Nosework

Below is an excerpt from the U.S. Military Dog Training Handbook by Department of Defense (DoD)

DFR BACKGROUND

This chapter discusses the DFR method for training substance detector dogs that is now in use in the MWD course at 341 TRS. Because Deferred Final-trained dogs are taught to look at source when indicating odor, rather than at their handlers, this method is frequently referred to as ―focus training.‖ This paper is provided to assist MWD users in understanding and trouble-shooting DFR Dogs.

In order to become proficient in substance detection, the dog must acquire two main conditioned associations, or lessons. First, the dog must learn to recognize the target odor, meaning that it must begin to expect reward when it smells that odor. Second, the dog must learn that, in order to receive the reward, it has to sit. This sit is, of course, called the ―final response‖ in DoD terminology. 

REWARD NOT FROM SOURCE METHOD

The traditional DoD method for teaching odor recognition and final response was called Reward Not From Source (or Reward NFS). The dog is encouraged to investigate an area or a scent box, commanded/helped to sit, and then given a reward (ball/kong). Over the course of many trials the dog learns to sniff a series of locations, and sit ―on‖ the location that smells of target odor, without sitting on any locations that do not smell of target odor. Although it appears procedurally simple and therefore practical for students and non-experts, the Reward NFS method in reality relies upon great experience and skill on the part of the trainers. Furthermore, Reward NFS often ―builds in‖ to the dog undesirable behaviors which present long-term challenges to trainers and users. This is because the method:

• Attempts to teach the dog odor recognition and final response simultaneously.
• Makes no attempt to associate reward with odor source. 

These two factors interact to bring the dog‘s attention to its handler during the crucial period when it should be learning to pay attention to odor. The process by which this learning occurs is as follows: During the initial trials, of course, the target odor has no ―meaning‖ for the dog. In Reward NFS the animal is persuaded to more or less accidentally sniff the odor, normally by having the handler use hand presentations to get the dog to place its nose near the training aid. Then, hoping that the dog has ―noticed‖ the odor, the handler manipulates the dog into a sit position, most often with a combination of verbal commands and physical prompts such as leash tugs and/or pressure on the hindquarters. In the course of this intervention, the dog‘s attention becomes focused on the handler and, once in the sit position, this attention to the handler is rewarded by presentation of the reward.

Because the dog is rewarded while looking at the handler rather than while sniffing odor, the dog is very slow to learn that odor predicts reward (often requiring more than 150 trials). In addition, it learns that handler behavior also predicts reward. It begins to watch its handler closely during detection problems; it tends to rely upon presentations to make it sniff; and it learns much about the cues (stutter steps, etc.) a handler gives when a training aid is nearby. As a result, the dog often develops a false response tendency based on handler cues, and very much time and effort is expended in trying (often unsuccessfully) to ―work out the cues.

In addition to slow acquisition of odor recognition and high false response tendency, the most common deficiencies in dogs trained with Reward NFS are handler dependence/lack of independent search behavior, weak change of behavior in response to odor (the dogs tend to simply sit when encountering odor, making it difficult for the handler to discriminate between false responses and ―hits‖), and poor localization (many Reward NFS dogs have a ―fringing‖ tendency). All of these weaknesses are consequences of a system that encourages the dog to look to its‘ handler for reward and guidance early in training, before the dog‘s behavior has come under strong control by target odor.  

DEFERRED FINAL RESPONSE METHOD

The DFR method differs from the Reward NFS method in three critical ways:

• First, during initial training of odor recognition, the final response is not required. Instead we defer teaching of the final response until later, when the dog has mastered some critical pieces of learning. On initial trials, the dog is prompted to search a given area, normally by pretending to hide the reward in that area. When the dog investigates, sniffing for the reward, and sniffs target odor instead, then the reward is provided. Thus we provide the dog with a very ―clean‖ and simple pairing of target odor and reward, without any cues from the handler to compete with odor for the dog‘s attention. As a result, the dog learns to recognize odor extremely quickly. Dogs trained with DFR normally exhibit strong changes of behavior in response to target odor after 5 or 6 trials.
• Second, DFR is a ―reward-from-source‖ method, meaning the dog is taught that reward originates from odor source. There are many ways to provide reward from source, including simply placing the reward with the odor, but in DFR we rarely place the reward with the training aid. Instead, we ―lob‖ the reward in from behind the dog when it is sniffing odor, so that the reward appears without any warning and falls as softly as possible directly on odor source. 
• Third, during early stages of DFR training, the role of the handler in making presentations and guiding the dog‘s search is greatly reduced. In fact, the handler gives the dog as little information as possible. The dog searches off-leash in a confined area, or works on leash with an absolute minimum of influence from the handler. We are not concerned if the dog ―walks‖ the training aid. We operate from the principle that the worst thing the handler can do in detector dog training is show the dog where the aid is by stopping the animal on the aid. It is the dog‘s job to stop the handler on the training aid, and if it fails to do so it learns two important lessons: 
•       No one will help it to find odor  
•       Leaving odor is a mistake to be avoided, because it results in more work and greater delay to reinforcement 

OVERVIEW OF THE DFR TRAINING SEQUENCE

The first step is to teach odor recognition by ―paying on sniff‘ without any requirement for final response. The dog is not paid unless it is clear that it is sniffing and that it has ―noticed‖ the smell of the training aid. Normally within 1 day of training and no more than 10 trials the dog exhibits vigorous changes of behavior when encountering target odor. These changes include bracketing to source, perhaps ―freezing‖ behavior, and sometimes scratching or biting (i.e. ―aggression‖) at the source. Aggression is not desirable and efforts are made to control it, but all recent experience in the Specialized Search Dog and other dog Courses indicate that, in a focus-trained dog, a degree of aggressive responding early in training is common and does not necessarily indicate that the animal will develop into a persistent ―aggressive responder.

During the first 6 to 8 days of training and perhaps 30 to 50 trials, the dog works on one odor, with a minimum of handler presentations and interference, with ―pay on sniff‖ rather than final response. Efforts are made to encourage the dog to look/focus/point at odor source, because when the dog focuses on odor it is not observing its handler and learning about handler cues. The dog is desensitized to people in the search area, and taught to ignore physical contact and interference from personnel while it is searching. At the same time, it is taught to focus on odor for 2 or 3 seconds at a time and taught to ―check back‖ to odor on command when focus is broken. 

Simultaneously with odor recognition training, the final response (sit) is pre-trained. The sit is not performed in ―obedience‖ mode, and corrections are not used. Instead the dog is induced to sit with food or a reward. Ideally the dog learns to sit when it is pointing at a reward (such as a kong held by the handler) that it cannot take because it is blocked by a barrier (in this case the handler‘s hand enclosing the reward). The dog is never forced to sit. Instead it is kept interested in the reward until it chooses to sit voluntarily, and then it is rewarded. Initially, the use of a ―Sit‖ command or any other cue to sit is avoided, so that the sit is not dependent on a command. Ideally, the dog perceives a blocked reward and learns to spontaneously sit in order to unblock the reward. Once the dog sits quickly and easily to unblock a reward, then we may attach a cue to the sit, such as a light touch on the dog‘s rump, or a slight tug on the collar, or even the verbal command ―Sit.

 When the dog recognizes target odor, stops on it and stares, can be made to ―check back,‖ and when the pre-trained sit is fluent, then we begin to require the animal to give the final response on odor. This stage is normally reached on day 6 to 10 of training. Initial final response training can be accomplished in a number of different ways, but in the most straightforward method, the aid is placed in a piece of furniture at about nose height. Rather than paying on sniff, the handler cues the dog into the sit using one of the cues (tug on leash from behind, touch on rump, verbal command ―sit‖) that was taught during pre-training of the sit. It is important to realize that this use of the word ―cue‖ is not identical to the common use of the word ―cue‖ in DoD terminology. When we speak of a ―handler cue‖ in Reward NFS, we are referring to a piece of information that the handler gives the dog to help it sit on/find the training aid (normally the ―handler cue‖ is a mistake or an accident). In DFR we never deliberately ―cue the dog to help it find the training aid. The dog must always find the training aid and stop on it without assistance. The ―cue‖ in DFR is merely help from the handler in completing the final response. One way to think of it is if the dog independently detects and localizes the training aid, then it ―earns‖ help with the final response.

The dog is introduced to additional odors, beginning with pay on sniff. Once the dog begins to show recognition of the new odor (normally within three to five trials or even less), then we add the final response to the new odor. The dog also is taught to perform search exercises in new training areas (vehicles, aircraft, etc.). Finally, as one of the last important lessons the dog acquires, the handler for the first time encourages the dog to look to him/her for help in finding the training aid (i.e. the dog is trained to accept guidance in the form of handler presentations of productive areas). 

LIABILITIES OF DFR

Above we reviewed the significant liabilities of the standard DoD (Reward NFS) method. DFR also has liabilities—it can produce certain undesirable behaviors. However, these undesirable behaviors are not as harmful and difficult to correct as the handler dependence, weak change of behavior and localization, and high false response tendency so common in Reward NFS dogs. In a word, the kinds of problems associated with Deferred NFS are the kinds of problems we would rather have. 

DFR-trained dogs can be excessively independent while searching, and may sometimes be reluctant to accept presentations. This shortcoming is offset by the fact that these dogs are often extremely accurate and effective when ―scanning‖ and working independently. 

The most common problem in DFR dogs is a tendency to stop/stand and stare at odor rather than sit. It is worth noting that this is not a functional problem from the standpoint of detecting narcotics or explosives. A distinct change of behavior followed by localization and then a stop and stare is every bit as recognizable to the handler as a sit. In fact, the DoD Specialized Search Dog Program officially authorizes the stop and stare as a final response. 

The first thing to realize about ―stop and stare‖ is that it is the result of physical and psychological tension rather than disobedience. For the DFR dog, an odor source is like a magnet that draws the dog‘s attention. The dog orients to odor in the same way it looks at the reward—with great excitement. As a result the dog carries much more physical tension in its body than a Reward NFS dog. This tension makes the final response more difficult because, in order to sit, the dog must relax major muscle groups in its back and hindquarters. 

The expression ―stop and stare‖ is used to describe a wide range of difficulties with the final response, ranging from an dog that delays its‘ sit for 5 or 6 seconds, to an animal that exhibits a classic ―locked up‖ stop and stare, in which its‘ body goes completely rigid. The classic ―locked up‖ stop and stare is a comparatively rare problem with DFR dogs. More common final response issues are: 

• Slow sit
• Not completely reliable sit (stops/stands and stares sometimes) 
• Refusal to sit unless cued somehow

SLOW OR RELUCTANT FINAL RESPONSE IN DFR DOGS

If the dog‘s sit is slow or delayed, but all that is needed is to wait a few seconds for the final, we advise that you leave well enough alone. There is no compelling reason to pressure the dog for a rapid or ―crisp‖ sit; so long as the animal stops on odor without any help from the handler, will not leave source, and sits on its own without assistance given a few seconds. 

If the dog sometimes does not complete the final response (this may happen when the footing is difficult, or the aid is very low to the ground, or the dog is tired, etc.), then it is important not to over-react. A stop and stare error is not equivalent to critical errors like missing an aid or false responding. Keep in mind that the dog has completed the most important part of the job: it has found and indicated the drug or explosive hide. Now we just need the animal to fulfill the statutory requirement for the final response. Strong verbal corrections (―No!‖) and physical corrections (jerk on choke collar) are normally not helpful because they increase the dog‘s stress and tension, and make it all the more difficult for the animal to relax enough to finish the final response. Likewise simply trying to force the dog‘s hindquarters down to the ground should be avoided, because it generates resistance and makes the dog lock its‘ legs. 

The first corrective action for a dog that stops and stares should be familiar, because it is a fundamental part of handling any DoD-trained detector dog. When the dog stops on the training aid, the handler should keep moving away from the dog and present the next location in the search area. If the dog neither leaves the training aid nor sits, then the handler should put more pressure on the dog with slight leash tension or a tug on the leash. If the dog is not actively working the odor, or sitting, then it needs to move on and actively sniff the next location. This procedure is no different than handling a Reward NFS dog, but still many handlers need to be reminded of it. In this situation, many dogs will react by completing the final because they do not want to be taken away from the aid. Some dogs will allow themselves to be taken away from odor but then they immediately realize this was a mistake. They begin trying to cut back to it, and if they are allowed to do so (on their own—the handler does not ―bounce them back‖ to the aid) they will complete final when they get to source. 

If the above procedure is not effective, and still the dog stops and stares, then the next thing for the handler to try is to simply give a ―Sit!‖ command from behind the dog. Very many animals will comply. In preparation for this procedure, it is a good idea to practice and ―fine tune‖ the dog‘s response to the ―Sit!‘ command during obedience training. 

If the dog does not complete the final response with the verbal command, then the handler can try two other types of assist. First, he/she should use the leash to administer light snaps or tugs on the choke collar, straight back towards the dog‘s hindquarters. If this is not effective, then the handler should try lightly tapping or pushing downwards against the dog‘s rump, just ahead of its tail and behind the hips. The handler should NOT push against the dogs back ahead of the hips, because this normally causes the dog to brace against the pressure. Most dogs know one of these assists for the final, or a combination of them may be most effective, and will respond by executing the sit. Once the dog has completed the final response, then the handler praises with ―Good!‖ and prepares to provide the reward. 

The reward should be given while the dog is focused on odor, from straight behind or some other blind spot, and it should be lobbed in as softly as possible. Ideally, it goes ―dead‖ at odor source, so that the dog can simply pick it up, rather than chase it frantically around the room. SSD trainers and other specialists in DFR sometimes arrange nets or even pillows near the training aid, or use a ―spiked‖ tennis ball or something similar, so that the reward lands at source and stays there without bouncing. If the assist distracted the dog, then focus on the odor source should be re-established prior to reward. 

If the sit is consistently assisted as described above, the dogs final response normally improves, and no other special efforts are required. Sometimes it may happen, however, that the dog becomes dependent upon the assist it will not sit unless it is given some cue. In this case, it is often effective to allow the dog to make repeated finds of the same training aid. This procedure removes the search element from the exercise and allows us to concentrate the dog‘s energy on the final response. Also, because the dog is asked to sit repeatedly in the same location, the animal‘s anticipation works in our favor to produce a ―fluent‖ (relaxed, easy, prompt) sit. The first time the dog finds the aid, it is assisted into the sit and paid. After the reward is recovered from the dog, the animal is allowed to go right back to the aid. On the second indication of the aid, if the dog does not complete the final response, then it is assisted, but the dog is not paid. Instead it is praised for the sit, then gently pulled backwards off of the training aid. (Keep in mind here that a well-trained ―focus‖ dog will not allow itself to be easily taken away from a training aid. Instead it will fight to ―check back‖ to the aid.) Once the dog is pulled away, it is encouraged by praising it, and then allowed to go to the aid again. If the cue is still necessary to produce the sit, then the dog is again praised for the sit verbally and physically, then pulled gently away from the aid, and allowed to go back again. The procedure is repeated as many times as necessary (within reason). When the dog eventually sits without the assist, the handler praises it and pays. Once they understand that they must sit without the assist in order to be paid, most dogs quickly improve their final response. 

A word of caution is necessary here. Unless the dog is well-trained in ―checking back,‖ pulling the animal off of the aid repeatedly can produce some undesirable side-effects. Some dogs may become discouraged or frustrated, and begin to leave the odor source. If a cage is used (after a bit of basic training to establish good focus on the visible aid), then the dog is less likely to leave the visible aid. On the other hand, the dog may not leave the aid, it may instead become more tense and resistant, and begin stopping and staring more. However, the resistance may be specific to the aid that it has encountered again and again. If we move the aid to another location and ―surprise‖ the dog with it, the animal may show one of its‘ best unassisted sits. Similarly, if we have a session one day in which we ―hammer‖ the sit with many repetitions, even though the dog may conclude the session looking as though it has regressed, sometimes we find that on the first aid of the next training session, the dog will give a crisp, perfect final. 

STOP AND STARE IN DFR DOGS

If the dog‘s sit is extremely slow, or unreliable, or if the dog has become very tense, ―locking up‖ and going rigid on the training aid, then we must take a more patient and multipronged approach to improving the final response. Probably the most important part of this approach is making sure that the dog knows how to sit in order to gain access to a blocked reward. This process is referred to as ―pre-training‖ the sit. It was mentioned briefly above, but now we will discuss it in detail. 

First, the dog is taught to sit when the handler holds a reward in the air above its‘ head. If possible, it is better to avoid use of the verbal cue ―sit.‖ Instead just keep the dog‘s attention and let it experiment with various behaviors until it hits the right one. When the dog sits, the handler gives the reward, preferably by simply handing it to the dog calmly or dropping it into the animal‘s mouth. The reward should NOT be vigorously thrown or bounced. With many dogs it is more constructive to begin pre-training the sit with food rather than ball/kong reward, because food produces a more manageable level of drive. In addition, in the process of taking and eating food, the dog can more easily be taught to ―respect‖ (i.e. not bite) its‘ handler‘s hands.

In the next step, the reward is held lower, where the dog can reach it, but the handler ―blocks‖ the reward by enclosing it in his/her hands. The dog can see the reward, smell it, even get nose and teeth on it, but not take it. We use the hands to block the reward rather than a physical barrier because most dogs respect their handler‘s hands— they do not ―aggress‖ a reward that is held in the hands the way they would a reward held in a drawer. In this situation, with the reward just inches away, nearly every dog falls into stop and stare behavior very similar to the behavior we see in detection. The handler does not try to hurry the dog into the sit. Instead he/she provides time for the dog to ―think,‖ just keeping the dog‘s attention on the reward by moving it slightly when necessary. When the dog eventually sits, it receives the reward (i.e. is ―paid‖). 

As time passes, the handler holds the reward lower and lower, trapping it against the legs or body, waiting until the dog sits and then paying the animal calmly. Once the dog is very ―fluent‖ with this behavior, easily and quickly offering a relaxed sit in order to unblock the reward, then the handler begins to hold the reward against the wall or against furniture, teaching the dog to give the sit whether the reward is held six feet up the wall or on the ground. Low placements of the reward, like low placements of a training aid, are particularly difficult for the dog. Note: Blocking the reward by holding it against walls and furniture should be performed sparingly because the dog can confuse this gesture with hand presentations, causing a tendency to sit or ―false respond‖ on hand presentations that will then have to be extinguished. 

Next the handler gives the reward to another party. The handler holds the dog on leash while the assistant shows the dog the blocked reward, and pays the dog for the sit. The assistant begins with the reward held high, but then quickly passes through all the steps of holding it trapped against the body and then legs, and then against a wall or furniture at various heights. 

Eventually, the reward is not held by a person. Instead it is ―trapped‖-- jammed between a piece of furniture and the wall, or between a car door and its frame, etc. The exact situation does not matter, only that the dog can see and smell the reward but not touch it or take it. The moment the dog sits, then the door is pulled open or the furniture pulled away from the wall to liberate the reward. If the dog responds aggressively rather than sitting, then you wait until the dog becomes discouraged, stops scratching/biting, and sits, and then you pay. The most important factor is not whether the dog is allowed to ―aggress‖ a reward or aid—it is whether the dog achieves some result with the aggression. With the vast majority of dogs, if aggression does not produce any result such as opening a drawer or moving furniture or making the handler react, then alternative behavior (i.e. sitting) will quickly take over if it is rewarded. Occasionally, it may be necessary to pay the sit not by liberating the trapped reward, but instead by lobbing another reward in from behind the dog. 

As a last step, the fluent and relaxed sit is placed under the control of some cue. This cue can be a voice command, a tug on the leash, or a touch on the hindquarters, whatever works best with that dog—but the cue must be given from behind the dog while it is facing a blocked reward. The exercise is set up as before, with a reward trapped in a suitable location. The dog watches the reward placed, and then is allowed to move forward on leash until it is stopped by the barrier blocking the reward. The handler gives the cue as the dog is assuming the sit position. This means that the cue does not cause the sit—the sit is voluntary. But by pairing the cue with the voluntary sit, you can gradually give the cue the ability to trigger the sit. Most importantly, the sit we obtain is the relaxed, fluent sit the dog has learned in pre-training, rather than a tense, resistant sit. 

RETEACHING THE FINAL RESPONSE

Once the sit is adequately pre-trained, with a cue that enables the handler to sit the dog from behind when it is facing a blocked reward, then we bring odor into the situation. We choose a location in which the dog has been extensively drilled on sitting to unblock a reward—a ―trap‖ between a cabinet and a wall, or between a car door and its‘ frame, etc. We place both the reward AND a training aid in the trap. We must be careful to place the reward and training aid at a height that facilitates the sit—normally at about nose height. 

The dog associates the particular location of the trap with sitting. It also associates a blocked reward with sitting. These conditioned associations help to overcome the dog‘s tendency to respond to odor by freezing and staring. 

The dog is allowed to see the reward and the training aid being placed in the trap. When the dog approaches and sniffs at the reward, it also catches target odor. Depending on the individual dog and the preferences of the trainers, we can then do one of two things. 

• Wait until the dog sits. We wait as long as necessary without giving the dog any input or cue. We watch for the sit and ignore anything else that the dog does, except that, when the dog becomes so frustrated it leaves odor, the handler may entice it back to the trap again. This process may take a very long time (up to 5 or 10 minutes) and it may get ―ugly,‖ in the sense that the dog may stand and stare for minute after minute, or begin looking around, even leave odor and return, or it may begin scratching or biting at the trap. However, nothing produces so much learning as a problem independently solved, and if we allow the dog to choose the sit in its own time and in its own way, just a few trials can result in dramatically less stopping and staring and a faster and faster final response. 
• Once the dog sits, then it is rewarded-- not by unblocking the trapped reward but instead by lobbing another reward in from behind. After a bit of rewardplay, then the dog is immediately sent back to the exact same trap to practice the final response again. In fact, the dog is sent back to this same trap several times in a row. This is not a search exercise, it is a final response exercise, and knowing the location of the aid only helps the dog. 

For the second session the exercise can be moved to another location, so that the dog has to search a little before finding the trap. This will revive the animal‘s sniffing behavior so that on the first trial or two we can be sure that it is smelling the training aid (and reward), rather than using eyes alone.

Cue the sit. If the dog‘s sit (to unblock a reward) is controlled by a cue, then the trainers can also cue the sit rather than wait for it. This method has the advantage that it results in a faster sit and less ―ugly‖ behavior like leaving odor or aggressing, but it has the disadvantage that the cue will eventually have to be extinguished, which may involve a surprising amount of repetition and work. 

As before, the dog is sent to a trap containing a reward and a training aid. The moment the dog sniffs, it is cued by the handler into the sit. The handler does not wait to see if the dog will stop and stare; he/she immediately cues the sit without any pause, almost hurrying the dog into the sit. Also, once the handler begins to ―ask‖ for the sit with the cue, they make sure that the sit is obtained, even if in the process the dog loses focus on the training aid. When the dog completes the sit a reward is thrown in so that it lands at odor source (training aid odor + trapped reward odor). 

This exercise is repeated many times, sometimes leaving the trap in a familiar location so that associations with that location help the dog to sit, but moving it often enough so that sniffing behavior is from time to time re-established. Every now and again, the handler tests the dog by waiting a moment to see if the dog begins to sit before being cued. Then for the next few trials, the handler returns to immediately cuing the sit again without waiting to see what the dog will do on its own. Once we find that during the test trials the dog has begun to reliably initiate the sit without waiting for the cue, then we move to the next step. 

Throughout the trials described above, the dog was paid once it reached sit position whether a cue was necessary or not. Now we raise the criterion—if a cue is required to obtain the sit, we will not pay the dog. The dog is allowed to find the reward/training aid, and the handler watches to see if the dog gives the final. If not, the handler cues the sit, then praises the dog verbally and perhaps physically while in the sit, then pulls the dog by the leash away from the aid 5 to 10 feet, and then allows the dog to go back to the trap. This procedure is repeated again and again, cuing the sit if necessary but never paying the dog for the sit if a cue was used. When the dog finally sits on its own without the cue, we give the reward. 

Eventually, once the dog is very fluent with the sit when presented with reward and training aid together in a trap, then we omit the reward and present the training aid alone. We follow the rules described in the last paragraph, giving the dog only praise and encouragement if it requires the cue in order to sit, but giving the reward if the dog sits without a cue. 

WAITING THE DOG OUT VS CUING THE SIT

Choice of strategy is guided by trainer inclinations and dog characteristics. If the dog has a very rigid, ―locked up‖ stop and stare, but will not leave the aid and does not tend to aggress, then the waiting approach is often best. Attempting to cue such a dog often stimulates more tension and resistance. Instead the animal just has to be given the time to learn that stop and stare does not produce reward, but sitting does. It is not unknown to wait 10 or 15 minutes.  

 If the dog has a very prompt, fluent, unresisting sit when cued, and if the dog is not extremely tense when indicating a training aid, then the cued approach is likely to be quick and effective, and has the additional advantage that it will prevent the dog from aggressing the aid, leaving it, looking around, etc.

THE USE OF A CAGE TO CONTAIN THE TRAINING AID

Rather than the trap, DoD‘s Specialized Search Dog Course often places the training aid inside a sturdy ―cage‖ of some sort that allows the dog to see the training aid. Although being able to see the aid may decrease sniffing somewhat, visual access to the aid helps to establish and maintain focus. This is helpful because maintaining focus can be very difficult while re-training the final response, because we may leave the dog ―on the aid‖ for long periods of time while we wait for the sit, or subject the dog to strong physical influences in order to cue the sit. 

The addition of visual information to the situation can have a beneficial effect on many dogs that have difficult habits when working ―hidden‖ training aids. Dogs that are very aggressive to aids in other situations fall into staring behavior when they are caged; dogs that walk away from hidden aids in frustration stay with them when they are visible inside a cage. 

Initially the cage is placed on the ground in an empty corner. A few introductory cage trials are performed to establish the dog‘s focus in this unique situation. The dog is paid ―on sniff,‖ which may seem puzzling in view of the fact that the goal is to cure a stop and stare problem. However, a few ―pay on sniff‖ trials will help us establish focus on the caged aid, and they won‘t make the stop and stare problem any worse than it already is. 

An assistant shows the dog the reward and pretends to place the reward on/in the cage. The dog is released from a distance of 15 or 20 feet away. As the dog moves in to investigate the cage the assistant ―fades away‖ slightly so that, just as the dog reaches the cage and sniffs, he/she can drop the reward in on top of the cage without the dog seeing where the reward came from.

After two or three trials like this, the dog‘s sense of sight will assert itself. The animal will begin sniffing less and looking more. It may do any number of things. The dog may stop short of the cage and stare at it, or it may go to the cage and check it and then turn away or begin to look around. In response, the assistant encourages the dog to investigate the cage, often by presenting it with his/her hand, and then drops the ball in on top of the cage when the dog ―checks‖ it. The most important consideration is to pay the animal when it is focusing on the caged aid. 

After several trials, when the dog‘s sniffing has been reduced because it has begun to look at the training aid/cage instead of smelling it, then you can ―wake the dog‘s nose up‖ again by moving the cage to a new location. You can move it to another corner of the room-- or you can place it inside a chest of drawers, replacing one of the drawers with the cage. In this location, the cage and training aid will be easy to find, but the dog will still have to sniff to find it. In addition, once found, there are plenty of visual cues to help the dog stay with the training aid. The dog will go to the familiar corner, find the cage is gone, and begin sniffing for it. When it finds the cage it will check the aid closely and sniff, and you can pay for this sniffing and checking. After a few trials like this, you should obtain good ―focusing‖ and ―checking back‖ and sniffing behavior on the cage. Now, if you withhold the reward, you should see the dog‘s stop and stare behavior in full force. At this point you can adopt one of the two strategies described above: 

• If you choose to wait for the sit, then the visual cues of the training aid inside the cage will help to keep the dog with the aid while it ―thinks‖ the problem over. In addition, the cage will protect the aid and prevent the dog from getting any result should it become frustrated and begin scratching and/or biting at the aid. 
• If you choose to cue the sit, then again the cage will help keep the dog focused on the training aid while you use the cue. This is especially important if the dog loses focus while you are cuing the sit. Once in the sit, with a little encouragement the dog can be induced to look back at the training aid and then be paid for checking it. You assume you will have to assist the dog many times. Initially you pay the dog even if you have to assist it with the cue. Later you pay only if the dog completes the final response without assistance. 

ISSUES WITH DFR DOGS

Disrupted Focus. In order to understand the importance of maintaining focus in the DFR dog, you must again consider the common shortcomings of the Reward NFS dog. Two of the cardinal faults of a conventionally-trained Reward NFS dog are that the animal:

• Tends to sit on fringe odor rather than going to source. 
• Tends to stop working odor/sniffing after the final response. 

These two faults, both rooted in the fact that the dog expects reward from the handler rather than from source, interact in the following way to produce final responding on handler cues rather than on odor. When the dog smells odor, it final responds and looks at the handler. If the dog is not as close to source as you desire, then the handler asks the dog to leave the final and continue to sniff/search. The dog normally does not do so effectively, because once it has smelled odor it quite rightly expects reward to come from the handler, and so it looks at the handler, its‘ mind on visual cues rather than sniffing. The handler normally makes a hand presentation or gives some other sort of encouragement; the dog makes a token head movement while looking at the handler, moves in the direction the handler has gestured, and then immediately sits again. If the dog has moved closer to source, the dog is then paid. However, from the moment the dog first gave final, it has not done any further sniffing or processed any more odor. It has merely oriented at and responded to a series of handler cues, and then been rewarded.

In contrast, a DFR dog needs absolutely no encouragement to go directly to source. In addition, if the dog is well-trained, it continues to work odor and investigate source even after first final. This is the basis of the ―checking back‖ behavior that is so valuable to the detector dog handler. When the dog ―checks back,‖ it points at, or crowds, or drives into the odor source, normally in response to some attempt on the handler‘s part to take it off of odor. This checking back behavior is extremely useful because it allows us to: 

• Refocus the dog‘s attention on odor after it has been disrupted by, for instance, cuing the final response. This enables the handler to pay the dog for sniffing odor rather than looking at its handler. 
• Ask the dog to ―confirm‖ a find. You ask the dog to confirm by trying to pull it away from something it is interested in, either before final or after final. A well-trained DFR dog must be physically dragged away from odor and will exert strenuous efforts to get back to odor and repeat the final. This makes the dog extremely easy to read, and simplifies the problem of telling the difference between interest in a distracting odor and a change of behavior in response to a bomb or drug hide

When a DFR dog has lost focus on training aids, when it no longer checks back when asked, this is normally as a result of ―damage done‖ while teaching the final response. However, note that many dogs lose focus while learning the final and have to be ―rebuilt‖ once the final is fluent. This is not difficult as long as the dog was taught to check back to odor prior to introduction of the final response.

• The first thing to do in re-establishing focus is to run a few ―pay on sniff‖ trials. It is very helpful to have a visible training aid (in a cage) so that the dog has a visual fixation point. Next, allow the dog to find the aid again but wait for the final response. Once the dog has completed final, then the handler can attempt a variety of different techniques to induce the dog to ―check back.‖
• Stand behind the dog and wait for a few seconds, encouraging the dog verbally to ―check!‖ Be extremely alert for a head movement towards the training aid, and be ready to pay on odor source. 
• Exert a slight backwards pull on the dog‘s collar, as though to pull the dog straight back off of odor. At the moment the dog feels this traction, it is likely to ―headpoke‖ back at the aid, and at this instant the handler or an assistant must pay by dropping the reward on source. 
• If slight pressure does not cause the dog to check back to source, then the handler can pull harder and actually pull the dog backwards out of the sit. Praise and encourage and excite the dog and then release it to go back to source. Initially, you pay the dog for checking back to odor without demanding the sit. Later you withhold the reward until the dog checks back and then completes the second final response. (Note: The decision about whether it is ―safe‖ to reward the dog on sniff without the final response depends on how fluent the dog‘s final response is—If the dog has a very fluent, relaxed sit, then it is normally not harmful to sometimes pay on sniff. If, on the other hand, the dog has a resistant, tense sit, then paying on sniff may quickly result in stop and stare.) If the dog completes the second final but loses focus again, then give the slight pull (as above) to try and re-establish focus during the final. However, it may be impossible to obtain both focus and final at the same time, so alternately reward one and then the other until the dog unites them both. 
• If the dog has completely lost focus on the aid so that none of the above produces the check-back, then the handler can attempt a hand presentation to bring the dog‘s attention back to the aid. Do not pay the dog while the hand is still on source. This merely rewards the dog for looking at a part of the handler‘s body. Instead tap on source to draw the dog‘s attention, and then attempt to ―fade‖ the hand out and pay for focus on source. 
• One trick used by the SSD Course is to place a radio with the training aid. When the dog is in final, the radio is keyed. The dog looks at source and is paid. 
• Another trick to re-establish focus is to use a ―reward system‖ to deliver the reward from odor source. There are very, very many of such devices—some complicated and elaborate and some very simple. In one version used by the SSD Course a training aid is placed in a box that opens towards the dog. Poised two or three feet above the aid is a reward in a PVC pipe, held there by a piece of thin bungee cord stretched from one side of the tube to the other. The reward rests on top of the bungee. Another piece of bungee cord or string is tied to the middle of the first so that if someone pulls sideways on it, the reward can slip past the first bungee and down the tube and land on the odor source. The important elements of such a reward system are that the reward is separate from odor prior to payment; that the dog is paid by someone other than the handler actuating the reward system, and that the reward lands as directly and softly as possible on odor source. 

As a final note, keep in mind the purpose of focus on the aid—It is to make sure that the dog does not learn ―handler cues‖ by watching the handler, and to make sure the dog can be rewarded on odor source for approaching source. If both of these conditions are met, then you are likely to produce a dog that shows strong independent search behavior, an obvious change of behavior, good bracketing to source before final, and a desire to check back to odor if walked or pulled off of it. If you have these elements in place, then exactly where the dog focuses while in the final response is not especially important as long as the animal does not look directly at the handler during payment. In fact, many DFR dogs learn a sort of ―superstitious staring‖ behavior in which they do not focus on source (especially when it is high overhead and well-hidden); instead they stare generally upwards or away from the handler. This behavior still serves the most important purpose of focus on odor source—it keeps the dog from being paid for focusing on the handler.

Refusal to accept presentations

• Because early in training they were never encouraged to look to their handlers for any help or information, DFR dogs tend to be very independent while searching. Many of them work very effectively on loose-leash ―scans,‖ and handlers find that this is often the best way to begin a search. They make presentations only when the dog misses a productive area. Eventually, if the dog is unable to make the find independently, or when it becomes tired and needs some guidance and encouragement, then the handler becomes more active and supplies more guidance. 
• If a DFR dog is excessively independent and will not accept its‘ handler‘s presentations, this is normally because the dog has simply not yet learned that the handler can be a help in finding the aid. (Note that this is to some extent deliberate-- you do not want the dog to learn to ―use‖ presentations by its‘ handler until late in training, after it has already become completely independent, aid-focused, and self-reliant.) In working with such a dog, the two most important factors are: The handler should give few presentations and the handler must make sure that these presentations are productive from the dog‘s standpoint.
• Set up a search problem with one training aid. Begin the search with the dog working independently, without presentations, but stay away from the area where the training aid is hidden. Wait until the dog tires substantially, and becomes more receptive to the handler‘s influence. Then move the dog into the area of the training aid and make a careful presentation in a productive area near the training aid, so that the dog catches the odor and completes the find and receives reward. Repeat this sort of search exercise a few times-- Run the dog on a long problem without any possibility of finding the training aid until the dog wears down somewhat, then take the dog into the vicinity of the training aid and make a presentation or two that put the dog ―into odor.‖ The animal will quickly learn to look for presentations, because they indicate an area in which odor can be found. Once the dog becomes very receptive to that initial presentation, then the handler can begin carefully adding a few more presentations here and there, very gradually building up the number of presentations the dog will eagerly accept before finding the training aid. 
• When teaching the dog to accept high presentations, it is especially important to make sure that the first few presentations are productive for the dog—that they help the animal find odor. Otherwise the dog quickly learns to merely rear up against anything that is presented, but without sniffing.