Posted by Steve Price on 01-01-2005 05:17 PM:

How reliable is the "educated eye test" for dyes?

Hi People

Nearly everyone who collects or sells antique rugs believes he/she can distinguish natural from synthetic dyes by eye. I think it would be fairly simple to present a large number of such people with samples of rugs, ask them to identify the synthetic dyes, and test all of the dyes by thin-layer chromatography. The object would be to determine just how reliable their eyes actually are. Does anyone know if such a test (not necessarily this specific one, of course) has ever been carried out? It would be useful to know just how good "educated eyes" are at dye analysis.


Steve Price

Posted by Richard Farber on 01-02-2005 04:10 AM:

dear steve,

if such a test were carried out . . . and all of the participants failed it will still not be a positive indication that some people are not able to differentiate between colors according to the origin of the dye stuff . . .

i know quite a bit about musical hearing and there individuals who can perform amazing feats that are quite on the order of one in a million people or one in ten million. i would assume that this is also true of the scene of sight.

in any case such a test would depend on which dyes are studied -- synthetics of indigo are pretty much the same as natural indigo except for the impurities . . . it is the radically different synthetics that are relatively easy to discern


richard farber

Posted by R. John Howe on 01-02-2005 08:54 AM:

Dear folks -

While I think that while the logic of Richard Farber's argument above is sound, the pressure of his conclusion moves in a direction different from the one I personally think is best for the messy business of rug collecting.

The possibility that there might be some who can recognize "synthetic dyes" is a bit like the claim that there are some who can recognize aesthetically superior rugs. I could give more credence to the possibility of this claim if those who claim it would also state (at least hypothetically) both the test situation in which it could be demonstrated and the outcome that would make them conclude that it is in error. Richard's post seems to resist such a test and especially such a conclusion.

For me, the better answer to Steve's question resides in a story I've told here before.

One morning, at a TM rug program, the then head of the Christie's rug department, Jim Ffrench, I believe, brought out a Caucasian rug with bold, jarring, rather unpleasant colors that many in the room would see as very likely the result of synthetic dyes. He said, "I took this rug in for auction recently and only as a favor to a friend. I don't like it for the same reason, likely that you don't, its colors." He then produced a Turkish silk fragment that he said had some age that had in it many of the same colors. He said, "I don't like this piece either for the same reason, but the inconvient thing is that all of the colors in this Turkish piece have been chemically tested and are the result of the use of natural dyes."

He then, held that is was likely that a far wider range of shades can be and were produced with natural dyes than those we usually consider to be within the natural dye palette. He said, "It seems at least likely that we need to be re-educated about the range of colors that can in fact be produced with natural dyes. We are likely currently often excluding perfectly collectible pieces incorrectly on the suspicion of their having synthetic dyes."

My own answer and sense of the correction needed with regard to Steve's question is that we should likely mostly give up on the possibility that any of us can visually recognize synthetic dyes accurately and acknowledge that the sounder posture in this arena is to be less certain about which pieces should be passed over on that basis without chemical test.


R. John Howe

Posted by Steve Price on 01-02-2005 09:23 AM:

Hi John and Richard

I don't doubt for a moment that some "educated eyes" are better than others, but I doubt that any are perfect and would be interested in knowing the extent of the imperfections in those who are, let's say, better than most.

I'm familiar with the fact that a much wider range of color can be produced with natural dyes than those usually found in antique rugs. It would be good to know how often such colors of natural origin were used in antique rugs.

Identification of a palette as all natural dyes or partially synthetic is probably the most commonly used criterion for estimating age of a rug. The real issue, the one that has consequences to rug collectors and dealers alike, is how reliable this criterion for judging age really is. To my knowledge, it hasn't been directly addressed (although I have a vague recollection of Paul Mushak comparing laboratory analyses with eyeball judgments, perhaps someone remembers the source and details).

I do disagree with John's proposal that the "educated eye" test should be abandoned altogether. First, because it is not likely to happen - it's too firmly set in the culture to be tossed out without a very good reason to do so. Second, because it may be very reliable, indeed; we just don't know at the moment. My guess is that it's about as reliable (or unreliable) as any other criterion that we apply.


Steve Price

Posted by R. John Howe on 01-02-2005 06:03 PM:

Hi Steve -

I don't think I quite said "abandon." I think my proposal was to acknowledge the uncertainty of the visual estimates we currently make.

Its result, I think, would be to increase the consideration of some pieces often quickly scorned because they have some shades that are outside the "natural dye" palette into which we have been socialized.


R. John Howe

Posted by Steve Price on 01-02-2005 08:06 PM:

Hi John

I misunderstood your post; I apologize. I agree that we should remain aware that the reliability of the "educated eye test" for identifying synthetic dyes is not yet documented.


Steve Price

Posted by Filiberto Boncompagni on 01-03-2005 07:56 AM:

It all comes to this: we know how good natural dyes look because we have examples of pre-1875 rugs in museums, private collections and so on.
We also know how bad some synthetic dyes - often applied on machine spun wool – look because we have plenty of modern rugs around.
We obviously prefer the natural ones so, when presented with some rugs, we identify the good-looking, vibrant colors with natural dyes and the flat or too harsh ones with synthetic dyes.

But, as John demonstrated in his example above, even natural dyes can look wrong.

In the archived discussion “dyes and ethnographic value” they were presented a couple of modern Anatolian yastiks with natural dyes on hand-spun wool.

In my opinion, the colors looked awfully harsh. Perhaps they will mellow with time, of course, but as they are now, I don’t like them at all.

On the other hand, synthetic dyes, if well applied, can produce excellent colors…

And here is a sort of example about that:
Mr. Horst Nitz sent me, some time ago, the scan of an old article from Hali (Vol I No 3 /1978 pp. 281-283 “The Dyes Of Turkoman Rugs” by Mark Whiting).
The author, a chemist and rug collector, chemically examined a number of Turkoman rugs from his collection. In short, here is his conclusion:

“If my contribution had to be compressed in a single sentence, it would be that more Turkoman rugs contain synthetic dyes, and therefore were woven after 1880, that is commonly imagined. Yet, many such pieces retain the same craftsmanship, fine design, good colour harmonies, and beauty that early Turkoman possess.
I regard many of them as valuable components of my collection.”

Meditate, folks, on the words “good colour harmonies” … Are those good looking colors in your rugs really all natural?


Posted by Horst Nitz on 01-04-2005 03:52 PM:


The „ educated eye test“ - brilliant term - beats them all.

Reasonable reliability, always at hand wherever you go, instantly applied, immediate outcome, no mess, no weight to carry, no cost, no auxiliaries, high input-output ratio with regard to training effort (efficacy), works upside down … what else is needed?

Optimum performance may require control of some variables: light, age range of textile, degree of being accustomed with type or provenance etc.

For special applications Thin-Layer-Chromatography (TLC) comes in: research interest, scientific substantiation of expert opinion etc. High reliability and objectivity is achievable but demands for a laboratory and a sophisticated reference library with regard to running properties of known dyes in various media if differential dye analysis is required.



Posted by Steve Price on 01-04-2005 04:14 PM:

Hi Horst

I believe it was the late George O'Bannon who coined the term, "educated eye test" for dye analysis.

It has all of the good properties that you mention, although nobody really knows what "reasonable reliability" means here. People bet serious money on it, rather unusual behavior when the risk of being mistaken is unknown. It suggests to me that they put an awful lot of confidence in it being reliable. I hope they (we) are right.


Steve Price

Posted by Horst Nitz on 01-08-2005 06:37 PM:

Hi Steve

„Reasonable reliability“ with regard to the discrimination of natural vs. synthetic dyes is to make no or very few mistakes under normal conditions and with senses aware. Judging and buying rugs should be no betting game - not outside eBay. Misjudgements in the case of rugs that have been tampered with belong to the latter category, i.e. having been bleached, artificially aged, painted etc.

I have taken chances at times and still do so. But if this happens I am usually aware that I am trying to bridge gaps of knowledge. So far I always reached the other shore, on occasions with wet feet. In those instances I surrender to the quotation by Mark Whiting in Hali, Vol I, No 3 pp 281-283, Filiberto has relayed for us in an earlier post to this thread. Although I prefer rugs with natural dyes, I am not obsessed with them. There are good tribal rugs with a few synthetic dyes on them around, and sometimes that is all that seems to be left of a particular tribal group’s heritage.

As to research into this issue, I don’t know of any. It is a complex matter. Researching complex questions calls for big samples and all that goes with it. I’d suggest an alternative approach by which the issues are taken one-by-one: testing a meaningful hypothesis like “high positive correlation in dye-discrimination can be expected between training input and performance with regard to early 20th century North-West Persian rugs ((specific and known palette of colours allows for save assessment))” - or something along similar lines.

Bye for now,