This post demos two alternative fingerloop braiding methods that a team of two Iron Age loop braiders may have used in making the 2,500-year-old, 15-loop fingerloop braid known as “HallTex 301” to archeologists. This braid, and a smaller one of ten loops (HallTex 306), are part of the large body of textile finds from the famous Hallstatt Salt Mines in the Austrian Alps. Joy Boutrup (co-author of European Loop Braiding) is one of the authors of a study of these two braids. The study found that they were most probably made by a loop braiding technique. The braids were of finely-spun, dyed wool thread, the smaller one in two colors and the larger one in at least three colors. This post is only on the larger of the two – the 15-loop HallTex 301. (More about the study in my endnotes.)
The paper provides one possible fingerloop braiding method for making “HallTex 301”. The method given is based on the method for a similar type of braid of only 7 loops described in a 15th C. A.D. loop braiding manuscript (Tollemach, “a lace broad parti” – link goes to another site’s description of the braid). Other fingerloop braiding methods would also be possible. The same braid can in fact also be braided with free ends, and the paper gives instructions and a schematic diagram for braiding it that way, however some of the micro details* Joy Boutrup analyzed would likely turn out a little differently if it were braided with free ends.
While I was procrastinating on trying out this braid, two braiding buddies of mine here in the San Francisco Bay Area jumped on it: Kim Davis of the Lace Museum in Sunnyvale, and Carol Wang of Chinese Knotting.
Together they came up with a slightly different method than the one suggested in the paper – still fingerloop braiding, but done with palms facing down. I have never heard of anyone making a team braid with palms-down braiding before! Neither Carol nor Kim had done any Slentre (palms-down) fingerloop braiding before this, and I hadn’t suggested anything to them about trying it that way, they just found it more intuitive/ easier to accomplish the necessary moves this way. Luckily, Carol’s son was available to make a video – and voila, I got an email the same day containing a video and pics of the resulting reproduction.
Then a couple of days ago the three of us got together and tried the method suggested in the published study. With Carol videoing on her cell, we got a second video – the vertical video below. So, with permission from Carol, Kim, and Brendan James (the other videographer), I can now present two possible methods for a team of two braiders to make HallTex 301, the 2,500-year-old Iron Age fingerloop braid!
The vertical video shows the method suggested in the published study, with my voice-over describing the moves.
Palms-up/ facing method (as described in the study):
Text description for the above moves – also a more authentic color pattern set-up, and notes about the color pattern of the original braid artifact – here (further down this page)
The horizontal video below shows Carol and Kim’s palms-down method. This video has no voice-over, but the moves are very clear. (Youtube videos can be slowed down, btw – click on the ‘gear’ icon in the video toolbar, and then under “playback speed” choose what speed to play the video.)
There is a slight structural difference between the original artifact and this version – I think Kim and Carol are doing the plain weave (“over/under one loop”) on the two edges of the braid, and the 2/2 twill (over/under two loops) in the center of the braid, whereas it’s the reverse in Hallstatt 301. This would be easy enough to change around – each braider could either switch their left and right-hand moves; or the left and the right braider could keep doing the same moves, but switch places with each other.
Palms-down method (an alternate possibility):
Whichever method is used, this is a different type of loop braid from any of the braids I have taught here so far, in that each loop functions as a single element throughout the braid. Both shanks of each loop stay next to each other – as if glued together – all the way through the braid, with the result that all the strands in the braid are doubled – each loop is a pair of strands moving together as one element.
To accomplish this, the braiders never put their ‘fetching finger’ through a loop, only over or under both strands of one or more whole loop(s). The fetching finger weaves up and down between loops, never through the two shanks of one loop.
To see the braid’s structure in photo above, follow the path of the lowest red loop from the right edge of the braid inward to the middle of the braid. Each ‘strand’ or ‘1’ is a loop, so a doubled yarn. Starting at the edge, the red loop goes under 2 (a blue and a green); over 2 (2 greens); under 2 (2 greens); over 1 (a green); and under 1 (another green). The next loop down is a blue one, performing the same order of interlacing, though offset by 1 strand from the over/unders of the red loop above. The interlacing is primarily 2/2 twill, with a section of plain or tabby weave (over 1/ under 1) at the center of the braid. If you are recreating HallTex 301 you can compare the thread passages to these. You may see the opposite over-under sequence (first two OVERS, then two unders, etc). That’s still the correct sequence, because ‘over’ and ‘under’ are two sides of the same coin. If you turn this braid over, or if instead of tracking that red loop you follow an opposing loop that the red one passes over, the ‘overs’ would all be ‘unders’ and vice versa.
I tend to agree with Carol and Kim that the palms-down method seems more intuitive somehow – for the basic moves as well as for the ease of NOT turning a loop when taking it (which is different for this over-under loop braid than for ‘normal’ through-loop braids, btw). According to the study, the two strands of each loop stayed in strict left-right order throughout the braid, never twisted around or crossed each other.
Not turning loops is a no-brainer when making this braid palms-down, it’s the most obvious way to take a loop. With loops held palms-down, it would be difficult to turn a loop while moving it. That’s not the case when holding the loops palms-up or palms facing.
In a nutshell, the palms-down method almost necessitates the correct “no-turn” loop transfer for this Iron Age braid, whereas the palms-up method allows the correct “no-turn” loop transfer, but doesn’t exactly encourage it. *More in endnotes*
(My own half-baked speculations about the palms-down braiding tradition here, in Endnotes).
Set-up, and Braiding Method:
Text instructions below are for the braiding method described in the published paper, but in my own words, and using slightly different abbreviations. (See notes on instructions in the study in Endnotes.)
I give two color pattern set-ups below. That’s because the study’s two schematic diagrams for this braid actually show a slightly different color-pattern than the one described in the text – it’s just a difference in the color of one strand (on Right Braider’s left hand in the set-up), but it does make the pattern come out differently. (more about that further down, in Endnotes). Unfortunately, the color pattern we are braiding in the vertical video is different from both of these! – I had been hoping for the pattern described in the text, but messed up in trying to set it up “on the fly” for the video.
a, b, c, d = index, middle, ring, and little fingers
La, Lb = Left index, Left middle; Ra, Rb = Right index, right middle, etc
LB = left braider
RB = right braider
Color set-up described in text of study:
7 bluish-green loops (green in video); 4 red loops; 4 dark loops (black or blue)
LB (left braider): Left hand holds 4 loops (a, b, c, d); Right hand holds 3 loops (b, c, d – Ra holds no loop). All 7 loops green in video; bluish-green in HallTex 301 braid
RB (right braider): Left and right hands each hold 4 loops, alternating blue and red (dark blue or possibly black and red in the HallTex 301 braid).
On both the left and right hands, RB should hold
a – red
b – blue
c – red
d – blue
Color set-up for the pattern shown in the diagrams:
7 bluish-green loops (green in video); 3 red loops; 5 dark loops (black or blue)
LB (left braider) – both hands as above
RB (right braider) – right hand as above.
only RB’s left hand has different color arrangement:
a – red
b – blue
c – blue
d – blue
The method in the vertical video is easiest if palms face somewhat upward, rather than facing each other. I believe it is necessary to always take a loop by reaching into it from below the loop, so as not to turn it over when pulling it off. (Note: this is the opposite of how you would take a loop without turning it when making a “through-loop” fingerloop braid like a ‘regular’ square or flat braid!).
(In the instructions below I made the left braider a “She” and the right braider a “He,” so the pronouns might help differentiate them.)
1. LB (left braider) uses her bare Right index to go through RB’s (right braider’s) left-hand loops in this manner:
Over one loop (La), under two loops (Lb and Lc), over 1 (Ld) to take RB’s (right braider’s) RIGHT little finger loop (Rd). She takes his loop from below, so as not to twist it.
SHIFT: RB (right braider) now shifts his remaining 3 right loops down one position, to free his right index finger for the following move:
2. RB (right braider) uses his bare Right index to go over and under the adjacent loops of his same hand (his right hand) in this manner:
Over 2 (Rb, Rc), under 1 (Rd) to take his LEFT little finger loop (Ld).
SHIFT: RB (right braider) shifts his remaining 3 left loops down one position, to free his left index finger for the following move:
3. RB (right braider) uses his bare Left index to go over and under LB’s (left braider’s) RIGHT hand loops in this manner:
Over 1 (Ra), under 2 (Rb, Rc), and over 1 (Rd) to take LB’s LEFT little finger loop (Ld).
SHIFT: LB (left braider) shifts her remaining 3 left loops down one position, to free her left index finger (La) for the following move:
4. LB (left braider) uses her bare Left index to go over and under the adjacent loops of her same hand (her left hand) in this manner:
Over 2 (Lb, Lc), under 1 (Ld) to take her right little finger loop (Rd).
SHIFT: LB shifts her remaining 3 right loops down one position, to free her right index finger (Ra) for the next move.
Repeat from Step 1.
Tightening: It’s not necessary to pull hard unless you want a very stiff, tight braid. Likewise, with short loops try not to pull too widely, that may also over-tighten the braid.
In the vertical video, Kim and I tightened mainly after Step 4, but if we had braided longer, we might have come up with a different tightening strategy. In team braiding, it’s usually more common/ natural for each braider to tighten their own loops – opening their own two hands apart to tighten, the way Kim and Carol are tightening in the horizontal video, rather than to pull against each other the way Kim and I did in the vertical video.
In the two samples in wool (navy, red, and turqoise) that I did on my own, I tightened whenever a loop passed between my two hands. My first sample was the ‘text’ version – upper braid in the photo below – and it came out a lot tighter than my second sample, when I was more relaxed. The braids compressed inward a bit during braiding. With both samples, after I finished braiding I pulled the two edges of the braid out width-wise (not lengthwise) up and down the braid to pull it into a flatter shape, with a clearer-looking pattern.
Solo-braider samples: I made the 2 solo samples with palms-up loop braiding. (Palms-down does not seem workable when you are carrying more than one loop on some fingers.) Working solo is slower than working with a partner, but develops a nice, predictable rhythm of moves. I hold the left braider’s loops on my left hand, and the right braider’s loops on my right hand. Both hands hold 1 loop per digit thumb-to-ring finger, and 3 to 4 loops on the little finger. (This is not the only possible way to hold the loops. I often end up changing a “too-many-loops” method after making a few more of the braids. So far, I have only made these two small samples.)
At the start of braiding, Left braider’s Left a, b, c, d are on my Lth, a, b, c. Left braider’s three right loops are on my Left d finger, Rb nearest the tip, Rc mid, Rd at the base (the left braider starts with no Ra-loop). Same general arrangement for the right braider/ my right hand loops, except that the right braider starts with 4 loops on his right hand, not 3, so likewise I start with 4 loops on my Rd finger. But this one-to-one correspondence doesn’t stay exactly the same throughout the braiding process, because I shift the loops as I braid to keep 1 loop per digit on thumb-to-c, with the remaining loops on d.
Click link to read and/or download the paper on two Iron Age fingerloop braids.
Ancient salt mines might not seem like a likely source of textiles, but apparently the mine shafts were also used as garbage dumps, and the unusual environment of the mines preserved discarded textiles, as well as all sorts of other material. These mines revealed a treasure trove of textile techniques from the Bronze age onward, showing dyed yarns and fabrics, woven color-patterns (tartans, etc) and very fine spinning as well as the two braid fragments in the study.
These two braids were excavated from the Hallstatt Salt mines many years ago, but for a long time were thought likely to be contamination or litter from more modern times, since they look so similar to shoelaces that are still used in the area. They were finally carbon-dated a few years ago and their true age was discovered – it turns out they are from sometime between 756 and 415 BC! Both braids are of fine wool yarn, one (HallTex 301) made with 15 paired strands, and the other (HallTex 306) made with 10 paired strands. Joy Boutrup – as one of the authors of the study – was able to deduce that they were almost certainly braided by loop braiding, based on the way each pair of strands stays in the same left-right configuration for the whole length of the braid, as well the way the pairs “turned the corner” on the edges of the braids.
I should have remembered about this study without being reminded, as Joy Boutrup had sent me information about it when she was working on it, so I am a little embarrassed but grateful to Réka Tóthné, a reader in Hungary, for bringing it up in the comments to my recent history post. Thank you, Joy and Réka!
Here’s a link to an interesting and well-illustrated article in an archeology periodical about even earlier textiles uncovered in these mines, from the Bronze Age. It has diagrams of the mines and gives a sense of how ancient they were and how long they were continuously in use. (one of the authors is Karina Grömer, who is also one of the co-authors of this Iron Age Fingerloop braids study): https://core.ac.uk/download/pdf/42372140.pdf
The study’s 2 schematic color-illustrations of this braid show a slightly different color pattern from the one described in the text. The text says:
The colour sequence is one dark (blue or black) followed by one red, repeated four times, then seven green-bluish ends.
But neither of the two diagrammatic illustrations (one of the braid itself, the other showing how to do the moves with free ends) show alternating blue and red loops (four of each) as described in the text above. Instead, both of them show five blue and three red loops – the red loops all separated by blue loops, followed by the seven green-blue loops. The difference is of only one loop – one red loop in the text description has become a blue loop in the illustrations. In the actual photos of the braid, the colors are so mud-stained they can barely be seen – the braid’s original color-pattern is completely obscured. I’m guessing that the color pattern described in the text was an earlier assumption by the researchers about the color pattern that was later revised (as shown in the 2 detailed schematic diagrams), but mistakenly left in the text of the article.
Following the instructions in the published paper:
In the instructions given in the published paper, the two braiders are described as Person 1 and Person 2. It isn’t mentioned how they are positioned relative to each other, but they must be (generally) side-by-side, and it can be deduced from the description of the moves that Person 1 must be on the left side and Person 2 must be on the right.
In the abbreviations given, the first symbol is an L or an R, which stands for the left or right HAND of a braider, the next symbol is a 1 or a 2 for Person 1 or 2, and lastly (if used) comes a lower-case letter to represent a finger or its loop. (a, b, c, d = index, middle, ring, little fingers)
L1 = Left hand of Person 1 (deduced left braider)
R1 = Right hand of Person 1 (deduced left braider)
L2 = Left hand of Person 2 (deduced right braider)
R2 = Right hand of Person 2 (deduced right braider)
Loops, index to little finger:
Person 1 (left braider)
left hand: L1a, L1b, L1c, L1d
right hand: R1a, R1b, R1c, R1d
Person 2 (right braider)
left hand: L2a, L2b, L2c, L2d
right hand: R2a, R2b, R2c, R2d
There is a minor typo in the instructions given in the published study.
In the first move, the loop that should be taken/ moved by Person 1 is Person 2’s right little finger loop, not his left little finger loop as stated.
She (Person 1) should take his (Person 2’s) Rd loop, not his Ld loop. This can be seen immediately afterward in the instructions, as Person 2 then shifts his right loops down to fill the empty spot (and free his index finger), not his left loops.
Palms-up version is not really equivalent to A-fell braiding, imo:
To make this braid while keeping the doubled strands in the same left-right orientation for the length of the braid requires a horizontal loop orientation for the purpose of the loop transfers. Even if two palms-up braiders choose to hold their hands in more of a “palms-facing” position than “palms-up”, they need to transfer loops as if those looops are on a horizontal plane, keeping the left strand of the loop on the left, and the right strand on the right, even when the loop crosses to another hand. To accomplish that, the ‘taking’ finger must be inserted into the transferring loop from below the loop for the “palms-up” method. This is essentially the exactly reversal, or flipped-upward version, of the “palms-down” method.
This horizontal loop orientation is very different from that of palms-facing, two-layer braiding methods like the A-fell and V-fell methods for making square braids. For both those methods, loops are held on an idealized vertical axis (regardless of whether the palms themselves happen to tip upward or downward at various points in the braiding process). Each loop has an upper and a lower strand, not a ‘left’ and ‘right’ strand, and those two strands interlace separately, one on the upper layer of the braid, and the other on the lower layer of the braid. Physically, the palms-facing hand position allows turning (in either of two directions) or not turning a loop just as easily during the process of transferring it. Again, the same “from below” loop transfer that turns a loop when you are making a two-layer braid like a square braid, results in NO turn to the loop when making this flat, Iron Age braid of horizontally-oriented loops.
The palms-down hand position is more limited when it comes to turning vs. not turning loops. A rather convoluted move is required in order to transfer a loop with a turn – in the case of this Iron Age braid – or without a turn for the flat or divided version of a Slentre square braid. It’s not impossible to do – I have made flat and divided braids using Slentre (palms-down braiding), but so far I haven’t seen or heard of this as a common practice, which I assume is because the necessary loop movement is so inconvenient.
In a nutshell, the palms-down method almost necessitates the correct “no-turn” loop transfer for this Iron Age braid, whereas the palms-up method allows the correct “no-turn” loop transfer, but doesn’t necessarily encourage it. That “from below” loop transfer is not the most directly obvious way to transfer the loop. It’s easy enough to do, but it’s a bit less direct than simply reaching through the loop “straight-on”, which would effectively turn it for this Iron Age braid.
I should add that the difference in the actual braid would be fairly minor, even if the braiders did turn the loops every time! The turns on the left would be cancelled out by the opposite turns (of the same loop) on the right. The two strands of each loop would cross each other and change places repeatedly in each cycle of braiding, but they wouldn’t end up twisting around each other. The basic structure would be the same, the only difference would be that the doubled strands probably wouldn’t lie as neatly side-by-side.
Side note: I just realized that the 15th C. instructions for the smaller solo-braider 7-loop analog to this braid specify for each loop transfer: “Reverse upward“! (“reverse” in those manuscripts meant “take the loop by turning it”).
“Upward” means “by inserting the fetching finger from below into the transferring loop” (as opposed to “by taking the upper strand of the loop from above the loop” i.e. “downward” turn.)
This is the only 15th C. braid instruction I’m aware of that specifies the direction to turn the loop! I believe all the other instructions just say “reverse” without specifying whether the loop should be turned from below (“upward”), or from above (“downward”).
This unusually specific line in the instructions means essentially “Take the loop from below AS IF TO TURN IT UPWARD.”
However in this braid alone of all the other braids described in the 15th C manuscripts, taking the loop from below will in fact result in the loop not turning! As I would bet the original writer knew well! However the average loop braider of that era would only think of the “from below” loop transfer as a turning transfer, so the direction is simply “reverse” – the writer does not go into complicated explanations the way I do, apart from specifying “upward” solely for this braid.
If these directions are followed, all the paired strands should stay in a fairly consistent left-right order with each loop transfer. However, I just did some trials with this – using bicolor black/white loops for the 4 loops that start on one hand, and 3 single-color gold loops for the other hand – and the results were somewhat messy and inconsistent when I tightened normally. Maybe partly because of the thick cotton yarn I’m using, and these long twill passages of “over-under 3 loops”. I had to ‘fussy-tighten’ and sometimes correct the order that the strands ended up in as they crossed the braid in order to achieve the neat left/right side-by-side positioning of the black and white shanks that I expected.
However, when I divided the loops at the bottom to braid two smaller braids (to form a loop), the results were much neater, and the doubled left-right shanks of the loops made a very smooth, neat 3-element flat gold braid; while the 4-element black/white braid came out with a neat color pattern because of the regularity of the left-right black and white shanks of the loops. When working a three-‘strand’ pig-tail-type braid with loops and holding palms upward (to facing), it seems very obvious to insert fingers into the loops from below in order to keep the left strand of each loop on the left, and the right strand on the right. Plus it becomes very clear that you aren’t reversing / turning the loop in taking it from below.
The two-person Hallstatt braid has several over-unders across the braid before the loops turn at the edge to go back across the braid. The Broad Lace Parti has only one “over” on each half of the braid, and that one long “over” goes around the edge of the braid to become the long “under” on the other side, so it’s inherently a less-flat, “bunchier” braid than the Hallstatt braid. Also the Hallstatt braid’s over-unders are shorter – over-2-loops compared to over-3-loops in the Broad Lace Parti. I think that’s why the Broad Lace Parti tends to have messier alignment of the left and right strands of the loops.
Historical speculation about palms-down braiding:
The idea that the two braids in the study could have originally been made with a palms-down braiding method is very intriguing to me, because the palms-down loop braiding method has so far only been known historically from one very tiny part of the world – the Faroe Islands. I’ve always wondered about this. Why there and (apparently) nowhere else in the world? Elsewhere in Europe the only reports are of palms-facing methods (which can also be done somewhat “palms-up”). But these 2 braid artifacts in the study are from the Iron Age, and are a good 2000 years older than first known European loop braiding report – that being the 15th C. English loop braiding manuscripts. That’s a lot of time for traditions and techniques to evolve or change.
The Hallstatt Salt mine area was peopled by Celts during the Bronze and early Iron Ages. Peoples and languages and presumably textile traditions have done a lot of wandering over the millennia in Europe. The Celts ended up in the British Isles and nearby, after occupying much of Europe for thousands of years. (Apparently, the Celts were already weaving tartan fabrics when they were in the Hallstatt area, long before they arrived in Scotland!)
The Faroe Islands is an isolated group of islands located between Iceland, the British Isles, and Norway, although politically it’s actually part of Denmark. Its population is Nordic – their language, Faroese, is descended from Old Norse like most Scandinavian languages. So not a very obvious connection to the Celts. But after watching the video of Carol and Kim braiding this Hallstatt braid palms-down, I can’t help wondering if Faroese palms-down loop braiding might be a relict tradition that was once much more widespread in Europe.
In my info page on A-fell, V-fell, Slentre, and other types of loop braiding methods, I wrote that I doubted that palms-down braiding could be used (easily) to make team braids. Kim and Carol easily proved me wrong on that! The move I had imagined would be difficult to do palms-down is the “loop-exchange” move, in which 2 adjacent braiders swap their two closest adjacent loops to connect their loops into one braid. With palms-down braiding, it would have to occur between two neighboring braiders’ adjacent little fingers. But that move never happens in this braid – instead the two braiders use another way to “exchange” loops, in which each braider takes the other braider’s loop as part of one of their own loop transfers, rather than in a separate loop exchange move. This obviously works fine with palms facing down!
© 2019 Ingrid Crickmore