My Dog Is Not A Dalmatian: Why Breed (Mis)Identification Matters

It happens all the time. Someone will say, “What is he? Dalmatian, obviously, but what else?”

I don’t blame them. That’s what I was told, when I stumbled across a squirmy white puppy covered in tiny black spots at an adoption event and asked what breed he was.

“Dalmatian/Cattle Dog,” the rescue worker told me.

OOH!  I thought. How neat!!! I had recently decided to get a dog and had been dutifully researching dog breeds that might fit my lifestyle. Australian Cattle Dogs were on my “Cute” list, my “Frequently Found in LA-Area Shelters” list, my “Probably Fits My Lifestyle” list, and my “Not Banned in Apartment Complexes (As Far As I Know)” list. Dalmatians, however, were on NONE of my lists; they weren’t even on my radar.

Unfortunately, they were on the radar of my then-boyfriend, with whom I was adopting a dog–in the worst possible way. His parents had owned a Dalmatian, before their son came along; apparently, this dog had been “crazy” and was ultimately put down because of his behavioral issues.

So for a couple minutes, we honestly considered walking away from this puppy. Could we really handle a Dalmatian?? But we watched him there, contentedly wriggling belly-up in the grass, all puppy-smiles at every new person who came up to him. He didn’t seem crazy.

We took him home, but we nearly didn’t. And this is important, because as it turns out…he doesn’t have a drop of Dalmatian in him.

Nope. This dog. Is. Not. A Dalmatian. And actually, not even a cattle dog.

So what is he?

The honest answer is, no one really knows. Mars Veterinary, the makers of the DNA-test Wisdom Panel, thinks he’s got a little bit of Catahoula Leopard Dog somewhere far back in his ancestry–but there are no purebreds of any stripe (or spot) at least three generations back.

And it’s this honest answer–no one really knows–that I think we should all be more comfortable saying when it comes to our rescue dogs.

It’s standard practice for rescues and shelters to guess and list breed cocktails for their dogs in an effort to make them more adoptable. A “Papillon/chihuahua mix” sounds more exotic than “tiny dog with shaggy hair,” and a “Staffordshire/lab mix” sounds less dangerous than “medium-sized dog with short hair and a boxy head and yeah it totally looks like a pitbull.”

And of course, it’s the very first question most potential adopters ask. The organization I volunteer with, Forever Fido Rescue, just saved a litter of seven-week-old puppies, and everyone wants to know what breed they are (honestly, at this point, they look mostly hamster). Sure, they want to get an idea of what they’ll look like as adults, and how big they’ll end up being, but the rescue bought them off some dude selling them in a Walmart parking lot in Bakersfield. They have no idea what these little babies are. Many rescues’ knee-jerk reaction would be to make what they feel is an educated guess and slap a label on them.

But I’m gonna go out on a limb and say we should really stop doing this. Sure, it’s fun to speculate, and it’s even more fun to connect with other owners of the same breed and swap war stories. I, for one, am particularly guilty of overusing “#catahoula” on Instagram. But I don’t think we should be emphasizing a dog’s assumed breed in order to get them adopted.

First, you’re probably wrong. Don’t feel bad–it has been scientifically demonstrated to be nearly impossible to accurately identify a dog’s breed heritage based solely on what they look like or how they act. EVEN DOG PROFESSIONALS CANNOT DO THIS WITH ANY REASONABLE ACCURACY. Seriously! I’ll prove it to you–just keep reading. Unless you DNA-test a dog, you’re (probably) lying if you say you know what breed(s) it is.

But second, and more importantly, I believe that assigning breeds to rescue dogs reinforces the cultural value that our society puts on “breed” in the first place. It’s a tacit affirmation that breed matters. If we want to work towards fewer people buying from breeders, we must devalue the caché of having a definable breed. Finding out your dog’s breed composition should be a fun parlor game you play after-the-fact; you should decide to get any dog based solely on its individual merits.

(If you have no patience for depth today, jump to the summary at the bottom of this post. But you’re gonna miss some FASCINATING stuff.)


You’re Probably Wrong (It’s Science!)

In 2012, the Maddie’s Shelter Medicine Program surveyed over 5,000 dog experts from various professions (“breeders, trainers, groomers, veterinarians, shelter staff, rescuers and others”) and asked them to attempt to identify the breeds of 100 dogs in a series of photographs. The results may surprise you: “The 5000+ responders were only correct – that is, named at least one of the breeds detected by DNA analysis – less than one-third of the time. And no profession did much better than any other. Every profession’s responses, in total, were correct less than a third of the time.”

Another study, this one focusing on identifying pit bulls, asked staff members at four different shelters (and a veterinarian, apparently) to speculate on the breed make-up of 120 dogs. Here’s a graphic from this report summary:

You can see how easy it is to be wrong. In fact, of the 120 dogs in the study, shelter staff identified 55 to be “pit bull type” breeds. Of those dogs, only 36% were actually pit bulls! There were 25 “true” pit bulls in the study (as determined by DNA analysis), and shelter staff completely missed identifying 5 and could only reach a consensus on 8!

It’s super-tough to guess a dog’s breed composition just by looking at it. There are so manyvisual aidsthat helpprove this point.

This isn’t new science. In 1965, John Paul Scott and John L. Fuller showed that sometimes dogs bear almost no resemblance to their mother and father, let alone more distant ancestors:

So maybe you’re thinking, “OK Carrie, if a dog’s looks aren’t a dead giveaway, how a dog acts sure is! You can tell a hound but his prey drive perhaps, or a cattle dog by his tendency to nip heels!”

Not so fast.

A 2006 Swedish study by Kenth Svartberg (and summarized here by the National Canine Research Council) “found that modern purebred dogs grouped according to [four groups based on historic function–terriers, herding dogs, gun dogs, and working dogs–] simply had nothing more in common in terms of behavior than dogs in general… [He] attributes this lack of conservation of historical traits to the practice over the last century and a half since the advent of organized dog shows of breeding dogs primarily for appearance.”

Put simply–while many dog breeds were at one time developed to emphasize specific behaviors, those dogs are rarely still used for those behaviors, and so the purebred representatives of the breed are selected more for appearance than the original behaviors.

““If you’ve got a dog who looks like a duck, and you encourage it when it quacks like a duck because you believe quacking is your dog’s genetic destiny, you may well end up with a quacking dog that looks like a duck.

But that doesn’t mean you have a duck.”

Furthermore, “if traditional traits have been so diluted as to be indiscernible in purebred dogs, we should certainly not expect to be able to predict them at all in dogs of mixed breeds.” That means that while the original Australian Cattle Dogs almost certainly were predisposed to nipping at heels, your watered-down animal-shelter cattle dog may not be!

Conversely, in “The Relevance of Breed in Selecting a Companion Dog,” author Janis Bradley points out that behavior for which Greyhounds are bred (“chasing a small prey-like object”) is actually expressed in only 70-80% of Greyhounds–“a rate that may not be much higher” in all dogs, regardless of breed. Which is probably why so many dogs love playing fetch–not just retrievers!

In yet ANOTHER study, this one just published in Scientific Reports, a discernible difference in dog behavior as a result of breed WAS found–but ONLY in the “working lines” of each breed! I can’t say it any better than Janet Fang over on IFLScience, so here’s how she summarized the study:


“The team found that border collies, on average, were more impulsive than Labradors – consistent with the original purpose of breed selection. However, the team only found a difference in levels of impulsivity when they compared the so-called working lines of the two breeds: There were no major difference between show lines of those breeds. When appearance, rather than behavior, becomes the primary focus for breeders, this reduces the differences in impulsivity between breeds. 

In fact, the differences within breeds can exceed the differences between breeds. These findings caution against breed stereotyping: Predictions about an individual dog’s tendencies can’t be made based only its breed.”


So if breed isn’t such a reliable predictor of behavior, why do some dogs seem to be such good herders, or retrievers? Bradley suggests that “those with a genetic predisposition to express certain behaviors will excel when they are put in a training environment that rewards and perfects those behaviors.”

I can easily see how, if I adopted what I thought was a Golden Retriever-mix, and he happened to one day grab a ball and bring it back to me, I would be influenced–by my own perception of his breed–to think he’s doing what he was bred to do, and encourage him to do it over and over again.

But I can also see how, if I thought I adopted a Rottweiler-mix, I may be horrified when he starts lunging and snarling at delivery men, but I might also be subtly prejudiced to believe this is a behavior innate to his breed–which could lead to only weak corrections on my part, as I would subconsciously believe it’s not going to work anyway.

And unfortunately, I can see, too, how I may adopt a teeny pouf of a Maltese, and dismiss its first Rottweiler-esque lunge/snarl at a delivery man as a total fluke, because that’s not what you expect from a Maltese. And I may not even hear the second time because I’m in the kitchen and let’s be honest that snarl sounds like a sneeze. And maybe the third time it’s kinda funny because it’s 3 pounds of cottonball hurling itself at the screen door. But before you know it…I’ve got a Maltese with uncorrected Rottweiler-like rage issues and tiny razor-like teeth.

So. If you’ve got a dog who looks like a duck, and you encourage it when it quacks like a duck because you believe quacking is your dog’s genetic destiny, you may well end up with a quacking dog that looks like a duck.

But that doesn’t mean you have a duck.


DNA-Testing And The Quest For Answers

The only reliable way of discerning your dog’s ancestry is through DNA testing, such as the Wisdom Panel test from Mars Veterinary.

Several articles have been written about the various tests on the market and their accuracy–two good recent ones are Colleen Kane’s “What you should know before you test your dog’s DNA” for Fortune and Gemma Tarlach’s “Can Doggie DNA Tests Decode Your Mutt’s Makeup?” for The Crux, Discover Magazine blog.

And I’ve weighed in with my own experience DNA-testing Chance on this very blog. Long story short–I found out essentially that Chance’s breed history is too muddied for me to ever be able to identify individual signatures, but he probably had a Catahoula Leopard Dog ancestor somewhere in the way-back. If your dog didn’t also come from Mexico, you’ll have a much better chance than we did of actually getting your dog’s genetic history decoded (read my post for the full story.)

This information is more helpful to me than assuming the rescue’s visual identification was correct (Dalmatian/Cattle Dog) because I have less reason to believe that Chance could be afflicted by the mutation for hyperuricosuria that affects all Dalmatians (that can cause urinary stones) or the mutation of the MDR1 gene that affects many shepherd-type dogs like Cattle Dogs (that can cause dangerous drug sensitivities).

This confirms what the National Canine Research Council has stated: “The DNA test is better than visual breed identification because it takes into account the pattern of genetic variation at many different regions across the dog genome to generate a ‘genetic snapshot’ of a mixed-breed dog’s ancestry. The resulting genetic evidence for what breeds make up a mixed-breed dog may or may not agree with visual observations, but they do agree with what scientists have discovered from two decades of sequencing and studying genomes.”

You can only tell if you’ve actually got a duck if a DNA test says you have a duck.


STOP TALKING ABOUT DUCKS, Why Is Speculating About Breed Problematic?

Good question, reader.

Inventing and assigning breed cocktails to rescue-dogs only reinforces the idea that a dog’s breed is a valuable piece of information. The cultural value we put on a dog’s breed is what makes people buy from a breeder, or seek out a “designer dog.” Devaluing the idea of “breeds” starts with the understanding that, in the modern era, a dog’s breed has much to do with their appearance and very little to do with their personality or behaviors.

Deciding to adopt a dog based on its perceived or assigned-by-the-rescue/shelter breed is as misguided as banning pitbulls from apartment complexes, as Janis Bradley explains: “Most behaviors associated with specific breeds are only tangentially related to desirable and undesirable qualities in pet dogs… Therefore, the practice of relying on breed identification as a primary guide in either pet-dog selection or dangerous-dog designation should be abandoned.”

You can’t tell what a dog is by looking at it, and it’s irresponsible to suggest you can without some disclaimer. It’s also irresponsible to suggest a given mixed-breed dog will most likely exhibit behaviors “typical” of the breeds you think comprise it. Instead, emphasis should be placed on an individual dog’s temperament.

Rescues should be re-educating the public about the relative insignificance of breed, rather than feeding into the idea that it’s one of the first things you need to know about any given dog. As Janis Bradley says, “the focus of predicting behavior should shift to the particular dog’s personality as developing from the interaction of genes and environment and to dogs as multifaceted individuals, bearing in mind that the guardian’s choices about how to live with a canine companion are likely to shape the dog’s behavior.”

What that means is that we should treat dogs more like we treat humans–as individuals. We frown on saying, “Oh, that person’s father is from Sweden, so they’ll probably act like _______.” So we should similarly frown on assumptions about dogs.

When we don’t, when we allow breed stereotypes to drive our assumptions about individual dogs, we pave the way for breed-specific legislation and other breed-based biases (like the hell of trying to rent an apartment in Los Angeles with a dog over 40 pounds). Even veterinarians are starting to realize that speculated-breed reporting on medical records has corrupted population data: “An article by two veterinarians and an attorney published in November 2012 in the Journal of the American Veterinary Medical Association (JAVMA) has considered the implications of these undisputed findings [of breed] for veterinary practice, and recommends that veterinarians stop attempting to assign breed labels to mixed-breed dogs whose parentage they do not know.”

So what can a shelter or rescue do? If they’re willing to invest the cash, Wisdom Panel has a program through which you can register your non-profit rescue/shelter and purchase bulk discounted DNA-tests that are fast-tracked; instead of waiting the typical month for results, Wisdom Panel will get back to you in 3-5 business days. This is also the only real way you’ll get an idea of what a young dog might end up weighing when he’s a mature adult–that’s pretty much the only thing Chance’s DNA test was right about from the start.

Alternatively, rescues can hedge their bets. Admitting the ambiguity in breed speculation is a good first step. You can still get your dog to show up in searches for “border collie” without calling your dog a Border Collie.

I like what this rescue has done:

I don’t think Roxx the Dog is any less appealing because the rescue is honest about not having any clue what breed he is. He looks like an athletic medium-sized dog with a big old smile. If I was searching for a Border Collie-type dog and came across him, I’d be interested. Unless I have my heart set on a long-haired dog that sheds A LOT.

I also like what this one did:

Listing possible breed combinations like this will make these dogs show up in searches for those particular breeds, but it acknowledges the fluidity of breed designation. But do you see how someone could just as easily say “German Shepherd (pointy ears)/Cocker Spaniel (size & coloring) mix”?

A description like this, however, leaves little room for interpretation:

Is this rescue trying to avoid the obvious “pitbull mix” designation this dog might otherwise get? Probably. But unless this dog has been DNA-tested, declaring him to be a “Border Collie mix” is misleading.

Remember those puppies that Forever Fido Rescue had for adoption? Here’s how they ended up advertising them:

I like that they were honest that it’s pretty unclear what this little guy is, and instead of mentioning specific breeds, they used groups that suggests more genres of dogs. It gives potential adopters a vague idea of adult weight and general look, without promising anything more.

Most importantly, a new study published in the academic journal PLOS ONE found that “dogs labelled as ‘pit bulls’ may wait three times as long to be adopted from shelters than differently-labelled lookalikes.” The authors’ solution? “Assigned breed labels can be inaccurate, based on sometimes misleading appearances, and this research may indicate that dogs could be inadvertently penalised when labelled as a pit bull breed. The authors state that removing breed labels [my emphasis] seems to be an easy way to improve the experience of pit bull type dogs in animal shelters.”



If you haven’t read my post detailing my experience with Wisdom Panel yet, spoiler alert: Chance is probably a teeeeensy bit Catahoula Leopard Dog somewhere in his distant past. His lack of purebred-ness has not stopped me from having a considerable amount of fun hashtagging all his Instagram photos with variations on #Catahoula. I’m not denying that there’s a certain “team loyalty” when it comes to your dog’s breed, and I’m not trying to take the fun out of it. I’m just for less-definitive, more casual approach to guessing the breeds of dogs in our care. Because, after all, #rescuedismyfavoritebreed.


TL;DR Summary

  1. When you look at a dog and guess its breed/mix, you’re probably wrong. Science says so.
  2. DNA-testing is the only way to get an accurate read, but that only works if the dog in question has purebred ancestors in the past three generations. It very accurately predicted my puppy’s adult weight, even though it could not identify specific breeds.
  3. Unless your dog is a purebred–AND from the “working line” of the breed, not the “show line”–knowing the breed of your dog is pretty useless information, anyway. Modern breeds generally aren’t bred for the behaviors their ancestors were bred for, so your German Shepherd will look but probably not act like the original German Shepherds–but he may be prone to hip dysplasia.
  4. But when we validate the idea that knowing your dog’s breed is one of the most important pieces of information, we reinforce the cultural value of “breed,” leading to stuff like breed-specific legislation and the caché of buying a purebred dog.
  5. Rescues should either DNA-test their dogs, or be upfront about the fact that they don’t actually know what breed each dog is. They should champion that idea that breed doesn’t matter as much as evaluating each dog on their own merits.



Leave a Comment