Dog Health, Responsible Breeding

Does linebreeding (and inbreeding) really flush out faults and concentrate virtues?

If you don’t know the basic principles of linebreeding, or what it implies, start up here in this section. If you know all that stuff and want to get to the good part, start here.

One of the most common pieces of advice given to new and experienced dog breeders is to linebreed their dogs. The reason this advice – which is about 150 years old and is a holdover from the livestock breeding strategies of the Industrial Revolution – has survived is that linebreeding, over time, creates a situation where your breedings will be more predictable.

Imagine that you are pulling from a giant jar of M&Ms. The jar is opaque; what you have grabbed will not be seen until you’ve pulled your hand out. You HATE the green ones and you’re not too fond of light brown. You really want the red ones and you’re neutral about yellow. Your first handful brings up a whole bunch of reds, some yellows, a few light browns, and one green.

From here you have a choice – you can either go back in for your next handful, or you can put the handful you already pulled out into a second jar. That second one is a magic jar, and it’ll instantly fill up with thousands of M&Ms, but they’ll be in the exact proportion of colors that you put in.

This is not a trick question – OBVIOUSLY, to get closer to what you want, you choose the second jar. Once the second jar fills, almost every handful you pull out will be pleasing to you. Sometimes, by accident of chance, you’ll get a lot of greens and light browns, but on the whole you’ve created a great way to get the M&Ms you want.

This is exactly the way it works in dogs. By starting with something you like, and then creating many more dogs closely related to the thing you like, you are stacking the odds in your favor.

The things that are very important to you – heaviness of bone, pretty head, or fill in a hundred other traits – will be statistically more likely to appear in a linebreeding, and over time this will lead to the creation of a “Line.”  We all talk about “line” to mean a pedigree, which is fine and works in a pinch, but the truest definition of a line of dogs is a group of dogs related over multiple generations that is incredibly similar and predictable.

This point – where you have the happy M&M jar – is usually where the explanation stops. Of COURSE every breeder would want to do this, and so OF COURSE it’s the best way to breed.

However, what is not communicated – often because it’s misunderstood, because those passing along the advice don’t really understand it and are just passing along advice THEY heard, and back it goes for generations of breeders – is that the M&Ms are not coming by themselves.

Picture that each M&M is attached to a string of invisible mardi gras beads. You have fifteen small pieces of candy in your hand, and dripping between your fingers are fifteen yards of beads. That means when you put your handful of M&Ms in the new jar, you’re not just putting candy in – in fact, most of the jar is not going to be candy. It’ll be a big tangle of mardi gras beads with some candy attached to it.

Those invisible mardi gras beads are all the things that make up the invisible dog under the dog that you see and touch. They are the immune system, the emotional response system, circulatory, nervous, lymph, muscle development, bone growth rates, hormonal secretion, digestive… it goes on and on. And, of course, when you make your M&M color more alike, and more to your liking, you’re also making the bead strings more alike.

But – and here’s where the problem comes – you can’t see those beads when you’re making your choice between jar one and jar two.  So NOW which one do you choose?

But – but – you say “Linebreeding concentrates virtues and flushes out faults!” That’s what we’ve heard, often from breeders advocating not just a moderate amount of linebreeding but of tightening and tightening as you go. You don’t just stick with jar two; you put a few M&Ms, just exactly the colors you like, from jar two into jar three. And from jar three you pick not just the colors you like (because now they’ll all be the colors you like) but the roundest and plumpest candies. Jar four, therefore, is all the right color AND very likely to be just the type of candy shell and just the consistency of chocolate you like. This is all GREAT, say the advocates of line breedings. And the very tightest of linebreedings is the best of all, because you have “concentrated virtues and flushed out faults.”

“Flush out faults” is a great example of people using words that sound good but are a disaster in real life. Flushing out genetic faults means making faults appear. So you have, for example, sister mom with no immune disorder, brother dad with no immune disorder, breed them together and you get three puppies with an immune disorder. This reveals (flushes out) an inheritable genetic disease in their familial lineage.

When breeders had no ability to outsource any kind of knowledge – in other words, 150 years ago – this could be useful. You’d then know that the family carried a inheritable immune disorder and you could stop breeding them, or breed them only to families whose breeders had done similar inbreedings and found that their families did not have that disorder. This practice was honed on cows, so the default answer was “eat them all.” You eat your failures, which are going to be the vast majority of your breedings.

Here’s the big problem nowadays: Now you have three puppies with an immune disorder.

This is where the whole strategy breaks down, when it comes to dog breeding. Nobody is (or should be) keeping entire litters to adulthood, and nobody is killing puppies anymore. You can’t make the problems you create disappear into the freezer. Those puppies go into pet homes and show homes, months and years before any of the “flushing out” is going to take place. In effect, you’ve given your pet homes the result of an experiment to see what bad stuff is in a family.

And this is where we, as breeders, have to stare this in the face. Is it ethical to try to make disorders appear, when we intend to put those disorders in homes?

I’m going to leave you with that – and look for some wise remarks in the comments.

Previous Post Next Post

You Might Also Like


  • Reply K April 27, 2015 at 7:20 pm

    To answer the title question, yes, yes it does. Now the real question, is whether there is a better way than linebreeding? Right? And I would guess that you would define “better” as “producing fewer sick dogs”? Do you have another way to both set traits and reveal/remove faults that has better outcomes on the individual and population levels?

    Let’s say that instead of breeding mother-son and getting the three sick puppies you breed both to unrelated dogs and they make healthy puppies. Yay! Life is good – but only for that generation. Everybody is more or less outcrossed for, say, 5-10 generations, then you have all of these practically unrelated dogs and some percentage of them carry those nasty genes. That is when you end up with, say, 10% and seemingly unrelated dogs all carry the nasty genes. Sure, only 1% of litters are pairs that could produce it and even then it’s only one pup per litter of, say, 5 or so, that really has it bad, but you’ve gone from 3 pups produced once and culling the line to 1 in every 500 dogs of the breed dying from this condition. That’s a dog a year in the smaller breeds, or 2000 dogs a year in the larger ones. You end up with more dogs sick and dying.

    Isn’t it better to do what the heritage livestock folks do and create several lines, each closely related to each other but not the others and outcrossed with one another judiciously to keep from being painted into a corner, than to let all of the good and bad spread out unpredictably?

    • Reply Abbe PB April 27, 2015 at 11:35 pm

      “Do you have another way to both set traits and reveal/remove faults that has better outcomes on the individual and population levels?”

      150 years ago, no. That was why linebreeding became so popular – it was the best way to find these traits so you knew if you needed to cull a line from your breeding program.

      These days we have genetic testing. Is it perfect? Certainly not, but neither is linebreeding – you could get really “lucky” and have a line with awful recessive disorders and just not pass them on enough to discover they’re there. Will it be an absolute predictor? No, especially because sometimes you don’t know what will crop up that we haven’t discovered yet, and we’re certainly just discovering the amazing world of epigenetics. But it CAN inform you of the known risks much further ahead of time and let you make an informed decision about creating a litter – long before you make and distribute those possibly sick puppies, or spread a line of 500 dogs with genes for disorders.

      • Reply K April 28, 2015 at 2:59 am

        I totally agree for the few things where we have a genetic test, but most of what we are concerned about are polygenic traits like temperament. There is not and never will be a genetic test for all of the factors that contribute to, say, final height at the withers, or whether the ear is an equilateral triangle, or all of the things that go into herding instinct. That is why we use linebreeding – to fix all of that stuff. DNA tests will make it safer to linebreed at the same or better rates than linebreeding will become no longer necessary because we have all these DNA tests for the traits we like.

        • Reply Elizabeth May 1, 2015 at 12:38 am

          Linebreeding makes health worse by increasing gene homogeneity.
          9 times out of 10 there’s no reason to remove x trait as long as you don’t inbreed, because you need two carrier parents (hard to get with two not closely related adults) to produce a child with the bad trait. Even if you DO remove a trait from your line, you have by definition made other bad traits more likely to appear again due to gene homogeneity.

          • K May 1, 2015 at 2:03 am

            Sorry, posted this accidentally below and not sure how to delete it…

            Gene homogenity is why Labradors make Labradors and don’t randomly birth German Shepherds or Pomeranians or Poodles. We WANT homogenity for the traits we like – that is how we get them and how we have breeds in the first place. Labradors are all solid because they are homogenous at the K locus. Is there a genetic health problem associated with being solid-colored?

            The question is, how do we increase the homogenity of the genes we want while not also concentrating stuff we don’t want or painting ourselves into a corner? Up to now, the only way we have of even getting *breeds* as fixed or semi-fixed populations is line breeding. Yes, that means that the white-footed herding breeds concentrated susceptibility to certain modern medications along with concentrating their herding ability, but until we have another way of concentrating herding ability (which requires homogenity) I don’t see that we can toss the baby out with the bath water.

            I would rather keep having Labradors while battling their diseases then go to the other end of the spectrum and have all dogs bred for maximum heterogenity and end up with only yellowish dingo-like things unsuitable for most pet or working applications.

  • Reply K April 27, 2015 at 7:47 pm

    Also, what about the protective function of genes? What if in two lines of Poodles one rarely has the genes for epilepsy and so doesn’t get the condition, one line has all sorts of genetic epilepsy but has a gene that suppresses it?

    So if you breed line 1 dogs to line 1 dogs, you get healthy dogs, and if you breed line 2 dogs to line 2 dogs you get healthy dogs, but then you breed line 1 dogs to line 2 dogs and all of a sudden everyone has epilepsy. A line not only concentrates what’s there, but it also gives you the chance to make sure that everything that is there plays nicely together (kind of like if you breed all yellow Labs you can only get yellows and if you breed all chocolates that don’t carry yellow you will have all chocolates, but if you breed them together you get all black puppies that can produce all three colors, only in this case the “black” color is some horrible genetic condition).

    • Reply Elizabeth May 1, 2015 at 12:41 am

      The idea of ‘a gene that suppresses the epilepsy’ sounds implausible – is there a scientific basis for genes ‘playing nicely together’? I doubt it. But even if it’s true what about traits other than epilepsy? Those will deteriorate due to reduced gene homogeneity.

      • Reply K May 1, 2015 at 1:49 am

        There are many cases. For an easy example, in both mallard and muscovy-derived ducks “white” color is autosomal recessive and completely blocks the expression of all other color genetics. All of them. Period. But if you breed a white Muscovy to a white Mallard you NEVER get white birds because the genes are at different loci (and don’t forget both are recessive). Thus, you get the “hidden” colors (many of which have been hidden for decades if not centuries) and since “color” is actually pigment proteins that are useful for other stuff, you have the potential for some messed up birds in more than cosmetic ways.

        So in say in dogs, imagine that the “disease” is double-merle. Line 1 is carefully managed so you never breed merle-merle, they are red dogs b/b (red or liver dilution). Line 2 are dogs that are e/e (yellow as in Labradors) which means that they cannot express merle, and even if M/M won’t be affected by it! The reds are light and the yellows are dark and they sure look like the same color. You have no idea that some percent of the yellow line 2 dogs are actually merle or double merle. So you breed yellow dog to single-merle dog because it’s “safe” – neither line has any problems with double merle, after all! It is easily prevented by testing line 1 and non-existent in line 2. Then 25-50% of the pups end up double merle and with major problems (since it turns out that daddy’s e/e is recessive so no one gets the protection (oops!). That is one very easy genetic example that I can come up with off the top of my head, you can see how this is quite plausible for things that are polygenic in nature. Even diseases we thought we had a handle on are turning out to be families of diseases (schizophrenia is called a single disease and researchers now think it has at least half a dozen if not many more distinct causes). Why is it so silly to think that a dog can have epilepsy if it is T/? and i/i and anything but s/s? So you can have a whole line of dogs that are T/T and i/i but also s/s that are completely healthy and another line that is completely healthy because they are t/t (but also i/i and S/S)?

        • Reply Lauren May 8, 2015 at 3:32 am

          Very poor example. Mallards and Muscovies are different species – not different breeds, and can only produce mules, because all of the genes are at different locii.
          If you breed a white Pekin and a white (called Silver) Appleyard, you get white ducks – same species
          Dogs are all the same species, and gene homogeneity to the degree you suggest isn’t useful or desired. You ask “The question is, how do we increase the homogenity of the genes we want while not also concentrating stuff we don’t want or painting ourselves into a corner?”
          The answer is, we don’t need to increase homogeneity of a breed once the breed is established. At that point, the desired point of homogeneity is established – as you said, we have Labradors producing Labradors, and not poodles. Past that, there is no useful function to further inbreeding. Simply breed desired traits to desired traits.

          From a livestock standpoint – and I breed livestock – this is all you need to do, once the desired traits are established. If I have Nubian goats (and I do) and I want them to produce a gallon of milk a day, I can breed to any Nubian goat that also has that level of production, and I will STILL get a Nubian goat, with all the Nubian traits, that produces a gallon of milk a day.
          Furthermore, if I cross in a LaMancha goat of good conformation, that milks a gallon a day (and I have) I will get well conformed goats (as shoulder angle, neck length, depth of ribs, etc, share the same gene locii, as they are the same species) that milks a gallon a day, while only losing some of the breed traits, like long ears, and roman nose (short ears are dominant in goats, as drop ears are dominant in dogs. Roman nose is impartially recessive, as brachycephalic faces in dogs are impartially recessive)
          What I won’t do is lose any trait that cannot be regained in 3 generations of breeding back to a parent breed, as they are the same species, so really, not all that different. And note I said a parent breed – not a parent. Any unrelated Nubians with the desired traits will return to me goats with all Nubian traits – that are, in fact, indistinguishable from purebred Nubians without studying pedigrees – within 3 generations.

          Furthermore, as the open studbook of certain horse breeds shows us, it is patently untrue that you can only solidify traits by breeding related individuals. Actually, the Border collie breed shows us this as well, and, since you mentioned white-footed herding breeds, lets touch on that in a sec.
          Trakehner horses all look very much the same, in spite of huge amounts of outcrossing. Why is this? It is because the goal is always for horses that can perform the same function – the modern test is 3 day eventing. To hold up that needs horses with a certain angle of shoulder, a certain depth of rib, a certain length of back and neck, a specific conformation of the rump – a very complex thing in horses!
          And by breeding animals of the same species, who are otherwise unrelated, but who can all perform this same, complex function, you get very prepotent conformity.
          Similarly – no one bred for white footed herding dogs. For a long time, according to the Border Collie breed association, a Border Collie was any dog that could do the job. And yet, somehow, with no other requirement than that, they got a recognizable breed of dog! Thirty five to fifty pounds, Irish spotting, rangy build, loose shoulder, strong eye, drop tail, some variety but semi-prick ears, black and white coloring and rough coat predominating.
          Why did that work with so many unrelated dogs?
          Sheep respond the fastest to dogs with the classic Irish spotting and semi-prick ears. A dog with that coloration and ear set will have an easier time performing that complex function than an equally skilled dog without – which is why so many herding breeds display these traits. Dogs of that size can best hold up to the miles asked of them. Dogs with that shoulder conformation are better able to catch a sheep’s attention with their strong eyes, etc.
          Breeding for specific traits set the breed, inbreeding not required – not saying it never happened in the breed, just that the history clearly shows it was not needed.

          Whatever needs created any breed, they are now created. All dogs of a certain breed are already closely related. Further inbreeding can only be detrimental – observation of any population of any species shows that. All that is needed is to continue breeding for the desired traits, which is just as easily done without linebreeding. Once a breed is established, any linebreeding is “breeding yourself into a corner”.

    • Reply K May 1, 2015 at 1:54 am

      Gene homogenity is why Labradors make Labradors and don’t randomly birth German Shepherds or Pomeranians or Poodles. We WANT homogenity for the traits we like – that is how we get them and how we have breeds in the first place. The question is, how do we increase the homogenity of the genes we want while not also concentrating stuff we don’t want or painting ourselves into a corner? Up to now, the only way we have of even getting *breeds* as fixed or semi-fixed populations is line breeding. Yes, that means that the white-footed herding breeds concentrated susceptibility to certain modern medications along with concentrating their herding ability, but until we have another way of concentrating herding ability (which requires homogenity) I don’t see that we can toss the baby out with the bath water.

  • Reply Kat May 6, 2015 at 2:52 am

    “Do we have another/better way?” – Well, I think yes.
    The things we want in the dogs – Labradors being solid, Poodles being fluffy, having a nose exactly this long and bone exactly this thick – they are things we can see, things we can measure. Those nasty recessive genetic disorders, unless they are the few that have been investigated in the breed, them we cannot see. And , statistically speaking, each dog will carry about 10 different lethal ones, and who knows how many more that are detrimental in the homozygous individual. So, instead of breeding brother to sister and playing Russian roulette as to whether or not you will have pups that suffer, what about outcrossing instead? You take that solid Labrador, and you breed it to another solid Labrador unrelated in the last 3-5 generations, then you take a solid pup and when she grows up breed her to yet another fairly unrelated solid dog. You can fix that solid variant in your breeding population with a greatly reduced change of fixing something like epilepsy by accident as well. You can do the same for all other visible and measurable features you have. Now that doesn’t solve the issue of problems being associated with or directly related to the physical feature – like merle dogs and pugs who cannot breathe, and some breeds do not have the genetic diversity to do this well. However this would be another way – perhaps a longer way – but why not?

  • Reply JMo May 7, 2015 at 2:31 pm

    Everyone has made good points so far. However, rather than talking about specific genetic traits we do or do not want to pass on, there is a bigger picture to consider. It is true that even in the absence of deleterious genes, linebred dogs eventually succumb to such things as smaller litters and puppies who fail to thrive or simply aren’t as vigorous as their outcrossed friends. I’m not a breeder but I’ve seen this personally and up close with a breeder friend’s dogs. It’s also pretty common that conformation-bred dogs tend to be more tractable and I believe this is not solely due to their being bred for friendly temperaments. These dogs also tend to be line-bred moreso than, say, dogs from “backyard breeders” and so this could be due to simple lack of vigor compared to puppies of complete outcrosses. Not so much that anyone would say, “That dog is sickly”, but by comparison. Lack of vigor and smaller litters are not apparent in any one dog but are only seen over time with repeated linebreeding or inbreeding. This is something for breeders to keep in mind; it’s a tough balancing act retaining type and the breed characteristics while also keeping the dogs as healthy as possible in a general manner as well as with regard to specific inherited problems we do not want.

  • Leave a Reply

    CommentLuv badge