Pattern puzzle: Scarf pt.2

We had some very good guesses in response to yesterday’s challenge. Without further ado, here is the scarf laid flat:


The designer of record is Vionnet -who else? I found this item courtesy of the Metropolitan Museum. This is how they describe the piece:

Madeleine Vionnet was a consummate technician, particularly known for her innovative use of the bias cut, mathematically precise construction, and use of geometric forms. Often interpreted in a rectangular form, Vionnet instead renders the outline of three stepped quadrilateral forms–2 squares and a central parallelogram–in a bias cut to give this scarf a highly unique shape. Its irregular form, with only one seam on each side, is an example of her economy of cut. The abstract shape, textural fagotting and bright color combination, reversed on the opposite side, give the piece a lively presence.

I don’t know that I’d agree with “economy of cut” considering the yield and how else would it be sewn down the center but with one seam -but that’s beside the point. Of course I had to try to recreate it, which was much easier said than done. Below is the pattern I came up with:


It’s not exact but pretty close.

[And now that I’ve made it rather easy to pick off and re-scale, I have little doubt this scarf pattern will soon be featured in our favorite home sewing magazines, on our favorite home sewing websites and the pattern for sale by various home sewing vendors. Without attribution of course. I should probably beat them to the punch.]

For one scarf, the yield is not so good but the waste isn’t as bad in a marker. In idle thought I wonder if it is possible to engineer the interchanging parts of this in such a way as to make it interlock more readily. For me, that’s the fun part. I suppose I can tackle it if struck by a bout of insomnia…

Get New Posts by Email


  1. Grace says:

    RE “economy of cut”
    Am I missing something here?

    Each scarf takes 4 pieces, 2 of each color.
    The pieces on each side are the same and
    the ones on the reverse are mirror images.

    You can nest the pieces and then stack the double
    nested unit.

    If you cut both the top and bottom pieces while nested like this,
    you can’t use them for the same scarf because they have C2
    rotational symmetry, but no reflection plane.

    But, you can use the nested pieces on different scarves.
    Just cut an equal number of pieces that are mirror images of them.

    If you are mass producing the scarves, this would be very

  2. Kathleen says:

    I made a marker. I’m showing 54.44% utilization on 60″ wide fabric. I suppose “economy” is relative. For us, that’s abysmal.

  3. Kathleen says:

    You know, there is one weird sewing thing that would be helpful to know and is not obvious that would be good to know in making this scarf. Nobody taught it to me, I figured it out from recreating so many Vionnet drafts (I have no doubt it was done in her shop). Threads likes weird sewing things particularly if couched in absolutes. As in, you will die as will all of your progeny if you don’t do X thing in exactly X way.

  4. Quincunx says:

    It IS two-dimensional. I’m disappointed.

    However, if you can tesselate it without regard to a grain line, does the yield of the marker improve?

  5. Quincunx says:

    Argh. Once again, posting without benefit of tea is a bad idea. I’m not doing this to inflate my comment count, I swear. . .

    Tesselate it without regard to the width of the pattern as given, but with respect to the fabric. Whatever the width of the fabric is, that is the length from one parallel end to the other parallel end, not the length from point to point.

    For an imaginary, selvage-free, nice round numbers for demonstration’s sake bolt of 60″ fabric, the finished (imaginary, no turn of cloth allowance or trailing decimals that break the decimal-to-fraction converter) scarf length from point to point would be a bit over 72″.

  6. Is there an inset between the two main colors? Looked like there was in yesterday’s photos, and maybe here. Both pieces could be sewn to a narrow ribbon before turning.

    Lots of pivoting when sewing, but straight lines.

    Is the back plain (ha, a coffin scarf)?

    Could we have more on this, please? Specifically making the pattern yield more efficient (cut in pairs, sure), and construction details. The original must have been slip-stitched by hand at the end.

    Vionnet’s clientele was into conspicuous consumption. I love taking gorgeous concepts like this and figuring how they could be done efficiently.

  7. LeslieW says:

    I am wondering if the inset is actually the “textural fagotting” from the description? And I think economy of cut is strictly cut only, not economy of fabric. Interesting.
    Nice job on the pattern Kathleen, looks very close to me. Why not submit yesterdays post to Threads? You could have a new monthly article with them called “pattern puzzles”.

  8. Julia says:

    Actually, the folks at our extension center have been teaching a one-color zig-zag scarf much like this one for years and years…

  9. Quincunx says:

    I don’t see an inset. What I see is the zig-zag topstitching, in black, wide, not quite close enough to be called satin stitch (less than 20 spi), and presumably done over both seam allowances so as to hold the whole thing flat. It’s clearer in yesterday’s photo, for sure. You can even see, on the right side just as the scarf starts to twist and throw the shadow of the turquoise ripple onto the purple, where it isn’t zig-zagged–guess that’s where the scarf was turned inside-out? and it isn’t on the same side as today’s photo, that’s for sure.

  10. Kathleen says:

    I sent Grace a screen capture of the marker with the hope she’d see something I don’t. I still plan to play with it some more.

    sdBev/Theresa: in my copious free time!

    Samantha: It’s not really 2D (x plane being however high two fabric thicknesses are) but I know what you mean. But your point teases me into wondering if/how it is possible to create a more defined 3rd dimension.

    I’m still not sure if stacking the pieces with respect to fabric width as opposed to grain (such as it is, it was a guess on my part) would improve yield much. It is one of those things where the pattern needs to be re-engineered to those constraints. Again, another lovely idea to mull as I fall asleep at night. I can tell you tho that 60″ wide goods isn’t optimal, there’s waste along one edge. Narrower goods would improve the yield.

    Carol/Donna/Sam: that’s a faggoting stitch. Much more evident if you follow the link to the Met’s site to see the full size image.

    Donna: I don’t think a “whole new career” is optimal for me. Mostly it’s a case of “not my customer”. I do plan to sell patterns but I need to figure out how to set the product apart from what’s already there. Pricing can resolve much of that. In short, I won’t be selling $10, $15 or even $20 patterns. Or if I do, only rarely. I used to sell patterns 15 years ago. Even 15 years ago, my least expensive pattern was $25, the most expensive was $45. I had to shut it down because demand was much greater than I anticipated and I wasn’t prepared for it.

    Julia: From the context of this post, it’s very clear I didn’t invent this. I’ve also seen zig zag scarves, the difference being that the zig zag is along the outside edge. Sewing wise, it’s not even in the same category as the Vionnet scarf. Outside edge zig zag is a snap to sew. Vionnet’s…not so much.

  11. Brina says:

    If you look closely at the image of the lain out scarf on the Met page–it enlarges quite a bit–you’ll see seams within the purple area on the right. One is from the edge at top right to the edge at bottom right–parallel to the faggoted edge–but some inches to the right. There looks like another on the underside judging from what looks like the ghost of seam allowance, that crosses the seam mentioned above, running, to the right, from the first angle of faggoted seam from the top to the outside edge. If we more of these to look at it’s be interesting to see if the seams were in different places.

    I can’t image that there was any machine sewing on this. And likely Vionnet worked with 35-36″, or the equivalent in metric, wide fabric. The faggoting is handworked and the small squares of fabric inset at the corners would be difficult to pull off with a machine, not to mention all the outside edges are on the bias. Vionnet used to weight the fabric that she cut on the bias in order to have a more stable material and make it easier to manipulate.

  12. Brina says:

    FWIW, I think the person who wrote the description on the Met site means by “economy of cut” that there are very few pieces to assemble, not that the pattern for this would result in an economical yield. So more art oriented than production oriented.

  13. I am gobsmacked.

    The highest-magnification views of the scarf on the mannequin look now like one piece of purple with drawn-thread work, with a piece of turquoise seamed to each side. You can see the tiny overcast stitches along the cut edges of the purple squares, and the very slightly lighter (silk?) thread that gathers each set of floating stitches that make the faggoting.

    The outside seams could be done by machine, but the inner ones would be by hand. This is consistent with couture work of that era.

  14. dosfashionistas says:

    I’m still thinking about that Threads article. It would be fun to do a rif on this with shaped bias scarves cut up in different ways. In what way did this shape of scarf enhance the drape around the neck? What else could you do? There are possibilities….

  15. Larissa says:

    Thanks for showing us this piece I have never seen it, despite being a huge Vionnet fan. It’s truly gorgeous. I may have to make a shibori silk version for myself. :)

  16. Matthew Pius says:

    Wow. Like Quincunx, I initially thought the “piecing” seam was zig-zag topstitched. Then, after reading the later comments, I followed the link to the Met’s site and zoomed in to the max. Not only is it drawn-thread work, but it’s done so that you see through both layers – meaning either it was assembled first and then both purple layers got the drawn-thread work together or it was matched up meticulously. This also means that the purple and turquoise pieces are not cut from the same pattern – the purple pieces must be slightly larger along the piecing seam. What Carol had originally thought was an inset is part of the purple piece.

    All that fascinating stuff came from trying to see where the grain goes. RE: yield – these could stack together efficiently if the short ends are on-grain, which would mean all parts of the piecing seam are on true bias. HOWEVER, the actual scarf is cut so that that outside edges are on the bias. The piecing seam must be on-grain to do that drawn work. I assume this is why Kathleen reports such a low utilization on her marker.

    It might be an interesting experiment to do 2 mock-ups (without the fancy drawn work seam) experimenting with the 2 different ways to use the bias and see what sort of difference that gives you in the drape.

  17. Kathleen says:

    The piecing seam must be on-grain to do that drawn work. I assume this is why Kathleen reports such a low utilization on her marker.

    Precisely. There is no way to do that drawn thread work if the center seam was not on 90 degree angles.

    We are discussing this at length here in the forum. Teijo posted a test marker of a variation of this style (very intriguing!). I posted the marker I mentioned above. I also just finished posting a series, step by step pattern modifications that could potentially increase yield to something on the order of 100%….

  18. Jennifer says:

    I think the “economy” refers to the fact that the two fabrics are initially cut identically. Cut the scarf from one layer of turquoise and one layer of purple. Place them together and sew half the perimeter, from one point of the lightning bolt to the other. Then turn the seam and align the remaining raw edges. Sew the second half of the perimeter. Then cut the lightning bolt through the turquoise layer. Turn the second seam and there is a scarf with finished perimeter but an exposed raw turquoise edge on both sides which line up exactly. Finish the raw edges of the turquoise while adding the faggotting to the purple. The faggotting is centered on the lightning bolt yielding a narrow turn down on each turquoise fabric.

    A toy version could be made with two bias cut squares of fabric where the on grain diagonal is ornamented with the faggotting.

  19. Brina says:

    What Carol said earlier is that the purple is one piece with the drawn work down the middle of it. So the turquoise pieces are sewn on either side–one on the ‘right’ side or face of the fabric and one on the ‘wrong’ or back to create the double layer. I was reading in the Kirke’s book that Vionnet didn’t really care which side of the fabric showed, she interested in the effect. It also said that one of her assistants made the color choices, as Vionnet was not gifted in that area, and tended to use neutrals by choice.

    That’s an interesting construction idea–sounds like it could work–although you’d need to do the faggoting before joining the turquoise to the purple at the faggoted edge–because otherwise you’d have to use a magnifying glass to sew the turquoise on the exact straight of grain of the purple–which is what you’d end up with the the drawn work. And I agree with you about the economy.

  20. Jennifer says:

    I have no idea if this is how the final scarf would be made, I just suspect this is how it was conceived. How can a piece of fabric be split so that the two halves ornament the two sides of the scarf? Economy of thought. I also don’t really know how to do the ornamental stitching or how much allowance is needed in skilled hands. But clearly it’s meticulous work and I’m sure I’d need more than a magnifying glass to do it and everything would have to be perfectly aligned and on grain. How the tiny squares of the original scarf have survived is beyond me.

    Here’s a drawing I made. I arbitrarily set the scarf length to 36″ along the short ornamented diagonal. I set the width of the faggotting stitches at 1/4″. The gray line is where the fabric would need to be pieced if the goods were 36″ wide as the on grain width of this scarf is ~43″.

    Thinking about this with the flat view is challenge enough, how could you offer only the arranged scarf for the puzzle? Thanks for pointing it out, such amazing work.

  21. Here’s a breakdown of how it was done (I share Kathleen’s rabid enthusiasm for Vionnet, and while I am not an expert on her, I’ve done considerable research, built many of her patterns, and done the drawn-thread work that gives these results).

  22. My web-fu is not engaging today. Got the graphic but not the accompanying text.

    1. Left: pattern pieces, probably of Vionnet’s frequently-used silk crepe, which looks much the same right/wrong sides. Both sides of the scarf will be seen.

    2. Turquoise pieces layered in correct final position. Drawn-thread work started on purple piece.

    3. Turquoise pieces sewn to purple layer along outside edges. Turned to inside and slip-stitched.

  23. Matthew Pius says:

    Thanks for pointing out what should have been obvious – that the purple is one piece. Even after seeing that it is drawn-threads, the initial impression of a piecing seam was distorting my thinking about the construction. :- )

  24. Brina says:

    Nicely done, Jennifer–the proportions looks right and it shows the seam where the purple is pieced. I wonder though if the turquoise would have been cut in one–because the seam I think I saw–it’s looks like the ghost of the seam allowance–runs perpendicular not parallel to the purple seam. When I do these puzzle I always want to get my hot little hands on the piece and really examine–because it’s all conjecture without live examination. Still it’s fun though–because even if you have the piece you can’t always know everything that way either.

  25. Alison et. al. – you’re welcome. I have to make a model or draw stuff like this out to get it to quiet down and out of my head. This morning, coming up to consciousness (?), there was an origami – Möbius – bias scarf floating and twisting thing going on, and I realized that the last drawing should have the colors reversed, as after the outside seaming the pieces flip to the other side.

  26. Jennifer says:

    Brina, good point, I see the ghost your talking about, but I’m doubting it could be a piecing seam as wouldn’t it be poor planning as that seam would collide with the ornamental work and very close to the corner, too. Could there be an advantage to that? The interior turquoise seams must be hand appliquéd would there be any help having a seam at that corner when you could put a piecing seam anywhere?. Also, I assume the piecing seam is never used in the construction process just completed before layout? The turquoise might still be pieced similar to the purple so that the piecing seam is entirely on one side of the garment and thus will be inside the finished scarf. But I can’t find it and it would be nice to see the real scarf. But if the turquoise pieces are cut separately from the get go, they are narrow enough to avoid piecing.

    I also don’t see why the entire thing could be cut rotated 90 degrees. The edges would still be on the bias and the fagotting would still be on grain. Then piecing could be avoided and you could probably cut three scarves from a width of fabric. But since the piecing is so nearly invisible and impossible to avoid on more elaborate garments, I assume everyone was completely comfortable with piecing and there was no effort to avoid it even in smaller garments.

  27. Brina says:

    If you look at pre-1940 vintage garments, it’s pretty common to see piecing. Looms, in general, were not as wide as they are now, so fabric were narrower. When I first started studying vintage garments, that surprised me and I thought it was either bad planning or artlessness. Now I know it was a way to get wide garment sections out of narrow fabrics and to reduce waste, when the fabrics were 24″ or 36″ wide. And as you say, good piecing could be almost invisible. Piecing is even more common in older antique garments, because the fabric was expensive and/or hard to come by. We modern folks are really spoiled in comparison.

    Putting the piecing seam of the turquoise where it is extending out from a corner (if the ghost is indeed piecing) doesn’t seem like a bad idea to me. However I’d really need to know more, like how wide the fabric they were working with was and how they needed to piece. Even couture houses don’t want waste. Also you wouldn’t want the layers pieced and stacked in the same place because it would add bulk and draw attention. Criss-crossing the piecing seams seems like a good way to lessen bulk and make them more invisible. I looked through the drafts in the Kirke book and most of the garment have some piecing–and it looks strategically placed.

    As far as rotating 90 degrees–I don’t know that the Met laid it out necessarily as it would have been cut. But let’s say they did. Then you’d really need to know how the bias is going to act to know whether rotating would make a difference or not. The stretch of the warp thread is different from the stretch of the weft–so that needs to be taken in account when patterning. Kathleen has a related post on bias stretch somewhere.

    Oh, and I don’t know about Vionnet, but it’s been common to pattern the pieced section(s) as separate pattern pieces that are joined as part of the construction process.

  28. Jennifer says:

    I wasn’t being clear I do understand that piecing was common and acceptable for the reasons you explain as much as my knee jerk is to avoid it. And I’d say this is a perfect example because even though the scarf is not that large it is pieced (I estimate on grain it requires a 26″ x 44″ piece of fabric). So clearly the maker had higher priorities than avoiding piecing and you’re correct hang of the cloth should be one. I was some how assuming the visible piecing seam here was parallel to the selvage but you’re right it may not be.

    If I’m seeing the same ghost you mean, I’m still skeptical that it is a piecing seam in the turquoise. I see the advantage of having it perpendicular to the purple piecing seam, to reduce overlap. But if we’re talking about the same thing, it looks like a continuation of the ornamental zigzag. I would think that would draw attention to the piecing seam and would also mean there would be an extra layer of fabric folded back exactly at the tiny square in the fagotting. I’d think that would be possible but to execute but not ideal. Those squares have to be so precise, any extra bulk on one edge could make even precise work look sloppy. At least that’s my inclination not that I’ve ever done hand work of this precision. I would just guess you would avoid that situation unless there is some advantage to having the seam there. On the scarf overall, I actually see very little show through, so I think it’s entirely possible the turquoise is pieced and it just can’t be seen. I just think that particular ghost may be spurious.

  29. Brina says:

    Jennifer, everything you say makes sense, and I agree. Thanks for clearing up what you meant.
    since couture and other fine sewing folks like to avoid extra fabric [bulk] unless it’s needed for weight/hang, piecing the turquoise somewhere else makes more sense, if it is pieced.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.