Based on private notes of inquiry and discussions on various historical needlework-related boards and forums of late, I see that people are still confused about the working logic of linear stitching. In specific, how to determine if a design can be worked entirely two-sided.
First off – the two most popular historical methods for working thin linear designs are double running stitch and back stitch. The big difference between the two is the appearance of the reverse. Done meticulously, with care paid to invisibly terminating threads, double running stitch is almost indistinguishable front and back. Almost because a few people do produce a slight difference due to differential thread tension on each of the two passes required to produce a unbroken line, but that difference mostly settles out over time. Back stitch on the other hand produces a public side very much like double running, but the reverse of the work is heaver, and depending on the stitcher can look like outline or stem stitch, or even like a split or chain stitch if the needle pierces the previous stitch as a new one is made. Of necessity in back stitch there is twice as much thread on the back of the work as there is on the front.
Double running stitch takes two passes to accomplish because it first lays down a dashed line, with the spaces between the dashes being filled in on the second pass. A back stitch line is completed in one pass, with no need to revisit areas previously stitched to complete the line.
Many people prefer back stitch because there IS no going back. They like the certainty of knowing exactly where they are at all times, over the pretzel logic of calculating how not to be caught in a cul de sac while retracing steps in double running. Personally, I prefer double running, and follow double running logic even if the piece I am working will not be seen on both sides. I find that path planning to be fun, and I appreciate thread economy, especially when working with more costly or difficult to source hand-dyed silks.
But for some one challenge of double running is knowing which designs can be worked in that stitch such that both sides can be made totally identical.
It’s easy. Any design that has no “floating elements” is a prime candidate. If true double sided is a total goal (including invisible termination of thread ends), any piece that has a floating element large enough to allow that burial is also a possibility. It doesn’t matter how complex a design is, so long as elements are all branches and detours off of one or more main baselines, they can be stitched double sided. And yes – there CAN be more than one baseline in a design. More on baseline identification is here. The logic of following detours and returning to the baseline is here. How to break up a large design into several smaller baselines is here.
Identifying floating elements
That’s easy. They are any ornament or detail that is discontinuous from the main line of the design.
Here are several that I’ve done in double running, based on one or more continuous baselines, with no floating element deviations. In these designs every part of every work is attached to every other part, at one or more points.
By contrast, here are several that have those “floating elements” called out.
The knot element in the all-over at left is not attached to the main pomegranate frame. It is however just large enough manage thread-end-hiding. So while its presence makes this a tedious and difficult pattern for double-sided double running stitch, it is not a deal breaker. However those little accent diamonds are deal breakers. Too small to hide the ends, and detached from the main design. The ladder element in the arms of the repeat at right is broken from the main design, and is too small for end-camouflage.
There are often short lines or sneaky little floating accents hidden in both simple and more complex repeats. Strawberry pips are notorious for this, although I haven’t any stitched examples to hand:
My dragonbeast, however lovely, has quite a few floating elements, making him a problematic choice for a fully double-sided work. (Eyes and faces are almost always difficult).
And this bit, stitched from a Lipperheide book, is the absolute poster child for discontinuity. I didn’t mark them all, but you get the idea. The spaniel and possibly that center bundle thing are the only bits large enough in which to bury the ends, if a fully two-sided result is desired.
Here’s a tricky one. Look closely at the bit on the left.
It looks continuous, but it’s not. There are in fact FOUR separate double-running baselines, AND a discontinuous element in the motif. He’s in the red circle on the right. Like the round knot in the first example this might be done double sided, provided that the stitcher was willing to terminate separate ends for that relatively large floating element.
So in short – it doesn’t matter how complex a design is, so long as all elements are continuous it CAN be stitched fully double sided, in double running stitch.
Accreted Section Double Running Stitch Logic
I promised to discuss a second logic for double running stitch. I call this one “accreted section” and use it for the more complex patterns, especially non-linear ones.
What’s a linear pattern? Pretty much any of the banded strapwork style strip patterns I’ve been using on the Do-Right and Clarke’s Law samplers. However the phoenix from Do-Right is distinctly non-linear:
I could work the phoenix using the baseline method by identifying one of a zillion possible baselines and following it in the usual way. The outline would make a good baseline:
If I were to use baseline, I’d start at a point on the outline, then work in the indicated direction, following the little detours as I came to them. But in this case there are LOTS of detours. It’s too easy to get lost. For example, If I were to start at the indicated spot at the base of the flight feather, then continue up to the wingtip, it looks like I’d be following the little striations on the first feather. Not too hard. Little lines and hatchings like these make the pattern easier to follow because they can be easier to count than straight runs of stitching with few reference points to use for location verification. BUT I have a lot of possible detours. It’s very easy to start the feather with the stitch from the feather base to the first striation, then verge off down that bit of shading to the lower part of the wing and from there get lost in the body.
I find it easier to break up patterns like this into logical units:
Purists will note that the blue segment might be considered a baseline, with the other elements as detours off of it, but that’s quibbling. For me at least, parsing the pattern into three units helps keep me on track. When I stitched this I started with the blue unit, working the bird’s neck and breast detail as departures from that line. I did have the luxury of not needing to do this piece double sided, so I did begin a new strand to work the green section, stitching the feather striations and other connected bits as I went along. The same with the orange section. I did that last, again working the feather stripe and flame detours as I came to them. The flame section at the pattern’s bottom left is a closed loop departure off of the orange line.
Baseline First/Hybrid Logic
So far I’ve established a visual baseline, then worked along it, stitching all departures from that baseline. The last step has been to stitch back along the baseline to complete the work. But sometimes it’s better to stitch the baseline first. Occasionally I work a pattern that way – making my first pass along the baseline to outline or otherwise establish the location and veracity of pattern placement, then working the details or fillings on the second pass:
You can clearly see that I did that on this strip from Do-Right. I started with a baseline that outlined the flower, then on the second pass, filled in the petal details. In this case I worked using a hybrid logic. Instead of establishing one baseline for the entire repeat, I worked it more along the lines of the accreted method above – isolating the flower, then the branch from which it buds, and then the branch segment that connects this flower to the next (flipped) repeat.
For some very wide patterns, this mixed approach works best, especially if you’re using an in-hand tambour style round frame. With a round frame the area that’s taut and ready to work is quite small. Large repeats easily occupy more than the space at hand:
Being limited to the frame’s real estate lends itself to this compartmentalized, hybrid approach.
Having the luxury of using a flat, slate, or roller frame that provides acres of tautness makes a sprawling approach easier:
In any case, this concludes the series on double running stitch logic. Please feel free to ask questions. I don’t pretend to know it all, but chances are I’ve faced some of the same stitching problems that might be challenging you, and I’d love to help.
Well, with luck this mini-series will be useful to someone out there. There is someone out there, right?
Continuing from the last post, now that we’ve figured out that our design can be worked double sided, and we’ve identified the baseline, how do we go about the stitching? What logic can we follow to ensure that no areas of the pattern are orphaned, and that all lines are covered?
The method most often followed is the baseline method. The other I’ll call “accreted section” and deal with it in a later post.
I find baseline to be pretty easy – nearly foolproof, provided the stitcher can remember the nested logic of detours-within-detours encountered along the way. Some stitchers keep a paper copy of the design, and overtrace it to keep track of where they’ve been. Some use a paper copy to try out their logic before committing needle and thread. I have to admit I do neither. I just go for it.
Let’s look at this pattern. It’s from my very first booklet on Blackwork, a hand-drawn photocopied piece done in my teens and distributed entirely within the SCA. I know I got this particular design from a historical source, but my original annotation wasn’t complete enough for me to include in later books or in fact to find the source again, so this design has sat on the shelf ever since. I’d consider this one to be a pattern of intermediate complexity, but well within the reach of most beginners (click on pix below for enlargements).
To parse out the stitching logic, let’s look at a half repeat. I’ll illustrate the entire stitching path for one half repeat. The logic to complete a whole repeat is very much the same. In the pix below, green indicates my first pass of double running, and blue marks a return path, in which I retrace my steps. The first stitch and its direction in any pass or return is marked by an arrow. Click on any of the drawings to enlarge. And please keep in mind that the method below is just one of a huge number of possible paths through this particular pattern. Path planning and trying out different strategies is what keeps this style of stitching fresh to me. Which is to say there’s no guarantee at all that I work every repeat in exactly the same manner. YMMV.
To work this design double sided, I’d start along the baseline leaving at least three inches of thread extra on the back (no knots). I find it helpful to wind the excess around a pin placed in an inconspicuous spot. I travel along the baseline (1) in double running stitch until I encounter a branch. My preferred logic is to then follow the branch to its end, then turn back and fill in the “every other” running stitch, to eventually return to the baseline (2). Then I continue on to the next branch decision and follow that detour (3).
In this case I’ve gotten to the first of the double bracelets on the main stem. Unless a branch is a turn left only branch, given a choice, I tend to turn right. Gamers, the mathematical and those who study behavioral sciences or robots/autonomous navigation will recognize this – it’s a classic. Any maze can be successfully navigated by putting one’s hand on the right hand (or left hand) wall, and following it, without taking one’s hand from the wall. The path traced may not be the most efficient, but sooner or later, the wall-hugging, maze-wandering mouse, robot, or high school topology student will emerge from the exit.
So here I am at the top of the brackets. I could continue down and wander around the bracelets, or I can turn right again and follow the main stem back to the half-heart motif on the left edge of the swatch area. I take a right hand turn from my line of travel, and stitch back up to the main motif (3). When I get there, I notice one little tiny detour – the single stitch between my current line and my starting point. (4) makes quick work of that. Then I continue around the necklace at the base of the heart motif. Again I turn right (5), then double back on my path and continue down and around the wing at the base of the heart (6).
After completing the first pass at the base of the heart and ending up at my “bounce line” – the centermost point of the strip repeat – I do a mini-step back to the heart’s outline (7), then I continue around the heart’s perimeter, eventually reaching the detour point to complete the small inset detail in the heart’s center (8). Again I stitch to the bounce point, and then return to the heart’s perimeter (8).
Once I’m back at the edge of the heart, I can do the antenna that sticks up from its top (9). Heading back from there turns out to be a long run all the way back to my baseline, filling in all of the “missing” stitches to complete the first half of the left hand heart motif (10). Now for a minorly tricky bit – one that folk unfamiliar with double running stitch logic occasionally miss – the little detours that fill in the bracelets around the stem. It’s easy to miss stitches in these, and very easy to get lost, not remembering which way to turn next. We’ll step through.
The first bit is to progress along our baseline. The initial stitch is marked with the arrow. I work it, then the two stitches along the bottom of the upper bracelet, followed by the stitch that completes the three that define the top of the bracelet (12).
Time to head back to the baseline, but it’s not very far away. One stitch brings us back to it (13). On the next step because it’s extra confusing, I’ve marked two stitches with arrows. First I head south from the upper bracelet, then work around the lower one (14). There’s now one stitch left to finish defining the box between the two bracelets. I take that one stitch (15).
Now I’m ready to return to the baseline again. A couple of quick stitches takes me there (16). If you look at the work now, you’ll see only one “unfilled” path through the two bracelets area. That’s the path of our baseline. All of the other stitches have been completed, and none are orphaned, unworked. Now to progress along the baseline again. I detour for the little side curl, worked there and back again style just like I did before (17, 18) landing me back on the baseline again.
The logic should be a bit more obvious by this point. I progress along the baseline, making a detour back up to complete the outline of the stem unit (19). And back again to the intersection just below the necklace at the base of the next heart flower (20), and up around it (21).
Now I move on to the wing section that defines the lower edge of the flower (22). As before, having hit the center point, I head back to the outer edge of the heart (23), then continue around the heart’s perimeter, and down into its center detail (24).
Almost done now, there’s just heading back out to the edge of the heart (25), and doing the first half of the antenna (26). Our grand finale is here! Starting at the antenna, we work all the way back around the heart’s edge, and then all the way back to the beginning of our pattern, following the established baseline. At this point there’s no more counting, just following the snail trail laid down before (27):
It’s done! The entire half repeat – worked 100% two sided in double running stitch, with no little orphaned areas left unstitched. We worked through the baseline concept on a pattern of moderate complexity, stitching along detours as they present themselves, always returning to the baseline before moving on, and leaving one long final unifying run along that baseline to finish off the pattern. Yaay!
O.k., some of you ask. “Smarty pants, that all works great for the half-repeat shown above, but what about the full repeat?” I answer – the logic is the same. With the exception of the antenna which needs to have both “ears” worked one after another the first time they’re encountered, the stitcher can follow the “to the center” logic above, verbatim, or can work each heart flower as an entire unit when it is first encountered, following around its entire perimeter up to the point of return to the baseline before doubling back around the heart to arrive at the original spot of departure from the baseline.
If you’ve got questions about this logic, please post them. I’ve already gone on long enough for one post. The next post will be on the accreted section method and when to use it or the baseline approach. The series will end with how to finish off ends invisibly for double sided work. Hope this is helpful!
There’s been a discussion of late on the Yahoo Blackwork embroidery discussion group about stitch order, direction and stitching logic in double running stitch – especially reversible (two sided) double running stitch. I contributed to the discussion with these thoughts, but the answer below is a bit of an elaboration on my original discussion group post.
Double Sided Double Running Stitch – Is it possible for your chosen pattern?
The first thing to do is to determine whether or not your contemplated design can be done 100% reversible. Those that can have every design element connected. There are no floating little diamonds or sub-motifs off on the side un-connected to the main design. This simple design is easy to do two sided:
This one, although vastly more complex, only presents a couple of challenges. The center diamond in the nodule at the base of the plume flower is one. Every other element is connected, but that one diamond stands alone. If I were to work this design double sided, I’d add a stitch to the top and bottom points of the diamond to connect it to the rest of the design. The visual impact of that modification would be minuscule. The other challenge is the presence of some detached stitches in the “bark” area of the branches – the little floating verticals unattached to the main body of the work. If I were to do this one double sided, I’d either omit them, or lengthen them to intersect with a segment of the branch’s outline. A pain, but not totally fatal, and both changes wouldn’t be very evident.
By contrast the column and wreath design below, though simpler, presents a greater challenge for two-sided stitching. Each of the small circlets in the centers of the wreath units stands alone. Attaching them to the rest of the work would diminish the impact of the design. Although the rest of the design can be worked entirely two sided, the circlets are problematic because they’re free floating and rather small. If they were worked independently, with their own lengths of thread, there isn’t enough real estate in each one to cleanly hide the thread ends.
Much of this mermaid panel can be worked double sided, but by now you can spot the facial features, fruit dimples, flower centers (and prominent nipples) as presenting problems that can’t be solved by modifying the design. BUT the small dolphins, although separate from the main pattern aren’t a problem. They’re big enough to stitch with their own threads
O.k. Now we’ve determined which designs can be done double sided without modification – the ones that have no isolated design elements. On to stitching logic.
Stitching Logic – Baseline
I use two methods for completing a double running stitch pattern – baseline and accreted section. I’ll tackle baseline first
In the baseline method, the stitcher identifies a line that travels the entire length of the pattern. That’s the baseline. It can be obvious, like a stem from which all of the pattern’s flowers grow, or part of an outline; or it can be less obvious. In this oak leaves and acorns border, one baseline is blindingly obvious:
In this pattern it’s slightly less clear. Any one of several options can be used as an effective baseline:
Here’s one possible baseline:
Every other element of the design can be worked as a detour off this main highway. You’ll note that the baseline needn’t march around the perimeter of the acorn. In fact the entire acorn is one nested set of detours. And this isn’t the only possible baseline. Here’s a more efficient though less intuitive one:
Both are perfectly logical. I might use the one at the top if I wanted to quickly establish the height of my piece. It’s just one unit shy of total pattern height. But the only reason to chose one or the other is personal preference. Please note that the logic of these to baselines applies equally well to the horribly complex plume flower:
I’d suggest folk new to double sided work start with patterns with easily identified baselines, and work up to some of the more daunting patterns.
Next post – stitching logic. Traveling along the baseline and its detours.