guide the research. Do not have a preconceived agenda which will only distort the result."
-Barbara Tuchman

You and me both!

_________________
Martin

"Tomorrow is the most important thing in life. Comes into us at midnight very clean. It's perfect when it arrives and it puts itself in our hands. It hopes we've learned something from yesterday." John Wayne

Ship Model Gallery

What I tried to do was take the 1941 Munsell notation & convert it into a modern notation. It's a tricky adventure to do that, but in this case fairly straightforward. Turning into a modern notation means we can use modern chips for comparison, as long as we know the Munsell notations of those as well.

The NBS guys would have been using the 1929 Book of Color supplemented by special chips that they had between them & Munsell. I checked tonight & indeed the /1 chromas were not included in either the 1929 or 1942 reprints, but were in a special collection called 'Odd Chromas' which were for special research study. The 'Old Munsell System' 5PB3/1 chip was measured not only in 1953 as I mentioned above, but I just saw it in a 1943 paper as well. (JOSA again). This time it measured 5.51PB3.11/1.22 in modern Munsellese, a little less chroma than the chip measured in 1953. It's better to use the 1943 measurement as a guide since it is closer to our NBS report date of 1941...most likely they were looking at this chip as reference. Again, subtracting 0.2 off the chroma, we are down to 1.0 in modern chroma.

Going back to old Munsell now, I found a chart dated 1939 & it was prepared by the NBS & published in Research Report 1239.* Recalling the chroma in the Munsell notation was /0.8, you can see it falls in the 'Dark Bluish Gray' box'. Not by a lot mind you...see that the range is from 0.5 to 1.5, so it's not the poster child for the dark bluish gray color, but there was probably just enough blue to know it's there.



I'm sure this was intentional, and it's the same kind of thing that causes a lot of arguments in the olive drab world too. By making these grayish colors borderline chromatic, the designers were creating a color the behaved differently under different lighting conditions, direct sun vs. overcast for example.

Just one more thing, the units of error for Munsell at this portion of the spectrum are Hue:1.25 , Value: 0.1 , Chroma: 0.15

So Value is the most sensitive attribute, followed by chroma, and hue much less so.

*the chart I posted yesterday corresponds to modern Munsellese

OK, now that I've had a bit of a chance to re-read this...

As far as my research has shown, there were two 5-D Dark Grays. There was a "pure" one, and one that was made by adding a conversion paste to stocks of the pre-war light gray. The Navy's goal was to use up the existing stocks of prewar gray by converting it first, and THEN shipping out the pure 5-D. The #5 Standard Navy Gray used prewar had a bit of blue to it, but I don't know enough about Munsell and color measuring to know if it would be strong enough to influence the results in the NBS tests

5-D and 5-D conversion paste formulas memo
5-D manufacturer difficulties memo

I have NOT seen anything that indicates what ship got what supplies, so this is one of the question marks with 5-D. The other question would be which formula NBS was measuring. I would presume the "permanent" and pure 5-D, but it's not a presumption I would place any confidence in.

_________________
Tracy White -Researcher@Large

"Let the evidence guide the research. Do not have a preconceived agenda which will only distort the result."
-Barbara Tuchman

Again, I know nothing of USN colors but I'll throw this out there for the more learned. A few years ago when I bought a 1942 Munsell Book of Color, there was a type written page that was placed into the binder* just in front of the 5PB page. In addition, there was a 5PB page included from a more current BoC, I'm guessing vintage 1970s. This 1942 set was from a library, and I'm going to call the book seller next week to see if he recalls which one, because they might have more resources that led to the creation of this typewritten page. I do recall him mentioning something about the Navy, but nothing else. Well, here is the page:



*not a trivial thing on the 1929/42 binder. the person really really wanted to have this included

I find the idea of Pearl harbor ships being painted in 5PB2/2 a tough cookie to swallow. The /.8 is a lot more believable to me

If I understand you right, you are saying that priority was given to take existing stock of 5PB7/2 and convert it to mutant 5-D & distribute that before commencing with pure 5-D. And the primary tinting agent was lampblack. It would make sense to get a 2.7/0.8 from the addition of black to 7/2. The addition of black would destroy both the value & chroma, but the lightness at a much faster rate.

If they had really wanted to get rid of the blue to get pure achromatic gray, the way to do that is to add the color on the opposite of the color wheel, this case being yellow. An ochre pigment would likely have been added if they had wanted to totally neutralize the blue. My guess is that the 2.7/0.8 was the mutant 5-D & the 2/2 was the pure 5-D which wasn't used. Does that make sense to you? Just a best guess at this point.

Looking at the NBS report again,

Light Gray 5-L 2PB 6.7/0.4
Dark Gray 5-D, one coat 5PB 2.7/0.8

vs. the Munsell notations on my sheet & seen elsewhere

Light Gray 5-L 5PB 7/2
Dark Gray 5-D 5PB 2/2

a) I like the NBS report a whole lot more because they are analyzing the actual paint, not the paint chips. And the NBS colors seem a whole lot more believable to me

b) The NBS 5-D is likely the 'pure' 5-D, not mutant, because if we assume that the 5-L is a 'pure' 5-L base, there is no way you can go from a 0.4 chroma to a 0.8 chroma by adding lampblack. That can't happen. Chroma can't be increased from adding black or white

c) The difference between the NBS measured chromas & the more commonly seen /2 stated chromas makes me suspect the /2 standards. I wonder if whoever chose /2 did so just because that was the closest chip in the book (remember /1 didn't exist) and all these /2 notations we see are indeed bogus.

I mean there is just no way somebody can call the 5PB2/2 a gray color. In modern munsell that comes out to even higher /2.3

"a) I like the NBS report a whole lot more because they are analyzing the actual paint, not the paint chips."

No, they were analyzing a particular batch run of paint produced at one of two manufacturing facilities. Is that significant?

Actually Russ, they might not have been - Mare Island and Norfolk were the paint manufacturing yards, but Philadelphia was the paint formulation yard (for lack of a proper term - they were the facility that new paints were developed and initially tested.

Sorry for being off the thread for a bit.. the last part of the week was ... bleah.



I don't really like being pedantic, but to keep my head from exploding too much I'm going to say, "no, priority was givin to convert stocks of #5 Standard Navy Gray." I'd just prefer to refer to the paint as the paint and the color of the paint separately, because we do have the possibility of differences

That said, from the documents I've posted, the 1941 camouflage changes are typically thought of introducing new colors, but the bigger change was that they switched to a synthetic alkyd resin formula. As your supplemental sheet shows, they Dark and Ocean Gray color values existed as an ideal before 1941... I see them mentioned in some earlier camouflage experiments. But the Navy was trying to come up with new camouflage using better, more durable paints, so it was more of a "kill two birds with one stone." There exists the possibility of remanufacturing paints to a certain degree, but I doubt it was economical to take the pre-war formulas and break them down enough to use what they could with the alkyd resins.

So, we see in January 1941 the Bureau of ships states:



Later on, In February, they issue the Formulas for both 5-D and a "5-D (Interim)" to just the paint manufacturing yards; note that even 5-D is a two part "white base plus tinting paste" formula as far as the paint manufacturing yards are concerned, but that it is to be issued to forces afloat as a complete paint (I.e., not two parts mixed dock-side or on ship):



I haven't seen anything else on 5-D (Interim), but I haven't hit the BuShips files personally yet. We see the fleet start to report their paint stocks and that it's not an insignificant amount. There is discussion of the "conversion paste," also referred to as "#5 B.P." (I presume B.P. is Black paste, but I'm running out of time this morning to write this and am going to let that go for now) in April from Mare Island that is also relevant:



So herein lies the problem. Was the Pacific fleet done with the converted (aka "mutant") 5-D before manufacture of all 5-D was ordered halted? Was *any* pure 5-D even applied to any of the ships? On one hand, we have a late April order that seems to hint that they might not have been. Paragraphs 3 and 6 hint that the first issue of the "new" paints will be converted/mutant. Paragraph 5 orders commanding officers to requisition new paint after August 1st and THEN turn in remaining stocks of converted/mutant 5-D. Since manufacture of 5-D was ordered halted on July 30, there's a possibility that most of the ships never saw application of the pure 5-D.



While it would have been "nice" for them to measure both, I doubt that the Navy or NBS would really care too much what an interim, cost-saving paint really matched. So I agree, the 5-D measured by the NBS is most likely the "pure."

_________________
Tracy White -Researcher@Large

"Let the evidence guide the research. Do not have a preconceived agenda which will only distort the result."
-Barbara Tuchman

I don't know the reason they made the measurements, there might be more documents we are not seeing which explain the circumstances.

But the measurements themselves are significant. The NBS had a limited color staff, only 3 scientists I'm aware of and some technicians. Plus Dorothy Nickerson over at the USDA. These were absolutely the A Team for color measurement in the world and their time was extremely valuable. They weren't in the business of making routine batch measurements. Even the B Team consisted of a few academics scattered across the USA. These people in the early 1940s had a lot on their plate which included mathematically redefining the entire Munsell color space which they unveiled in 1943, and they squeezed in support requests from the services when they could. Keep in mind precise color measurement is still in it's infancy at this time. An academic exercise. Whatever they're doing has substantial purpose to it

Doing precise Munsell notations from visual data was a tedious task only about a half dozen people in the world could do reliably. It wasn't a routine thing. And 1939-42 is when the spectrophotometer is just getting known and would eventually replace visual Munsell notations, but it also never became routine. Color measurement never really became routine until computers came to be and came down in price which wasn't that long ago. I could probably find out a true date, but my gut is telling me 1990ish. Even now a decent bare bones portable is $5K.

Getting back to visual Munsell, there were two ways to do it. The typical novice way was to simply match your color to a chip in the Book of Color & say that it looks like so & so. So at least you can communicate to somebody what your color kind of looks like.

Precise visual Munsell notations required plotting on large table size charts, doing extrapolations & interpolations rigorously between the hues, values & chromas. very tedious.

The best data is of course machine generated from the General Electric Recording Spectrophotometer, of which only a few even existed in 1941. It was invented just a few year prior I think by a prof at MIT, and some data was generated on these in WW2, but it becomes more common after. With that data, you can get precise Munsell notations which work with a modern Book of Color, so that' really cool. But even that machine was rare, and needed trained (expensive) personnel to run which plagued it really up until computerization. The best data ever derived from these machines is back from WW2 when the A Team was using them. As technicians became more prevalent in the 50-60's data can get spooky at times.