Merle in Chihuahuas Is Not New: What the DNA Says

By Rev. T. Stringer, Lead Researcher, Stringer Integrated Pedigree-Genomic Study (SIPGS).

People say merle is a newer addition to Chihuahuas. They say it came in through a recent cross with another breed.

We did not want to argue about it. We wanted to test it.

So we asked the DNA.

The short answer

Merle came into the Chihuahua once, a long time ago.

Our best estimate is that the Chihuahua merle copy split from its closest outside relative about 115 years ago, around 1910.

That is not a 1990s cross. It is the era of the breed's founders.

How merle is inherited

Before the study, the basics. This is where most confusion starts.

Merle is dominant. One copy is enough to make a dog merle.

Every dog has two copies of the merle gene (PMEL), one from each parent. Each copy is either merle (M) or non-merle (m).

Parents	Puppies
mm × mm	0% merle. Always.
M*m × mm	About 50% merle, 50% mm
M*m × M*m	About 25% mm, 50% M*m, 25% two merle copies. Risk depends on allele length.

Merle cannot skip a generation.

A merle puppy must have at least one parent with a merle copy. Two true mm parents cannot produce merle.

So why does it seem to skip?

Because the merle parent did not look merle. This happens when:

• The dog is a cryptic merle (Mc or Mc+). These short lengths show little or no pattern.

• The pattern is very small, or hidden in a white area.

• The coat color hides it. Merle shows poorly on red, fawn and cream.

• The dog is a long coat, which can blur the pattern.

The merle copy was there all along. Only a DNA test shows it.

What a "merle-free line" really means

A pedigree cannot promise a merle-free line. Only testing can.

A dog that tests mm cannot pass on merle. A dog that was never tested might.

Allele length can also shift from parent to puppy. That is why we Embark test every puppy at birth. Embark shows if a puppy has merle. Length needs a separate test.

What we tested

• 380 Chihuahuas with Embark raw DNA data

• 177 dogs with a known merle result

• 61 merle dogs (M*m) and 116 non-merle (mm) dogs

• Merle dogs from more than one breeding program

• A reference panel of 5,406 dogs from purebred breeds and village dogs worldwide (Shannon et al. 2015)

This is part of our Stringer Integrated Pedigree-Genomic Study (SIPGS) and the Chihuahua History Project.

The science, step by step

Step 1: What merle is

Merle is caused by a small piece of extra DNA inside the PMEL gene on chromosome 10. The extra piece is called a SINE insertion.

A merle test checks for that insertion. Embark tells you if a dog has it, but does not measure its length. A length test, such as the one from UC Davis VGL, measures it. Those lengths are the Mc, Ma, M and Mh you see on results.

The test does not tell you how old the insertion is. Allele length can change from one generation to the next, so it is not a clock.

How old is merle itself?

No one has dated it yet. Merle has not been found in wolves, so it most likely arose in domestic dogs.

Our educated guess is that it arose in European working dogs, likely herding types, a few hundred to about 1,000 years ago.

By the time breed clubs formed in the late 1800s, it had already spread into breeds across Britain, Germany, France and Spain.

Our study asks a narrower question: when did it come into the Chihuahua?

Step 2: Genes travel in chunks

A puppy does not get single genes from each parent. It gets long pieces of chromosome.

So when merle is passed down, the DNA right around it is passed down too. Scientists call that surrounding stretch a haplotype. Think of it as the merle gene's "neighborhood."

Step 3: The neighborhood shrinks over time

Each time a dog makes eggs or sperm, its two copies of each chromosome swap pieces. This is called recombination.

Every swap can cut into the neighborhood around merle. Over many generations, the shared neighborhood gets shorter.

• Dogs that got merle from a recent common source share a long neighborhood.

• Dogs whose merle copies split long ago share a short one.

Step 4: Measuring in centimorgans

We measure the neighborhood in centimorgans (cM). One cM is the length of DNA that has about a 1 in 100 chance of being cut each generation.

If two copies share a stretch of L cM on one side of the gene, they split about 50 ÷ L generations ago.

The front of chromosome 10, where merle sits, recombines slowly. About 9.5 million DNA letters there equal only 2.7 cM. We used a published dog genetic map to convert this correctly.

Step 5: Sorting the two copies

Raw DNA data does not show which copy of a chromosome each letter is on. So we phased the data. Phasing is a standard method (Beagle 5.4) that works out which copy each letter sits on.

That let us find the merle copy in each merle dog and measure it directly.

Step 6: Comparing to other breeds

We searched for the Chihuahua merle neighborhood in 5,406 dogs from other breeds and village dogs. For each dog we measured how far it matched without a break.

What we found

1. One source

All 61 merle dogs carry the same merle copy. We found no second version.

Merle entered the breed one time, not over and over.

2. It came from outside the breed's other lines

The merle neighborhood is almost absent from non-merle Chihuahuas. In the 14 reference Chihuahuas it appears about 2% of the time.

So this stretch of DNA came into the breed along with merle.

3. The same merle seen in other breeds

The Chihuahua merle neighborhood matches the merle background found in many merle breeds, including Australian Shepherds, Catahoulas, Border Collies, Shelties, Great Danes and Dachshunds. That fits merle being one very old mutation shared across breeds.

The closest single matches were one Australian Shepherd and one Dachshund, nearly tied.

4. An old split

The closest outside copies share 1.2 to 2.1 cM with the Chihuahua merle copy.

• That works out to 24 to 42 generations.

• Middle estimate: about 33 generations

• At 3.5 years per generation: about 115 years

• That places the split around 1910.

A cross made in the 1990s or 2000s would be about 10 generations ago. That would leave about 5 cM or more shared. No breed came close.

Methods at a glance

• Data: Embark raw genotypes, chromosome 10, 380 Chihuahuas

• Merle status: Embark results (M*m or mm)

• Reference: 5,406 dogs, Shannon et al. 2015 public dataset

• Shared markers: 4,297 on chromosome 10

• Phasing: Beagle 5.4 (Browning et al. 2021)

• Genetic map: centimorgan positions supplied with the Shannon 2015 dataset

• Dating: shared haplotype length (Gandolfo et al. 2014)

• Generation time: 3.5 years

What this study does not show

• It does not name the breed merle came from. The closest matches were nearly tied.

• It dates when the Chihuahua merle copy split from other breeds' copies. The year merle entered the breed may be later than the split.

• Many of the merle dogs tested come from a small number of programs. More lines will sharpen the date.

How this fits the history

• The AKC standard has said "any color, solid, marked or splashed" since 1904. It never barred merle.

• The Chihuahua Club of America was founded in 1923.

• Our archive research found that one of the club's original members had merles.

Breeding then and now

The timing makes sense once you know how different breeding was in the early 1900s.

Then: an open door (1900s to 1930s)

• A dog could be registered on its looks. Its parents were often unregistered or unknown.

• "Chihuahua" on early papers was a label based on looks. It was not proof of breed or origin.

• New dogs could enter the breed, and new genes came with them.

Now: a closed door

• A puppy can only be registered if both parents are registered.

• Nothing new can enter without breaking the rules.

What that means for merle

• In the early 1900s, a dog of mixed or unknown background could still be registered as a Chihuahua.

• That is when a gene from outside the breed could come in. It was allowed, and it went on the papers as Chihuahua.

• Our DNA date for merle falls in that same window, around 1910.

• Today, a dog can only be registered if both parents are registered. For merle to come in recently, someone would have had to cross in another breed and hide it.

• A recent cross leaves a long matching stretch of DNA next to the merle gene. Our merle dogs do not have that.

• So merle did not come in through a recent hidden cross. It came in during the early years, when the registry was still open.

What this means for breeders

Merle is part of the Chihuahua's history. It still needs to be bred with care.

• Test allele length before pairing two merles. Length decides the risk. A cryptic merle (Mc or Mc+) bred to a merle is low risk. Two full merles (M x M) give about 25% double merle (M/M) puppies, which are at high risk for deafness and eye defects. This is how merle breeds pair merles safely.

• Test every breeding dog for merle. Embark finds merle but does not report length. For length, use a lab that sizes the allele, such as UC Davis VGL.

• Watch for hidden merle. Some merle dogs show little or no pattern (cryptic merle). We Embark test every puppy at birth, so we know if one is a hidden or minimally marked merle.

• Keep up full health testing. Heart, patella and eyes, plus DNA.

Contact

Questions, or Embark data to share: healthresearchataa@gmail.com

Methods, scripts and summary tables are available to researchers on request.

We thank the breeders and owners who shared their dogs’ Embark data.

Sources

• SIPGS merle haplotype analysis, A & A's Chihuahuas, October 2026.

• Shannon LM et al. 2015. Genetic structure in village dogs reveals a Central Asian domestication origin. PNAS 112(44):13639-13644.

• Browning BL, Tian X, Zhou Y, Browning SR. 2021. Fast two-stage phasing of large-scale sequence data. Am J Hum Genet 108(10):1880-1890.

• Gandolfo LC, Bahlo M, Speed TP. 2014. Dating rare mutations from small samples with dense marker data. Genetics 197(4):1315-1327.

• Clark LA et al. 2006. Retrotransposon insertion in SILV is responsible for merle patterning of the domestic dog. PNAS 103(5):1376-1381.

• Langevin M et al. 2018. Merle phenotypes in dogs: SILV SINE insertions from Mc to Mh. PLoS ONE 13(9):e0198536.

• Murphy SC et al. 2018. Length variations within the Merle retrotransposon of canine PMEL: correlating genotype with phenotype. Mobile DNA 9:26.

• Chihuahua Club of America. Merle Coat Pattern. Chihuahua, Vol. 4, No. 2, July 2005.

• Australian Shepherd Health & Genetics Institute. Understanding Merle and the Science Behind It.

• © 2026 A & A's Chihuahuas and Rev. T. Stringer. All rights reserved.

  This educational content may not be reproduced without permission.