Snake Avoidance: A Positive Reinforcement Approach

Snake-avoidance training can be a lifesaver for dogs that live among venomous snakes. A common way to train snake avoidance is to pair a strong aversive, like an electric shock, with the appearance of a snake to instill in the dog a fear of snakes. People who live in areas where venomous snakes are prevalent have legitimate concerns about the safety of their dogs. A positive alternative has eluded many trainers.

I have been working on a positive reinforcement protocol for teaching snake avoidance to dogs since 2009. This year, prior to Covid-19, I completed the first phase of a project designed to test the effectiveness of this protocol. The trainers of the Southwestern Desert Dog Club in Nevada helped me with this testing phase.

Project goals

We had four primary goals:

• The dog should avoid snakes of any kind. Many venomous snakes do not have a defensive sound like a rattle. So, the dogs should respond to visual, olfactory, and auditory cues.
• When the owner is not present, there must be a way to reinforce the dog or to maintain the behavior.
• Once a snake is identified, dogs must be prevented from going out again and reencountering the venomous snake.
• The focus must be on the use of positive reinforcement techniques.

Preparations before training begins

The protocol I developed needed to address all four of the stated goals, as they are all challenges dog owners face. There are several separate components that must be in place before snake-avoidance training can begin:

• A well-trained recall.
• A well-trained target behavior. The dog should be able to push hard on a stationary target.
• A well-established kennel behavior. This is another key component, particularly if you need to isolate the animal to prevent him from going back to revisit the snake after the recall behavior. Like the recall, the kennel behavior should already be a well-trained and reliable behavior.
• A self-locking gate latch for your kennel. This is only necessary when you want to prevent the dog from going back to revisit the snake after alerting to its presence. There are several self-locking gate latches available at Home Depot, Lowes, or other hardware stores. If your dog is an escape artist that knows how to open a self-locking gate latch, there are more expensive devices that lock securely and require a key to open.
• A remote feeding device. If you want to reinforce the dog for responding to a snake in your absence, some type of remote feeder is required. A Pet Tutor can be rigged to be triggered by the press of a target, or a weighted lever can be used so that when it is pushed by the dog it drops reinforcers into the kennel. The same lever can be designed to trigger the door of the kennel to close and lock.
• Rubber snakes of various sizes and colors that can be used as visual cues for the presence of snakes. Realistic-looking snakes are available cheaply online. Fishing line or other thin string can be used to pull the artificial snakes and simulate snake movement.
• Canisters of snake odor. We used snake skins (that were shed), snake defecation and urine, and dead snakes. We worked with a local zoo to obtain most of this material and focused on snakes native to the region where the dogs lived. (Our study focused on Sidewinder, Mohave, Speckled, and Western Diamondback rattlesnakes.)

The dogs

I had lined up 18 dogs to test the protocol in Nevada. Although all the handlers and trainers knew the requirements, five of the dogs did not have a truly reliable recall or kennel behavior. The two behaviors must be well-established and reliable before you can begin to use this protocol. Two other handlers did not have the right mindset and dedication to commit to the protocol, so their dogs were eliminated from the test as well. For the actual training, we were down to 11 dogs. These dogs and their handlers were committed to testing the protocol. We were successful training snake avoidance with ten out of the eleven participating dogs. The one failure was with a dog that feared the type of kennel we used for the project. Unfortunately, we did not have time to switch out the kennel during the three days we had allocated for the training.

The ten dogs that completed the project included six males and four females. Breeds included three “purebred” dogs: a Belgian Malinois and two border collies; three dogs bred by their owners: an Australian shepherd mix, a rottweiler, and a springer spaniel; and four shelter rescues: two pit-bull mixes, a Chihuahua mix, and a terrier mix. I had used the protocol successfully with four different dogs previously, but that version of the protocol was less structured and not fully developed. It became the proving ground for the protocol; the early tests were with two Labrador retrievers, a pit bull, and a rat terrier.

The training plan

We had a dedicated and motivated group of trainers who agreed to follow the protocol precisely. I was able to implement and oversee every step of the training program for the first three days. I then monitored and advised the team on progress for the next three months. The protocol we followed is outlined below.

1. Combine the kennel behavior with the push-the-target behavior to dispense the reinforcer. When the dog enters the kennel, he pushes the target, and gets reinforced. We rigged the target by imbedding a remote control inside the target. When pushed, it dispensed a huge, high-value food reinforcer, simultaneously triggering the kennel gate to close. Get the dog comfortable with the kennel closing and locking behind him after he enters the kennel.
2. Combine the recall and the kennel/target behavior into a reliable linked sequence. Sound the recall. As the animal returns to the trainer, cue the kennel behavior, which leads to the target behavior that triggers the reinforcer. Do not move to the next step until this sequence is very reliable on just the recall cue. With animals that already have reliable recall, kennel, and target behaviors, this step should be established easily and quickly.
3. Transfer this behavior sequence to outdoor locations where snake encounters would be realistic. An outdoor trail would be ideal. Schedule training so that there are likely to be minimal outside distractions at first (no other people or dogs). All remaining steps should be trained in real-world environments. We were able to establish the first three steps of the protocol with nine of the ten dogs in our test in one day of intensive training. The tenth dog completed the third step on day two.
4. Use a cue-transfer procedure to switch the dog’s current recall cue to a snake movement visual cue. This requires rigging the fake snakes on a fishing line, preferably on a long line so that the person who is controlling the movement of that snake is well out of the dog’s sight. As the dog approaches, the snake is made to move across the dog’s path. This takes practice, and I suggest practicing without the dog numerous times until the movement of the snake looks realistic. Once the person on the line is ready, bring the dog to the trail and allow him to explore, search, and smell the environment for a while. Don’t cue the recall too early; dogs will respond better to a recall if they have had plenty of time to get used to the new environment. After five or ten minutes, lead the dog to the vicinity of the fake snake. As the dog gets within visual range of the snake, start making the snake move. Immediately after the dog notices the snake, sound the very reliable recall cue. Over the next few days, practice this sequence several times in new places and after varied lengths of time on the trail. It usually takes four or five repetitions for the dog to respond reliably to the snake’s movement as the new cue for a recall.
5. You want the dog to recognize that any visual appearance of the snake is a cue to recall.
   Introduce variety in the snake’s appearance as soon as possible, using different sizes of snakes and assorted colors. Varying speeds of movement are important as well. You want the dog to recognize that any visual appearance of the snake is a cue to recall. Seven of the dogs in our test group learned the visual cue with five repetitions on day two. Two dogs learned the cue with six repetitions, and one quick learner understood the new cue after four repetitions.
6. Once the visual cue is well-established, begin using the cue-transfer procedure again with canisters of odor. The goal is for the dog to respond to the smell of a snake as another recall cue. As soon as the dog sniffs a canister with snake odor, sound the original recall cue. Place lots of random canisters in the field that do not have snake odor in them. Only one canister should contain snake odor at each trial. It may be necessary to guide the dog toward the target canister if he does not approach it on his own. When the dog approaches the canister with odor, watch closely and note when the dog notices and smells the canister (give him a few seconds to really smell it), then immediately sound his original recall cue.
7. Begin to vary types of snake odor, sometimes using snake skin, other times fecal/urine samples, and other times a fresh carcass. Work at this step until the dog responds to the odor with a strong recall response regularly. For our test group of dogs, four dogs required three repetitions, two dogs needed four repetitions, three dogs required five repetitions, and one dog took seven repetitions to respond reliably to the odor as a cue. It is worth noting that the same dog always took longest to achieve each step, which appeared to be related to his learning history.
8. Begin to intersperse outings with visual cues and outings with odor cues with outings that do not have any snake-cued recall associated with them.
9. During initial training, the dog will be sensitized to the idea that recall behaviors are being cued. This will increase the likelihood of success at the early stages. As the training proceeds, it will be important to introduce more and more outings that do not include a recall. This is a critical step that will ensure that there will be no false alerts as the training progresses.
10. Continue to practice with different and varied versions of the snake-triggered recall so that the dog responds strongly to visual and olfactory snake cues in various versions. Continue to generalize the animal to snake cues that are never quite the same as previous versions of the cue. With our experimental group, we paired rattle sounds with some of the visual cues and with some of the odor cues, since these dogs had a high probability of encountering rattlesnakes. If a dog did not respond to the visual or scent cues right away, we were prepared to pair these cues with the dog’s original recall as a reminder. We only needed to do this with two of the dogs and the reminder was only used on one trial with each dog. We discovered that the rattle was unnecessary, as the odor and visual cues were so strong. However, we felt that the rattle was a critical distraction that the dogs needed to get used to so that they were not thrown off if they encountered a real rattlesnake.
11. Continue to use the normal recall from time to time.
12. Give access to the kennel and target when they are not rigged to reinforce, so that the dog has the opportunity to discover that kenneling and touching the target on his own, when he is not being cued by the presence of a snake, results in no reinforcement. This is an important step that is separate from the initial training. It is essential to completing the snake-avoidance behavior successfully, as it reduces the likelihood of false alerts.

Follow-up

For the more than six months after the initial training, all ten trainers continued to follow the protocol and test it regularly, and all of the dogs’ responses remained strong. During that time, six of the dogs had a total of nine actual snake encounters and all responded immediately with a strong recall. I have consulted with several other clients who used the same protocol for their dogs. They also report successful results. The only unwanted side effect is that most dogs do not discriminate between venomous and non-venomous snakes. The training protocol does not differentiate between them, so dogs respond to all snakes the same way. But that is a minor nuisance that owners who have dealt with dogs bitten by snakes are happy to deal with.

Positive alternatives work

I hope more trainers will try this protocol. It is not difficult to follow. At its center, it is a strong recall that is cued by seeing or smelling snakes.

Most importantly, this protocol demonstrates that snake avoidance can be trained positively. It is clear why many people are drawn to aversive tools. The draw comes from a desperate desire to protect their pets, and many trainers searching for a solution are simply not aware that positive options exist. We still have a long way to go to help the average pet owner recognize that the best solutions are not found by concentrating on the unwanted behavior, but by focusing on a desirable or acceptable alternative. This snake-avoidance protocol is an example of using positive reinforcement in a creative way to deal with a serious problem that dog owners who live among venomous snakes often face. It is worth the extra effort to search for reinforceable options to problem behavior. These solutions are not always easy to find, but as creative trainers we owe it to our dogs to keep searching.

Happy Training,

Ken