1. Merle Fundamentals
Merle Fundamentals opens with a single rule this module keeps returning to, because it is the rule breeders get wrong most often, and getting it wrong has real welfare consequences: merle is never carried. A dog either has a merle allele at the M locus and is a merle, or it does not and is not. There is no third, hidden state where a dog quietly holds one copy of merle without being one — the way a Bb dog quietly holds one copy of brown without looking brown. That distinction is the spine of this whole module.
Merle is caused by a variant at the M locus, the gene PMEL. Unlike the two-allele loci covered in Level I, the merle variant is not a simple substitution but a mobile genetic element, a SINE insertion, that has landed inside the gene and disrupted it. What matters at this module's depth is not the deeper mechanics of that insertion — the site's engine actually models a fuller seven-class, length-based system for its dedicated merle tools, and the full classification and pairing-risk matrix stay Level III material — but the basic consequence: a dog either carries at least one copy of a disruptive insertion at this locus, or it does not. The site's general calculator and its own Merle Reference Centre both lean on this simplified frame, M for a merle-capable allele and m for none, precisely because at this depth the visible-versus-not question and the never-carried rule matter more than which exact length class produced the result.
A merle dog's coat shows a mottled, patchy pattern: irregular patches of diluted pigment sitting next to patches of full-strength eumelanin. No two merle dogs pattern identically, because the process that produces the mottling is partly random during coat development. Merle also frequently affects eye color, producing blue eyes, blue-flecked eyes, or heterochromia (one eye of each color), because the same gene plays a role in eye pigmentation as well.
That visible pattern is exactly what makes ‘never carried’ counterintuitive at first. It is tempting to think a subtly patterned or barely patterned dog is ‘just a mild carrier.’ It is not. A dog with any merle-capable allele is a merle, full stop, regardless of how much or how little of the pattern shows. Two situations make this concrete.
The first is cryptic merle: an M-locus variant short enough that it produces no visible pattern at all. A cryptic merle dog looks, by eye, exactly like a solid-colored dog. It is not a carrier of merle — it is a merle, and it passes that allele to roughly half its offspring at the normal rate, exactly like any other merle allele. Bred to another merle dog, cryptic or visible, it carries the same double-merle risk any merle-to-merle pairing carries.
The second is a merle allele sitting on a phaeomelanin-only (red, apricot, or cream) coat — a dog that is ee at the E locus. Merle disrupts eumelanin; an ee dog produces no eumelanin in its coat at all, so there is nothing for the merle allele to visibly pattern. The dog looks like an ordinary solid red or apricot dog. The allele itself is completely unaffected by this: the dog is still genetically merle, still passes it on at the normal rate, and can still produce a double merle if paired with another merle dog. This is one of the most common routes to an accidental double-merle litter, precisely because both parents can look entirely clear.
Both cases point at the same operational consequence: a coat, by itself, can never rule merle in or out. The only instrument that resolves a dog's actual M-locus status is a DNA test.
The reason ‘never carried’ is not just a vocabulary preference is the health stake sitting behind it. When two merle dogs, of any kind — visible, cryptic, or ee-masked — are bred together, each puppy has roughly a 25% chance of inheriting a merle-capable allele from both parents at once: a double merle. Double merles carry a substantially elevated rate of deafness and eye defects, because the same pigment-producing cells involved in coat color are also required for normal inner-ear and eye development; without enough of them, those structures do not form correctly. This is why responsible merle breeding rests on one unconditional rule: never breed merle to merle, in any combination, and DNA test every dog whose merle status is not already confirmed before making that decision.
Coat-color naming for merle stays colour-first, exactly as it does everywhere else on this site: a black-based merle is the pattern on a black-based (B_D_) coat, a blue merle is that same pattern on a dilute (dd) coat and is the true dilute merle, a brown-based merle patterns a bb coat, and a lilac merle patterns a bb dd coat. If progressive clearing later acts on a merle dog's coat, the dog is not renamed into something else — it keeps its merle name, with the clearing possibility noted separately.
Worked example
Two Standard Poodles are being considered for a pairing. The sire has been M-locus tested and returned a clean, non-merle result: mm. The dam has been M-locus tested and returned a positive result with a visible classic merle pattern in her coat.
Because only one of the two parents carries a merle allele, this is a single-merle-to-non-merle cross, not a merle-to-merle cross. Each puppy inherits one M-locus allele from each parent: an m from the sire, and either an m or the dam's merle allele from the dam, with roughly even odds of each. That gives an expected 50% non-merle (mm) and 50% single merle across the litter — no double merles are possible from this pairing, because only one parent can contribute a merle-capable allele at all.
Contrast that with a second, hypothetical pairing: the same dam, paired instead with a sire whose owner describes him as ‘clear, definitely not a carrier,’ based on his solid apricot coat and a clean-looking pedigree, but who has never actually been M-locus tested. That description is exactly the failure mode this module exists to prevent. An apricot coat is ee, and merle does not show on an ee coat at all — a merle allele could be sitting there completely invisibly. Without a DNA test, this sire's status is not ‘probably clear’; it is genuinely unknown, and pairing him with a confirmed merle dam without testing him first risks an accidental double-merle litter neither party intended. The correct next step is not to proceed on the pedigree description — it is to M-locus test the sire before any breeding decision is made.
Uncertainty case
A prospective puppy buyer messages a breeder: ‘My last dog's litter had a merle puppy show up out of nowhere, from two solid-looking parents. Someone told me that means one of the parents was “carrying” merle. Is that right?’
Working through what is actually known versus assumed matters more than reaching for the easiest label.
KNOWN: both parents in that prior litter looked solid-colored, and at least one puppy in the litter visibly showed the merle pattern.
INFERRED, reasonably, from established M-locus biology: because a merle pattern only appears in a puppy that inherits a merle-capable allele from a parent, at least one of those two parents carried a merle allele at conception. That much genetics supports directly — the pattern did not appear from nothing.
Where the buyer's phrasing goes wrong is the word ‘carrying.’ The parent in question did not carry a hidden, non-merle copy of merle the way a Bb dog carries a hidden copy of brown. Whichever parent transmitted the allele either had a cryptic merle result (an insertion too short to pattern its own coat) or was ee and masking a visible-length merle allele entirely on a red or apricot coat. Either way, that parent was not a carrier of merle — it was a merle, fully, in the same sense the visibly patterned puppy is a merle. The only thing hidden was the pattern, never the fact of the allele itself.
UNTESTED, not unknowable: which of the two parents it was, and what class of M-locus allele is involved, are both resolvable questions — a DNA test on each parent would answer them directly. The breeder simply has not run that test yet. This is different from something like the poodle greying gene, where no validated test exists at all regardless of effort; here, the tool to resolve the question already exists and has simply not been used.
The honest correction to give the buyer: not ‘one of the parents was a merle carrier,’ but ‘one of the parents was already a merle, and it just was not visible on that particular coat.’ The distinction is not pedantic. It changes what testing is owed to every other dog descended from that parent, and it changes how the next pairing involving that line should be planned.
Summary
Merle is caused by a disruptive variant at the M locus (PMEL) that patterns eumelanin into irregular, mottled patches and can shift eye color toward blue, blue-flecked, or heterochromatic. The rule this module keeps at its center is that merle is never carried: a dog either has a merle-capable allele and is a merle, visibly or not, or it has none and is not a merle at all — there is no hidden intermediate state. Two situations make merle easy to miss by eye alone: cryptic merle, where the insertion is too short to pattern the coat, and merle on an ee (red, apricot, cream) coat, where there is no eumelanin for the allele to pattern in the first place. In both cases the allele is fully present and fully heritable; only a DNA test can resolve it. That matters because breeding two merle dogs together, in any combination of visible, cryptic, or ee-masked, carries a real chance of producing a double merle, with a substantially elevated rate of deafness and eye defects — which is why responsible merle breeding rests on testing every dog before pairing, and never breeding merle to merle.