A lot of misunderstandings about dog breeding start with language. Basic genetic terms are often used loosely, and different ideas are treated as if they were the same thing, which makes it harder to judge risk and harder to ask good questions about the dogs people live with and breed from.
This post unpacks a few of the terms that get most commonly tangled up, because clearer language makes better decisions more likely.
1. Is selective breeding the same thing as inbreeding?
Selective breeding and inbreeding refer to different aspects of how dogs are bred, even though people often use the words as if they describe the same process:
- Selective breeding is about which dogs are chosen to reproduce, because someone is deciding what kinds of traits they want to see more of in the next generation.
- Inbreeding is about how closely related the parents are, because related parents share ancestors, and shared ancestry increases the chance that a puppy receives the same version of a gene from both sides of the pedigree. That matters because some harmful variants have little or no effect when a dog carries one copy, but can cause problems when a dog inherits two copies.
Once you separate those two mechanisms, it becomes easier to see why selecting for traits does not automatically require close breeding. A breeder can make strong choices about which dogs should reproduce while still avoiding close pairings, although doing that usually means putting effort into finding suitable dogs outside the most popular lines, and sometimes accepting slower progress on a desired trait. Inbreeding can also increase even when there is little deliberate selection, because repeated use of the same families and popular sires can increase relatedness across the population over time, particularly in closed breeds where the set of potential breeding dogs is already limited.
However, human-free breeding does not guarantee low relatedness. In free-ranging dog populations, related dogs can mate when breeding happens within stable local groups and the set of available mates is limited by geography and social structure rather than by a registry. Many village or free-breeding populations show moderate to high genetic diversity overall, but local or semi-feral groups can still become substantially inbred when numbers are low or movement is restricted. Genetic studies of some such populations have found elevated inbreeding under these conditions, reflecting constraint and repeated local breeding rather than deliberate selection. Avoidance of close breeding usually depends on familiarity, so when relatives are not familiar to each other, mating can occur through simple opportunity.
So the question is not whether a dog was “selectively bred”, because all intentional breeding involves selection in some form. The question that matters for genetic risk is how related the parents are, and how quickly ancestry is becoming concentrated in the breeding population as a whole. That distinction becomes even more important when people start talking about “closed gene pools”, because closure changes what is possible, but it does not dictate that every mating is close.
2. What do people mean when they say a breed has a “closed gene pool”, and what risk does that create?
When people describe a breed as having a “closed gene pool”, they are usually referring to registry boundaries, not to how individual matings are carried out. A closed breed is one where new dogs are not routinely added from outside the recognised population. That definition says nothing by itself about whether dogs are bred to close relatives, how often particular lines are used, or how genetic diversity is managed over time.
“Closed” also does not mean static. Even in closed populations, genetic composition continues to change from one generation to the next as some dogs reproduce more than others, some lines expand, and others disappear. What closure does is limit where new variation can come from: once a population is closed, all future diversity has to be maintained from within the dogs already there.
The risk this creates is not automatic inbreeding, but management pressure. In a closed population, the way dogs are chosen for breeding matters more, because there is no external source of new genetic material to offset repeated use of the same families. This is where population size starts to matter in a specific way. A breed can have thousands of registered dogs and still behave genetically like a much smaller population if only a fraction of those dogs are actually used for breeding, or if a small number of individuals contribute a disproportionate number of offspring.
This is why population genetics distinguishes between the number of dogs that exist and the number of dogs that actually contribute genes to the next generation. These two numbers are often very different. A breed might have thousands of registered dogs, but if only a small fraction are bred from, or if the same individuals are used repeatedly, the genetic behaviour of the population is much closer to that of a much smaller group.
That smaller, gene-contributing number is called the effective population size (often abbreviated as Ne). It is not a headcount. It describes how many dogs are meaningfully shaping the genetic makeup of future generations, and it determines how quickly relatedness accumulates over time. Practices such as heavy use of popular sires, repeated breeding within a narrow subset of lines, or strong preferences for certain types reduce the effective population size even when the breed looks large on paper. From a genetic perspective, it is effective population size, not registry size, that drives long-term risk.
3. Does shared ancestry automatically mean harmful close breeding?
Shared ancestry is not inevitable in all breeding populations. In very large or loosely connected populations, especially where dogs are drawn from different regions or continents with little recent exchange, the probability of shared ancestors can remain low for many generations. In those cases, dilution works in the way people intuitively expect: ancestors fall far enough back that their genetic contribution becomes negligible.
By contrast, in closed or tightly interconnected populations, shared ancestry accumulates quickly. When the same lines circulate within a limited pool, ancestors reappear more often and more recently, and their genetic contribution does not have time to dilute. In that context, relatedness is not a coincidence. It is a structural consequence of how the population is organised and how breeding choices are made.
All dogs in a breed share ancestors – that fact alone does not tell you whether breeding has been close or risky. Any population that has existed for more than a few generations will show some shared ancestry, simply because individuals reproduce, lines branch, and pedigrees overlap.
What matters is how inheritance actually works across generations. Each dog inherits half of its genetic material from each parent, roughly a quarter from each grandparent, and progressively smaller fractions from more distant ancestors. As generations pass, the genetic contribution of any single ancestor becomes diluted. An ancestor several generations back may still appear in a pedigree, but contribute very little genetically to the dog in front of you.
Risk increases when the same ancestors appear recently and repeatedly, because their genetic material has not had time to dilute. If the same dogs or the same close relatives are used again and again within a few generations, their genetic contribution becomes concentrated rather than spread out across the population.
Linebreeding is one way shared ancestry becomes concentrated. It means breeding dogs that are related to each other in order to increase the genetic contribution of particular ancestors. The intent is usually to make certain traits more predictable by increasing the chance that offspring inherit the same variants repeatedly.
Because each parent contributes half of a dog’s genetic material, breeding to relatives increases the chance that the same variants are inherited from both sides of the pedigree. When this happens occasionally and remains local, its effects may be limited. When it happens repeatedly, or when the resulting dogs are then used widely across a population, the same genetic material starts to appear in many dogs at once. This matters because harmful variants are not evenly distributed. Some do little when present in one copy but cause problems when a dog inherits two copies. Repeated breeding to the same lines increases the probability of that happening and reduces the range of alternative genetic material available in the population.
So the big issue is not that ancestors are shared in general – it is when recent ancestry becomes concentrated in a small number of dogs and then spreads through the population through repeated use. That is the point at which shared ancestry turns into a population-level risk.
4. If shared ancestry is normal, what exactly has gone wrong in some dog breeds?
If shared ancestry by itself is not the problem, then the question becomes what turns it into one. In most cases, the answer is not a single decision or a single breeder, but a pattern of choices that concentrate recent ancestry faster than it can dilute.
One of the main ways this happens is through repeated breeding to related dogs, often described as linebreeding. In genetic terms, the distinction between “linebreeding” and “inbreeding” is rhetorical rather than biological. Mating relatives is mating relatives – the difference is one of degree, not kind. What matters is how quickly relatedness accumulates and how widely the resulting dogs are then used.
Linebreeding is usually defended on the grounds of predictability: the aim often is not only to reproduce visible traits like size or colour, but to stabilise characteristics that are harder to see directly, such as behaviour, working style, or performance. By increasing the chance that offspring inherit the same variants from both sides of the pedigree, breeders try to make outcomes more reliable.
The problem is that this mechanism does not distinguish between variants that are desirable and variants that are harmful. Inbreeding increases homozygosity across the genome as a whole: that is why it can increase uniformity, but it is also why it increases the probability that harmful recessive variants are paired up. Every dog carries many such variants, most of which are invisible until two copies meet.
The risk increases further when linebred dogs are then used widely, rather than locally. This is where popular sire effects come into play: when a small number of males produce a disproportionate share of the next generation (e.g. 100-300 litters/300-1500 puppies), their genetic material spreads rapidly through the population, displacing alternative lines. From a population perspective, this collapses the number of dogs that are effectively contributing genes, even when the breed appears numerically large.
At this point, outcomes that look like individual success start to produce collective cost. Selection decisions that make sense in isolation, such as using a proven performer or a dog with a strong track record, begin to accumulate into a population-level pattern of rising relatedness and shrinking diversity.
Concentration also affects how traits are expressed: when the same behavioural or physical tendencies are reinforced over multiple generations, they can drift from being useful or functional into being excessive (also known as hypertype). Characteristics that are advantageous in moderation, such as speed of arousal, intensity, or drive, can become problematic when amplification replaces balance. Over time, this is how populations end up with dogs that are harder to regulate, less resilient, or more prone to stress, even when no single breeding decision seemed extreme at the time.
An alternative way of thinking about breeding is not to try to “lock in” a narrow set of traits, but to consider how two dogs balance each other. Every dog has limitations – the question is which limitations can be tolerated, and which need to be offset by complementary characteristics in a mate. That approach tends to distribute genetic material more widely, rather than concentrating it around a small number of admired ancestors.
These dynamics are easier to miss in large, uncoordinated populations. In some working breeds, including working cocker spaniels**, breeding decisions have historically been made at the level of individual kennels or professional users, with little or no shared population management. There is no agreed ceiling on sire use, no mechanism for tracking how representation is distributed across lines, and no collective responsibility for what happens several generations down the line. What works now is rewarded, and what accumulates later is nobody’s job.
This is where size can be misleading. Large populations can tolerate inefficient management for a long time, because the effects are gradual and uneven. Smaller or newer breeds, by contrast, are often scrutinised precisely because their margins are thinner. Yet a small population with deliberate coordination can retain diversity more effectively than a large population in which reproduction is highly skewed.
A related issue is how genetic risk is communicated. It is possible for a mating to have a very low or even zero pedigree-based coefficient of inbreeding, while still drawing from heavily linebred lines one generation back. That number can sound reassuring, especially to people without a background in population genetics, but on its own it says little about how concentrated genetic material already is in the population. This is not a question of bad faith. It is a limitation of how single metrics are often used and understood.
Taken together, these patterns explain how breeds can accumulate genetic risk without any single decision looking obviously irresponsible at the time. Shared ancestry becomes harmful not because it exists, but because it is repeatedly concentrated, widely propagated, and rarely managed as a collective problem. The next question, then, is what responsibility looks like once these mechanisms are understood.
** This difference in structure is one of the reasons I focus on spaniels I know well, rather than making claims about all breeds. I did not set out to audit breeding practices – I learned about these dynamics while trying to make sense of mate choice for my own dog, and once you understand how concentration happens, it becomes difficult to unsee it elsewhere.
So how should this be done, if people are trying to do it responsibly?
Everything in life carries risk, and “responsible breeding” is no different. We can’t avoid risk altogether but we can decide which risks to accept, which to reduce, and which to refuse to compound further.
The first step is recognising that every breeding decision is a trade-off. Selecting for type, performance, temperament, or predictability always removes genetic material, because dogs that do not breed contribute nothing to the next generation. Each generation is a genetic sample, not a copy. The practical question is therefore not how to produce a certain kind of dog, but how to avoid losing the genetic material that supports health, resilience, and long-term function while doing so.
That shifts the focus away from individual pairings and towards patterns over time.
Responsible breeding pays attention to how genetic material is distributed across a population, not just whether a planned mating “looks safe” in isolation. This includes asking which dogs are already over-represented, which lines are carrying most of the reproductive load, and which dogs are rarely used despite being healthy and functionally sound. When the same dogs or the same families dominate reproduction, diversity is lost quickly, even if each individual mating appears reasonable.
It also means thinking in terms of balance rather than perfection. No dog is free of faults, so responsible breeding involves deciding which limitations can be accepted and which should be offset by the choice of mate, rather than trying to amplify a narrow set of desirable traits generation after generation. When selection focuses too narrowly, whether on appearance, performance, or temperament extremes, it tends to pull other traits out of balance. Over time, this can produce dogs that are impressive in one dimension but compromised in others.
A personal example to illustrate these choices
As an example, my decision about whether or not to breed with Grace was not just about whether I thought she was the best possible, perfect dog for me. I also needed to consider objectively and rationally whether her genetic contribution was, at least theoretically, valuable to the future generations of the breed.
Choosing to breed a dog is effectively multiplying them – we don’t add just one, but several dogs into the world, and ultimately they will become someone else’s problem, if something isn’t quite right. I have strong opinions about people breeding their dog just because THEY think their dog is lovely – a dog can be worthy and lovely without multiplying their genes.
I needed to be brutally honest with myself not just about her strengths, but also about her flaws, whether they are acceptable in case they are passed on to the next generations and the probability of those traits being passed on (or whether they might actually be something I had shaped as her guardian!). I then needed to think how I could mitigate against or soften those shortcomings by choosing a mate that complemented and balanced her.
In the spirit of radical transparency, I have shared my decision process openly elsewhere on this blog. For convenience: “With Grace, I wanted to balance her spirited nature with a male who was more relaxed and calm – not doing so might risk accentuating the extremes of each parent, and the goal is for the next generation to always be better than their parents.”
Just to make this explicit for anyone who hasn’t read my other posts: making the next generation better than their parents isn’t about some kind of performance that benefits humans at the expense of the dog’s wellbeing. I chose a male who exhibited working characteristics I wished to have – calm focus, which was evidenced by achieving a full score and 1st place in a hunting test when he was only 12 months old. These characteristics also serve the dogs I brought to this world throughout their whole lives.
This is where trade-offs become unavoidable: choices that increase predictability for the breeder often reduce flexibility for the population. Practices that make outcomes easier to control in the short term tend to increase long-term risk, even if that risk is not immediately visible. Responsible breeding does not eliminate these trade-offs, but it acknowledges them and makes them explicit.
Health testing belongs in this picture, but it does not replace population management. DNA tests can identify specific known mutations, but they cannot see the accumulation of small, unknown risks that come from rising homozygosity (“genetic uniformity”) over generations. Testing can help avoid predictable harm, but it cannot justify repeatedly narrowing the genetic base and assuming the consequences will remain manageable.
Responsible practice therefore relies on measurement, restraint, and revision as well as coordination within a breed community:
- Measurement means tracking relatedness, not just named diseases.
- Restraint means limiting the use of any one dog, no matter how successful or appealing.
- Revision means being willing to change direction when outcomes suggest that a line, a strategy, or a norm is producing costs that outweigh its benefits.
- Coordination means working together to build a better future for the breed, which sometimes means setting aside one’s ego and interpersonal politics – something that humans often find difficult!
None of this guarantees good outcomes. It does, however, make improvement possible. Breeding without this level of attention does not merely accept risk; it passes it forward and compounds it.
This is why access to information matters. People cannot ask better questions or make better decisions if the structure of risk is invisible to them. Making that structure visible does not solve ethical disagreement, but it gives it something solid to rest on.
The next post looks more closely at health risk and uncertainty, and at why phrases like “the science isn’t there yet” are often used in ways that obscure what we do, in fact, already know.
Closing: what this does and does not settle
This post has stayed with language and mechanisms. Its purpose has been to clarify how selective breeding, shared ancestry, and population structure actually work, and to show why some common assumptions create more confusion than insight.
It does not settle whether breeding is morally justified, which practices are acceptable, or where responsibility should ultimately sit. Those are real questions, but they are not questions that genetics can answer on its own.
What this clarification does change is what it makes sense to argue about. Once the mechanics are clearer, it becomes easier to see where risk is genuinely created, where it is often overstated, and where it is simply misunderstood.
The next post moves from structure to consequence by looking at health risk, genetic testing, and uncertainty, including why DNA tests cannot “solve” breeding risk, why measures like COI are often misused, and how apparently reassuring numbers can obscure rather than reveal what is actually going on.
For now, the aim has been simpler: clearer language, fewer shortcuts, and a shared starting point for disagreement that is grounded in how inheritance actually works.
Next post: Risk and uncertainty of dog genetics and health testing
Please note: I have not made detailed in-line references in this article simply because of the time it takes – my blog does not usually have a huge volume of readers, so I balance the time and effort it takes to write these longer educational pieces with the number of people who will actually benefit from the additional time of being academically precise with referencing. If there is a particular claim that you’d like me to point you to, please contact me and I’d be happy to do so.
For further reading, I recommend this open access source that has been hugely helpful to me: Institute of Canine Biology (Blog) They also offer free and paid courses. Much of the detail in this blog post is thanks to what I have originally learned from this site.
BusyDoggie 