Melange yarn looks like the simplest purchase in the world. You see a heather grey tee, you point at it, you say “that, but in a 30s cotton,” and you assume the rest is just paperwork.
Then the bulk shows up in four bags, four slightly different colors, and someone upstream is asking why the last drop looks like it came out of a different factory. Short answer: it kind of did. Melange is a blended-color product made from dyed and undyed fiber mixed before spinning, which means every single thing about it — shade, count, evenness — is a range, not a number. Buyers who don’t know that get eaten alive.
This is the stuff that usually only gets explained after you’ve already paid.
What melange actually is, and why it changes how you buy
In a normal dyed yarn, the color goes on after the yarn exists. Consistent, predictable, boring. In melange, you dye fiber (or buy pre-dyed fiber), then blend it with undyed or differently dyed fiber, then spin. The color lives inside the yarn as a random distribution of colored and uncolored fibers.
Consequences you need to internalize before you send a single inquiry:
- Shade is created by ratio, not by a dye recipe. “10% black melange” is a completely different color depending on fiber fineness, dye depth on the colored component, and how well the mill blends.
- Two mills can both honestly sell you “10% black melange” and hand you two visibly different yarns.
- Neps and specks are not defects. They’re the product. Fighting them to zero means you don’t understand melange.
- Reproducibility is the hard part, not the first run.
Shade: the part nobody warns you about
Stop treating shade as a single value
Ask any experienced buyer and they’ll tell you the same thing: a melange shade is a distribution. You are approving a range, and if you don’t define that range in writing, the mill defines it for you — usually as “whatever came out.”
So your spec needs actual numbers, not a swatch photo:
- Target color values (L*a*b* or equivalent) measured on the finished yarn.
- A tolerance window, and what happens if a batch lands outside it.
- The illuminant(s) you’ll evaluate under — at minimum two, so you catch metamerism before your customer does.
- How many measurements per lot, and on what instrument.
A better tool than a single tolerance: a shade band. Get three to five approved cones spanning the acceptable light-to-dark range, sign them, and keep them. Now “is this in spec” is a physical comparison instead of an argument.
The lab dip trap
Here’s the uncomfortable one. Lab dips are frequently made by hand-weighing small fiber quantities and blending them in a little lab mixer. Bulk runs through hoppers and chute feed with completely different blending efficiency. A lab dip can be a beautiful, honest, totally unreproducible preview of a color you will never receive.
So don’t approve on a lab dip alone. Ask for a sample off production machinery, even a couple of kilos, and treat that as the real reference. If a supplier can’t produce a production sample, that’s information too.
And never, ever approve a shade from a photo or a phone video. Auto white balance, screen calibration, and compression will all conspire against you.
Ask what the blend ratio actually is, and how it’s controlled
“Melange” tells you nothing about percentages. Ask directly: what’s the percentage of dyed fiber, is it weighed per batch or dosed continuously, and what’s the batch size? A mill doing 20 kg hand-weighed batches and a mill doing 500 kg hopper batches have very different variation profiles — neither is wrong, but you need to know which one you’re buying.
Count: nominal, actual, and the weight trick
Nail down the count system first
A spec sheet that just says “30s” is a payment waiting to happen. Cotton count, metric count, and tex/denier are different universes, and the ambiguity almost always resolves in the seller’s favor. Get the system, the nominal count, and the tolerance written down.
Check what the tolerance is hiding
- Count deviation: a couple of percent either way is normal. Beyond that you’re getting less yarn per kilo than you paid for.
- Twist and twist direction: affects hand feel, pilling, and how the melange reads visually.
- Evenness and IPI: ask for the mill’s baseline neppy level for that specific shade, then hold them to their own baseline. Don’t invent a standard they can’t hit.
- Splicing vs knots: matters enormously if you’re knitting. Knots become holes.
- Waxing or lubrication if it’s going onto a knitting machine.
Buy on conditioned weight, not scale weight
This is the quiet one. Yarn weight changes with moisture. If you’re invoiced on the raw scale weight at the door, you’re paying for water on a humid day and getting shorted on a dry one. Ask whether invoicing is on actual weight or conditioned weight at a stated moisture regain, and get the answer in the contract. It’s a small line item that compounds across a container.
Supplier checks that separate real mills from repackagers
Tear a sample apart
Grab a length, untwist it, and separate the fibers. True melange gives you individually colored fibers mixed throughout. What you sometimes get instead is two differently colored plies twisted together, or a yarn with color printed on the surface only. Those look convincing in a swatch and behave totally differently after washing, abrasion, or dyeing. A quick scratch or rub test exposes the surface-only version fast.
Ask about fiber substitution — in writing
Substituting a cheaper fiber mid-lot happens more than anyone admits, and melange hides it better than solid shades because the eye is busy. Require fiber composition verification per lot if the blend is part of your product claim.
Test reproducibility before you commit to volume
Ask the supplier to show you measured shade data across their last three production lots of that shade. A real melange spinner has this record. A repackager has a story. If the shade is a custom blend, ask what happens when their dyed fiber lot runs out — because that’s the actual reason colors get discontinued without warning.
Certifications need to be shipment-specific
Generic certificate PDFs are decoration. For recycled-content or chemical-safety claims, ask for the transaction-level documentation tied to your specific shipment or order ID. If they can’t produce it, the claim doesn’t exist, no matter what the brochure says.
A workflow that keeps you out of trouble
- Define the target in numbers: color values, tolerance, illuminants, count system, count tolerance, twist, composition.
- Get a lab dip and a production sample. Approve based on the production sample.
- Build a signed, sealed shade band with retained standards for both sides.
- Run bulk with per-batch measurement data supplied, not requested after the argument starts.
- Measure every incoming batch and chart it. Drift shows up as a trend long before it shows up as a claim.
- Keep one reference cone from every lot in a dark drawer. It’s the only evidence that survives a year.
Quick pre-order checklist
- Count system named, tolerance agreed
- Dyed fiber percentage and blend control method confirmed
- Color values with tolerance and two illuminants
- Physically signed shade band held by both sides
- Conditioned weight basis in the contract
- Baseline neppy/evenness levels documented
- Reproducibility history reviewed
- Shipment-level certification documents confirmed
- Splicing vs knotting specified
- Retained sample from each lot, stored
The bottom line
Melange is a product that punishes buyers who treat it like a solid color. Nothing about it is a single fixed value — the shade is a spread, the count is a tolerance, and the supplier’s real capability only shows up on the second and third order. The people who buy it well aren’t doing anything exotic. They’re just defining the range in writing, holding a physical standard, and refusing to approve a color from a screenshot.
Do that, and the four different bags in the shipment stop being a crisis and start being a Tuesday.