Halfway writing this article I considered not calling the extant piece a hairnet. Because yes, it’s done with a netting technique, but no, not all done with this technique were hairnets. So what do we have here? What is the striped, butterfly shaped net that is called the Tartu hairnet?
I’ll examine this in a couple of chapters:
- The challenge
- The extant hairnet
- Taking a closer look
- The shape and size
- The netting – stripes and diamonds
- Plaiting and gathering the net
- Wearing the hairnet
- Conclusions and additional thoughts
Chapter 1: The challenge
I don’t think I have the mindset for reproductions, I truly don’t. But I do enjoy a good research, reporting my findings and thinking about the whys, wheres and hows. But making exact reproductions is nerve wracking and opens your work to quite strict scrutiny. It’s much more fun to do research, make my own version and open the work for others to make their own conclusions.
The Tartu net poses a challenge that has occupied some corner of my mind ever since I saw it in a book five years ago. Two years later I found out that I’ve literally seen just a half of the net. How often is this the case? It would be ideal to see and hold the piece in your own hands, but most of the time they are too fragile or far away. Maybe someday I’ll make a date with the extant Tartu hairnet in the City Museum of Tartu in Estonia, just 274 km south from my home town of Helsinki.
Chapter 2: The extant hairnet
The net was described in Riina Rammo’s 2016 article Silk as a luxury in late medieval and early modern Tartu (Estonia) and you can read more about the context it was found in her doctoral thesis Textile finds from medieval cesspits in Tartu: technology, trade and consumption (2015). This chapter relies heavily on the findings in the book and article.
Let’s start with a quote of most interest from the article:
“The remains of two knotted mesh hairnets have been found in Tartu. The first item (15th century) is made of S/zz yarns: one light (0.2 mm) and the other dark brown (0.4 mm) that results in diagonal bands (Fig. 3). The net is edged with a loosely plaited cord of four paired dark brown yarns (S/zz). The mesh squares are of various sizes (1.6 mm and 2 mm) that form a rhomboid pattern. Another example is a small fragment of a mesh (with mesh square measuring 3 mm) made of lightly plied yarns. Hairnets of a similar design are known from London (4 pieces), Amsterdam (1), Lübeck (2) and Schleswig (1), and are dated from the end of the 13th (London and Schleswig) to the 17th century (Lübeck) (VonsComis 1982, 154; Tidow 1992, 249; Crowfoot et al. 2006, 145 ff.; Müller 2008, 364). These nets were at least partly produced on the spot from imported silk yarn (Crowfoot et al. 2006, 145 ff.). The two copper alloy netting needles found in medieval deposits from the former market place – currently Tartu’s Town Hall Square (Fig. 4; TM A-26: 106; 463) – support this hypothesis.”
During the middle ages Tartu could be described as a medium-sized Hanseatic town with a lot of commerce. The richest classes were German and this is of some interest to us who follow hairnet fashions through the ages as they brought with them Hanseatic fashions.
The net was found in a cesspit excavated 1988-89/1992. A cesspit is a lined hole in the ground where everything discarded was thrown away. One cesspit was in use about 30-50 years before it was covered over and a new one dug or it was emptied. The textiles in the pits are dated mostly by the pottery and glass shards joining them. How the net ended up in the cesspit with food scraps, human waste, pieces of cloth and pottery? Maybe it was broken and old and used in cleaning and then thrown away. Maybe someone tilted their head too much and it fell in. Best not to think about it too much.
Doctor Rammo considers all silk finds from Tartu imported and the items of pure expensive luxury. The secrets of silk production started to open to Europeans in the 14th century and especially silks of Italian origin were sold by Hanseatic traders. Silk fabrics and yarn were imported and sold in Tartu and its use was restricted to higher and richer classes. Silk thread was used in sewing visible seams, buttons and buttonholes, but also in tablet weaving and hairnets.
There are actually two hairnets mentioned in the article, but only the famous striped one is photographed. Well, half of it anyway. As mentioned above, it came as a great surprise to me and other hairnet enthusiasts, that there was more. I found out about it when I met doctor Rammo and she asked if I wanted to see pictures of the whole net. And she very kindly directed me to contact the Tartu City Museum to ask for more pictures. Not only did I get pictures of the whole net, “the butterfly”, but they also took some measurements for me. Unfortunately I’m not allowed to publish the pictures, but I’ll do my best to describe them and I have drawn a diagram.
This is something that I have pondered before: were the nets imported or were the materials imported? Netting as a technique isn’t that difficult to learn, although there were some French netting guilds that could have guarded their secrets. But in Tartu we have proof of imported silk yarn and netting needles. Only thing missing is the gauge or mesh stick, but even very small ones can be homemade from bone, wood or metal.
The rhomboid or diamond pattern is easy to see and reproduce. It is a wider version of a Swiss Diamond net that I’ve used and written about in a previous article. The eyes or loops are quite small, but not tiny like in some older hairnets with embroidery. And if we can trust the museum measurements, the eyes are bigger than mentioned in the article quote. The net has stripes that coordinate with the pattern. Original colours are lost to us. But maybe we can find some hints of the fact that most light brown fabric fragments contain red and blue dyes. Usually when both are present it indicates dark purple or black.
I had an opportunity to meet with doctor Rammo in October 2016. She urged me to contact the museum and ask for measurements and colour pictures. We also talked little about her work and she mentioned that the present colours of the finds are unreliable. The soil changes the colours strongly. The first pictures of the hairnet used in this research paper were black and white. In the description on the article the colours of the extant hairnet are dark brown and light. The real colours of the hairnet could be researched with samples in a laboratory. As this is not possible for this short research, the colours in the interpretation are picked mostly from the fashionable colours of the 15th century.
According to Pastoureau’s fascinating book: Black. The history of a color, black colour could be reached by dyeing with oak apples or bark with iron oxide mordant. This produced a true deep black,but could also be very corrosive to the fibers. Dipping the yarn in a woad or madder bath before the black dye also gave a better result. If the black dye faded, it’s possible that an underlying blue or red might come to the surface.
I chose dark brown as the darker and thicker yarn. In the picture the dark yarn is more damaged than the lighter yarn, this might be an indicator of the more corrosive dyes that were used to achieve a true black. But any other dark colour like the dark brown in the extant hairnet could have been chosen.
Grey was popular in the fifteenth century, being the symbol of hope and joy. The colour could be obtained from the same dyestuffs and mordant as black. The combination of black and grey was considered aesthetic and ethical as the separate colours and the combination was seen more virtuous than others. Also purple, green and white were often used in combination with black.
The shape and gathering is fascinating. And I personally had no clue how this butterfly was supposed to work as a hairnet. But I received help from fellow netters through Hairnet hub and now I can see the coif pieces they mentioned. The net is obviously done in two identical hexagon shaped pieces that should be sewn together with the plaited cord. It lies quite flat on the edges and bottom, but is pulled together at the top. I can’t see any cords for tightening the net, maybe they’re gone or the hairnet is supposed to sit like a hat, a bit loose.
Chapter 3: Taking a closer look
For a closer look, I have compared the black and white picture in doctor Rammo’s article, which is a little bigger and the colour pictures l received from Tartu City Museum. In the B/W photo you can see the structures of the knots, increases, stitches and the cord better. In the colours photos the shape, both wings and dimensions better.
There are obviously some problems with using these pictures and this diagram. How can you be sure that my measurements are correct? How can I be sure the museum measurements are correct? We’ll just have to live with a bit of uncertainty for now. There are some problems comparing my measurements to the measurements mentioned in doctor Rammo’s book. My measurements are a lot bigger, almost thrice the size. If I were making a reproduction, I’d probably be tearing my hair out at this point. But luckily we’re just doing research, are allowed questions and most important of all, allowed to make mistakes.
My diagram was drawn on clear plastic from the screen of my laptop. (There, now you know all my hightech secrets.) Seriously, before that the quite small pictures I received from the museum have been brightened up and clarified. After drawing the diagram, I blew up the picture to 1:1 scale on my screen and started measuring loop length on all the three different shapes in the net. This was done with a ruler and the numbers were rounded to the closest millimeter.
You can see from the black spaces on the diagram that there are tears in the net and maybe some pieces missing. But the left wing is pretty intact and you can easily count both the amount of loops in a pattern repeat, 8, and the number of stripes, 10. The stripes start with the dark colour and end with the lighter colour. I counted the amount on repeats at the top and at the bottom on both wings and it seems that they are symmetrical. At the top, which I believe is the beginning of the net, there is just one repeat (8 loops). At the widest point, on stripe 8, there are 8 repeats. And at the bottom are six repeats. This means that the shape of each wing is a rough hexagon.
The knot typically used in silk hairnets is the sheet bend knot used in Silkewerke’s tutorial. The same knot is used in Victorian era netting books and instructions. Sheet bend knot produces a very distinctive type of knot, where the old loop and new loop lock together symmetrically. You can see the effect especially well when changing colour and l believe that I can see this type of knot in some places in B/W photos. There are other netting knots, but they are asymmetrical and meant for coarser yarns like hemp or linen on fishing or hunting nets. Silk yarn is so slippery that we need to use a more complicated knot.
Chapter 4: The shape and size
The extant piece has the feeling of those fancy yellow and black striped German hairnets, as in the 1507 portrait of Barbara Schellenberg by Hans Burgkmair. But making it from two pieces never crossed my mind. I’m certain that you could achieve almost the same shape from just one piece. But then the stripes would sit a bit differently.
Experimenting with the shape cut out of silvery tulle fabric with 4 mm eyes. It was the fabric I had at hand and that mimics netting pretty well. I decided to try a symmetrical pattern and the pattern for both wings. It makes very little sense to net the piece asymmetrical and I’m sure I can see the shapes that form when you always increase at the end of the row. The paper pattern gets its height from the average height of rows times the amount of rows (1,7 cm X 10 rows). At the top the width is one average diamond pattern repeat, 2,1 cm. Widest point is on stripe 8 with 8 pattern repeats and then it narrows down two pattern repeats, so that the bottom is 6 pattern repeats wide. Why didn’t I trust the shape and measurements in the colour photos? That’s because netting has a strong tendency to take on any shape we like. It stretches every which way, but also shrinks down. It is just holes, isn’t it?
I attached the edges according to the diagram first, using the measurements from the colour photos: top c. 21.2 cm, short side 13,5 cm and bottom 20 cm. This butterfly shape was then just pinned into a coif shape, bringing the left wing without the cord together at the short sides. This produces a pretty net that does resemble some Italian renaissance hairnets called reta. And then the bottom is joined with the corresponding sides with the cord. This left the top with the cord at both sides open. The result is a very small peaked cap, that will only fit my hairbun inside it. The opening is just 42 cm, so it probably isn’t meant to encase the whole head. But how big were heads then? Most extant hairnets are quite small and the versions I make can be 10-20% bigger. When you see small hairnets, they are usually written off as belonging to children. But maybe they were worn differently than we think or see in paintings.
The shape and size of the gathered net reminds me more of a net kept over a church or guild reliquary bag. Maybe it was recycled? In that case it was done really well, because we can’t see obvious signs of cutting, i.e. ends of threads or misshaping of the net.
Chapter 5: The netting – stripes and diamonds
In the world of netting, stripes are easy, especially when working with a flat net. Rows are simple to see and count. Diamonds on the other hand are trickier, with their double and long loops and including those in the increases. This I have discussed in previous articles, trying to teach my readers how I make them with ease. There aren’t that many diamond patterned early medieval nets, but you can see examples of simple patterns in Renaissance paintings. Most early medieval hairnets are of single colour, striped and/or embroidered.
I’ll use the silk yarns I already have, because hunting for the exact same strength of yarn gives me a headache. The thinner white yarn is Nirdar (50 tex X 2, 450m= 50 g) and the brown is Maharaja (350 m = 50 g). Both yarns are spun in the same direction as the yarns in the extant net: S/zz. Meaning there are two z-spun (clockwise) strands in the white, and three in the brown yarn, that are s-plyed (counterclockwise) together. Both yarns are a bit thicker than their extant partners and this will change the result. The knots will be bigger in comparison to those in the original.
Before I started the whole net that took a fair share of time, I did some experiments with the pattern on paper and with yarn. First experimenting with the wide version of the Swiss Diamond net pattern, with eight central loops like in the original and with striping. I’ll call it the Tartu diamond. And this diamond needed some serious thinking, especially how the rhythm of the netting would be easiest and to best incorporate the increases. Luckily the maker of the extant net had battled with the same questions. Looking at the picture, it’s not quite clear if the pattern starts with an opening or closing triangle, e.g. half of the diamond. But it seems to me that it starts with an opening diamond and this makes the pattern easier, much easier.
I started with an extra chain of eyes with about 20 loops. You cut away the spare rows when the work is finished. This gives you an added benefit for not wearing out your top loops while working as the cord for attaching the work to a cushion or a post goes through the ones you cut away. I’ve used this method with a Victorian era chenille hairnet. I find it the easiest way to begin a flat net. You can find out about circular nets in my instructions. In this experiment I never minded the original gauge and used my smallest antique bone gauge size 1/16″.
I increase at the end of each row during the first six stripes, starting at the very first row and using regular loops or double loops for increasing according to pattern. In the beginning of a row I do the same, following the pattern as well as I can. You could make the increases and beginnings using just regular loops. This makes attaching things much easier. But you can see clearly in the pictures, by the sides, that they’ve used both lengths of loops in the extant net.
I know the pattern looks exhausting, but actually after a couple of repeats you can see a pattern emerging and you don’t have to follow every written row. They’re there in case you need them. I made the pattern to a separate PDF because it’s so long and it will be easier to follow it printed out or with your own markings.
Turn work after every row, even when starting a new colour. Attach a new colour next to the last knot made in the old colour, this way you don’t lose loops.
You can find a masterclass for diamond netting in my blog. The videos will show how I do the special loops. But if you have your own way of making the longer loops or flipped double loops, please feel free to use them.
Before you start gathering the net, block the pieces. After netting they are long and narrow, but for the work you need it wider, especially at the edges. Pin the net as wide as possible with straight pins, along the edges. Then spritz the net with a tiny amount of water and let it dry. You can use steam too, but be careful, silk doesn’t really like heat or moisture.
Chapter 6: Plaiting and gathering the net
According to the article Silk as a luxury, the net is edged with a loosely plaited cord of four paired dark brown yarns (S/zz). Looking at the pictures it seems that the yarn is twice the thickness of the one used in the net. And the colour looks throughout the plait lighter than the dark yarn in the net. So possibly the plait or cord was done in third colour, let’s say dark brown. It was sewn in place with the white yarn used in the net. In some places the stitching has come a bit undone and you can see that the plait is quite loose. And if I’m not mistaken, it tries to twist around itself, which is a trademark of a classic flat four strand plait. The width seems to be approximately 4 mm.
I chose Jaipur Fino silk yarn (50 g = 300 m) for the plait, in a dark pink colour that isn’t the same as in the extant net, but follows the colour values. You need approximately 80-90 cm of finished plait for the gathering and edging. Make the plait too long and cut off the rest, it’s much easier than trying to continue it. For each strand I took two 150 cm long yarns and then coiled them around bobbins to make the plaiting easier. It was so quick and easy to work with my rarely used lace bobbin pillow. Basic flat four strand plait is easy, but if you’ve never done it before, practice first with this tutorial and thicker yarn. Keep it loose. Finish one end of your plait by folding it twice and sewing it in place as neatly as humanly possible. This will sit at the forehead.
If you look at the picture with the whole net, you can see that the plait goes all the way around just in the right wing, but only along one edge in the left wing. I have a theory that the left parts without the plait were sewn together with the right. This will give you a nice, logical net or coif shape. You can actually put the pieces together in couple of ways and ignore my theory and measurements completely.
Now a bit of measuring! I must thank my husband for helping me with the ratios, math and looking at endless schematics of how on earth am I supposed to sew this thing. If you want to make this fit your head, use the math here. But you can also just use my measurements or the enclosed measurements from the museum.
- Around the head: Measure quite snuggly around your head, from just above the hairline on the forehead to the nape of your neck. Then divide that measure with two to get the front length for both pieces of the net. For example, my head is 56 cm, so one front part is 28 cm. (In the extant piece, according to the museum measurements, the front part is approximately 22 cm. Therefore mine is 27% bigger!)
- Over the head: Measure over the head from just above the hairline on the forehead to the nape of your neck, leaving room for a thick crown braid. This figure is divided 40/60% so that the shorter part is the forehead to crown measurement and the longer is for crown to neck. For example, my over the head measurement is 42 cm, divided into 17 cm and 25 cm.. (In the extant piece, it is, according to the museum measurements, 33,5 cm divided into 13,5 cm and 20 cm. Mine is again 27% bigger!)
Lay your blocked net pieces side by side, the first row is up and the inner sides are touching. Mark your measurements with small safety pins or spare bit of yarn on both net pieces:
- Forehead to crown (17 cm) is measured along the inner sides starting from the first row.
- Forehead to neck (28 cm) is measured along the outer sides starting from the first row.
- Crown to neck (25 cm) is the rest of the net gathered onto the plait.
Again, you can follow my instructions or figure out your own way to attach the plait. My version has the plait going around as one piece. Start by pinning the plait in place on the right net piece starting from the forehead, keep the net flat. Forehead to crown in my net is from the first row along the left side to four and a half stripes. (In the extant net it continues to the bottom of the seventh stripe.) Crown to neck is bunched up along the left side from the top of crown, around the corner and along the bottom of the net and once more around the corner and up to the right side to the place where you have measured the front to end. My measurements come down to seven and half stripes. (In the extant net it comes down to the beginning of the tenth row.) Pin the plait to the front, keeping the net piece plat. This should take you around the right net piece, ending up back to the forehead. Continue with the same measurements, mirroring them on the left piece, going down the left side to the end of front measurement. Don’t cut the plait just yet, just in case.
The original net is sewn to the plait with overcast stitches using the white yarn used in the net. This came out quite ugly in my first trial, so I decided to sew it with the same yarn as in the plait. Gather the loops of the first row tightly with a few stitches. Sew along the plait neatly. Reaching the forehead, sew the beginning of the plait to the plait in the front part. When you come to the end of the front of the left piece, you need to start pinning and sewing the net to the plait previously sewn on to the right piece. Again, mirror measurements and make sure your rows match. Finish the end of the plait with neat stitches and attach to itself in the neck. Your work might now look weird and wrinkly, but sew on.
Chapter 7: Wearing the hairnet
My finished hairnet has a very particular shape and hints that you need a particular hairdo underneath. First experiment with my lovely practice head Anastasia showed a couple of things to bear in mind. This hairnet needs to be fitted on the wearers head, as it doesn’t have any handy cords to tighten it around the head. My head circumference is 5 cm bigger than my models so that shows. Also, you need padding, lots of padding underneath. Maybe not the full wulsthaube of the 16th century, but a padded false braid at least.
I’ve seen an extant version of padded linen tubes braided together and found this excellent tutorial from the German Renaissance of Genoveva blog. A padded cloth braid was also found close to the 15th century Lienz hairnet [linkki!] and the net has the same sort of shape as the Tartu hairnet.
My zopfe or false braid for this project is made of brown cotton as it was at hand and came closest to my hair colour. For this experiment I ended up using some polyfill I have for stuffing toys. For Anastasia or for more regular use I would probably use braids made of silk. The tubes can be stuffed with wool, cotton or in a pinch, with polyfill, but that has the tendency to push through the fabric.
I followed the above mentioned tutorial, but made the falsies quite a lot longer to account for my one meter long (but very, very, very fine) hair. These braids were then piled a top my my head, but not in a milkmaid fashion close to the hairline, but further back. They were then lightly pinned in place and the hairnet was manoeuvred over the whole thing. It looked quite good with just this combination, but I decided to tie the net over the braids to get that tight look you see in paintings of these striped hairnets. For tieing I used the same red silk plait as in the edging.
Chapter 8: Conclusions and additional thoughts
Honestly, I would prefer to make this hairnet with a bigger gauge, with a narrower pattern (5-6 central loops) and try to achieve the same height of the stripes as in the original. Or make it wider with a bigger gauge and work just seven to eight stripes (rows 1.-72.). Or work just one wider piece to simplify things and to better suit a low bun.
One piece has 3891 loops, both have 7782 all together. This pattern is slower to work than just regular loops. In this size and pattern I can knot approximately 380 loops in an hour, which means the netting alone takes 21 hours.
The pattern has been read through and tested by other netters. They came up with some very good points and improvements. So I would like to give my heartfelt gratitude to Perline, Rita Bartholomew from Knots indeed and her daughter Carmen.