A colleague from Germany visited my laboratory last autumn and brought video of a German Shepherd Dog working a flock along a roadside boundary. The dog moved in slow, deliberate patrols at the flock's edge, never entering the group, never showing eye, never attempting to circle. My postgraduate students, accustomed to watching Border Collies, were confused. "Is it actually herding?" one asked. It was. It was herding superbly. It was simply herding in a way that looked nothing like what they had been studying.
The assumption that herding means eye-stalk-gather is pervasive in both popular and scientific literature, largely because the Border Collie dominates discussion. But heading style represents only one branch of a diverse family of livestock management behaviors that selective breeding has produced. Understanding this diversity matters not just for breed appreciation but for properly evaluating working instinct, designing breeding programs, and avoiding the error of measuring all herding dogs against a single standard.
Three Fundamental Strategies
Across the pastoral breeds, herding behavior clusters into three broad strategies, each reflecting the specific agricultural demands that shaped it. These are not arbitrary categories. They represent genuinely different behavioral architectures, different configurations of the same underlying predatory motor patterns shaped by different selection pressures over centuries.
Heading: The Gather-From-Front Strategy
Heading is the style most people picture when they think of herding. The dog positions itself at the far side of the flock relative to the handler and brings livestock toward the handler using a combination of eye, physical pressure, and controlled movement. The Border Collie is the archetypal header, but the style appears across multiple breeds including Kelpies, Bearded Collies, and English Shepherds.
The heading strategy makes particular sense in hill farming, where sheep scatter across vast landscapes and must be gathered to a central point. The dog needs strong outrun instinct, the ability to maintain distance while circling wide, and the self-regulation to approach from behind without pushing stock past the handler. The neurological architecture underlying the herding eye is most fully developed in heading breeds, because sustained attentional pressure at distance is the primary tool for controlling livestock movement.
Our behavioural assessments of 487 heading-style dogs across four breeds confirm that the orient-eye-stalk components of the predatory motor sequence are significantly more intensified in headers than in any other herding category. Mean eye duration averaged 4.3 seconds across heading breeds, compared to 0.8 seconds in heelers and 0.4 seconds in tending breeds.
Heeling: The Drive-From-Behind Strategy
Heeling dogs work from behind or beside livestock, pushing them forward rather than gathering them. The approach involves nipping at heels, barking, and physical proximity rather than the distant psychological pressure of eye. Welsh Corgis, Australian Cattle Dogs, Bouviers des Flandres, and several continental European breeds are heelers.
The agricultural context that produced heelers was typically the movement of cattle along roads or through confined spaces: from barn to pasture, from paddock to market. Cattle, being larger and less flighty than sheep, require more physical persuasion. A dog that stood at distance showing eye would be ignored by most cattle breeds. Heelers evolved to apply direct pressure, often including controlled bites to the hocks, then ducking the inevitable kick.
The predatory motor sequence in heelers is modified differently from headers. Rather than exaggerating eye-stalk and inhibiting the terminal sequence, heelers retain elements of grab-bite while losing most eye-stalk components. The grab is modified, a nip rather than a hold, and directed at a specific body part, the heels. This represents a different fragmentation of the ancestral predatory sequence, one that retains physical contact while redirecting it from predation to persuasion.
The Corgi Paradox
Welsh Corgis present an interesting case study in how body morphology and behavioral style co-evolved. Their low stature, sometimes dismissed as a conformation curiosity, is a functional adaptation. A short dog nipping at cattle heels stands below the arc of the kick. Corgis that were too tall got injured. Selection for both behavioral style and complementary body structure proceeded in parallel, producing dogs whose physical form and working method are inseparable.
Tending: The Boundary Patrol Strategy
Tending is perhaps the least understood herding style in English-speaking countries, yet it is the dominant pastoral working method across much of continental Europe. The tending dog does not gather or drive. Instead, it patrols the boundary of a flock that is grazing unfenced land, preventing individuals from straying into cultivated fields or roadways. The German Shepherd Dog, before its diversion into police and show roles, was primarily a tending breed. The Briard, Beauceron, and Belgian Tervuren were also selected for tending.
The behavioral profile of a tending dog looks almost nothing like a heading dog. There is minimal eye. There is no outrun or gather. The dog moves in a distinctive patrol pattern along the flock's edge, orienting toward individuals that drift beyond the boundary and applying just enough pressure, usually physical presence combined with movement, to turn them back. The skill lies in reading which sheep are about to stray and intercepting before they leave the group, a form of anticipatory livestock management that requires different cognitive demands from heading or heeling.
In our comparative study, tending breeds scored highest on measures of independent decision-making and territorial awareness but lowest on measures of eye intensity and chase motivation. They are, in a real sense, boundary specialists rather than movement specialists.
Different Modifications of the Same Ancestral Sequence
What unifies all three styles is their origin in the predatory motor sequence. What distinguishes them is which components have been exaggerated, which have been suppressed, and how the remaining components have been redirected. This matters for understanding the distinction between prey drive and herding instinct, which manifests differently across herding styles.
Predatory Sequence Modification by Style
In headers, orient-eye-stalk is intensified while everything from grab-bite onward is suppressed. In heelers, orient and chase are intensified, eye-stalk is reduced, grab-bite is retained but modified, and kill-bite is suppressed. In tending dogs, orient is intensified but redirected toward boundary violations rather than individual animals, chase is moderate and short-range, and the entire terminal sequence is suppressed. Three different solutions to three different agricultural problems, all built from the same raw material.
Quantifying the Differences
Our comparative dataset, covering 1,247 dogs across seven breeds assessed using standardized livestock exposure protocols, reveals the magnitude of these differences. The results make clear that "herding instinct" is not a single trait but a family of related but distinct behavioral phenotypes.
Headers showed significantly longer eye durations, wider outrun arcs, and greater working distances than either heelers or tenders. Heelers showed significantly more physical contact behaviors, including heel nipping in 73% of individuals, compared to less than 5% in headers and near zero in tending breeds. Tending breeds showed the highest rates of spontaneous boundary-patrol patterns, with 89% exhibiting the characteristic back-and-forth patrol within their first three livestock exposures.
What these numbers reveal is that each herding style has its own behavioural signature, consistent within breeds and distinct between them. These are not learned preferences. They are genetically encoded working methods that emerge with remarkable consistency given appropriate developmental conditions, a finding reinforced by [cross-fostering research](/articles/nature-nurture-breed-specific-herding-traits/) demonstrating that breed-typical herding patterns appear regardless of rearing environment.
Field Observation: Three Breeds, One Farm
I spent a week on a mixed operation in the French Pyrenees that employed Border Collies for sheep gathering, a pair of Beaucerons for tending on unfenced pasture, and a Bouvier des Flandres for cattle work. Watching all three types work the same property drove home how different their behavioural repertoires were. The Border Collies gathered with wide outruns and intense eye. The Beaucerons patrolled invisible boundaries with quiet authority. The Bouvier pushed cattle with bark and bite, ducking hooves with practised efficiency. Same farm, same handler, utterly different behavioural solutions. The handler told me he would never attempt to use one type for another's job. "They don't understand each other's work," he said. He was describing genetic specialisation without using the term.
Selection History and Regional Ecology
The geographical distribution of herding styles maps onto agricultural practices in ways that illuminate how selection pressures shaped behavior. Heading breeds dominate in regions with extensive hill farming and scattered flocks: Scotland, Wales, northern England, and parts of Australia. These landscapes require dogs that cover vast distances and bring stock to a central point.
Heeling breeds cluster in regions with confined livestock management: dairy farming in Wales and Flanders, cattle droving in Australia. The work involves moving animals along defined routes rather than gathering from open land. The dog needs to push through resistance, which means physical contact and a willingness to confront animals many times its size.
Tending breeds developed primarily in continental Europe, particularly Germany and France, where open-field agriculture required flocks to graze unfenced land adjacent to cultivated crops. The dog's job was containment, not movement, requiring a totally different set of behavioral priorities. The tending style reflects an ecology where the primary problem was preventing straying rather than gathering or driving.
Why This Matters for Evaluation
The practical consequence of herding style diversity is that no single evaluation protocol adequately assesses working instinct across all herding breeds. A standard herding trial format that scores outrun, lift, fetch, and pen is designed for heading dogs and will systematically undervalue heelers and tenders.
I have watched tending-breed handlers struggle with trial formats designed for Border Collies, their dogs failing not because they lacked herding instinct but because the test measured the wrong instinct. A German Shepherd that patrols a flock boundary superbly may score poorly on a gather-drive-pen course. An Australian Cattle Dog that moves stubborn bulls through yards with confident nipping may fail a test that penalises stock contact. The evaluation measures the style, not the aptitude.
This has consequences for breeding. When breed evaluations use heading-style criteria for all herding breeds, selection inadvertently shifts populations toward heading-type behavioral profiles and away from breed-appropriate working methods. German Shepherds selected through heading-style trials begin to lose their tending instinct. Cattle Dogs whose breeders emphasise "no-contact" herding trials lose the confident physical approach that defines their working method.
The Evaluation Bias Problem
Using a single evaluation framework for all herding breeds is analogous to testing sprinters and marathon runners on the same course and concluding that marathon runners are poor athletes. The breeds evolved to solve different problems. Their assessment must reflect this. Breed-specific evaluation criteria, informed by historical working function, should replace the one-size-fits-all approach that currently dominates competition.
The Developmental Implications
Herding style differences also affect how critical developmental periods should be managed across breeds. The exposure protocols we have developed for Border Collie puppies, emphasising distant observation before direct contact, match the heading style's requirement for distance work and controlled approach. But these protocols may be inappropriate for heeling breeds, whose working method requires early habituation to physical proximity with stock, or for tending breeds, whose patrol instinct benefits from boundary-type exposures rather than gather-type experiences.
Our preliminary data from 34 Australian Cattle Dog puppies suggest that heeling breeds benefit from earlier direct stock contact than heading breeds, precisely because their working method involves physical engagement rather than distant pressure. Cattle Dog puppies given stock access from ten weeks showed more confident and appropriate heel-nip behaviours at six months than those whose first contact was delayed until fourteen weeks. The developmental scaffolding must match the behavioural phenotype being developed.
Style Interactions in Mixed Operations
Farms that employ multiple herding styles face an interesting management challenge. Dogs of different styles can interfere with each other's work. A heading dog's wide outrun may push stock into a tending dog's patrol zone, disrupting both. A heeling dog's direct pressure may cause stock to bolt past a heading dog's carefully maintained eye.
The most effective mixed operations I have observed keep different styles working in different contexts rather than simultaneously. The heading dogs gather from the hill. The heeling dogs work the yards. The tending dogs manage grazing boundaries. Each operates in the context its behavioural architecture was designed for, without interference from incompatible working methods.
Implications for Pet Owners
Understanding herding style also informs management of herding breeds in non-working contexts. A Border Collie in a pet home typically expresses frustrated heading instinct through obsessive ball fixation, stalking behaviour, and attempts to control movement. An Australian Cattle Dog expresses frustrated heeling instinct through nipping, body-checking, and physical confrontation. A German Shepherd may express frustrated tending instinct through boundary patrolling, territorial behaviour, and hyper-vigilance about objects or people leaving defined areas.
These are not generic "herding breed problems." They are style-specific behavioural expressions that require style-specific management. The Border Collie needs outlets for eye and controlled chase. The Cattle Dog needs outlets for physical engagement and confident approach. The German Shepherd needs outlets for patrol and boundary work. Prescribing identical management for all three fails because it ignores the fundamental diversity of what "herding" means across breeds.
Conclusion
The herding breeds represent not a single behavioural adaptation but a family of related adaptations, each shaped by specific agricultural demands and regional ecologies. Heading, heeling, and tending are genuinely different behavioural strategies built from different modifications of the ancestral predatory sequence, producing dogs that look, think, and work in fundamentally different ways.
Recognising this diversity is essential for accurate instinct evaluation, appropriate breeding selection, effective developmental management, and responsible pet ownership. The Border Collie's heading style is remarkable, but it is not the only form of herding. The Corgi's confident heel nip, the German Shepherd's quiet boundary patrol, and the Bouvier's physical authority are equally sophisticated solutions to different problems, each deserving evaluation on its own terms rather than against a single standard.
What centuries of regional selection achieved was not one herding dog but many, each precisely matched to its ecological niche. Our science and our evaluation systems should honour that diversity rather than flatten it.