Gear

A flat collar's job is to hold identification. That is the whole job, and it is an important one.

A plain buckle or snap collar with tags is the right default, sized so you can slide two fingers under it comfortably. Check it as puppies grow, which is faster than you think.

Do not attach the leash of a dog who pulls to a flat collar. A dog's neck is a crowded place: the trachea, the oesophagus and the thyroid gland all sit near the front, and steady pulling puts a surprising amount of pressure through them. Dogs pull hard enough for this to matter, with measured leash forces well into the tens of newtons (Shih et al., 2021), and reviews of walking equipment treat neck pressure as a real welfare concern (Townsend et al., 2022). You will also see specific claims online about collars raising pressure inside the eye or damaging the thyroid; those come from small studies that get quoted far past what they showed, so we are not going to lean on them. The plain version is enough: this is not about the collar being cruel, it is about a dog who leans into it for twenty minutes twice a day. Clip to a harness instead, and let the collar carry the tags.

A martingale, or limited-slip collar, tightens only to the width of the dog's neck and then stops. Fitted properly it cannot close down further, and it exists for dogs who back out of flat collars, sighthounds especially. It is a containment solution, not a training tool, and it has to be sized correctly to be either safe or pointless rather than harmful. If you find yourself relying on it to control pulling, you want a harness and a plan.

Often confused with

Slip leads and choke chains, which have no stop point at all. The presence or absence of that stop is the entire difference.

A body harness spreads leash pressure across your dog's chest and shoulders instead of their throat. For any dog who pulls, any dog with a short muzzle, any small dog, and any dog on a long line, it is the comfortable and obvious choice.

What to look for: a Y-shape at the chest that sits on the breastbone and leaves the shoulder blades free to rotate, straps that clear the armpits, and no strap running horizontally across the front of the shoulders. Two clip points, front and back, give you more options. Get it fitted while your dog is standing and walking, not just standing.

A front clip changes your leverage rather than your dog's understanding: pressure on the chest ring turns your dog toward you instead of letting them lean straight ahead. That makes an unmanageable walk manageable while you teach the walk. It is useful and it is honest about being management.

Some dogs need the harness itself taught, particularly anything that goes over the head. Feed through the neck opening, then on and straight off again, before you ever fasten it and walk out of the door. There is a plan for this, and it is worth the four sessions it takes.

A review of twenty-one studies on walking equipment came out roughly where this entry does: there is no one right device for every dog, and non-tightening front-clip harnesses strike the best balance between reducing pulling and keeping the dog comfortable (Cavalli et al., 2025).

Often confused with

A tool that stops pulling. No harness stops pulling. It stops pulling from hurting, which is what buys you the weeks it takes to train the alternative.

A halter that fits over the muzzle and behind the ears, so that gentle pressure on the leash turns your dog's head rather than pulling against their chest.

It belongs in this part of the glossary rather than in the section below, and the distinction is worth being precise about. A head halter does not work by causing pain. It works mainly by steering, the same principle as a halter on a horse, and a dog wearing a well-conditioned, well-fitted one can walk on a slack line all day feeling nothing at all. To be precise, the pressure on the nose when your dog pulls is mild, and easing off to avoid it is a mild form of negative reinforcement. This is why the conditioning and the slack line matter so much: they keep that pressure rare and small. That is a genuinely different tool from one designed to be uncomfortable when your dog pulls.

It does need respect, because the head and neck are involved:

• Never jerk the line, ever. A sharp pull can wrench your dog's neck. If you cannot walk your dog on a loose line with it, this is not the tool for today.
• Never attach a long line or a retractable to it, and never a tie-out. A dog hitting the end of a long line at speed is exactly the injury this tool can cause.
• Clip to a harness as well, on a double-ended leash, so the harness takes any real load and the halter only ever does the steering.
• Fit carefully. The nose loop should sit high, snug enough not to slide down toward the eyes, loose enough for your dog to open their mouth, pant, drink and take treats. A dog who cannot pant cannot cool down.
• Condition it before you use it. Most dogs dislike a nose loop initially and will paw at their face. Treat it exactly like a muzzle: food through the loop, one second on, straight off, over days. A halter forced onto a dog who has not been taught to like it is a miserable walk and an easy way to teach a dog to hate their walking gear.

Then the honest part about when to reach for one. Some dogs are simply stronger than their handler. Some handlers have a shoulder, a hip, a wrist or a balance issue that makes a lunging dog genuinely dangerous. And some dogs need to be walked safely today, while loose leash walking is a project that will take two months. In all of those cases a head halter is a reasonable, humane answer, and refusing to discuss it is how people end up in a pet shop buying something far worse.

One caveat in fairness: the same equipment review that favours front-clip harnesses urges caution with head collars, mostly on the grounds of the forces involved and how many dogs dislike wearing one (Cavalli et al., 2025). That is not a reason to rule them out, it is the reason for every bullet above.

It is still management, not training. It does not teach your dog anything about walking, and it should be paired with an actual plan and faded out as the training takes over. And some dogs never accept one, in which case the answer is a well-fitted front-clip harness and more patience.

Often confused with

A muzzle, which it is not, and which is worth knowing because strangers will assume it is. And prong or choke collars, which it is emphatically not: those depend on discomfort to work at all, and this depends on steering.

A leash is a safety line, not a steering device or a communication channel.

For most dogs and most walks, a plain flat leash of about six feet is right, clipped to a harness. Long enough that your dog can move and sniff without dragging you, short enough to manage near roads. A shorter four-foot leash gives more control in tight spaces; a lightweight leash matters for small dogs.

Hold it with slack in it. A leash that is permanently taut teaches your dog that pressure is simply the condition of being outside, and a leash that only tightens occasionally is one your dog can actually notice.

Retractable leashes are not aversive, but they are easy to misuse, and the reason is mechanical: the line stays under constant tension, so a dog who has not learned loose leash walking learns instead that pulling produces forward progress. That is reinforcement, working perfectly, in the wrong direction. They also give you almost no control in an emergency, and a dropped handle chases your dog. A long line does the same job better.

If you prefer one anyway, use it only when your dog is already trained to walk on a loose leash, is on a harness, is not a flight risk, and is not near traffic, other dogs or other people.

Often confused with

A training tool. Nothing about a leash teaches. What you do while holding it teaches.

A lightweight length of rope or webbing, typically fifteen to fifty feet, clipped to your dog's harness and never to a collar or head halter. It drags behind your dog and gives them something close to off-leash freedom while you keep the ability to stop them.

It is the single most useful piece of equipment for recall training, for sniffy decompression walks, and for any dog whose off-leash reliability is unfinished, which is most dogs for a long time.

Safety, because this one genuinely can hurt you:

• Tangles. The line wraps around your legs, your dog's legs and other dogs. Practise the long line dance: keep half an eye on the line near your feet and step or hop over it as it moves. Avoid running two dogs on long lines at once.
• Rope burn. Wear gloves. A sudden sprint pulls line through your hands fast.
• Being pulled over or dragged. Never wrap the end around your hand, never clip it to your waist, and never stand on the end with both feet. To shorten it, let the excess drag behind you or fold it into a loose coil that will unravel cleanly.
• Wet and dirty lines. Biothane costs more and stays clean and does not soak up water. Worth it if you use one often.

It is not necessary or productive to motivate a dog with pain, fear or coercion. Any tool or method that works by causing discomfort carries the same risks: physical injury, fear, aggression, and a dog who trusts you less.

The research here is consistent enough to be worth citing. Dogs trained at aversive-based schools showed more stress behavior, higher stress hormones afterwards and a more pessimistic outlook in an unrelated test, with the effect growing the more aversives were used (Vieira de Castro et al., 2020). Owners using confrontational methods reported their dogs responding aggressively at striking rates (Herron, Shofer and Reisner, 2009). Reward-based professional trainers got faster and more reliable responses than professional e-collar trainers following best practice (China, Mills and Cooper, 2020). European veterinary behavior specialists reviewed the evidence and concluded the welfare risks of electronic training devices outweigh their claimed benefits (Masson et al., 2018), and reviews of walking equipment discourage choke, prong and electronic collars on welfare grounds (Cavalli et al., 2025; Townsend et al., 2022). The practical experience of trainers who moved away from these tools points the same way.

The tell is always the same. Ask what makes this work? If the honest answer is that your dog will stop doing something because it brings an unpleasant thing, you are looking at positive punishment. If it is that your dog will do something to make an unpleasant thing stop, or to avoid it, that is negative reinforcement. Most aversive tools do both, whatever the packaging says about communication, information, tapping, pressure or leadership.

Choke chains and slip leads

Both tighten without limit when your dog pulls. Slip leads deserve their own mention because they are everywhere: handed out at vet clinics, shelters and rescues, sold as a simple loop of rope, and used by people who have never been told they are a choke collar without a stop. Convenient for thirty seconds of handling in a clinic; not a walking tool, and not a training tool. A dog's throat is not the place to put steady pressure.

Prong or pinch collars

Designed so that pulling produces enough discomfort to outweigh the reward of getting where your dog wants to go. That is the entire mechanism, and it is the reason we do not use them.

Shock collars, e-collars, remote collars

A shock collar only changes behavior if the sensation matters enough to avoid, which means it works because it is unpleasant. Every euphemism in the marketing exists to obscure that. They also fail in ways other tools do not: batteries die, contacts malfunction, and a dog can be shocked for something the handler never intended, which teaches them that pain arrives unpredictably.

Invisible and electric fencing

A shock collar wired to a boundary, so it carries all of the above plus three problems of its own. A determined dog will take the shock and leave anyway. Dogs form negative associations with whatever happens to be passing when they get warned or shocked, which is often people and other dogs. And it does nothing to stop anything coming in to your garden.

Bark, spray and citronella collars

Punishment triggered by a machine that cannot tell why your dog is barking. Most problem barking is boredom, alarm or distress, and suppressing the sound leaves the cause untouched.

Methods, not just gear

The same test applies to things that cost nothing:

• Leash corrections, pops, or "check" jerks — positive punishment applied to the neck.
• Alpha rolls, pinning, scruffing, "dominating" a dog — frightening, unnecessary, and built on the "alpha" or pack-leader model of how dogs relate to people, which the science abandoned decades ago. Your dog is not staging a coup.
• Shouting, hitting, kneeing a jumping dog, squirt bottles, shaker cans, throwing things — all work, briefly, by startling your dog, and all cost you exactly what you least want to spend.
• Forcing a frightened dog to face the thing until they stop reacting. See Flooding and Learned Helplessness.

What to use instead

For pulling: a harness and a loose leash plan. For jumping: pay four feet on the floor, and manage the door. For barking at the window: close the curtain and teach a different job. In general, management so the behavior gets no practice, plus differential reinforcement so a better option pays well. It is slower to start and enormously faster to finish, and nothing you build this way has to be maintained with fear.

If you are currently using one of these tools, you are not a bad owner. Almost everyone who does was told it was the responsible choice, usually by someone confident. The way out is not guilt, it is a plan.

And this is not simply the house style. Our Terms of Service say it outright: we "do not recommend, endorse, or instruct in the use of aversive equipment, including prong collars, choke chains, slip leads used as correction tools, or electronic collars," nothing an AI feature says in answer to a question about them "should be construed as guidance on its use," and plans that recommend aversive equipment or methods can be hidden or removed. If you ask Coach how to use one, it will decline and offer you the reward-based version of the same problem. That is the product working, not Coach being difficult.

Yes, sometimes. That is why this question deserves a real answer rather than a slogan, and why the people who sell these collars are not wrong about everything. If you have watched one stop a behavior in an afternoon, you saw what you saw, and an answer that starts by telling you otherwise is no use to you.

Punishment works. It reduces behavior, and it often does it fast. Any trainer who tells you it never works is being sloppy. What follows is not a denial of that. It is what happens next.

"But it works every time"

This is the most common thing owners say in defence of a remote collar, and it is worth looking at closely, because it is not the endorsement it sounds like.

If the collar had changed why your dog chases, you would need it less as the weeks went by. A month in you would be reaching for it occasionally, then rarely, then not at all. A tool that is still needed every single time has taught your dog something narrow, and not what you wanted. It has taught them that the behavior hurts while the collar is on, so the collar itself has become the signal, which is why the classic report is a dog who is flawless in the collar and gone the moment it comes off. The reason to chase is not changing. The chase is being held down, and you are the one holding it down.

"The lowest setting" is not the reassurance it sounds like either. The reason a low level interrupts a dog mid-chase is that it is unpleasant enough to override what they were doing. If it did not feel like something your dog wanted to stop, it would not stop them. Low does not mean neutral. It means efficient.

The conditions, and why a house cannot meet them

Behavior science is clear about what punishment needs in order to suppress a behavior effectively. Read the list as four ways it comes apart, because that is what it is:

• One chase that gets through still pays. It has to land every single time, because a chase that is only sometimes punished is still being reinforced the rest of the time, and an intermittent payoff makes a behavior especially hard to shift.
• A second late and you have punished the recall instead, or the other dog who happened to be beside them, or nothing at all. It has to arrive while the behavior is happening, not afterwards.
• "Just the lowest setting" never stays the lowest setting. Escalation is built in: start low and climb and your dog habituates as you go, so you end up needing far more than you would have at the start.
• Tell a dog what not to do and you have closed the only door they had. Punishment carries no information about what to do instead, so if chasing was the best thing in their life, nothing has replaced it.

Those conditions come out of a laboratory. European veterinary behavior specialists make the same point about ordinary households with ordinary reflexes: people almost never meet them (Masson et al., 2018). Which gives you the shortest version of the whole argument: the skill it would take to punish well is the same skill that makes punishing unnecessary.

Now the situation people most often buy a collar for: a field, forty yards out, behind a bolting rabbit.

A shock can interrupt a chase that is already running. The question is what happens afterwards, and there the evidence is thin and mixed. In the one small controlled study to test chasing head on, e-collar training stopped six dogs chasing a lure within a session or two (Johnson and Wynne, 2024). What happened next is unclear: the paper's summary says two thirds of those dogs chased again in an unfamiliar arena, while its detailed results describe four of the six as not chasing there. With six dogs either way, it settles little.

The older field work is small and mixed too. Of thirteen dogs trained off kiwi, only one avoided them in a new place, and hunting dogs shocked near sheep were no different from unshocked dogs when they met a lone sheep somewhere new (both cited in Blackwell et al., 2012, whose survey of 3,897 owners also found lower reported success for recall and chasing with e-collars than with reward-based training). Against that, a later New Zealand study found kiwi avoidance that did carry over to a new place, without the collar on, in 48 of 55 dogs a year later, though the dogs were tested with stuffed and frozen kiwi rather than live birds, and the authors say its value against real kiwi is not yet established (Dale et al., 2013). And when industry-approved professionals trained real referred dogs, most of them referred for chasing, the collar gave no advantage over reward-based training, while the e-collar dogs showed more tension and yawning and vocalised in proportion to the level used (Cooper et al., 2014; China, Mills and Cooper, 2020). Neither of those trials tested the dogs with the collar off, so neither settles durability on its own.

So the honest summary is narrower than the sales pitch, and narrower than the easy version of our own argument. E-collars can suppress chasing, sometimes for a long time. But in the trials that compared methods head on, the collar gave no advantage over reward-based training, and it came with stress the reward-based dogs did not show.

Why it comes apart is not a mystery, though this part is learning theory and the settled view of behavior professionals rather than something measured in dogs on a hillside. You would have to catch every chase, and a dog who hunts out of sight guarantees you will not, with every miss paid at jackpot rates. A chase is a chain rather than a single act, so by the time you can see it and press, you are punishing a link whose predecessors were just reinforced again. Arousal blunts pain, so the level that registered in the garden is not the level that registers in the field, which is where the climbing starts. And none of it removes the wanting, so the motivation is still there on the day the collar is not.

One honest caveat, because it cuts against the easy version of our own argument: in that lure study the reward-based protocol did not stop the chasing either. Chasing is genuinely hard, and no method here is quick. The route the evidence supports is the unglamorous one below.

The cost you do not see coming

A shock does not only follow your dog's behavior. It is also paired with whatever is in front of them at the moment it lands, which out on a walk might be the rabbit, or might be the other dog, the horse, or the child on the path. This is how dogs become frightened of things that were never a problem, and it can surface much later, looking unrelated.

The rest of the evidence is in Positive Punishment and Gear and Methods to Avoid: more stress behavior and higher stress hormones, no better obedience in a controlled comparison against professional e-collar trainers working to best practice, and confrontational methods drawing aggressive responses at striking rates.

"I only had to shock him once, and now I just use the beep"

This is the most common thing people say about an electric fence, and it is almost always offered as reassurance. It is worth taking apart, because it describes something worse than it sounds.

The beep works because it predicts the shock. That is the entire mechanism, and there is nothing else it could be. Your dog is not being handed information and choosing to respect a boundary; your dog is being warned, and stopping in order to avoid something that hurt. The tone has become the thing: it now brings up a learned version of the same fear.

And "only once" is not evidence that it was gentle. It is evidence of the opposite. An event a dog learns from in a single trial was usually a big one. Learning from a single experience is how frightening events get stored, not mild ones, and once a dog has learned to avoid something, the avoidance keeps them from ever finding out whether it would happen again. Someone telling you their dog only needed it once has usually told you how much it hurt.

Two things follow that owners rarely see coming. A warning that never gets backed up can keep working for a long time, and keep the fear going with it, because a dog who avoids the line never finds out the shock has stopped. When it does start to fade, the fix everybody reaches for is another shock, which is why "just the beep" tends to have a shock behind it every few months. And a beep is a high electronic tone much like a microwave, a reversing van, a watch alarm or a smoke detector with a flat battery, so the fear can spread to sounds you never chose and cannot switch off.

The rest of the trouble with boundary fences is in Gear and Methods to Avoid: a dog in full drive runs through the line and is then outside with a collar that shocks him for coming back, and nothing at all is stopped from coming in.

The false choice

Most people asking this have been handed a false choice — two bad options offered as though they were the only two. Shock the dog or crate him all day. Shock the dog or lose him to a road. Shock the dog or give him up. If that is genuinely the comparison, of course the collar wins. It is almost never the real comparison, and the useful conversation is the one behind it — the fence you cannot afford, the week you have to get through, the hour a day that nobody has. Those have answers, and they are usually cheaper than the collar.

What actually produces control is unglamorous: an environment where failing is not available while you build, and a long reinforcement history for the thing you do want. Management buys the time, reinforcement does the teaching, and neither has to be maintained with fear. See Management and Differential Reinforcement.

If your dog will not work for food or toys, you are not out of currency. For a hunting-bred dog the paycheck is permission to hunt: recall, mark, and release them straight back to the cover, so coming to you becomes the thing that starts the fun instead of the thing that ends it. A chase paid for with a chase is the only currency that competes.

In Push Drop Stick

Coach will explain why an aversive tool fails and will never explain how to use one. If you are using one now, say so. Nothing happens except that plans get chosen more carefully. See My Dog Already Wore a Shock or Prong Collar.

Often confused with

Reliability. A dog who performs in the collar is demonstrating that the collar is on, not that the behavior is trained, and the test of a trained behavior is what happens when the equipment is off.

Suppression is not reliability, and the difference can cut the other way. A dog who has learned that pain arrives unpredictably outdoors can become harder to predict, not easier, because fear has been added to the situation without anything being taught. And a dog whose growling and stiffening were punished out has lost the signals you relied on to read them, which is the most expensive kind of unreliability there is. See My Dog Already Wore a Shock or Prong Collar.

A basket muzzle prevents biting while letting your dog pant, drink and take treats through it. Conditioned properly, it is one of the kindest pieces of equipment there is.

It belongs immediately after the section above as a deliberate contrast, because a muzzle is not a punishment device and it does nothing to your dog. It is a seatbelt. And like a seatbelt, its real function is to expand what is possible: a dog who is comfortable in a muzzle can go to the vet without being pinned, can be walked in more places, can meet a wider world with a safety net under everyone.

Basket muzzles only for anything longer than a moment. Fabric sleeve muzzles hold the jaw shut, which means your dog cannot pant, and a dog who cannot pant cannot cool themselves. They have a narrow place in brief veterinary handling and no place on a walk.

Condition it long before you need it, and treat it as a training project rather than an item you put on. Food smeared inside the basket, your dog putting their own nose in, one second, off again, over days and weeks. The goal is a dog who pushes their face into it when they see it, and that is genuinely achievable for most dogs. See Start Button for the same idea applied to handling.

Two more things. Every dog benefits from muzzle training, not only dogs with a bite history: injuries, emergencies and unfamiliar vets happen to friendly dogs too, and the worst time to introduce a muzzle is during a crisis. And a muzzled dog in public is a responsibly handled dog. The stigma is backwards. That handler is doing the most considerate thing available to them.

One honesty note: there is very little published research on muzzle training or on how muzzled dogs are perceived. Everything in this entry is the settled practice of behavior and veterinary professionals, and it is close to unanimous, but it is experience rather than evidence.

If you need a muzzle because your dog has bitten or has threatened to, please also get a certified positive-reinforcement professional or a veterinary behaviorist involved in person. The muzzle keeps everyone safe. It does not address why the biting happened.

A small box that makes a distinct click, used as a marker: it tells your dog the precise instant they got it right.

The proposed advantage over a word is consistency and clarity. It sounds the same every time regardless of your mood, it does not resemble anything else in your dog's day, and it cuts through background noise. The disadvantage is that it is one more object in hands that are already holding a leash, a treat pouch and a phone. And to be fair to the word: studies comparing them are mixed, and the controlled ones have mostly found little or no difference in how fast dogs learn, so this is largely a matter of preference.

You do not need one. A marker word works, and "Yes!" is always in your pocket. Choose whichever you will actually use consistently, and if you start with a clicker, condition it first: click, treat, twenty or thirty times, nothing else asked.

Do not click at your dog to get their attention, and never click near a sound-sensitive dog without checking how they feel about it first. Some dogs find it startling, in which case a soft tongue click or a quiet word is kinder and works just as well.

Food is usually the most practical and versatile reinforcer in animal training, and using it is not a shortcut.

Most of the behaviors we want are not intrinsically rewarding to a dog. Sitting still while a stranger walks past is not fun. This is similar to you requiring payment to go to work: no matter how much you love your job and respect your boss, you will not work forty hours a week for free, or even for praise. Potent consequences are what change behavior. Besides, you have to feed your dog every day anyway. You can put it all in a bowl for free, or you can set part of it aside and let your dog earn it.

Treat values

The good stuff matters more as the work gets harder. Low value is fine at home; a new distraction or a big ask needs better. The best training treats are small and quick to eat, so you can keep the rate of reinforcement high.

High value: cooked meat, freeze-dried treats, canned meat, cheese
Medium value: dehydrated treats, food rolls, soft and chewy training treats, peanut butter
Low value: kibble, biscuits, fruit and vegetables

These are a starting point, not a ranking. Your dog's opinion is the only one that counts, so audition things and write down what actually works. And remember the ladder shifts: a treat that is high value in your kitchen can drop to low value in a car park full of dogs. See Motivating Operation.

Adjust meals to account for training food, keep any allergies in mind, and skip anything your dog has to chew for a while.

FAQs

My dog won't take treats. What's wrong?

Start here, because the reason matters enormously:

• Your dog is frightened. This is the most important possibility and the most often missed. A dog who will not eat is telling you they are over their threshold, and no treat is good enough to fix that. Do not go looking for better cheese; add distance, lower the intensity, or stop. It also works in reverse, and this is worth memorising: a dog who IS eating is not necessarily comfortable. Eating tells you arousal is low enough to eat, and nothing more. Watch the body. See Stress Signals.
• Your dog is too excited or too distracted. Same answer, less urgent: more distance, less going on, a quieter start.
• Your dog isn't hungry. Train before meals, or train using the meal.
• Your dog doesn't rate the treats. Try something better. Most dogs who "aren't food motivated" are dogs who have been offered kibble in a difficult situation.
• Your dog may be unwell. A dog who suddenly stops eating food they normally love deserves a vet, not a training solution.

Whether you have to use food at all, and whether you will always need it, are answered in the FAQ.

A food toy designed to be stuffed, smeared and usually frozen, so your dog has to work to get the food out.

Dogs descend from animals who worked for every bite, and a bowl of food is over in ninety seconds. A stuffed toy turns the same calories into twenty or forty minutes of licking and problem-solving, which is genuinely tiring in the way physical exercise often is not. It is also the single easiest tool for the practical problems of daily life: crate training, guests arriving, time alone, settling in the evening, and giving a bored dog something legal to do.

Difficulty

Match it to your dog's experience and persistence. Do they keep at things, or give up quickly?

Easier: lick mats, puzzle bowls, and toys with large openings such as a Toppl or Tux; or small treats in a Kong that fall out readily. Serve unfrozen.
Harder: Kongs and toys with small openings, packed tight and served frozen.

Start easier than you think you need. A dog who succeeds at an easy one comes back to a harder one.

Stuffing

Anything your dog loves, and vary it, because most dogs like novelty. Common options: canned tripe (deeply unpleasant, wildly popular), pate-style dog food, raw food, peanut butter, cream cheese, apple sauce, banana, bone broth, soaked kibble, or simply your dog's daily meals.

The cherry on top: something sticking out of the opening gets a hesitant dog or a puppy started. Beef lung, jerky, a bully stick, sweet potato chew, a sardine, cheese, or just overfilling it.

Cleanliness

Most dogs finish a frozen toy before it thaws enough to leak, but expect drool and a small mess. Feed them in a crate, on a mat, or outside.