Carbon Dating for Andean Textiles

On February 27, 2010, Amy Gould and Matthew Polk

gave a Rug and Textile Appreciation Morning program on “Carbon Dating for Andean Textiles.”

Maryclaire Ramsey, the CEO of The Textile Museum, introduced Amy and Matthew,

saying that

Note: We will say a little more about this last item at the end of this virtual version of their program.

Amy began with “A Short History of Andean Textiles.”

She started with the first Andean textile that she and Matthew bought together.

She described this piece as a painted and resist-dyed Chancay textile with interlocking images that seem to have anticipated MC Escher by almost 1,000 years.

Amy then sketched Andean textile history with a few examples making the point that we can often determine the age of a textile quite accurately simply by looking at the design.

Her first piece was from the  Chavin culture (1,000 to 400 B.C.).

The menacing faces, fangs and claws of Chavin art are unmistakable and when we see them we can say with confidence that the textile is amongst the very earliest decorative Andean textiles.  Some Chavin images can be quite unsettling  as in the  piece above.  It is carbon dated to about 500 B.C.

On the South Coast of Peru Paracas culture followed Chavin and produced textiles with distinctive surreal floating figures.

As we look at the other-worldly “floating shamans” in this  Paracas textile, it is easy to start believing in Aliens.  Or, at least that these people had access to some very potent drugs.  Yet this powerful and distinctive style tells us both how old the piece is and where it is from.   This textile is carbon dated to about 50 B.C.

Moving into the Common Era we have  the piece below.

The “proliferous” style graphically combines images of eyes and mouths to create figures within figures which places it in the proto Nasca era. This  piece is carbon dated to about 280 A.D.

A classic “temple step” design suggests the piece below is solidly in the Nasca period.  It is carbon dated to about 320 AD, only slightly later than the previous piece .

Next we have the “jagged edged serpents and jaguars” of the Moche, carbon dated to about 820 AD.

The “war-like abstractions” of the Huari are instantly recognizable in the “Skull” tunic below and of about the same age as the Moche piece above.

The Chancay civilization returns us to graphic drug-induced visions as in this powerful painted piece below, estimated to have been woven around 1000 AD.

Amy, now drew attention to a ceremonial ensemble from the Chimu civilization.

The ghostly elegance of these finely woven gauze pieces is uniquely Chimu and places these pieces in the range of 1200 to 1400 A.D.

The Inca are the last of the pre-conquest civilizations and brought “corporate uniformity” and incredibly fine tapestry weave to the design of their garments.

This classic Inca “Checkerboard” tunic is an example of what was being worn by high ranking Inca’s at the time of the conquest.  We can be highly confident that it dates to the first half of the 16th century.

However, design and technique are often not enough to tell us much about the age of a textile.  Below are two textiles, both tie dyed and both with elements that make them tough to pigeon hole.   One might think they are of similar age, but which one is really older?

The piece above has colors that could be identified as Nasca or Huari.   It has been carbon dated to about 650 AD.

The second piece is a four color tie-dye loosely woven in the Nasca style.  But this one has been carbon dated to 450 B.C more than 1,000 years older than the first tie-dyed piece.

Amy concluded her part of the presentation with the observation that while graphic design and analysis of technique have helped build a general framework for the Andean textile time-line, carbon dating is often needed to correctly place individual pieces within that framework.

Amy then introduced her husband, Matthew Polk, to speak about the radio-carbon-dating technique, especially as applied to a series of Aymara men’s tunics and women’s Iscayo’s in their collection. (See “Andean Textiles”, Hali #162 and “Valley of the Spirits”, Alan Kolata, 1996 for more information on the Aymara and their textiles.)

Amy and Matthew had brought seven Andean textiles, 5 Tunics and 2 Iscayo’s (womans mantle) They were arrayed (with an ID number on each one) on forms in the front of the room.

Matthew said that all the pieces were warp faced Andean textiles, believed to be Aymara or Aymara-related and all had been carbon dated.

The Aymara, he said, are people who, to this day, live at very high altitudes mainly on the Andean Altiplano that includes parts of Peru, Bolivia and Chile.  The origins of the Aymara are not known, but they were known to be troublesome to the Inca who forcibly  resettled many Aymara to make them  easier to control.

The Aymara practiced a form of ancestor worship which involved many rituals which may seem quite odd to us.  The desiccated remains of ancestors were periodically disinterred and dressed up in carefully preserved heirloom textiles for the purpose of advising the living.  The Spanish discouraged such practices as they converted the indigenous population to Catholicism.   But many of the rituals continue to be practiced amongst the reclusive Aymara to this day.   The tradition of heirloom textiles has come to take the place of digging up dead ancestors and the pieces preserved in many villages may be of surprising antiquity.

Matthew  noted that the care and reverence applied in the making of these apparently simple textiles is quite remarkable and not immediately apparent.

For example, in completing the piece the warps are never cut off the looms.  Instead the piece is carefully removed from the loom with the warp loops intact and painstakingly finished off-loom to create a complete textile which is in effect a single continuous sacred piece of thread.

The sophistication and mastery of the art are also evident in the subtle techniques employed.  Alternating s and z spun warps make monochromatic fields come alive without ever introducing a change of color.

The development of the Aymara aesthetic was also influenced by the Spanish.   The emergence of large scale silver mining in Bolivia in the early 17th century led the Aymara to follow the Spanish custom of including  silver threads in their finest garments.

Matthew’s presentation was participative.

He said he would try to give us a good short course understanding of the radiocarbon dating technique and would discuss in detail the radiocarbon results obtained for these seven textiles.  He also equipped the members of the audience with a matrix form with which to make their own rank-order estimates of their relative ages.

He pointed out that this is not a false task.

That “while Andean decorative textiles span over 3,ooo years…the powerful and distinct iconography associated with each of the Andean civilizations is often enough to tell us to which civilization a textile belongs and approximately how old it is.”

And “when design alone is not enough, technique and materials can often place a piece in time.”

So he invited us to estimate, based on whatever indicators we could employ, the relative age of each of these seven Aymara textiles.

The complexity of this task soon became clear and Matthew quickly suggested a simpler but no less daunting challenge.

The new task was to identify (only) which of these pieces seemed oldest to us.   (Many of those who arrived early had had an opportunity to handle these pieces and examine them closely before the presentation.) Matthew then facilitated our choices by working through the pieces one at a time and asking for a show of hands for those who thought that piece might be the oldest of the group.

In each case he:

1)  provided some general information about the piece such as likely geographical origin, where collected if known, etc.

2) commented on some of its visible  physical and tactile characteristics such as design, weaving techniques and handle.

3)  looked to see how many in the audience felt that particular piece was the oldest.


You can test your own sense of  which of these pieces seems oldest to you, by taking a sheet of paper and numbering from 1 to 7 and then taking a few notes on what follows, before making your own choice. (You will also get a chance to look back once Matthew’s more conventional review of them has been completed.)

I am going to give you several images of each piece, plus Matthew’s indications about it.

Piece #1 was on the far left and looked like this.

Matthew said that this is a “green-gray-blue Iscayito – a woman’s outer garment woven in two parts.  Somewhat loosely spun yarns and a nice floppy handle”

Here are some additional images of textile #1.  The first, below, is how it looked on the form.

Here are some closer details of it.

Tunic #2 is the one below.

Matthew said that number 2 is a “frayed striped man’s tunic (really old looking) with some small blue supplementary yarns said to indicate the family ownership.”.

Here is how it looked on the form.

And again some closer details of it.

Tunic #3 was the one below.

Matthew describe this piece as “a pink and black striped man’s tunic collected in Puno on the shores of Lake Titicacca in the mid 1970’s.”  He said “the stains are from chicha, an indigenous beer-like drink that has been produced at least since Inca times.” “tone on tone stripes within the pink areas are created by alternate groups of z and s spun warps.  Very finely spun yarns.”

Here are some additional views of tunic 3.

Tunic #4 looked like this.

Matthew described this piece as “a very fine multi-colored striped man’s tunic collected in Bolivia, 1984.”

Here is how it looked on the form.

And again, some additional detail images.

Below is an Isacayo #5 in this series.

Matthew described this piece as “a pink-brown woman’s iscayo – silver threads are included at the edge and in the center.” “very finely spun and woven, a wonderful patina of age and surprisingly light, kashmir-like handle.”

Here is how #5 looked on the form.

And some closer detail images.

Tunic #6 looked like this.

Matthew said that this is a “black-violet tunic/poncho, woven in two parts, sides joined only at the lower corners.” He commented specifically that what appears to be a solid black field is in reality a variegated violet.

Here is how it looked on the form.

Here are some closer details in which I was trying to capture some of the purple highlights in this piece.

My monitor shows a purple cast in the image above.

I don’t think I see them in this second detail.

Matthew said that the purple aspects of this piece are often only visible under the type of full sunlight found in the Andean highlands and he took this piece outside after the program to let it be seen in that light.

I took four additional images of piece 6 outside.

Here they are below, without intervening comment.

On my monitor, the purple cast of piece #6 IS enhanced by the bright natural light.

Tunic #7 looked like this.

Matthew gave this description: “A red-blue-green striped man’s tunic woven in two parts, possibly archeological but in virtually perfect condition.”  He commented on how the adjacent green and blue stripes are so close in hue that they trick the eye into seeing a third color.  (Look at the detail images below to see the blue and green next to each other then look back at the full scale image to see this illusion.)

Here is how tunic #7 looked on the form.

And here are some additional detail images of it.

This ended Matthew’s review of the seven pieces in terms of more traditional indicators of age.

He now asked the audience to look back over the seven examples and to pick by show of hands the one they thought was oldest.

Note: If you are interested in “playing” this game, this is the point at which you should (without looking ahead) make your own selection.

Because you can look back and have close-up images, you are, in fact, more favorably placed to use conventional indicators of age than were those in the audience.

Now not nearly everyone responded to Matthew’s “which seems oldest of these seven pieces?” task, but here is the result of those who did:

Now that he had our likely ignorance captured, Matthew turned to explain a bit about carbon dating and some of the questions that frequently surround it.

He said that radiocarbon dating was invented in the 1940s, but that only since the early 1990s has it been possible to use it with small samples.  This has increased its usefulness greatly.

The scientific concepts underlying radiocarbon dating go something like this:

BUT, in fact the “miracle” is not exactly what we have hoped for.  Matthew provided the slide above again with notations on the right side about what actually is the case.

This leads some, Matthew said, to ask whether carbon dating is, in fact, completely useless.

Matthew said that although carbon dating does not necessarily deliver the precise indisputable results we might want, the results are still very helpful.  He then explained what is behind the results that the labs deliver.

In spite of the many possible errors in radiocarbon dating we can get useful results by matching the c-14 results we get from our sample to a calibration curve.  The calibration curve is a collection of c-14 results from samples where we already know the age.  Tree ring samples alone have given us a calibration curve  that extends back 10,000 years.  And, the accuracy of this calibration curve is being constantly refined.

In addition, modern AMS radiocarbon dating can be done with less than 10 mg of material which is more than 50 times less that what was required 25 years ago.

In general, the older the sample the more useful the c14 method is in giving an indication of age.    And, the method is absolutely limited to samples older than 1950 due to the changes in atmospheric carbon isotopes produced by atomic bomb testing after that date.

As we’ll discuss in a moment, C-14 testing often produces ambiguous results because the results from our sample may match the calibration curve in several places.  However, these results may also be useful when combined with more traditional observation based methods.  I’ll give an example of this in a moment.

With this background on carbon dating and its limitations, Matthew described how they use it.  He said that there are a number of labs that do carbon dating and that he and Amy have some that they think provide particularly accurate and reliable results.  He said that it costs about $400 now to get a sample from a textile carbon dated and that this cost is usually achieved by submitting a number of samples together.

So you send your sample(s) off and when the testing has been completed this is what you get back.

Let’s look at this report on a more readable basis.  First, an overall look at the report in the image below (some print too small to read).

But you CAN see that the report has two main parts.

At the top on the left there is an underlined title and a blue arrow that points to a result for the “Conventional Radiocarbon Age” of this piece.  The arrow points to the actual values reported.

In the lower part of this report is a statement on the left indicating that these are the “calibrated results.”  Two arrows point right, one to the numerical “calibrated results,” and the other to a graphic presentation of them.

Here, below, is a slightly larger version of this same report.

Before we try to understand what these results actually mean, here’s an even closer view of the “conventional radiocarbon age” reported at the form’s upper part.

You can see that this part of the report indicates that the “conventional radiocarbon age” (CRA)  for piece NZA 325 15 is 420 plus or minus 25 years BP.

This means that if we ignore all those variations and possible errors in the c14 curve the conventional radiocarbon age or CRA would be 420 years +/- 25 before present (BP).  The “present” is always defined as 1950.   So, the uncalibrated age for this piece is 1950 minus 420 = 1530 AD +/- 25.

Now to the lower part of the report.  Note the two blue arrows originate (you can see it in the overall image above) in the designation “calibrated results.”

Here is a closer image of the bottom part of this report with the two arrows.

The top arrow points to a highlighted result indicating that, for this same sample, the “calibrated” result is a “91.4% confidence interval from 1431 AD to 1496 AD.”

The bottom arrow points to the graphing of that result clustering around the year 1450 AD.  What this means is that with a 91.4% probability the age of this piece is within the shaded area from 1431 AD to 1496 AD.  Notice that this is about 100 years older than the uncalibrated age.   The calibrated age is always more accurate than the CRA.   In this case the calibrated age falls within a fairly narrow range and shows that this particular Aymara piece is pre-conquest.

BUT, Matthew cautioned, we are not always so lucky with our results.  He gave the following example.

Now you can’t read much of the two reports below, but that’s not necessary here.  Just compare the character of the two graphic results at the bottom.

Notice, as the labeling says, that the one on the left is the “lucky” report, since a date within a narrow range is indicated.  But the report on the right is decidedly unlucky, since any date all along the shaded area is possible.

Now, having armed us for understanding radiocarbon dating a bit, Matthew moved back to the seven Aymara pieces.

He reminded us of the results of our conventional estimates of which of these seven tunics was oldest,

then like a good “Academy Awards” presenter, he opened the “envelope” to show the ages that radiocarbon dating has actually assigned to these seven pieces.

It turns out that tunic 6, the black-purple one is the oldest,

with a calibrated 94% probability of being between 966 AD and 1210 AD, even older than pieces 1, 4 and 7 each of which are dated before 1500 A.D.

And five of those in the audience who made estimates DID chose #6 as the oldest of these tunics.

The largest number in the audience (8) selected tunic #3 as the oldest,

However, it turns out that this is the second youngest of the group, having calibrated results that waver between “before 1700” and “before 1800.”

No one in the audience thought pieces 1 or 2 were the oldest. The frayed tunic #2 turns out to be quite respectably old with a calibrated 70% probability between 1515 and 1599 AD and a 25% probability between 1618 and 1649 AD.  But #1, the gray green Iscayo is the second oldest of the group with a calibrated 95% probability range from 1409 to 1480 AD.

This exercise was great fun.  Matthew thanked everyone for being  good sports and observed that the task was very difficult.  Few people are familiar with Aymara material and the styles and techniques of their textiles have  not changed much in a thousand years.  However this points out how we can use radiocarbon dating in difficult situations to “second guess” the conclusions that we might reach from conventional indicators and human intuition.

In particular, Matthew said, this exercise suggests some conclusions:

Matthew had one last wrinkle about their experience with carbon dating to share with us.

He put up these two reports and asked whether could they be for the same piece?

Here are somewhat larger images of these two reports.

Here is a larger image the report on the left.  Look down at the graph at the bottom.

You can see that this report provides a carbon dating age pretty close to 1450 AD.

Now here is the carbon dating report on the right above.  Again, look down at the graph.

These results spread a little more, but the calibrated age is between 118o AD and 1262 AD, a result that diverges by 200-300 years from those of the report on the left.

Well, Matthew said, it turns out that these two reports are, in fact, from samples taken from a single piece.  This one.

Here are some additional images (some closer) of this tunic.

Looking for possible explanations of this divergent carbon dating result on this piece, one naturally gravitates toward its center patch.

And, in fact, one tested sample was taken from this patch and the other was taken from elsewhere in the tunic.  The sample from the patch tested considerably older.

Matthew gave the following explanation:

So, in this case when our human intuition told us something was not right carbon dating allowed us to uncover the true story.

Amy and Matthew then took questions

and brought their session to an end with one last indication.

They said that they have created “The Historic Textile Research Foundation (HTRF) dedicated to building an on-line database of radiocarbon dated textiles.

HTRF is a 501-c3 foundation.  Its long-term goal is to build a database of several thousand dated textiles for use by scholars, collectors or anyone interested in the topic.

They indicated that if anyone has any radiocarbon dated textiles, they would love to be able to publish your data in the HTRF database.  If you have data you would like to share please contact:

Matthew Polk

Amy Gould

Historic Textile Research Foundation

PO Box 226

Gibson Island, MD 21056


Amy and Matthew adjourned their session and the fulsome crowd surged forward.

Matthew’s sister Capie, a docent at the TM, models the pink and silver thread Iscayo.  (see below)

I want to thank Amy and Matthew for permitting me to construct this virtual version of their very informative and enjoyable RTAM session.  Thanks to them also for the considerable technical and editorial assistance they provided after their session.

My thanks, additionally, to Amy Rispin for another excellent set of notes.

Amy (Gould) and Matthew have provided carbon dating reports on the seven tunics in their program.  I have included them here after my signature below without further comment.

I hope you have enjoyed this solidly grounded, interesting session in this virtual version.


R. John Howe

Carbon dating reports below:


Report on Gary Green Iscayo # 1

Report on Frayed Striped Tunic #2

Report on Pink and Black Striped Tunic #3

Report on Multi-Colored Striped Tunic #4

Report on Pink and Silver Thread Iscayo #5

Report on Black-Violet Tunic #6

Report on Red, Blue-Green and Natural Striped Tunic #7


Comments are closed.

%d bloggers like this: