6 September 2019

Tutorial: creating Herjolfsnes-style garments using a ratio system

I've made a few Herjolfsnes-style garments in the last few years. All of these have been based on my or other's own measurements, rather than using a pattern (such as those in Medieval Garments Reconstructed). Obviously, using one's own measurements results in garments that are a different shape/size to those of the originals.

I felt it was important when doing this to try to preserve the proportions of the original garments. In particular, the proportions of the bottom hems (they are huge!) and the proportions of the side gores (they definitely do not form princess seams). Here's how I did it...



Starting note: I was initially reconstructing Herjolfsnes 42 (aka D10584) for me, so I will use that as an example throughout this tutorial. If you are also trying to make a Herjolfsnes 42 (or you are not so fussy on your Herjolfsnes numbers or you struggle with maths) you can use my ratios -- just skip to part 6.

However, if you are reconstructing a different Herjolfsnes garment (or you want to do proportional sleeves/armscyes/garment lengths) you will have to follow all the steps and calculate the ratios yourself.

In any case, this entire tutorial works best if you have a copy of both Woven Into the Earth (henceforth WitE) and Medieval Garments Reconstructed (henceforth MGR).


1) Find the original garment's measurements in the books.

This is rather obvious - to preserve the original proportions of a garment, you need to know what those original proportions are. However, going about finding those proportions can be a little tricky. Some measurements are listed for most of the garments in WitE - typically length and hem circumference. Other measurements are occasionally mentioned for some of the garments in WitE, MGR or both. Refer to both books as some things are in one but not the other.

This will, however, not get you that many measurements.

However, MGR has patterns in it for a selection of the garments (including Herjolfsnes 42). These are drawn to scale on a grid. Now, the patterns are not what you want - those are modern things extrapolated from the original garments by the writers of the book. What you want are the grey shapes that the patterns overlay. These indicate the sizes/shapes of the original panels, based on museum measurements. Granted, these are a) approximations and b) not going to be 100% accurate as many of the garments have places where they're not preserved. However, unless you're a ridiculously lucky sod who can beg a private viewing of the originals from the National Museum of Copenhagen and take measurements from them, this is the best you're going to get.

All of the patterns in MGR are on a 1:5 scale grid. For those who struggle with maths or who aren't experienced with scaling up patterns, this means that the drawings are five times smaller than the original garments. To get measurements from them, you use the scale which states "2 mm = 1 cm". If you use a ruler or tape measure, you'll find each square is 2 mm. Thus, you can get measurements from the drawings by one of two ways:
  • Count the number of squares. This is the number of centimetres those squares represent. E.g. 5 squares = 1 cm x 5 squares = 5 cm. 
    • This only works accurately for straight vertical and straight horizontal measurements, not diagonal or curved ones.
  • Take a measurement (in millimetres) on the pattern using a ruler, tape or piece of thread and multiply this number by five. You can then convert this to centimeters by dividing it by 10. E.g. if you measured a distance of 6 mm on the pattern, the same distance on the original garment would be 6 mm x 5 = 30 mm. 30 mm / 10 = 3 cm. 
    • This works for all measurements - straight or curved / vertical, horizontal or diagonal / etc.
    • You can use a dressmaker's measuring tape held on its edge to follow the shape of and thus measure the length of curves. 
    • Alternatively, you can lay a piece of thread over the curve of a pattern and cut it to the length of the curve. You can then place this piece of thread onto a ruler to measure the length of the curve.


2) Make a labelled sketch showing all the measurements you can find.

Don't worry if you "can't draw" (I can't either). This does not have to be a masterful feat of art. You just need to a) show all the pieces (and all the seamlines) and b) cover the drawing in as many clearly-labelled measurements as you can find or extrapolate from the books. Drawing it to scale doesn't matter - that's why you've got the measurements on it! (Though, sure, do that if you find it easy / enjoy the challenge.)

Here's my labelled sketches of Herjolfsnes 42:

You can really extrapolate a lot from those handy grey
shapes overlaying the patterns in MGR. (Note: I've since spotted a
few errors on here - the measurements listed below are more accurate.)


3) Make a list of the key measurements

On my diagram, I wrote down all the measurements I could find or extrapolate. (You never know when they might be useful!) However, for the purposes of reconstruction I was most concerned with a) the torso:hem circumference ratio being a similar to the originals (to get the correct silhouette) and b) the side gores:panels ratio being similar to the originals (to avoid the dreaded princess seams).

Comparisons between horizontal/circumferential body measurements and length measurements are, in my opinion, less useful. The ratio of height:torso circumference varies greatly among different people (and even on the same person depending on how much gym vs. pizza they've been doing or comparing them at age 20 vs. age 50). The Herjolfsnes garments have the added problem that a) we don't know which ones are men's or women's (or, indeed, adults' or older teens') and b) they may well be short compared to mainland European clothing (or at least women's mainland European clothing) of the same time. Hence, using vertical measurements may make things a bit wacky (at least from a mainland European perspective).

You could, if you wish, compare armscye, sleeve-length and sleeve-circumference proportions. I did not, since I was predominantly concerned with torso/skirt proportions (and also I know I have crazy-long arms compared with most women - and, indeed, many men who are my height). However, the below methods should work equally well for generating these proportions.

Here are the key measurements I took from Herjolfsnes 42:
  • Length of garment (front): 119.5 cm (WitE, p. 171).
  • Length of garment (back): 128.0 cm (WitE, p. 171).
  • Circumference at hem: 333.5 cm (WitE, p. 171).
  • Width of front panel at height side gore is inserted at (point d on pattern): ~31.5 cm (MGR, from pattern on p. 68).
  • Width of back panel at height side gore is inserted at (point k on pattern): ~36.0 cm (MGR, from pattern on p. 69). 
  • Width of front panel at base: ~ 40.0 cm (MGR, from pattern on p. 68, ignoring centre slit).
  • Width of back panel at base: ~ 39.0 cm (MGR, from pattern on p. 69, ignoring centre slit).
  • Width of left side gores at tip: ~3.0 cm (front one) and ~4.0 cm (back one) (MGR, from pattern on p. 70).
  • Width of right side gores at tip: ~3.0 cm (front one) and ~4.0 cm (back one) (MGR, from pattern on p. 71).
  • Width of left side gores at base: ~38.0 cm (front one) and ~40.0 cm (back one) (MGR, from pattern on p. 70).
  • Width of right side gores at base: ~40.0 cm (front one) and ~40.5 cm (back one) (MGR, from pattern on p. 71).
  • Width of centre-front gores at base: ~24.0 cm (left one) and ~23.0 cm (right one) (MGR, from pattern on p. 68).
  • Width of centre-back gores at base: ~22.5 cm (left one) and ~22.5 cm (right one) (MGR, from pattern on p. 69).


4) Use these measurements to get any more measurements you need

For me, the key things I wanted to know were:
  • How many times bigger is the hem circumference compared to the circumference at side gore insertion height?
  • At side gore insertion height, what percent of the torso circumference is:
    • Centre-front / centre-back panels, vs.
    • Side gores?
  • At hem, what percent of the circumference is:
    • Centre-front / centre-back panels, vs.
    • Side gores, vs.
    • Centre-front / centre-back gores?
To do this, you need the total circumference at both of the two points listed. Total circumference at the hem is easy, since it's listed in WitE as 333.5 cm (see above). Total torso circumference at side gore insertion height can be calculated by summing the width of the centre-front and centre-back panels at this height plus the width of the tip of each side gore. So:

Front panel + back panel at this height = about 31.5 cm + about 36.0 cm = about 67.5 cm
Left side gores' tips = about 3.0 cm + about 4.0 cm = about 7.0 cm
Right side gores' tips = about 3.0 cm + about 4.0 cm = about 7.0 cm

Total torso circumference at this height = about 67.5 cm + about 7.0 cm + about 7.0 cm = about 81.5 cm

Or, you could cheat and instead use the "chest width" measurement for the size small pattern for Herjolfsnes 42 given in MGR (p. 66), which is 79 cm. The small pattern sizes are, supposedly, equivalent to the original garments.


5) Calculate ratios (be prepared for some rounding/fudging)

First we will tackle the armpit-insertion-height-of-torso circumference:hem circumference ratios.
  • The total circumference of the torso at armpit insertion height is around 79 cm or 81.5 cm. Let's call it 80 cm for the sake of a nice round number. 
  • The total circumference at the hem is 333.5 cm. Let's round that to 330 cm.
So, how many times bigger is the hem than the chest circumference?

Percentage of chest circumference in hem = (330 / 80) x 100 = 412.5 %

~:~

Next, we will tackle the gore:panel ratios at armpit insertion height.
  • There are two gores per side (one at the front and one at the back). The front ones are each around 3 cm at their tips, the back ones are each around 4 cm here. So, let's call it 3.5 cm on average. 
  • (Yes, I said there would be a lot of rounding and fudging! You can be more precise if you wish!)
So, what percentage of the chest circumference is each side gore's tip?

Percentage of chest circumference accounted for by each side gore = (3.5 / 80) x 100 = 4.4 %

There are four side gores in Herjolfsnes 42, so all the gores make up 4.4 x 4 = 17.6 %

So, the front and back panels together make up 100 - 17.6 = 82.4 %

So, the front or back panel each makes up 82.4 / 2 = 41.2 %

~:~

Lastly, we will tackle the hem ratios, which is a bit more complex since there are more bits (and they're not all as similar as side gore tips are to each other).
  • There are two gores per side and their base measurements are around 38 cm, 40 cm, 40 cm and 40.5 cm. So, let's call that about 40 cm on average. 
  • There are also two gores at centre-front and two at centre-back and their measurements are around 24 cm, 23 cm, 22.5 cm and 22.5 cm. So, about 23 cm on average. 
  • There's also a front and a back panel, which measure around 40 cm and 39 cm at the hem. So, about 40 cm on average.
So, what percentage of the hem circumference is each side gore's base?

Percentage of hem circumference accounted for by each side gore = (40 / 330) x 100 = 12.1%

And the centre-front and centre-back gores?

Percentage of hem circumference accounted for by each centre gore = (23 / 330) x 100 = 7.0%

And the front and back panels?

Percentage of hem circumference accounted for by each panel = (40 / 330) x 100 = 12.1%

You can then check your calculations by seeing if they add up to 100% (i.e. a full circumference). Four side gores (4 x 12.1%) plus four centre gores (4 x 7.0%) plus two panels (2 x 12.1%) equals 100.6%, so we're pretty close!


6) Convert these ratios into actual measurements for a garment that fits you

All you need to do this is one measurement from your body. Yes! Just one.

(Ok, slight lie - you need other measurements to work out the height of the panels and for all of the sleeve malarky, which I'm not going to really go into here. To do those, you either a) do them as you would when making a normal dress with your normal patterning-from-measurements techniques or b) you use the above methods to work out length and sleeve proportions. However, for doing all the circumferential measurements you need just one measurement off your body.)

So, go measure the circumference of your torso plus a little ease, just about where the bottom of your armscye would fall.

My torso circumference at this height, with a little ease, is about 98 cm.

We know that the hem circumference should be 412.5 % of the torso circumference. So, based on the above ratios:
  • My hem circumference = 98 cm x 4.125 = 404.25 cm
  • Let's call that 400 cm for ease of future calculations.

We can now calculate all the (horizontal) sizes of our panels/gores!


Each of the four side gores should have a tip that is about 4.4 % of the torso circumference. So...
  • My side gore tips = 98 cm x 0.044 = 4.3 cm 
  • I'd probably call that 4.5 cm for ease of cutting.

Each of the two panels (front/back) should be about 41.2 % of the torso circumference. So...
  • My panel widths = 98 cm x 0.412 = 40.4 cm
  • Always best to round this one UP a little, as if the method is going to go wrong this is where it's most at risk of going wrong. We want to add any extra ease or extra in case of error into the front/back panels not the side gores to avoid the dreaded princess seams. So, perhaps call it 42 or 43 cm. You can always remove this extra cm or so in the draping stage if/when it becomes apparent it's unnecessary.

Each of the four side gores should have a base that is about 12.1% of the hem circumference. So...
  • My side gore bases = 400 cm x 0.121 = 48.4 cm
  • I'd probably call that 50 cm for ease of cutting.

Each of the four centre gores should have a base that is about 6.9% of the hem circumference. So...
  • My centre gore bases = 400 cm x 0.069 = 27.6 cm
  • This could be rounded to 27.5 cm or to 30 cm, depending on what you feel is easier.

You could, if you wish, cut the front and back panels as very slight trapezoids, as indicated by the patterns in MGR, and use the above method to find the base width of these panels. Personally, I didn't bother as I felt that the few cm difference between the top and bottom of these panels was probably down to one of the following a) stretch in wear, b) slight tapering in/out by adjusting seam allowances (purposely or not) when sewing, c) the centre slits potentially confounding things a bit.


7) Put these measurements onto a cutting diagram

As a recap (and not including any seam allowances, barring the allowances stated for the false seams):
  • My panels are each going to be about 42-43 cm wide.
  • My side gores are each going to be about 4.5 cm at the tip and 50 cm at the base.
    • One pair of these is going to be made by separating one larger piece using a false seam, as per the original Herjolfsnes 42.
    • So, the pair cut as one and separated by the false seam will measure at the tip 9 cm + seam allowances (for the false seam) at the tip and will measure at the base 100 cm + seam allowances (for the false seam).
    • The other pair will each be 4.5 cm at the tip and 50 cm at the base.
  • My centre gores are each going to be about 27.5 cm at the base.
    • If I cut them as separate gores, I will use these measurements.
    • If I cut the front or back as one large piece separated with a false seam the base of this large piece will measure 55 cm + seam allowances for the (false seam).
    • I always cut "triangular" gores as slight trapezoids, with the upper tip measuring two seam allowances in width; see my gores with M-shaped tops tutorial for more details on these gores.
To this, I added vertical measurements based on my measurements/how long I would like the garment to be.

Finally, I added seam allowances. I was then ready to cut out the fabric.

Here's what my cutting pattern sketches looked like:

Yes, I use inches when dressmaking. No, I have NO IDEA why
I thought 1 inch seam allowances were necessary when I made this pattern.


8) A little draping

If you have looked at the patterns in MGR you will see that whilst the front and back panels are very rectangular and the centre-front and centre-back gores are very triangular, the side gores are somewhat shaped. This aspect of the garments is convincingly argued by Robin Netherton in her article in Medieval Clothing and Textiles 4 to be the product of draping. I.e. pinning the garment in whilst it is on the wearer.

The important thing when draping the side gores is a) the shaping is mostly in the torso area, with the 'skirt' area being pretty much the original triangle proportions and b) you must avoid princess seams.

Avoiding princess seams is simply a matter of where you remove fabric from. For Herjolfsnes 42, it's apparent from the grey pattern shapes in MGR that fabric is taken out of the sides of each pair of gores, but not from between the two pairs of gores or from the front or back panels. So, like this:

Dots = where seam lines sit after draping the side gores based on Herjolfsnes 42.
(The 'wine bottle shape' of the pair of side gores is somewhat exaggerated to make things clearer.)

Some of the other garments (e.g. Herjolfsnes 38) also remove fabric from between the pairs/sets of gores. However, none remove fabric from the front or back panel -- this (plus getting the panel:gore ratios correct as described above) is how you avoid those side seams creeping towards the front of the garments and becoming like princess seams.

So, there you have it: a Herjolfsnes-style garment (minus sleeves), patterned to your measurements but retaining the key proportions of the original garment.


Please do let me know if you try this method. I'd love to hear how you get along!

1 comment:

  1. Amazing! I am 100000% new to Medieval sewing, and have been captivated by the Herjolfsnes finds. I just did something similar to this ratio thing to make a practice dress for my daughter, just to see if I could do it. This really clarifies it for me! I don't have either Woven into the Earth or Medieval Garements Reconstructed, so I've just been working from info on Marc Carlson's site. It's hard not knowing the gore widths, so this is helpful :) Thanks for such a thorough post!!!

    ReplyDelete