Heritability of Herding-Related Traits

C. Denise Wall, PhD and Mellissa DeMille, PhD

Summary of Previous Studies

As most people will agree, herding behaviors observed in Border Collies are influenced by heredity. The herding instincts exhibited by dogs are thought to have originated from the wolf, where certain members of the pack surround game and then close in for the kill. This instinct has been modified by selective breeding to encompass the range of herding behaviors we see in the modern Border Collie. These behaviors have been bred for and strengthened primarily by selecting and crossing specimens of the breed that are considered the most desirable. Though it would seem beneficial to determine the exact role genetics play in the inheritance patterns of the individual traits, there have been few rigorous studies to this end. Difficulties involved in such studies include evaluating certain herding traits after the dog has been trained, since training can strongly influence the expression of these traits. As well, casual observation of the differences in herding ability within any given litter of Border Collies would lead one to believe that these traits are complex in their mode of inheritance and expression, therefore requiring considerable time, effort, and controlled conditions to define. The few informative studies in which investigators have taken on this task are not extensive or definitive, which leaves the average breeder struggling to find their own way in producing the type of herding dogs they desire. We will attempt to summarize these studies in hopes that the tendencies noted will give breeders more information for their breeding programs, as well as giving future buyers an idea of what to expect.

Herding ability is generally described as the ability to round up and control livestock. The intensity of the desire to herd is called keenness. For evaluation purposes, it is useful to further break down herding ability into individual traits that presumably can be inherited separately. However, as will be discussed later, some of these behaviors commonly occur together. One important trait, the behaviors collectively known as "eye", can be further separated into two parts. The first, which is the tendency to watch and follow moving objects, is not unique to Border Collies. Other breeds such as Greyhounds also have this tendency. The second part, where the dog appears to check itself at a distance and "set" the object of interest with its eye, is also exhibited in other breeds such as Pointers and Setters. Another important trait in Border Collies can be defined as "style" in which the dog works with either a clapping (crouching down flat when checked by their handler or resistant sheep) or upstanding style (remaining on their feet while working). It is interesting to note that Pointers and Setters who also set objects with their eye, exhibit these two divergent forms of style. Perhaps most important, but most difficult to define, is the trait "power", which is commonly described as the ability to move livestock.

Kelley, 1949

In addition to clearly defining important traits, scoring systems that help to quantitate these behaviors are very beneficial. Scoring systems help in identifying a criterion for desired traits as well as aiding in determining heritability of traits involving many genes. Described in a book written by Kelley (Kelley, 1949), were crosses he performed in which "eye" was graded on a scale from 0-6 (0 = no eye and 6 = over eye). The following passage is a quote from his book regarding his scoring system: "The author prefers to think of this most important characteristic [eye] as being capable of further definition and of point scoring. Thus "over eye" with a score of 6 points exists when the dog becomes so firmly fixed that it is immobile. A young dog would set its sheep but not run around them. A good approach is natural in the next three grades. A "strong-eyed" dog, score 5 points, has great control over sheep which it "stares out", but the dog remains flexible under command. A pup of this nature will run around and clap or stand and stare. The term "medium eye", with a score of 4 points, indicates that the dog can control sheep by eye. It fixes first one and then another. Young dogs run around and develop eye by being worked on a few sheep. They may "hold a point", but they move on. "Free-eyed" dogs, score 3 points, have eye and will sight sheep but, in a group of 5 sheep, the dog's eyes will move over the whole five one after the other. Young pups of this class will show little eye but may pull up and sneak, then run on. Light-eyed dogs, with a score of 2 points, will barely drop their heads. Young pups run around freely and seldom pull up. Weak-eyed dogs, score of 1 point, will seldom drop their heads or pull up. Young pups of this kind may bark. Dogs with no eye, score 0, are hoolem-up dogs which bustle their sheep into position."

Kelley's scored dogs were bred and the resulting 28 offspring demonstrated an average score on eye that was midrange between both parents. Therefore, in this study, eye appeared to be inherited as an incomplete or co-dominant trait, suggesting a complex mode of quantitative inheritance involving more than one gene. It is of interest to note, however, that the dogs he bred from in this study scored in the midrange areas of 3, 4 and 5. The scores also tended to be the same as one parent or the other, averaging to a midrange score. It would be important to determine whether similar results could be obtained with breedings from dogs at the extreme ranges of 0 and 6. Moreover, since most dogs with eye would score in the midrange, with very few at either extreme, it might have been more beneficial to score dogs using a finer scale in the more common, midrange area. In addition to his specific work on eye, Kelley further demonstrated with his breeding studies that herding ability in general is indeed inherited and that outstanding characteristics can be fixed through linebreeding.

Burns and Fraser, 1966

In a genetics book, Burns and Fraser (Burns and Fraser, 1966) cite previously unpublished work where crosses were done using the same Border Collie sire for both another Border Collie bitch and a Pointer bitch. Pups from both litters were observed at two week intervals from six weeks until six months for the development of herding ability and eye. By five months all the purebred Border Collie pups were showing herding ability and eye whereas none of the crosses did. Interestingly, although purebred Pointers have been known to "point" sheep, none of the Pointer crosses demonstrated any of the characteristic pointing behaviors at the livestock. This study would suggest that herding ability and eye are inherited as recessive traits. In contrast, other studies done by Burns crossing Border Collies lacking eye with Border Collies showing eye yielded pups that all had some degree of eye. The differences in results between the two Burns studies may indicate that determining heritability of herding patterns by crossing different breeds with Border Collies may prove futile. Style is another trait that appears to have a complex inheritance pattern, as Burns found that crosses between upstanding dogs and clapping dogs rarely resulted in fully upstanding dogs. However, the degree of clapping was lessened in some of the pups, with most pups being in the intermediate range. Therefore, when using Border Collies for these studies, eye and clapping appeared to be partially dominant.

Burns, 1969

In a landmark study (Burns, 1969) originated to determine whether behavior patterns bred into British sheepdogs were compatible with those traits required to successfully herd sheep in Ghana, Burns observed the heritability of several herding type behaviors. While evaluating different herding characteristics, Burns made several interesting assumptions regarding these behaviors that are summarized as follows: The trait that most separates Border Collies from other types of herding dogs is the behavior known as "showing eye", which is usually combined with low tail carriage and a crouching, stalking-like posture. These dogs are said to move sheep by the "power of eye". However, dogs with excessive eye often appear weak, lacking power, and are often extreme clappers. They may be reluctant to get up and tend to rush in and grip sheep when made to move in close. The author states that it is generally accepted that dogs with great power tend to be upstanding dogs with moderate eye. Other observations regarding power are that strong dogs can make sheep move simply by walking straight towards them and can stop sheep just by looking at them. A strong dog is also less likely to grip. In addition, a powerful dog can push sheep through openings and obstacles they are reluctant to go through and has the ability to control a single sheep, holding it off from the rest of the flock. There are other types of sheepdogs that have the strong tendency to gather animals even though they lack "eye". These "non-eyed" dogs often bark when faced with a defiant sheep and are favored over silent dogs in areas where the terrain makes viewing the dog and sheep difficult. The author states that a barking dog is much less prone to grip than a silent dog. Without the eye to hold them off sheep, these dogs show little resistance to moving in close to the sheep and are never extreme clappers. Their inability to feel the pressure and keep the proper distance from the sheep makes non-eyed dogs less desirable for shedding single sheep and working ewes with lambs. These dogs can sometimes move stubborn sheep more effectively than an eyed dog by using their body movement instead of their eye. Because this is accomplished with body movement, the author terms this type of power "force", with the term "power" being reserved for dogs using eye to move sheep. The behaviors of barking, no eye and upstanding style were thought to be inherited together, as were the behaviors of silent working, strong eye and clapping. Taking all these assumptions into account, it was thought that a barking-type dog would best suit the needs of the shepherds in Ghana, given the bush grazing conditions.

Using both barking dogs and the typical British, silent working, eyed sheepdogs, data was gathered from four crosses involving six different dogs. These crosses yielded some interesting findings regarding inherited tendencies. The traits being evaluated were: strength of herding instinct (broken down into four categories), defiance barking versus silent, clapping versus upstanding style, eye versus no eye (broken down into four categories) and power (broken down into four categories). The first cross was between what was described as a non-eyed dog that barked freely at sheep (it is not noted whether this dog was a Border Collie), and a pedigreed, eyed, Border Collie bitch. The dam and three pups from this litter were imported along with an unrelated pedigreed, eyed, silent, Border Collie dog. The resulting crosses suggest that defiance barking, which appears unrelated to social barking, is inherited as somewhat of a dominant trait. This barking trait proved to be the only behavior where a definite pattern of heritability emerged. In all cases except one, dogs who barked also had no eye. Only two dogs from all four of the crosses completely lacked herding instinct, suggesting a recessive mode of inheritance for this trait. There appeared to be some correlation regarding the strength of herding instinct and an early age of onset of this desire to herd. Conversely, those dogs with only fair or weak herding instinct tended to demonstrate this instinct at an older age than those with strong instinct. The inheritances of eye, power, and clapping style were more variable. The following summary describes the findings regarding these traits:


Cross #1, where the eye of the parents was graded as one slight and one strong, produced one dog with strong eye (1a) and no others listed for that cross. Cross #2 (parents -- none X strong eye) yielded one dog with slight/no eye (2a), two dogs with no eye (2b and 2c), and two herding dogs ungraded (2d and 2e). Cross #3 (parents -- medium eye X none) produced two strong eyed dogs (3a and 3c), one with little eye (3d) and one ungraded due to lack of herding instinct (3c). Cross #4 (parents -- medium X strong eyed) resulted in four strong-eyed dogs (4a, 4b, 4d, 4e), one medium eyed dog (4f), one herding dog not graded (4g) and one dog not graded with no herding instinct (4c). These data are difficult to interpret unless one can assume that one of the original dogs (the parent with no eye in cross #2) was not a Border Collie (this was not noted one way or another by the author). If this was the case, these results agree with previous studies suggesting that eye is recessive in offspring from Border Collies X non Border Collie crosses. In cross #3, the parent with no eye was an offspring of cross #2 (dog 2c). This dog produced two strong eyed dogs when mated to a dog with only medium eye, further supporting the theory that eye is recessive in these types of crosses.


In cross #1, the style of the parents was upstanding X clapping and produced one clapping dog (1a). Cross #2 (parents -- upstanding X clapping) resulted in two upstanding dogs (2a and 2c), one clapping (2b) and two herding dogs not graded (2d and 2e). Cross #3 ( parents -- clapping X upstanding) yielded two upstanding dogs (3a and 3c), one herding dog not graded (3d) and one non herder (3b). Cross #4 (parents -- clapping X clapping) produced three upstanding dogs (4b, 4c and 4d), three clapping dogs (4a, 4e and 4f) and one herding dog not graded (4g). The data from these crosses, especially cross #4, suggest that clapping is somewhat dominant in its mode of inheritance.


In cross # 1, the power of the parents was great X great. This cross produced one dog with very great power (1a). Cross #2 (parents -- very great X great power) resulted in three dogs with great power(2a, 2b and 2c) and two herders not graded (2d and 2e). Cross #3 (parents -- medium X great power) yielded one medium (3a), one strong dog (3c), one herder ungraded (3 d) and one non herder not graded (3b). Cross #4 (parents -- medium X very great power) yielded three dogs of very great power (4a, 4e and 4g), one dog of medium power (4f), one dog with great power (4d) and one non herder (4c). Of note is the fact that all dogs used in this study had at least medium power, presumably because they had to have functional dogs for their actual work. However, it is apparent that even though some dogs lacked herding instinct entirely, those that did herd had at least as much power as the least powerful parent and one had more than either parent (dog 1a). It is also interesting that Burns graded several dogs with no eye as having power, given her definition that dogs with no eye did not possess true power but rather force.

True to her previous assumptions, Burns noted in her study that behaviors such as eye, clapping and silent working style tended to occur together. Lack of eye, upstanding style and barking were also noted to occur together. However, power was not limited to dogs with eye as was previously suggested by the author. Nor was there any apparent pattern to the relationship between power and style. Although it was thought that the behavior patterns of strong eye, power and defiance barking were mutually exclusive, one dog from these crosses had all three of these traits. Because of the terrain in Ghana, this dog was considered the most desirable by the author.

Taken together, these studies show that while heritability patterns for herding behaviors appear to exist, the mode of inheritance for most of these behaviors is fairly complex.


Burns, M and Fraser, M.N. 1966. Genetics of the Dog. The Basis of Successful Breeding. Oliver and Boyd, Edin.

Kelley, R.B. 1949. Sheep Dogs. Their Breeding, Maintenance and Training. Angus and Robertson, Sydney.

Burns, M. 1969. The Mutual Behaviour of Sheep and Sheepdogs in Ghana. Trop. Agric. 46:91-102.

Copyright 1996 by Denise Wall and Mellissa DeMille