David W. Fraser on “Vertically Twined Plateau Bags”

Dear folks –

On August 21, 2010, David W. Fraser

(David, in the center in the above photo, talks with his wife Barbara and Bruce Baganz, the president of the TM board)

gave a Rug and Textile Appreciation program here at The Textile Museum here in Washington, D. C., on the topic “Vertically Twined Plateau Bags.”

Wendel Swan, introduced David,

but digressed at first to indicate that there was a significant empty seat in the room.

He said Harold Keshishian,

who was instrumental in the  founding of these free Saturday programs in the 1960s and who was active in them ever since, both as a presenter and as a, sometimes irrepressible,  audience participant, had died recently, and so the chair at the side in which Harold often sat was unoccupied today.

Wendel said that he thought Harold knew more about a wider universe of rugs and textiles than anyone else he has known personally, and that Harold’s participation in RTAM sessions will be seriously missed.

Wendel said next that,” If Harold were in the room, he would be signaling  me, now, to ‘get on with it,’ with a ‘cut’ gesture across his throat, and in that spirit I need to turn to introduce David Fraser and today’s program.”

He said that David is trained as a medical doctor

and has had a distinguished career in medicine.  He headed the work to identify and deal with “Legionnaires’ disease,” and that demonstrating that the tampon was implicated in “Toxic Shock Syndrome.”

David has served as president of  both Swarthmore College and The Textile Museum Board, and continues to serve the TM as a Research Associate.

David is the author of the standard work on “weft-twined structures” and, with his wife Barbara, of a second  book on Chin tunics and other textiles, ethnic textiles with complex structures, woven in mainland southeast Asia.

They have done important field work in these areas, as well as in that of today’s topic, ‘Plateau bags’, a format twined by U.S. Indians in parts of the Columbia River basin.

(Ed.: In what follows I have drawn on both David’s TM presentation and on the article in American Indian Art Magazine in which the research on which David was drawing was also reported.)

David began by sketching some features of the “Plateau bag” format.

He said that they were functional bags, with front and back panels and top openings.

They vary in size, but are usually a little taller than they are wide.

( Ed.:  In the measurements I have seen as I prepared this post, the shortest Plateau bag I saw was 5.5 inches tall and 5 inches wide, and the tallest was 17 inches; its width was 13.5 inches.)

Mostly, but not always, the designs on the back are different from those on the front.  Here is the back of the Plateau bag above.

David said the the “Plateau” reference is to an actual geological plateau area in the northwestern U.S., between the Cascade and Rocky mountains near the Columbia River.

Plateau bags are sometimes described as “Nez Perce” bags, since Nez Perce did weave many of them.  But some Plateau bags were woven by members of other tribes.

David indicated that the Nez Perce, Umatilla, Yakama, Kliktat, Tenino, Walla Walla and Palouse also wove Plateau bags.

The map above indicates where the various tribal makers were located.  The one below indicates how the reservation areas were reduced.

David also showed some images of various Native American plateau bag makers.  If you look closely, all of these photos include Plateau bags in use.  Each also has a caption that you should be able to read.

Plateau bags are also sometimes referred to as “corn husk” bags, and David indicated that this, too, is a partial truth.  Corn husk was a frequent material, but Plateau bags can also contain Indian hemp, cotton, wool and even “bear grass.”

One of these photos shows a woman making a Plateau bag.

While some twining has been done on looms (twined Chilkat dancing blankets are made on warp-weighted looms), in the case of Plateau bags, no loom is necessary.

Look again at the photo above.  Notice that this lady holds the Plateau bag that she is making loosely on her lap and is twining it (moving left to right) working with the warps ends away from herself.  The warps on which she will continue to twine hang down in front of her.

David next moved to the various structures used in Plateau bags and how they were made.

As the photo below suggests, David is someone who often finds textile structures exciting, and he can make others feel his enthusiasm.

He used four Plateau bags in the room, that are part of the Textile Museum collection, to explain some of the basic variations in construction and weave.

First, is a Plateau bag was woven by the Nez Perce in about 1880.   It is made from Indian hemp, bear grass and corn husk.

(Ed. Esther Methe, the Textile Museum, Curator of Conservation, holds this boxed bag as David speaks to it)

Here is a little better look at its front.

And here is an image of its back from the Museum card.

The underlying twining is composed of two strands of Indian hemp, twisting each other as they interlace the warps.

The design is the result of the additional use of corn husk strips or bear grass in a “false embroidery” structure.

The drawing below shows how such a strand  of patterning corn husk is introduced into the underlying twining as it proceeds.

Notice in the drawing above that the patterning corn husk strand (striped) wraps the twining strands alternately, but only on the front side of the warps.

One result of this is that, when this specific structure is used, the design is visible on the outside of the bag, but does not show on the inside.

This structure is called “false embroidery” because in real embroidery the decorating thread is inserted after the weaving and moves between the front side and the back of a ground cloth (here the “ground” is the two twined cords).

The decorating thread is seen on both sides in actual embroidery.  The “embroidery” of the piece above is “false” in the sense that the corn husk strand decorates the front, but does not move to the back side of the warps.

Here, again, is one of the drawings above to make one additional point.

Notice that the striped cord COULD be taken around both cords when it wraps.  If  it did (and sometimes that variation is used) the design would be visible both inside and outside the bag.

David said that the  piece we have been discussing above (repeated here for convenience),

was made in the following way.  The twiner began at the bottom

with plain twining.   The work proceeded horizontally and spirals upward row by row.  There are no seams.

After several rows of plain twining, the maker shifted to false embroidery and introduced patterning, using bear grass and corn husk.  The weaver twined two different designs, one on the area of the bag that would be its “front,” and another on what would be its “back.”

The work spirals up shifting back to plain twining at the top.  The only opening is at the top.

Here is what the Textile Museum card on this piece says:


Cornhusk bag


Nez Perce

Purchase: Mary Elizabeth King, Washington, D.C.

Materials: dogbane fiber/cornhusk/bear grass

Construction:  tubular, no seams

Patterned with false embroidery in both yarn and fiber such as cornhusk or beargrass.  Some of these are darker than others, either dyed or slightly different in nature.  Plain twining at top and bottom.  Red “hourglass” motifs on one side.   Red zigzag motifs with brown, green, and blue dots on the other side.

David said that, while some species of weft twining are not difficult, his own experience with attempting the false embroidery variety suggests that that variation is.

One has to twine (often) with two basic strands that are of one material, say Indian hemp or cotton, and simultaneously do the false embroidery patterning using another strand of another material, say corn husk.  The  corn husk must not only be introduced so that it is visible on the front but not the back of the twining being worked on, it must also be kept moist as one works.

Here is David attempting a species of twining under the guidance of a famous, expert twiner, Irene Wyena Cloud.

Very often, he said, he felt like he didn’t have enough hands.

The next three examples of bags were also from the Textile Museum collection.  David said that they were made by members of the Wasco tribe, who are west of most of the tribes that make the Plateau bags.

The first of these was the piece below.

David said that this piece begins with twining at the bottom.  The fibers used are cotton and Indian hemp.  There is use of a patterning yarn but no false embroidery.  Design is seen both on the outside of this piece and the inside.

Here is what the Museum card says about this piece:




Possibly Wasco

Bequest of Irene Emory

Materials: fiber/cotton

Twined fiber and cotton

The second Wasco piece was this one.

My notes do not say much about this piece.  Only that indigo wool is used as a patterning strand on Indian hemp.

Here’s what the Textile Museum card on this piece says in part:

Twining:  pattern not of false embroidery but twining.  Bottom begins with crossed elements that radiate  out.  Design of striped lattice in blue.  Top bound with buckskin strip, thong drawstrings and handle.

A third piece is the one below.

Here is the bottom of this piece.

Here is the Textile Museum card on this piece.

A fourth  and last Textile Museum collection piece was the one below.

A little closer detail.

The aspects that strike one about this piece is the loose, open structure and the lack of any designs embellishing it.

Here is part of the Textile Museum’s card on it.

David had brought several Plateau bags and someone in the audience had one too.  Here is what they looked like.

We’ll label the one below, PB7

Comment on PB7: This bag is vertically twined, with a seam visible along the left side.  The false embroidery yarns are wool and cornhusk. The design includes prominent use of thin vertical lines

A closer look at a detail of this front view of PB7.  It’s PB7a.

Comment on PB7a: Note the steep diagonals and thin vertical lines in the pattern.

Here is the back of PB7

Comment on PB7 back view: Steep diagonals in the pattern are prominent on this side.

A closer look at a detail of this back.

PB7 back detail

The next Plateau bag in the room was the one below.


Comment on this side of PB8:  This bag is also vertically twined, with a single seam visible along the right side.  Relatively little of the surface is covered with false embroidery (done in cornhusk), suggesting that this utilitarian bag is of considerable age.  The four orange marks at the bottom have been tentatively identified as maker’s marks.

Here is the other side of PB8

PB8 other

Comment on PB8 other side:  Steep diagonals are prominent in the pattern on this side.

The next piece shown was the one below.


Comment on PB9:  This is a horizontally twined bag, with no seam.  It has been attributed to Ida Blackeagle, a well known Nez Perce weaver.

PB9 other side

Comment on PB9 other side:  The false embroidery is done with wool and cornhusk.  Note the shallow diagonals in the pattern.

The last Plateau bag shown in the room was the one below.


Comment on PB10: This lovely bag is horizontally twined.

PB10 other side

Now, although I don’t think most of us took it in, initially, David’s presentation to this point was actually context for the specific focus of his talk.

The expression “vertically twined” in his program title indicated that he intended to focus most sharply on a particular and quite small group of Plateau bags, that have interesting, even mysterious, features.

In fact one of David’s questions when he began his exploration was whether vertically-twined Plateau bags were a deliberately made Plateau bag format, rather than the result of re-working of some sort.

Reworked bags like the one below, exist,

so the question whether some plateau bags were deliberately made with a vertically twined structure is a real one.

But David has been able to demonstrate that vertically-twined Plateau bags were sometimes deliberately made and the results of his study were published in an article in the magazine American Indian Art.

The balance of David’s program was devoted to reporting key aspects of this research and article.

Here are the questions David established as he began his study of “Vertically Twined Plateau Bags:”

1.  Do they exist?

2.  How common are they?

3.  Who made them?

4.  Why were they made?

5.  How can they be detected in historic photos?

Let’s first notice the side seam that is a characteristic of  vertically twined Plateau bags.

The bag below has such a side seem.

VT21 (VT for “vertically twined)

A Plateau bag with twining rows aligned vertically cannot be created by twining in a continuous circle.  At the mouth (at the top of the bag) the twiner must turn back or end her row and start again in an adjacent row.  The work begins along one side and ends at the other side, where a gap is left that must be closed with a seam.

So one sign that vertically twined Plateau bags are a deliberately created format is the presence of this gap and the subsequent seam.

Another sign of deliberation is to demonstrate how the mouth can be created using vertically twining.

David said that the mouth in a vertically twined bag can be achieved in two different ways.  He provided a drawing.

The drawing on the left above shows how the twiner can move in one direction and then turn back at the mouth (shown at the left side in these diagrams) and twine back in the opposite direction.

The drawing on the right above shows the second way in which the opening that the mouth can be produced.  In this method, the twining begins circling one warp at the mouth lip and continues to the opposite lip where it terminates (with some terminating edge stitch not shown).  In this second method there is no turning back.  All of the twining moves in the same direction.  In both cases, the seam will be created at the top of these diagrams, after the last row of twining.

A second indication that vertically twined Plateau bags are a deliberately created format is the completeness of the designs on both sides.

David treated another characteristic of vertically twined Plateau bags that involves continuous twining (rather than stopping each row at the far lip of the bag mouth).  It is that the false embroidery is employed every other row (the intervening row is twined using the materials of the basic twining).  One advantage of using false embroidery every other row is that it is easier for a right-handed twiner to use false embroidery moving left to right.  The “skipped” row is twined right to left (or left to right but working with the inside of the bag facing the weaver) without the false embroidery.  This practice makes it possible for the twiner always to be twining the false embroidery rows moving left to right.

Here are some images of an incomplete Plateau bag in which (1) false embroidery is being used, and (2) every other row is done in plain twining.

You will be able to see the skipped rows as the images get closer.


Comment on VT1a:  This bag has cornhusk false embroidery on underlying twining of Indian hemp.  It is worked with the warp ends facing away from the weaver, as shown.


Comment on VT1b:  The mouth of the growing bag is shown on the left side of this image.


Comment on VT1c:  Notice in the photo above that one result of using false embroidery every other row is that only half of the surface in the areas in which it is used are covered with false embroidery.  The mouth edge is smooth, reflecting the fact that the Indian hemp wefts turn back each time they reach the mouth.  The latest row of twining (without false embroidery) has not yet reached this section of the bag.


Comment on VT1d:  The top row shown in this image is the last one to be worked.  It has been worked from right to left and has no false embroidery.

David said that having demonstrated that “vertically twined plateau bags” are a real Plateau bag variety, he would move to examine how often this format occurs.  The statistical term for this is “prevalence.”

In order to determine the prevalence of “vertically twined Plateau bags,” David (and, in the case of the Yakima Valley Museum, his wife Barbara) examined all available Plateau bags in the following collections:

o  American Museum of Natural History

o  University of Pennsylvania Museum of Archeology and Anthropology

o  Smithsonian’s National Museum of Natural History and the Museum of the American History

o  Yakima Valley Museum

o  Burke Museum

o  Natalie Linn’s The Plateau Bag

David analyzed a total of 432 Plateau bags.  Of these 19 were vertically twined.

Sixteen of these were excluded from the prevalence survey because they were in collections examined only because the collections were known to have vertically twined Plateau bags.   The 416 Plateau bags included in the prevalence study were from collections for which there was no prior knowledge of the presence in them of vertically twined bags.

David’s prevalence findings were:

o  Of the 416 Plateau bags included, only 5 bags, 1.2%, were vertically twined.  So although vertically twined Plateau bags are a deliberately created format, they are quite rare.

The next question was “who wove the ‘vertically twined Plateau bags?'”  This question entailed also answering the question  “who wove the ‘horizontally twined Plateau bags?'”  These questions are hard to answer because attributions of Plateau bags in museum collections are often incomplete or wrong.  In particular many such bags are reflexively attributed to the Nez  Perce.

David reported that the small universe of vertically twined plateau bags in collections other than local Plateau museums were attributed to twiners of the following tribes:

Nez Perce         5

Yakama              1

Umatilla            1

Cayuse                1

Unattributed 5

The much larger universe of horizontally twined bags were attributed to twiners of the following tribes:

Nez Perce           189

Yakama                  41

Umatilla                34

Cayuse                        1

Other                        37

Determining the provenance of given vertically twined plateau bags is made difficult because of:

1) the wide-spread tendency to attribute all plateau bags to the Nez Perce,

2) the fact that plateau bags were often given as gifts or used in trade, so the owner of a piece  may not be of the same tribe as the maker,

But David indicated that the thrust of his studies suggest that “…one, perhaps the, locus of vertically twined plateau bags was among the inter-linked tribes of the Wanapum, Walla Walla, Umatilla and Cayuse of southeastern Washington and across the border in northeastern Oregon…”

David’s next question was “Why were vertically twined Plateau bags made?”

He said that it is easier to think of reasons to twine a Plateau bag horizontally than it is to pinpoint reasons for doing so vertically.  For example, “…the seamlessness of the horizontally twined bag makes it stronger and less likely to lose its contents.”  David also cited reasons why a horizontally twined bag is easier to make.

But his study also identified some possible advantages of vertical twining.

First, David said, “Twining a bag vertically does create a neater mouth…”  Many earlier vertically twined Plateau bags are done with false embroidery every other row and, the continuous weft rows of this usage, makes for a neater mouth.  David conjectured that early, more utilitarian Plateau bags may have required a neater mouth.

But at the turn of the twentieth century, vertically twined Plateau bags are mostly made with false embroidery on adjacent rows and discontinuous wefts, so their mouths are not so neat and other reasons for twining vertically are needed.

David said that older vertically twined Plateau bags had simpler designs, but that those of the twentieth century became more complex.  So perhaps this increased complexity of design provided a reason for using vertical twining.

To explore this possibility David asked two questions  of the information he had collected on the plateau bags in his study universe.

These were:

o  How frequently do twined Plateau bags with thin vertical lines in the designs on their faces turn out to be vertically twined?

o  How frequently do twined Plateau bags with steep diagonals in their designs turn out to be vertically twined?

Beginning with the first of these questions David examined a preliminary question: what proportion of horizontally twined bags had thin vertical lines on their faces?

Here is an example bag face with thin vertical lines.

The by-tribe results were:

o  Nez Perce                          4.3%

0 Yakama                               6.1%

o  Umatilla                           16.4%

o  Other specific tribes  1.5%

o  No specific tribe            7.0%

So thin vertical lines in the designs of horizontally twined Plateau bags is fairly infrequent, but may be more common among the Umatilla.

David, further, found that the overall prevalence of thin vertical lines was 73% on faces of vertically twined bags and 6% on faces of horizontally twined ones.  So the association between the direction of the twining in a bag’s face and the orientation  of thin lines in its pattern is “very strong.”

The preference of the Umatilla tribe for designs with thin vertical lines suggests that this tribe may have had a design-sourced reason for using vertical twining.

David now moved to the question how frequently do twined plateau bags with steep diagonals in their designs turn out to be vertically twined?

In general, David said, Plateau bags tend to have a higher density of weft rows per inch than of warps per inch.  “…Because the diagonal goes out more quickly that it goes up,” the diagonals in designs on most Plateau bags are relatively shallow.

So David checked 423 twined Plateau bags to see what proportions of vertically twined vs. horizontally twined face diagonals  would be pitched more or less than 45 degrees to the horizontal.

He found that among horizontally twined faces

with measurable diagonals 99.2 % had angles relative to the horizontal of less than 45%.

“Among vertically worked bags, ten of the sixteen had angles relative to the horizontal of greater than 45 degrees.

He corrected this finding for examples in which the degree of steepness is affected by moving up only every other row,

and found that in general “a weaver wishing to create designs that have smooth steep diagonals would tend to favor vertical twining, whereas a weaver who wished to make designs with shallow diagonals would favor twining horizontally.”

The last issue that David treated was that of how one might identify vertically twined Plateau bags in historical photographs.

Of course, the presence of a side seam is nearly conclusive visual evidence of vertical twining, and the use of thin vertical lines and of steep diagonals in face designs also indicate Plateau bags that are likely vertically twined.

But also David noticed that Plateau bag designs woven with horizontal twining tended to twist up and to the left

and that those twined vertically often twist up and to the right.

His specific investigative question was:

o How frequently do twined Plateau bags in which the pattern(s) in the designs twist up and to the right turn out to be vertically twined?

“Of the 250 bags examined in which there was evidence of pattern twist, that twist was up and to the left in 98.7 percent of horizontally twined bags, and up and to the right in 92.9 percent when the twining was done vertically.”

David applied prevalence tests and concluded that “the most valuable clues in historic photographs regarding vertical twining (other than seeing the side seam) are likely to be the presence of steep diagonals and twisting of the pattern up and to the right.  Thin vertical lines offer additional evidence of vertical twining but are sufficiently frequent in horizontally twined bags to be less conclusive.”

David summarized his conclusions,

took questions,

and adjourned the session.

The Textile Museum examples, and others that had been brought in, were placed on tables in the Myers Room

and an adjoining one

for closer (no touching of the Museum pieces) examination.

People moved to see this material close-up, and the questions and conversation continued.

I want to thank David and his wife Barbara for sharing this fine piece of work with us, and for permitting me to fashion this virtual version of it.

Thanks to them, too, for some alert editing that saved me from reportorial error.

I hope you have enjoyed this look at a good example of the sort of empirical research that is still within the range of serious students of textiles and weaving.


R. John Howe

Comments are closed.

%d bloggers like this: