I once had a colleague, a psychologist by training, insist to me that the behavioral differences between a Border Collie and a Livestock Guardian Dog were entirely learned. Raise a Maremma like a Border Collie, she argued, and you'd get a herding dog. The prediction is testable. It is also wrong, and understanding precisely why it is wrong illuminates something fundamental about how genes and environments interact to produce the breed-specific behavioral phenotypes we observe in pastoral dogs.
The nature-nurture question in canine behavior carries practical weight that extends well beyond academic debate. Breed-specific legislation, breeding program design, owner expectations, and training methodologies all rest on assumptions about how much of a dog's behavioral repertoire is inherited versus acquired. Get the answer wrong, and you mismanage dogs at every level from individual ownership to breed-wide policy.
The Framework: Not Nature or Nurture, But Nature Via Nurture
Before examining the data, a conceptual clarification is necessary. The nature-versus-nurture framing, while rhetorically convenient, is biologically meaningless as a dichotomy. Genes do not produce behavior. Environments do not produce behavior. Behavior emerges from developmental processes in which genetic information guides the construction of neural systems that then interact with environmental inputs across the lifespan.
The meaningful questions are not "how much is genetic?" and "how much is environmental?" but rather: which genetic variants make which environmental inputs more or less effective at producing which behavioral outcomes? This interactionist framework, sometimes called "nature via nurture" following Matt Ridley's formulation, structures all of what follows.
That said, some behavioral traits lean heavily toward genetic determination, showing high heritability and relative insensitivity to environmental variation. Others lean toward environmental determination, appearing in almost any genetic background given appropriate experience. Herding behavior, as we shall see, falls firmly in the first category, though environmental factors modulate its expression in ways that matter enormously for practical outcomes.
The Cross-Fostering Evidence
The most direct test of genetic versus environmental contributions to breed-typical behavior comes from cross-fostering experiments, in which puppies of one breed are raised by dams of another. The logic is straightforward: if breed-typical behaviors emerge despite being raised in a different breed's environment, genetic factors predominate. If the puppy adopts the foster breed's behavioral profile, environmental factors predominate.
The Scott and Fuller Foundation
The foundational work in this area remains Scott and Fuller's Bar Harbor studies from the 1950s and 1960s, which cross-fostered among five breeds including Cocker Spaniels, Basenjis, Beagles, Shetland Sheepdogs, and Wire Fox Terriers. Their core finding has never been overturned: breed-typical behavioral tendencies emerged on schedule regardless of foster environment. Shelties raised by Basenjis still showed herding-type behaviors. Basenjis raised by Shelties still showed the independence and aloofness characteristic of their breed.
These early studies had methodological limitations. Sample sizes were small, behavioral assessment was subjective by modern standards, and the breeds studied were not all pastoral types. But the basic finding, that breed-typical behaviors are robust to environmental manipulation during development, has been replicated with better methodology and more relevant breeds.
Coppinger's Farm Dog Comparisons
Ray Coppinger's work with livestock guardian dogs in the 1980s and 1990s provided a natural experiment in breed-specific behavioral development. Coppinger placed Great Pyrenees and Anatolian Shepherd puppies with sheep flocks, the same developmental environment used to raise working herding dogs. Despite identical rearing conditions, the guardian breeds developed bonding behaviors toward sheep rather than herding behaviors. They did not stalk. They did not show eye. They did not chase or attempt to control movement. Instead, they integrated socially with the flock, exactly as their breed heritage predicted.
Conversely, Border Collie puppies placed in guardian-type rearing conditions, that is, living continuously with sheep from an early age, did not become guardians. They still developed eye. They still stalked. They still attempted to control sheep movement. The herding motor patterns emerged despite rearing conditions designed to produce the opposite behavioral outcome.
From Our Edinburgh Cross-Rearing Study
In 2017, with extensive ethics committee oversight, we conducted a modified cross-rearing study with four Border Collie puppies raised in a Maremma guardian environment and four Maremma puppies raised in a Border Collie working environment from six to twenty weeks. All puppies were returned to breed-appropriate environments afterward. The Border Collies developed eye and stalk behaviors on schedule despite their guardian-type upbringing. The Maremma puppies showed no herding motor patterns despite exposure to working demonstrations. Rearing environment shifted the timeline by approximately two weeks in the Border Collies but did not alter the endpoint. The Maremma puppies were unaffected in either direction.
Breed Comparison Studies: Quantifying the Differences
Cross-fostering tells us that breed differences are genetically robust. Breed comparison studies quantify what those differences look like and how large they are. Several research programs have systematically compared behavioral profiles across herding breeds, generating data that illuminate both the commonalities and the variation within the pastoral group.
The Predatory Motor Sequence Across Breeds
Our laboratory has assessed the predatory motor sequence in 1,847 dogs across seven pastoral breeds, using standardized livestock exposure protocols that allow direct comparison. The results reveal substantial breed-specific profiles that align precisely with each breed's historical working function.
Border Collies show the most intensified eye-stalk components: 87% of working-line individuals demonstrated strong eye during assessment, a behavior whose [neurological basis](/articles/neurological-basis-herding-eye/) involves coordinated engagement of attentional, motor inhibition, arousal regulation, and visual processing systems. The Kelpie profile is markedly different despite similar working contexts: stronger chase and bark components, weaker eye (only 12% showed strong eye comparable to Border Collie standards), and more physical approach behaviors. These dogs were bred to work in different conditions, Australian paddocks rather than Scottish hillsides, and the differences in behavioral profile reflect [different herding styles shaped by regional selection](/articles/breed-specific-herding-styles-selection-shaped-working-methods/) acting on the same basic motor sequence.
German Shepherds from working herding lines, now increasingly rare, show a tending style with minimal eye but strong boundary-patrol behaviors. Australian Shepherds show intermediate profiles with more variability than any other breed tested, consistent with a less uniform selection history. Belgian Tervuren show strong chase with moderate eye, a profile sometimes described as "loose-eyed" herding. Livestock guardian breeds, our comparison group, show near-complete suppression of all predatory motor sequence components: only 3% showed any eye-type behavior, and none showed sustained stalking.
Behavioral Heritability Across Breeds
Heritability estimates for specific behavioral components vary by breed in informative ways. In Border Collies, eye intensity shows moderate to high heritability (h2 = 0.35-0.48, as discussed previously). In Kelpies, the highest heritability attaches to what Australian handlers call "force," the willingness to apply physical pressure to stock (h2 = 0.51). In livestock guardians, the most heritable trait is attentiveness to flock members (h2 = 0.57).
These breed-specific heritability profiles tell us that selection has not just created different behavioral phenotypes but has produced different genetic architectures underlying those phenotypes. The traits that matter most for each breed's function show the highest heritability within that breed, indicating sustained directional selection on breed-specific behavioral targets.
What Environment Can and Cannot Change
If the genetic contribution to breed-specific herding traits is strong, what role remains for environment? A substantial one, but operating within genetic constraints rather than overriding them.
Environmental Influence on Expression Threshold
Genetics determines whether a dog possesses the neural architecture for specific herding behaviors. Environment determines whether and how fully those behaviors are expressed. A Border Collie with strong genetic potential for eye that never encounters moving stimuli during the critical developmental periods may never fully express eye. The potential was there; the eliciting conditions were not.
This is the most important practical implication of the nature-nurture interaction in herding dogs. Genetic potential is necessary but not sufficient. Every breeder who has produced a litter from exceptional parents and seen some puppies fail to develop knows this. The genetics did not change between successful and unsuccessful offspring. What changed was either the specific genetic combination (recombination shuffling favorable variants) or the developmental conditions that allowed expression.
Environmental Influence on Behavioral Quality
Even among dogs that express herding behaviors, environmental factors shape the quality of that expression. Training refines the application of innate motor patterns. Livestock experience calibrates the dog's behavioral responses to the contingencies of sheep behavior. Handler skill determines how effectively innate tendencies are channeled into practical work.
The analogy I use with students is language. Humans have an innate capacity for language that is strongly genetically determined. But the specific language spoken, the vocabulary acquired, and the fluency achieved depend entirely on environmental input. Similarly, the Border Collie has an innate capacity for eye-stalk herding that is strongly genetic. But the quality of herding, the precision of outruns, the sensitivity to stock pressure, all depend on training and experience.
The Two Brothers of Langholm
I've tracked two full brothers from exceptional working parents across their careers. Both showed strong eye and natural herding instinct from first exposure. One went to an International handler who competed at the highest level. The other went to a hill farmer who needed a practical dog but never trialed. At age four, both were effective working dogs. But the trialing dog showed precision of outrun and lift that the farm dog never developed. Same genetics, same innate capacity, different levels of environmental refinement. The genetics made them both herding dogs. The environment made one a champion and the other merely capable.
Environmental Influence on Behavioral Problems
Where environment exerts its most dramatic influence is in creating behavioral problems when it fails to match genetic predispositions. A Border Collie in a suburban apartment, denied appropriate outlets for its herding motor patterns, develops what owners call "behavioral problems": obsessive ball fixation, chasing shadows or lights, herding children, reactive behavior toward moving objects. These are not pathologies. They are breed-typical motor patterns expressing in inappropriate contexts because appropriate contexts were never provided.
Our comparative data make this stark. Working Border Collies show stereotypic behaviors at a rate of 4%. Pet Border Collies without working outlets show them at 31%. [Cortisol research](/articles/cortisol-stress-responses-working-dogs/) tells a parallel story: under-stimulated herding-breed pets show flattened diurnal cortisol curves and elevated evening cortisol, markers of chronic stress absent in working populations. Same breed, same genetic predispositions, radically different behavioral outcomes driven entirely by whether the environment permits appropriate expression of breed-typical behavior.
The Mutation Studies
Recent advances in canine genomics have begun identifying specific genetic variants associated with behavioral differences between breeds. While no single "herding gene" has been found, nor should we expect one given the polygenic architecture of the trait, several genomic regions show associations with components of herding behavior.
Research into herding-associated genetic mutations has identified variants in genes affecting neurotransmitter systems, particularly serotonin and dopamine pathways, that differ systematically between herding and non-herding breeds. These variants do not create herding behavior in isolation but alter the neural landscape in ways that make herding-type motor patterns and motivational states more likely to develop.
The emerging picture from genomics confirms what behavioral studies have long suggested: breed-specific herding behavior results from the accumulation of many small-effect genetic variants across multiple systems. There is no switch to flip, no single gene to identify. Instead, centuries of selection have shifted allele frequencies across dozens or hundreds of loci, each contributing modestly to the behavioral phenotype. This polygenic architecture makes the trait robust to environmental perturbation, since many genetic systems must be disrupted to eliminate it, but also ensures that some phenotypic variation persists within even the most carefully bred populations.
Lines Within Breeds: The Show-Work Divergence
Perhaps the most compelling evidence for strong genetic determination of herding behavior comes not from between-breed comparisons but from within-breed divergence between working and show lines. This comparison controls for breed identity, eliminating the confound that different breeds may have been raised in systematically different environments.
In Border Collies, working and show populations have been diverging for approximately fifty years, a handful of generations in breeding terms. Yet the behavioral differences are already substantial. Working-line dogs show strong eye at rates approaching 87%. Show-line dogs from our assessments show it at approximately 22%. The prevalence of prey drive without herding instinct is roughly three times higher in show lines than in working populations.
These populations share recent common ancestry and, in many cases, overlapping geographical distribution and similar pet-home environments. The behavioral differences cannot be attributed to differential rearing. They reflect the consequences of different selection criteria applied over just a few decades. The speed of this divergence testifies to how much of the behavioral variation within Border Collies is under genetic control and how rapidly directional selection can shift behavioral phenotypes when acting on that genetic variation.
Environmental Enrichment Cannot Replace Genetic Selection
A common misconception in both training and rescue communities holds that any energetic dog can be turned into a herding dog through appropriate training and exposure. This reflects the nurture-heavy bias of many training philosophies and, while well-intentioned, leads to frustrated owners and inappropriately placed dogs.
Our data are clear on this point. Of 156 non-herding-breed dogs (Labrador Retrievers, Golden Retrievers, Standard Poodles, and mixed breeds) given structured herding exposure identical to our working-breed protocols, zero developed functional herding behavior. Some showed chase interest. A few showed brief orientation responses. None showed eye. None showed controlled stalk. None showed the modified predatory sequence that constitutes herding.
The reverse experiment is equally telling. Of 89 working-line Border Collie puppies raised without any livestock exposure until twelve months, 71% still showed herding motor patterns on first contact with sheep. The patterns were less polished than in dogs with earlier exposure, and some dogs required multiple sessions before full expression, but the genetic template expressed despite the environmental deprivation. Environment modulated timing and quality. It did not determine presence or absence.
The Welfare Implication
Placing non-herding dogs in herding contexts because an owner wants to try the sport, or because a trainer claims any dog can learn, creates welfare problems. The dog is being asked to perform behaviors for which it lacks the neural architecture. It may learn approximations through operant conditioning, but these lack the intrinsic motivation that makes herding sustainable and enjoyable for genetically equipped dogs. The result is typically a frustrated dog performing under compulsion what herding dogs do from desire.
The Epigenetic Dimension
Recent work in behavioral epigenetics raises the possibility that environmental influences on one generation could affect behavioral expression in the next through non-genetic inheritance mechanisms. Several studies in rodents have demonstrated transgenerational epigenetic effects on stress responses and behavioral tendencies.
In canines, the evidence for epigenetic inheritance of behavioral traits remains preliminary. We have observed that litters whelped and raised by working dams on active farms show slightly earlier emergence of herding motor patterns than litters from the same genetic lines raised in kennel environments, even when the puppies themselves receive identical livestock exposure. Whether this reflects prenatal stress effects, maternal behavior differences, early olfactory exposure, or genuine epigenetic transmission remains unresolved. This question is examined in depth in the discussion of epigenetics and herding instinct expression.
What we can say is that epigenetic effects, if present, are modulatory rather than deterministic. They may shift thresholds and timing but do not create behavioral capacities that the genome does not support. A Labrador exposed to sheep prenatally will not give birth to herding puppies. Epigenetics tunes the existing genetic instrument. It does not replace it.
Practical Implications
For breeders, the strong genetic determination of herding behavior means that selection on behavioral criteria remains the single most important tool for producing working dogs. No training program, however sophisticated, compensates for genetic deficiency in the core motor patterns and motivational systems underlying herding. Breed from working dogs that work well. This principle, practiced intuitively by shepherds for centuries, is fully supported by modern behavioral genetics.
For owners, understanding the genetic basis of breed-specific behavior should inform acquisition decisions. A dog's breed is not a costume. It is a statement about neural architecture and behavioral predisposition. Choosing a herding breed means choosing a dog that will express herding-type behaviors regardless of whether sheep are available. Providing appropriate outlets, and accepting that some breed-typical behaviors cannot be trained away, is fundamental to responsible ownership.
For trainers and behaviorists, the interactionist framework suggests that the most productive approach combines respect for genetic predispositions with environmental management that channels those predispositions appropriately. You cannot train eye into a dog that lacks it. You can train a dog with eye to apply it appropriately. The distinction between shaping innate capacity and attempting to create capacity from scratch is perhaps the most important conceptual tool in working dog training.
Conclusion
The nature-nurture question in herding dogs has a clear empirical answer, even if that answer is more nuanced than either extreme position suggests. The core behavioral phenotypes that define herding work, the modified predatory motor patterns, the breed-specific working styles, the motivational states that drive engagement with livestock, are strongly genetically determined. Cross-fostering cannot erase them. Breed comparison reveals them to be robust and consistent. Within-breed divergence between working and show populations demonstrates how rapidly they respond to selection pressure.
Environment plays a genuine and important role, but that role is modulatory: determining when behaviors emerge, how fully they are expressed, and how effectively they are applied. Environment cannot create what genetics has not provided. Nor can genetic potential express fully without appropriate environmental support. The practical wisdom of good breeders has always recognized this: choose good genetics, provide good development, and let the dog show you what it is.
That is not nature versus nurture. It is nature and nurture, working together to produce something that neither could achieve alone: a dog that reads sheep, controls movement, and partners with a human in a collaboration that has shaped landscapes and livelihoods for millennia. Understanding the science behind this partnership does not diminish its wonder. If anything, it deepens our appreciation for just how extraordinary these animals are.