Just for fun, I tried a breed identification app that was mentioned in another thread. I know these apps are not as reliable as a DNA test. Posting here because I don't want to distract from the situation with a found dog in the other thread. I tried this photo of Luna ( she passed in 2018 ) and a couple others. Every time, the app described her as Australian Shepherd, Australian Cattle Dog, and about 25% Jack Russell Terrier. According to the app, Saint is pure Chihuahua - 98% or more. I wonder because her coat has an almost Shepherd-like feel. This is not one of the photos that I used, but maybe it will show what I mean. My guess is a grandparent was not a Chihuahua. Rose's results were: half Maltese, about 25% Bolognese, 25% Havanese. The half Maltese part, I already guessed. I sometimes think half Shih Tzu.
It's not that they aren't AS reliable as dna tests, it's that they aren't reliable at all. I posted the results of Tornado-dog on the other thread. It was wrong with all three photos. Here are the various results for Moose-dog: rottweiller and beauceron Primarily beauceron - other matches entlebucher mountain dog and appenzeller sennenhund Purebred beauceron And Bat-dog: Purebred beauceron Perro de presa mallorquin, shepherd, swedish elkhound, and jack russell Catahoula leopard dog - another match australian kelpie Moose-dog (larger) and Bat-dog: And Dog: Canaan dog and jindo Dog: The app only guessed somewhat correctly once - for Bat-dog as shepherd. And in addition, the results were for breeds that are impossible for them to be here in the States. Heck, some of those breeds are imposssible anywhere. Moose-dog was thought to be a rottie-shepherd as a puppy by the rescue. As he grew up tall and skinny, we figured the rottie was really dobie. DNA came back as primarily shepherd with poodle and newfie. That made sense. The poodle gave him the lankiness and the newfie gave him the white blaze down his chest and much of his personality. Bat-dog was a comglomerate. When I tried guessing her mix, I knew there was shepherd. Then I thought maybe beagle because of her body shape and talking. Then I thought maybe boxer because of her barrel chest. Then I thought of corgi because of her rule making and trying to herd the cats. Then I thought maybe shar pei because of the velvety feel of her fur around her mouth. Then I thought one of the Asian breeds like chow because of the black on her tongue. Her dna came back as mostly shepherd with beagle, boxer, shar pei, chow, corgi, and german shorthair. I was proud of myself on that one. The only breed I didn't see was the pointer. Dog was pre-dna testing, but she was basically a shepherd and lab mix. The problem with these apps is that they are looking for similarities to any known breed but are not taking into consideration how other breed traits may combine or show. For example, and black & tan shorthaired large dog like Moose-dog used to automatically be considered a shepherd or rottie or dobie mix. They were common enough breeds to have been breeding indiscriminately to produce a mixed breed puppy. But the AI in the apps, don't consider the actual odds of any given purebred having mated with another random dog. As of now, per the beauceron breed rescue, there are THREE beaucerons available for adoption in the entire U.S. One in California, one in Texas, and one in Florida. The odds of an entire litter finding its way to a random rescue is miniscule. But Moose-dog looks a lot like a beauceron so that must be what he is. Or, far more realistically, he could be a german shepherd mix. Then, we get into the odds of TWO of my rescue mixes being beaucerons (they were aged 4 years apart). That makes either identification even that much more impossible. Again, far more likely that they are both german shepherd mixes than either or both are beaucerons.
Good point! Beautiful dogs, all of them. Jindo and Canaan dogs are two of the breeds that I rarely hear about.
With your dogs, I would guess Luna to have shepherd, lab and jack russell from the photo. The freckles are JRT, the size and general build is more of a shepherd - lab mix. Lab mixes often get that chunkiness up front. For Saint, my guess would be chihuahua and corgi at minimum. For Rose, I'd guess maltese and poodle and possibly shih tzu. But those guesses are based solely on the photos. Meeting them and seeing how they walk, hold themselves, etc, as well as their temperaments could very well change my guesses. And that's where these apps fail completely. They don't consider the temperament, personality, or behaviors of the dog. To look at Tornado-dog again. Border collies are very obsessive - so are russell terriers (jack and parsons). But border collies tend to have a seriousness to their obsessions. They see it as their job and they are very good at their job. They are the very best at their job. No one can do better - if someone else does better, the border collie sees it as their own personal failure. They are a Type A personality. Russell terriers, on the other hand, are all about the fun. If it's not fun, they won't bother. If it is fun, you can't stop them. They are extremely talented and smart and clever - just like the border collie - but they don't care about doing the job, they just care about doing fun things. The easiest way to see this difference is to watch an agility competition. Both breeds excel at agility. But watch how they approach the course. The border collie is alert, focused on the handler, and ready to perform to their very best ability. The russell terrier is bouncing around like a kid on Christmas morning. They can't wait to be let loose and PLAY on that course. And they are very likely to detour from the course just to spice things up a bit. A border collie who detours from the course is a poorly trained border collie. A russell terrier detouring is just him having a blast running around. An app can't distinguish that difference. It can't look at the photo and say "well he is all about fun, he must be russell terrier not border collie". The app can ONLY base its answer on the physical appearance viewable in that selected photo. It can't think outside of the parameters it has been given. Now, I will give the app some credit. I did a photo of Looney1 and it did come back as CAT. And I did a photo of Cat-dog covered in brown powdery dirt and it did come back as german shepherd. So it isn't always wrong. Of course, it suggested I use their Cat ID app to learn the breed of Looney1. And that is ridiculous. Even DNA can't do that because the vast majority of cats are considered " polycats". Basically a naturally bred mutt in the world of cats. Even a cat with siamese markings may only have very minimal siamese dna. And I did dna test Looneys 1 and 2. They are polycats.
OK. I did two more dogs. First, Tornado-dog's mom. She came up as australian cattle dog, rat terrier and toy fox terrier. Funny how her verified offspring is all border collie Or border collie, canaan dog, english springer spaniel and english pointer OR fox terrier smooth, borzoi, and saluki. Only one of those is even remotely possible based on her results. The forum won't let me upload her photo. The other was my childhood dog. It got her half right stating she was a cross between an american bulldog and a boxer. In fact, she was a purebred boxer.
And I doubt many Americans have ever heard of Perro de presa mallorquin, entlebucher mountain dog or appenzeller sennenhund. In fact, it seems that the sennenhund is at risk (endangered) with a narrow but stable gene pool, so the chance that Moose-dog was such a mix is truly laughable.
How many dogs does it take to change a lightbulb? Border Collie: Just one. Then I’ll replace any wiring that’s not up to code. Labrador: Oh, me, me! Please let me change the light bulb! Can I? Huh? Huh? Jack Russell Terrier: I’ll just pop it in while I’m bouncing off the walls. Cocker Spaniel: Why? I can still pee on the carpet in the dark. Boxer: Who needs light? I can still play with my squeaky toys in the dark. Australian Shepherd: First, I’ll put all the light bulbs in a little circle… Golden Retriever: The sun is shining, the day is young, we’ve got our whole lives ahead of us, and you’re inside worrying about a stupid burned-out bulb?
German shepherd - of course I'll change the light bulb because it makes you happy and my goal in life is to make you happy.
I'm surprised that the Beauceron's toes weren't mentioned as being a defining feature. There are very few of them in the UK, so I wouldn't expect them to be a common cross-breeding. The likelihood of finding a breed in a particular locality has to be recognised.
Regional consideration is a huge factor. If you spend $1000 on a puppy and another $1000+ to have the puppy shipped from one of the few breeders, you aren't going to put him in your yard and ignore him. And the breeder is unlikely to sell him to you unless you are planning on showing him or competing with him. So the odds of that puppy making an unplanned mating are extremely slim. On the flip side, there are so many shepherds out there, so many backyard breeders selling them to anyone, so many being left neglected and intact in yards 24/7, and so many running the streets and ending up in the shelters, that it's extremely likely there will be shepherd mixes. Twenty-five years ago, I was working with a rescue group and we got a pair of husky shepherd mixes. It was very unusual. Now, husky mixes are everywhere. Because the backyard breeders have overpopulated the breed.
Coincidentally, this was posted online yesterday. It is from the KC information series but I've had to crop, it wouldn't let me copy the full page. What would your breed identification app make of these two?
I do wish I had had my clever but prey-driven Tweed's dna tested, but at that time the programme only had a 200 breed database, and people were getting some very odd results.
I have just put a picture of Tikva when she was 3-years-old into an AI dog breed identifying site, and it came back with Rhodesian Ridgeback!
I tested your first pair @CaroleC . The top came back as German spitz. With other matches American Eskimo dog, Volpino Italiano and the Japanese spitz. The lower came back as Samoyed. Then I did Tweed and he came back as Berger Picard. Another match is Irish Terrier. I am hard pressed to guess his mix not being as familiar with your common breeds there, but I would have guessed a terrier, possibly even an airdale, and likely a pointer of some sort - he just has that look in his eye to me. He was definitely a one-of-a-kind dog! And I still adore that name for him - it is perfect.
@Toedtoes - this is the picture I used There are various other sites but they can only be used by downloading an app onto a smart phone, not a PC - and I do not possess a smart phone. You can also use Google Lens but all the pictures of Tikva, including the one shown here, that I put in Google Lens, just came out as Dachshund, although of course she is only half Dachshund. Incidentally, she is still wearing the same style and size harness as she was in that picture - size XS. She is due for a visit from Ram next week for her Parkworm shot, Bravecto pill etc, so will be weighed then, but I do not think she has lost any more weight.
I just did it with the phone app and it came back doxie. At least it got her partly right. Glad to hear she's not losing more weight!!!! My boxer used to cringe every time someone called her a pitbull. She knew it was an insult (no offense to pitties, but she was always called a pitbull when people were afraid of her and wanted me to take her away). She had a major complex about it. My Mom and I joked that we were going to make her a sweater that said "I am a boxer not a pitbull" on the sides. Tornado-dog gets it too. Everyone thinks he's a border collie. "oh we have a border collie just like him", " my brother has a border collie, great dog", etc. If he weren't a no care in the world parson russell terrier, I'd expect him to get offended by it. But true to his breed, as long as you play with him you can call him anything...
Apart from Juli's result, I am quite impressed. The KC example was a Jap. Spitz and a Samoyed. Of course there is a size difference, but the Sammy is differently consructed too. One person commented with a photo of her White German Spitz, which would be closer match, but far more rare. Tweed will always be a guess. She was 21 - 22 inches, and I think she probably had a Whippet/Collie parent. We would call that a Lurcher. The other parent is likely to have been a Terrier - and Irish Terrier Rescue were convinced that that terrier was one of theirs! I was asked if I wanted to sell her a couple of times, by members of the travelling community. One of them told me that she was a Norfolk Lurcher - an old type of Drover's dog from the days when cattle were driven into the London stockyards. I don't think we have to go more than 150 years back for most of our breeds to be locally constructed varieties. Exceptions being the hunting, racing and toy dogs of the aristocracy.
Well, it IS a lot easier to ID a purebred. To me, the samoyed was easy - there's just something in their face structure that distinguishes them from the others regardless of the rest of the body. The other was a bit more difficult if just because I don't keep up on the spitz breeds. If the dog was seen in this area, I would have guessed eskimo simply because it's the only one I know to be somewhat common around here. With the mixes, it's more difficult because you have no idea how many breeds are in there. Tweed could have been a bit of all of those. I've read that the lurchers were bred for various reasons. They used sighthounds, herding and terriers in the mixes for their intelligence and speed, etc. But there was no standard and there could be two identical groups of dogs in different areas that had no relation to one another. Basically they were just common mixes though purposely bred. In the US, I think the closest we get to that are the regional hunting dogs in the south. Mixes were and are purposely bred to use as hunting dogs - the mixes chosen depending on the preferred hunting target. Elsewhere in the states, most mixes were randomly developed but just as common. For decades, shepherd labs were everywhere. They're great mixes - most had the easy-going attitude of the lab rather than the shepherd's attachment to one owner. They were a great family dog.
The programme seems pretty good at identifying pure bred dogs, but I wouldn't trust it for a mixed breed. I nearly opted for a Berger Picard when I lost Tweed. Certainly it was the nearest match in a pedigree form. A second Beagle was a better match for our family structure though, and there is a greater classification for them at breed shows.
