I'm re-posting this wonderful article here because it seems to have disappeared from the web. Written nearly twenty years ago, this is still an excellent discussion of livestock guard dog breeds. This article and personal discussions with Dr. Sponenberg have been important in my own research and writing.
Dr Sponenberg is a professor of genetics at Virginia-Maryland College of Veterinary Medicine, author of numerous books and articles on genetics and livestock preservation, and serves as the Technical Director of the Livestock Breeds Conservancy.
"This article is a set of thoughts that I have been pondering for the last few years, concerning breeds, livestock guard dogs, and the interaction of those two subjects. My usual sphere of activity is with conservation of livestock genetic resources, and dogs differ from this in many regards. At the same time, though, the issues of breeds and breed conservation of dogs have many features in common with those of livestock. This article is going to wander a bit, and then will come back and hopefully tie all the loose ends into some sort of neat package.
The development of all species of domesticated animals first arose as a partnership of humans with the species in question. In no case has this partnership been closer than has that of dogs and humans. Soon after domestication it was easily appreciated that not all dogs had equal talent for all tasks. As human endeavors became more complex, dogs were selected to be specialists for various tasks. This is basically the process of breed development, with profound consequences for the breeding of dogs or any other species. The important concept, at least in early stages of breed development, is that function guides the process, and external form simply follows along however it can.
A breed can basically be viewed as a predictable genetic package. To be useful, breeds need to be predictable. That is the way that they can fit certain niches with a high degree of success. This matching of a breed to a niche is something that has largely gone from purebred dog breeding as dogs have moved from being essential partners in performing tasks, to becoming companions and companions alone. As the functional abilities of dogs have diminished in importance for human endeavors, so has the emphasis on these in breeding programs. As a result, the predictability of dog breeds for specific tasks is something that is generally under appreciated by the general dog owning, or even dog breeding, public.
Breed development usually follows a fairly consistent pathway. The first stage of the development of most breeds is that people simply use what is locally available and adapt it to the task at hand. The resulting breed is therefore shaped by what is locally available (the founder effect), and the subsequent selection of this to suit a specific task. Since the goal of such breeding is function, the animals within the group are usually somewhat variable as to looks, but reasonably consistent as to function. This type of population is best termed a "landrace", which basically means a local or regional breed simply springing up and becoming uniform by virtue of local selection for a specific purpose. Any external consistency is a spinoff from a combination of founder effect or human selection for function. Border Collies are a reasonably good example of a landrace of dogs- they are consistent in behavior (the key element of selection), and most of them are visually similar enough to be recognized as Border Collies. However, variation does persist and some Border Collies by heritage, pedigree, and behavior are not all that easily recognized, even though they are still genetically predictable for the essential component of the breed (in this example, behavior).
The next stage of breed development is that of standardization. Standardization can occur through two main routes. One of these is local or regional, and more or less can be viewed as standardization "from within" as the breed is made more uniform but in its original niche. The other, aptly termed "gentrification", was coined by David and Judy Nelson, who neatly summed up this important process in a single word. Gentrification occurs when the landrace is taken out of its original site and then standardized remote from its original niche. This is standardization "from without". Either mechanism can result in a functional, predictable breed. Gentrification does have a certain inherent risk, though, in that removal of animals from the original niche can impose changes in the breed that deviate from the original purpose.
Landraces occur as populations by accidents of history (founders) and selection, and geographic isolation. Standardized breeds take that isolation a step further by specifically only allowing breeding within the group, and also limit variability by deciding on a range of variation that is acceptable. The result is that the breed becomes much more visually uniform. The level of uniformity varies from breed to breed as the breeders' associations decide what to include and what to exclude. For example, the occasional brindle or black and tan Labrador Retriever shows up in a litter, but is excluded from the breed which only allows black, chocolate, or yellow. Golden Retrievers are even more restricted, while something like the Cocker Spaniel is allowed to have more variation for color. The important issue is that the range of variation in a standardized breed is arbitrarily narrowed by the breeders, and really may not reflect the original state of the population when it was simply functioning as a landrace.
Gentrification has been an important refiner and definer of many livestock guardian dogs. When a breed is removed from its original location it is easy for the selection philosophy that guides its development to likewise vary. This poses a threat to many dog breeds, but especially to the livestock guardians whose task and ability are based on thinking patterns and not on external type. One way these breeds can change is simply selection for size. Most are large to begin with, and larger dogs are more impressive to the eye. At some point, though, bigger is not better and the moderate dog is more likely to succeed for years of hard use than is the oversized dog. This depends on breed, but breed differences for size are important and need to be fostered to maintain distinctive and useful breeds. The Spanish Pyrenean Mastiff, for example, has gone in this century from a somewhat plain, moderately sized, somewhat flat coated dog to a huge, huge impressive fluffy dog. For hard guardian work this change may not be beneficial. Especially if the change comes from crossbreeding, the dogs are also changed. Some Russian Ovcharkas may have increased size from outcrosses to St Bernards and other non-livestock giants - the result being impressive dogs, but not reliable guardians. The confusion of large size with inherent guardian ability is a very real threat to the livestock guardian breeds.
Some breeds are deliberately and somewhat artifially created, and circumvent the landrace stage. Such breeds are arbitrarily developed as standardized breeds from the outset. Doberman Pinschers are one example of such a deliberately standardized breed. These breeds can be expected to have even less variation that the breeds that were standardized from landraces. Few if any livestock guardian breeds fit into this type of breed, since most are regional breeds that spring from a given geographic area.
The process of standardization, including gentrification, may or may not matter biologically, depending what was left behind in the process of standardization. It likewise may or may not matter politically, since each breed has a specific heritage. The important issue in breeding and maintenance of breeds is to be consistent with the heritage, so that the breed can continue in harmony with its heritage. Breeds do not pop out of the heavens fully formed - each one has a heritage. Selection for consistency with heritage is especially critical for breeds that still have functions to perform, since ignoring the historical function can result in eventual inability for the dog to perform. This is critically important in breeds such as herding dogs or livestock guard dogs, or bird dogs. It may be less critical in Irish Wolfhounds (no more Irish wolves, basically), or in dogs historically used for fighting one another, or various other tasks that seem to have largely gone by the wayside. In such cases of obsolete (or hopefully obsolete) function, perhaps it is logical for breeders to opt for selection for companion animals in a sound, safe, visually pleasing package. Nothing wrong with that - as long as critical functions are not being compromised in those breeds for which such functions are important.
The question with livestock guard dogs is basically what sort of form does this genetic resource take? How is the genetic resource organized, and how should breeders breed and select dogs within the general framework of livestock guard dogs?
One basic question is the issue of breeds - what are they and how many do we need? This gets to be an issue of lumping versus splitting. In livestock breed conservation the American Livestock Breeds Conservancy is generally guided by the principle that it is best and reasonable to split if each subsequent population has a good chance of continued existence, selection, and function. Lumping makes the most sense when populations are related, similar, and unlikely to survive as separate populations. In each case the issue of lumping versus splitting can be tricky. One basic guideline is whether or not two populations are more like one another than any other genetic resource, and whether they can be expected to be vital and viable if split. Put another way, split when you can, lump when you must. Geographic origin and selection history are more important in this exercise than are external similarities, a point which is easily missed especially with the large, white guardian dog breeds.
Livestock guardian dogs are a fascinating genetic resource of great value and utility, and safeguarding them as breeds is of vital concern to dog breeders as well as agriculturalists. Having these as predictable genetic packages is essential to a host of livestock owners. Livestock guardian dogs need to be consistent and predictable in order for the livestock industry to have rational choices for different situations. Different dogs are needed for different situations, and this is where breeds and breeders come in. No one breed can do it all - if that is the case then the predictability has been replaced by variability and picking a dog gets more difficult. This is not to deny that the variation within a breed can be nearly as great or greater than the variation between these breeds, but it is to state that predictability and "subspecialization" within the general livestock guardian dog breed group is a good thing, and should be encouraged rather than discouraged.
One characteristic of these dogs is that they occupy somewhat neighboring ranges throughout a huge geographic area. Each geographic area can be expected to fine tune this resource to what was needed, and to what worked. This seems to have resulted in a number of related but distinct gene pools, from the Pyrenean Mastiff of Spain (spotted) to the white breeds (Great Pyrenees, Maremmas-Abruzzese, Kuvaz, Komondor, Polish, Russian, Akbash), colored breeds (Kangal, Kars, Shar Planinetz, Tibetan, Central Asian Owcharek).
Questions for breeders working with these breeds include some idea of the original range of variation before standardization. What is amazing from a breed development standpoint is the relative consistency of type and visual appearance among these breeds. Many are white, which seems to have been imposed on these breeds at a very early stage of development. White guardian dogs were already well known in Roman times. This is largely due to deep seated conviction that such guardians stand in stark contrast to colored predators, and make keeping track of friend or foe an easier task for the shepherd. Equally important to some cultures is that white dogs blend into white flocks. In most regions white dogs also stand out against the landscape, again contributing to ease of detection.
Against the obviously widely held preference for white dogs stand the colored breeds. These occur throughout the range of livestock guardian breeds as exceptions to the general rule of whiteness. The reasoning behind these being allowed to be variable for color would be an interesting study, of only because the preference for white appears to be so ancient and so firmly held.
Breeders of livestock guard dogs are doing a great service for the livestock industry - if their charges remain faithful to the original purpose for which they were originally developed. The breeders' work and how they do it is essential. Since the breeds appear to have different propensities for behaviors critical to guarding livestock it is important to maintain these so that livestock owner can have choices peculiar to their situations. Making all of these breeds similar is to deny livestock owners choices that they need. Small flock owners in suburban (or sub-rural) areas have very different needs than range flock owners. Different dogs will be needed in each situation. This is an especially critical factor when considering "outlier" breeds that do not fit the usual livestock guard dog model: Kangal, Kars, Castro Laboreiro. These, and no doubt other, breeds need to be developed as their own unique gene pools and not crammed into the usual model.
The challenge to all breeders of livestock guardian dogs is to reflect on the character and origin of their breed. This will guide the future development and selection of the breed, hopefully to retain its unique characters. The uniqueness and predictability of all of these breeds can then effectively serve livestock owners as they search for a practical solution to flock and herd safety under a wide range of conditions."
D. Phillip Sponenberg, DVM, PhD Professor, pathology and genetics
Technical director, American Livestock Breeds Conservancy
Virginia-Maryland Regional College of Veterinary Medicine