Comparing the effectiveness of citronella with unscented odours in the anti-barking spray collar

This study is concerned to compare the influence of two different odours, citronella and unscented, with barking problem dogs referred to the Canine Behavioural Clinic. The bark activated spray collar (Aboistop) emits a spray each time the dog barks. All dogs selected were older than six months and weighted more than 10 pounds. Eligible owners were being asked to fill out a behavioural questionnaire. The protocol excluded dogs with aggressive tendencies and/or separation anxiety, but we included dogs displaying multiple behavioural problems. Three questions were raised about the anti barking collar:

1. Is the unscented spray equivalent or less effective than the citronella spray with excessive barking problem dog ?
2. Do they bark less after the removal of the spray ?
3. Is there a factor of habituation and do dogs bark as much as before ?


Thirty-two dogs were tested and subdivided in two groups. One group started with the citronella odour during week 2 and 3. They ended the study with the unscented odour during week 6 and 7. The other group started with the unscented odour at week 2 and 3. They ended with citronella odour during week 6 and 7. The breakdown of no spray or odour, for either group, was similar during week one, four, five and eight. Dogs were kept in their natural home environment.

Owner brought the collar with the automatic counter at the office on the same day for eight consecutive weeks. The Bark Counter Circuit uses a 12C508A PIC (Programmable Integrated Chip) with a 24LC16B memory chip. Upon detection of a bark, the Aboistop circuit activates its solenoid valve to release the aerosol. One of the input pins on the PIC micro controller is connected to the solenoid valve and is thereby capable of detecting its activation hence a bark.

To access this memory, we call upon another circuit that reads the memory in sequence. It has a five-digit readout; 3 for memory number and 2 for number of barks (up to 99 barks an hour). This same module is capable of reinitialising the Bark Counter Circuits. F test between the two groups were none significant (P < 0.05). The correlation coefficient is significant (R=0.927) for the group starting with the citronella during week 2 and 3, and R=0.813 for those starting with the unscented odour. Therefore there is no significant difference between the types of odour used to interrupt the barking dog.

Therefore 80.9% of all dogs responded well to both sprays. We concluded that the element of surprise by the disruptive stimulus is necessary in the control of the barking problem dogs. The device seems to be a complementary tool in behavioural modification therapy. Specially barking related to environmental stimulus, play, territorial, protection and social facilitation.

However, the device can also be used in order to help out owners of dogs displaying separation anxiety, fear or stereotype barking. The results also show that 67.7% of dogs, in either group, kept a lower number of barks after the removal of the spray and not the device. We did not found a major recurrence in our results probably because there were too many interfering events, which did not permit the dogs to habituate. However some owners reported a slight increase of the number of barks after the removal of the spray at the end of the study.

Introduction

Canine behavioural problems are widely recognised. Epidemiological data on dog problems support that assertion. Among various behavioural problems, barking seems to be one of the most common complaints by owners (Diaz & Houpt 1996). Results obtained by telephone survey at the Canine Behavioural Clinic did classify excessive barking among the five major complaints. However it should be considered that these complaints differ from reported problems examined in referral behaviour practice. Beaver 1994 reported that problems with excessive barking ranked high for complaints on the survey 12,9%, but ranked low in cases evaluated 3.2%. This author explains this difference. Behaviours that are not considered public nuisance, destructive or potentially dangerous, the owners were not seeking professional advice. Public nuisance can also be reported differently by J. P. Samaille (1997). He mentioned that 35% of dog barking behaviours are considered as public nuisance.

It was observed by Campbell (1976) the estimated rate of failure to educate the barking dog to quiet down is situated between 20 to 40%. These failure are in direct relation with the choice of the behavioural modification technics, the lack of patience of the owners that tries to modify the behaviour, the time required by all members of the family household or the lack of proper information in order to modify the barking behaviour. In the mist of all these collected data, anti-barking devices commercially available to assist owners and to disrupt barking could be used to control the public nuisance.

The exact meaning of dog's auditory communication signals are not completely understood; these auditory signals have been identify in different situations. It was reported that barking is seen in selected breeds, to greet people, to warn the owner the presence of an intruder, to make contact during play periods, to gain attention during separation, to respond whenever the external stimuli occur, to the social facilitation or to fear. The latest was observed by Diaz (1997). The author is mentioning another unintentional mistake on the part of some owners by using either verbal or physical reassurance. It should also be pointed that excessive vocalisation is not a problem seen in purebred dogs exclusively and gender does not make a difference.

Remote punishment can take many forms. Commercially available anti-barking devices can use loud or high-pitched sound audible to humans, ultrasonic sound that are not audible to humans, electronic shock collars, and the new generation of citronella spray collar is available since 1995.

This study is interested in comparing the effectiveness of the citronella spray with the unscented odours within the anti-barking spray collar. The goal is to assess the effectiveness of the citronella odour in reducing the number of barks in different problem dogs. Three hypotheses were verified. The first hypothesis was to identify the potential differences or similarities between the two odours in their effectiveness to reduce the number of barks display by the excessive barking problem dogs. The second hypothesis was to compare the number of barks after the removal of the spray in the anti-barking collar and to verify the increment learning at the end of this present study. The last hypothesis was to identify the presence of habituation to the odours after the eight-week period.

Subjects

Owners with barking dogs were referred by different veterinary clinics to the Canine Behavioural Clinic. Every owner were asked if their dog was older than six month and weighted more than 10 pounds since these two parameters are being identify by the manufacturer. Owners eligible received by mail or fax the 14-pages behavioural questionnaire. The questionnaire included questions related to dog history, care, environment, reaction to play, reaction to grooming, general level of activity, past and present behavioural problems, type of the relationship with owner and strangers. It also includes questions about barking per say, the differential diagnostic table of barking (to validate the owner answers in other questionnaire sections) and the aggressiveness score sheet elaborated at Cornell Behavioural Clinic University (Diaz & Houpt 1996). It should be brought to the reader' attention that we excluded all dogs with aggressive tendencies or separation anxiety. We did not exclude barking dogs with multiple behavioural problems. Table I represent the letter attributed to each dogs, their age, sex and the breed.

Material

Subjective owner reports were not used to obtain their appreciation for the effectiveness evaluation of the device. However we did rely strongly on a new bark counter. The Bark Counter Circuit uses a 12C508A PIC ( Programmable Integrated Chip ) with a 24LC16B memory chip (Fig. 1). Upon detection of a bark, the Aboistop circuit activates its solenoid valve to release the aerosol. One of the input pins on the PIC micro controller is connected to the solenoid valve and is thereby capable of detecting its activation hence a bark. The memory chip is divided into periods of an hour. At a certain time after initialisation, the memory pointer is pointing to a specific memory for a period of one hour. If a bark occurs during this timeframe, the memory content is called and incremented by one. It then is stored back in memory. In this way, the 2 chips are capable of accumulating data on barking behaviour. To access this memory, we called upon another circuit that read the memory in sequence (Fig. 2). It has a five-digit readout; 3 for memory number and 2 for number of barks ( up to 99 ). This same module is capable of reinitialising the Bark Counter Circuits. All the above circuits were design by engineers.

The unscented odours used in the anti-barking collars during this study contain the HFC-134a (hydrofluorocarbon). This product is environmentally safe, therefore, it has an ozone depletion potential of zero. The citronella odour is also approved by the FDA. The odour is a mixture of HFC and diluted citronella natural essence.

No other material was necessary because all dogs remain with their owner at home. However an individual score sheet was used to plot the number of barks per hour, per the day and the week. Table II is an example of the individual scoring sheet.

Method

We did not control the dogs' environment and did not wish to influence this setting in any manner. Thirty-two dogs were randomly selected to participate in two different groups. Twenty dogs started with the citronella spray during the second and third week. They ended with the unscented spray during week six and seven. Twelve dogs, in the other group, started with unscented spray at week two and three, and the citronella spray during week six and seven. The breakdown for both groups with no spray was during week one (first baseline), four, five (second baseline) and week eight. To avoid any cancellations, all behavioural observations were discussed during a weekly appointment at the office at every seven days on the same day for approximately 15 minutes per owner.

The number of barks were entered on the individual score sheet. The number of barks was added in order to obtain the number of barks per week. Means of barks per day were also calculated to facilitate statistical analysis (Table III). Two-baseline week were used in this study (week 1 and 5). It is important to note that the baseline scores obtained at week 1 and 5 should be superior to the weeks where spray was used to interrupt barking (week 2, 3, 6, 7). If the baseline scores were inferior, the subject was not included in further analysis.

For the group starting with the citronella odour (week 2 and 3) dog # (C), (E), (Z) and (A1) were not included for statistics. As for the group starting with the unscented odours (week 2 and 3) dogs # (I), and (L) scores were also not included. Subject # (X) in this last group was excluded because owners did not respect the appointments at the office. This was the only dog. All the other owners did follow the agreement.

The previous means were used to calculate the percentage of improvement. This manipulation was necessary to determine which dog was barking less or more while using one of the two-spray odours. Subjects with a higher score than one of the baseline (week 1 and 5) were excluded and the difference between odours were verify in percentage. The percentage for each individual subject was needed to identify the effectiveness. Table IV shows the percentage of improvement or recurrence for the twelve subjects starting with unscented odour (week 2 and 3) and then citronella odour (week 6 and 7).

It should be pointed out that group 1 had three groupings (1-3-5) of subjects but they all started with the citronella odour and group 2 had only two groupings (2-4) of subjects. However they all started with the unscented odour. Mean percentage for each grouping was calculated. Table V shows the results and the differences between the two sprays.

The correlation coefficient was performed to identify if both odours were equal or different in their effectiveness to interrupt barking dogs. The F test was also used to identify if one spray was superior to the other type of spray.

Results

The percentage representing how much the subjects were affected in receiving the unscented spray (week 2 and 3) and citronella spray (week 6 and 7) was calculated for each dog. Table IV shows the results or the percentages that compares the effectiveness of one spray over the other odour. Of this group of twelve dogs, three dogs were excluded for further analysis: One dog # (X) was eliminated from the study because owners did not follow the procedure. The other two dogs did not respond to whatever spray used to disrupt their barking. Nine remaining dogs were included for further analysis, 81.82% were responding to both sprays. Dog # (I) was actually barking more at the end of the study. Dog # (L) did not seem too aware that he was wearing the anti-barking collar. However the owner of the dog # (L) reported that he was on vacation during week 6 and 7, and the dog was living with another member of the family in the same environment.

The same method was used with the other group receiving the citronella odour (week 2 and 3) and the unscented odour (week 6 and 7). Four dogs were excluded for further analysis. They were barking more at (week 7) compare with the baseline (week 5). This finding shows that some dogs are affected by the citronella spray and not to the unscented odour. We did not want to influence the statistical analysis and they were simply excluded. However it should be pointed out that dog # (E) and # (Z) had very long fur and owners did not follow our instructions to cut the fur short enough to permit the spray to reach the dog snout. The two other dogs # (C) and # (A1), owner reported some intervention. The type of intervention was to shout back at their dog while they were present at home. It seems that these dogs were waiting for the right time to bark. The disruptive stimulus (citronella and unscented odours) was effective at 80.00% in this group.

Groups are homogeneous from the previous calculations. The differences between odours per groups were then calculated. Table V show these differences in percentage numbers. The coefficient correlation for the total of dogs in the group of 16 dogs receiving citronella odour (week 2 and 3) and unscented odour (week 6 and 7) was 0.927. The correlation in the group receiving the unscented odour (week 2 and 3) and the citronella odour (week 6 and &) was 0.813. Both odours were significantly related or equivalent in disrupting the barking dogs. Then the F test was performed in both groups between the numbers of barks (week 2 and 3) and (week 6 and 7). As for the group receiving the citronella first, results were 0.923 and not significant. For the group 2 starting with the unscented odour, results obtained was 0.980 and not significant. It was also noted that the citronella was a little more efficient than the unscented odour for group 1 (Fig. 3). As for group 2, grouping 2 responded more to the unscented odour (Fig. 4).

The second hypothesis was to compare the number of barks after the removal of the spray odour in the anti-barking collar and to verify the increment learning. We have found that 17 on 20 dogs in the group receiving citronella spray (week 2 and 3) and the unscented spray (week 6 and 7) where barking less than the initial baseline obtained at week 1. However only 4 of 11 dogs where barking less with the unscented spray (week 2 and 3) followed by the citronella spray (week 6 and 7) than the initial baseline. It should be pointed out that 67.7% of barking dogs did show some increment learning. The coefficient correlation was 0.148 and the F test was 0.001 for the group starting with the citronella odour and the correlation of 0.218 for the group starting with the unscented spray odour during week 2 and 3. The F test was 0.006 for this group. These results suggest that some learning persist after the removal of any spray in the anti-barking device. It should be noted that the previous percentage represent the residual learning and not the dramatic drop of the number of barks with the collar. Table VI represent the raw data of the grouping 5. The drop is considerable in most instances.

Discussion

Communication displays are similar among dogs and wolves and they are critical to the establishment of social relations or activities. Domestication of the dog by the process of selective breeding had led to changes in certain behaviour patterns. Dogs seem to appear to be more vocal, although many of these differences are breed related (Fox 1971). Barking is only one out of many sounds being produced by the dog. It shows the level of arousal or excitement and it varies from one breed to another. These vocalisations are usually reinforced non-intentionally by the owner. However excessive barking can be often related to the pathological or serious behavioural problem (Pageat & Tessier 1997). Muting of a bark is important when it becomes continuous. It is consider a public nuisance and it will influence the relationship between neighbours and owners (Blacker 1986). The most clearly established health effect of excessive noise or barking is hearing impairment. Especially to long-term exposure to noise above 90 dBA that was identify by the Environment Health Department (Charles, Senn, Jeffry & Lewin 1975). These authors had noticed in their study that some barking could be as high as 104 dBA. However the barking per se was not considered continuous but was noted as series of peak noises of short duration.

Dogs' owners commonly solve the problem of neighbour complaints by confining their dog in areas farther from the neighbour home or by keeping the dog, at night in their home. Other options involve trying to control barking where some owners considered acceptable to use medication, muzzles or behavioural therapy for this purpose. However canine practitioners are drawn into problems with barking when they are asked for their advice in helping an owner. It is usually too late to take the necessary time to modify this behaviour and owners are requested to perform debarking operation or euthanasia (Hart 1978). Other option such electronic shock collar which can be used with behavioural therapy is not well accepted by most owners and behavioural practitioners (Juarbe-Diaz 1977). A different option is offered by the citronella spray collar. The dogs may be distracted by the odour and try to locate its source, may be calmed by the odour, may be startled by the noise of the spray or may object to the citronella spray odour. One of our hypotheses was to see the differences between citronella odour and unscented odour in their effectiveness to reduce the excessive barking dogs. Results obtained in this study tend to suggest that odour is not an important issue since we had significant correlations in both groups. Dogs were barking considerably less than the initial baseline and the secondary baseline. In a study at Cornell University there was 88% satisfaction compared to about 44% satisfaction with the shock collar. We did found that barking was considered eliminated by owners although some occasional barks were still present. The percentage of satisfaction, in this study, tends to support previous studies. 85% or 27 owners reported that they were satisfied with the citronella odour and 80% or 25 owners reported their satisfaction with the unscented odour.

Although the results of previous studies have shown a reduction in the barking dogs with the citronella collar, they were relying on owner reports on the frequency of barking or on behavioural observations, instead of using a bark counter. It is the first time that we are able to obtain objectively some numbers that we can analyse.

It was shown that most dogs would bark less because they get distracted, calmer and startled by any kind of spray odour. However it was also noted that the citronella was a little more efficient than the unscented odour. We did not found a significant difference. These results reinforce the findings of J.P. Samaille (1997). Taken individually we have subjects that were objecting both spray odours. Especially with subject # (J) where he got mad when receiving the spray and in return he was barking much more. This finding goes with the behavioural observations where some exceptional dog may object any kind of spray odour (Juarbe-Diaz 1977).

Many authors do suggest that any learning will remain if we use intermittent reinforcement. Skinner (1969) specify the strengthening of response or learning through intermittent reinforcement being called ''interval and ratio schedule''. Reward can be scheduled according to the passage of time, the number of responses or some combination of these. Continuous reinforcement seems to be less effective than variable schedule. It is important to note that owners where instructed to leave the collar on their dogs which is similar to the continuous reinforcement. However it was also suggested to remove the collar during the night. Because of the probabilities of barking is almost non-existent. The removal can have the same effect of interval or ratio schedule propose by Skinner.

The second hypothesis was to verify if dogs were barking less than the first week baseline. It is interesting to note that 67.7% of barking dogs did show some increment learning. It should be suggested in further studies to elaborate on the effectiveness of different schedule of reinforcement in order to maximise the effect of the anti-barking spray collar. The correlation was not significant and the F test does show some learning increment at the end of the study.

The last hypothesis was to identify the effect of habituation when using the spray collar. Results of previous studies have shown that without the appropriate behavioural therapy, some recurrence of the barking is possible (Pageat & Tessier 1997). We did not found any major recurrence in our results. Habituation depends on the complexity of the stimulus to which habituation occurs. It should be noted that the learning is easily disrupted or dishabituated by interfering events. Events in this study were not controlled and they vary from one moment to the next. We could not found a control pattern. Dogs could bark in different rooms, time of the day, or in different location. However owners did report that their dogs were barking little more by week 7 or 8 than during week 2 and 3 where the spray odours were first introduced.

Summary

Excessive barking can be a mild or severe problem if the owner is under pressure from neighbours. Different motivation behind excessive barking is difficult to isolate. These dogs referred to the Canine Behavioural Clinic had more than one underlying motivation. Practical advice and recommendations to clients could improve the difficulty. However most owners will stop the education midway or they will reinforce the behaviour either positively or negatively. In order to help these owners with their dog's education, behavioural consultants or veterinarians should help owners as soon as possible in suggesting either medication or the spray anti-barking collar. Owners will built their confidence and they can establish different exercise, suggested in the behavioural therapy, at their own pace.

In this study we can conclude that the spray collar is effective in reducing the number of barks with citronella odour or unscented odour. The odour does not play a major role in reducing the number barks per week. We also found some residual learning 67.7% after removing the spray from the collar. However more studies are needed to find the effect of the variable schedule of reinforcement. Finally it is interesting to see that owners were very satisfied with the results. The percentage of satisfaction, in this study, tends to support previous studies. 85% or 27 owners reported that they were satisfied with the citronella odour and 80% or 25 owners reported their satisfaction with the unscented odour.

Acknowledgements

This project was supported by a grant from Multivet International Inc. We are very grateful to Mark Garon (research and development), to all veterinary clinics who referred these excessive barking dog cases and to the owners who participated in the present study.