This post also appears on my shiny Wordpress blog.

The reason that the Peacock Dress was always a joke, and the reason I pledged to go ahead and do it anyway if you sent me to Haiti, was because it was obviously a maddeningly complex project. It teased me from inside the glass case: "Betcha can't make me." It has lived up to its reputation: we are now approaching the three year mark since I first pledged to do this.

At first, I wanted all the glory myself, but after a year of embroidery, I had to admit defeat and try Plan B.

Plan B - drafting in help from all of you - was also far too complex to be likely to succeed, and so we went to Plan C, and as soon as the decision was made to go to India for the embroidery, it seemed to fall into place. Of course we should do it the same way as Mary Curzon and M. Worth arranged it.

The first connection in India turned out to be less than enthusiastic enough to provide a good service, even in pursuit of a sample, and so we went to Plan C(ii). And by this time, the momentum had faltered.

In 2014, I want to see this thing done. I have been apprehensive about saying that, because of course we don't yet know what the cost will be, but Barry has always advised me to plan for success, and not failure. So here it is: progress renewed.

First things first
The first priority is the small feathers I promised to make for the biggest donors. Not all of them have been made and delivered, I'm embarrassed to say; that is about to change. The next one is pictured above.

Goldwork embroidery
As for India, there have been regular emails back and forth as we have attempted to communicate fully about exactly what I need and exactly how to do it. An initial sample was not even accurate enough that the company showed it to me; they knew it wasn't right, and asked me to send more pictures. I sent them a sample feather and an envelope of materials, and now they have a better idea. Having done some of the embroidery myself in that first false start has been a hugely useful exercise, now that I have to explain it to others.

The latest sticking point is that apparently at least two elements of the design cannot be found in India (huh??), and so we are casting around for a place that will not want to charge us to produce a minimum order of 25kg of goldwork materials. I can't believe that I might end up having Benton & Johnson (who provided the stuff I was using) send their stuff to India for me. We soldier on - and meanwhile, to the rest of the outfit:

Silk flowers
The hem of the Peacock Dress is circled in white silk roses. I have a source for these that is quite exciting - but first, I don't think I have told you the whole story about the roses, so this subject will wait for another day.

Underwear
I am still slightly unsure of the exact underwear layers, but I'm a mathematician and actually I'm just frustrated that someone didn't write down the exact, only, one, single, correct answer to that question sometime in early January 1903. :D But obviously the general rule was not necessarily slavishly followed by everyone. I will finish reading The Cult of Chiffon (the famous manifesto on women's underwear of 1902) just to make sure there isn't some kind of definitive guide written down, but for the moment the layers, as I understand them, are as follows:

Combinations
Petticoat
Corset (embroidered)
Camisole (ie corset cover)
Bust bodice (I have emailed the V&A for an appointment to see this)
Hip padding
Silk petticoat (underskirt)

I have learnt since the Oak Leaf Dress that the Edwardian silhouette involved gratuitous padding; I'm going to need it. howlgirl has recommended a book, which I now have, that will instruct me on the finer points of construction since I have some fairly fancy ideas about making underwear that's more than just an afterthought. I am delighted to have found some impossibly soft, sheer cotton and cotton/silk muslin in the LA Garment District last summer, and though I'd love to buy antique lace from Lacis, I don't think I'll get the combinations of matching lace that will be convincing, so I've found a supplier of new, almost-100% cotton Mechlin lace in a variety of widths and matching patterns that are right for the period.

Jewellery
I have re-emailed the incredible Andrew Prince for a quote on the Curzon Jewels that he knew immediately when I showed him Lady Curzon's Peacock Dress portrait; he never replied before, so I'm not sure whether he realised I was dead serious. We will see.

Beaded shoes
*sigh*. If only there was a historical shoe maker who was planning to release a 1900-1905 beaded evening shoe later in the year.

And that's the state of play. I hope you're still along for the bumpy ride!

The new sample from India

Photos of the sample have arrived from Indian Embroiderer mkII, and it's near-as-dammit perfect! I love that not only do the two photos above represent a very, very good approximation of the Peacock Dress itself, but that they also look like pictures I could have taken of my own version of the embroidery. When the email arrived, I screamed maniacally at a stormy, lightning filled sky, "MWAHAHAHAHA!! I HAVE SUCCESSFULLY CLONED MYSELF!!! IT LIIIIIVES!!!" And of course it looks like both the original dress and my embroidery because I provided pictures of both for reference, as well as an actual sample of my embroidery that I sent to India in the mail. (The difference, of course, is that they will have done this in about ten minutes flat, whereas I would have taken days.)

The next step is materials; Sweta is having trouble tracking down just the right bits and pieces, and doesn't yet know where to find beetle wings. Ebay is your friend, I shall tell her....

Pink feathers

Meanwhile, there is also progress on solving the mystery of the pink feathers. If you recall, the original dress features just a few feathers, randomly distributed over the skirt, that look reddish from a distance, and pink up close. What is the cause? It's not all that easy to see in a photo, but what's different here, in the feather in the center of this image?



When you look a lot more closely, it appears as though the passing thread (the white thread with a metal strip wrapped loosely around it in a stripy-looking effect) is the culprit - it's this that's pink.



Why does this level of detail interest me? Because I'm not producing a replica of this faded, tarnished beauty, but a replica of the dress that left the Worth workroom in late 1902. The pink caste seems to be evidence of something decayed, something that looked different when the dress was first produced, and I want to know what it was that the embroiderers intended so that I can reproduce it accurately.

Nowadays visitors do marvel at the Peacock Dress, gently lit in its glass case in the bowels of the Hall, but I want to tell them that they don't see it at all! What they are seeing is a pale, fragile, tired, heavily altered, elderly shadow of the dazzling dress that Mary Curzon wore on the night of January 6th, 1903 at the Coronation Ball in Delhi. I want to take them back in time.

Previously, I have had a goldwork expert at the British Museum tell me that he could not draw any definitive conclusions, but suspected that it might not be the right shade of pink to indicate copper content. But this week, I have had an email from a Senior Textile Conservator at the V&A:

...it is difficult to tell if the core thread is pink or if it is just the metal foil that is wrapped around it.

"It is possible that the metal has tarnished and the tarnish has a pink caste, I have seen that before and presume the effect is caused by the alloy used - and it maybe the tarnish has discoloured the core thread.

"It is very hard to say anything with certainty without being able to examine the piece. If it was possible to have the metal foil analysed that may provide an answer, as it could be that one particular batch of metal thread has a higher copper content.

Well, I would love to have a piece of the metal taken from the dress and analysed, but judging by the amount of time it took to convince the Hall staff to open the glass case and let me even look inside, I think it's highly unlikely they'll let me take anything away from it.

Since I have seen the dress in person, I would say that first of all I do believe that whatever is in that metal wrap is what has discoloured the thread. I looked very closely at the time and satisfied myself that it was indeed white thread discoloured to varying shades of pinkish hue in different areas, and not a pink thread. I think that shows in the photo, but of course the conservator wouldn't want to draw definitive conclusions on a photo alone.

So I would go with the alloy being the culprit. Secondly, those metal threads were made in a number of recipes; by this date various alloys were being tried out to reduce costs, and it was common for a gold or silver outer appearance to disguise a copper core. (Reference). It's perfectly possible that different batches might have been employed on different parts of the dress, given the sheer quantity of embroidery on this thing.

HOWEVER. Why would a different batch be used for specific feathers, dotted randomly around the dress? I can see one embroiderer working on a section with a different batch of passing thread, and the appearance there decaying slightly differently, leaving a patch like a large red wine stain, but this is a different batch being used for specific, individual, clearly defined feather motifs, beginning at the top of the feather and ending at the bottom. Surely this is deliberate.

And if it is deliberate, it further indicates that not only was the alloy different, but the appearance of the thread was different even then, and was used for effect. And therefore, I don't think it's unreasonable of me to have concluded that some of the feathers may have been made with a copper passing thread - or some other metal that has decayed to pink. In the absence of any alternative suggestion from the British Museum guy, even though he thought it wasn't copper, and in the absence of any other option in goldwork supplies on the market today (short of green, blue, etc), I'm going to instruct the embroiderer to make every tenth feather or so with copper passing thread, and call that a reasonable final assumption.

Does my reasoning make sense to you, research boffins, or do you think this is too much of a leap?


This *is* my own embroidery from 2011.

Earlier in the week, we reached a moment of truth that I’ve been anticipating ever since September 2012, when I realised that India was my best bet for realising my Peacock Dress dream.

The quote for the embroidery, by far the most expensive and complex part of the project, arrived in my inbox.

Facebook followers guessed that the project might take from three weeks to eight months, and cost anything from $1400 to $25,000. My own wild guess just about correlates with the average of all the Facebook guesses - I anticipated 3-6 months and $8000.

Ready for the real quote?

Another embroidery sample arrived from India on Friday. As I compared it to the photos of the original dress in the museum, something began to dawn on me.

Isn't it funny how you can look at a thing for years and assume it was done one way, but when you put it away for a while and get it out again, a lightbulb goes on and the answer suddenly becomes clear...



Remember how I was angsting about the seams when I first studied the dress? I couldn't convince the custodians of the dress to let me study it flat on a table, so I had to try to figure out the skirt pattern just by looking at the dress in the glass case (above).

The seams delineating each panel shouldn't be too difficult to see: the embroidery is dense, but I knew that the feather pattern would give it away. According to long-accepted tradition in embroidered costume, the pieces of the skirt would have been embroidered before assembly, so the edges of the feathers should run neatly along the seams.

So I looked and looked, but although there were plenty of feathers and plenty of edges and I had a good idea where they should be, I got very confused because I couldn't find any seams. They were deliciously, perfectly invisible, except for the darts around the waist and hips, around which the pattern has been distorted to fit, and the extraordinarily clumsy center back seam, where the feathers didn't even come together in line with each other (below)



The only evidence I could find of any other seam was this tantalising end-of-a-seam in the hem, which did not match up in any way with the embroidery into which it disappeared.



I wrote it off as a mystery. Either the ladies in the Worth workroom knew something that I don't know about invisible couture seams, and someone clumsily altered the back seam at the last minute (none of us have ever made a change or finished fitting the dress five minutes before the ball, have we, right?), or the whole thing was embroidered in one piece, which was impossible because the embroiderers would have needed freakishly long arms. Indian zardozi embroiderers traditionally work at long, narrow frames, and yes, normally you could embroider a large piece by rolling the parts not being worked on over rollers at the edges of the work, but you'd need to roll them very tight to keep the working area taut, and this gorgeous three dimensional goldwork would be crushed by doing that.

No, my embroidery would be done on long, narrow pieces corresponding both to zardosi tradition and to the the long, narrow pieces I'd found on a suitable Edwardian skirt pattern of 1902.



I would have to painstakingly join the finished, embroidered pieces of the skirt by hand at the edges (because they mustn't be crushed under a sewing machine foot), and then I'd have to finish some of the embroidered feather veins and eyes over the top, to hide the seams as invisibly as Worth did. Labour intensive, but that's what embroidery is about, isn't it?

BUT... the feather pattern doesn't fit into the pieces of the skirt. If I try to draw correctly sized feathers to fit onto the skirt pieces, I end up with sections of wide feathers and sections of narrow feathers that make the skirt panels stand out distinctly from one another. The join in the skirt fabric that disppears invisibly into the work was machine sewn, and DIDN'T correpsond to the feathers. The feather pattern starts off beautifully at the front of the dress, and veers drunkenly off course as it travels in an unbroken progression towards the back, ultimately failing to meet properly at the centre back seam.

As I turned Sweta's sample over in my hands, it dawned on me what's happening here. This is about attitude - my attitude to my work, and their attitude towards theirs. To me, this is art. Painstaking months of art on my Holy Grail dress, priceless little stitches with lovely, delicate materials, to be minutely sewn by hand and really impress everyone with my insane handwork. It must be treasured, carefully stored, and must not be crushed during its creation.

To them, it's a job that'll take about three weeks. And actually, if you pad this sample embroidery reasonably well, I don't think this stuff will crush as you wind it around a fairly big roller at the edge of an embroidery frame.

I think I've been completely overthinking this.

Sure enough, the skirt pattern I've developed from the Edwardian pattern and my observations of the Peacock Dress has curiously straight seams (excuse the quick sketch).



The skirt of the Oak Leaf dress, a gown made also by Worth in the same year for the same lady, laid curiously flat when the panels were joined... and the joins in the Peacock Dress skirt fabric don't correspond to the embroidery? Suddenly it's blindingly obvious.

Three or four widths of fabric were joined together in India by machine, and then embroidered as one.

I have asked Sweta what she thinks, and I have a funny feeling she's going to say "Of course it would be done in one piece, idiot, what did you think you were going to do?"

So if I'm right, no angst-filled weeks lie ahead in which I wrestle the feather pattern to fit into the skirt pieces, and there'll be no painstaking hand-joining of eleven narrow panels, except for a few darts and the back seam. It all becomes clear...