The German Approach to Breeding
The approach taken by Germans towards breed development rests on a system of standardized tests that are used as the basis of a performance based breeding system unlike that in United States. Although this approach is often taken when breeding livestock, it's not the approach used in America in the sporting field which tends to focus on the champion/line-breeding approach. Here, I might mention, for instance, dog field trials or the competitive approach to breeding based on champions that has led to our excellent thoroughbred horses, which excel at the specialized task of running a race, usually over a relatively limited range of distance. However, the German approach, which has produced some of the most world famous dog breeds, when applied to horses has produced the finest dressage horses in the world; horses that are capable of a variety of complex equine tasks.
Indeed, the testing approach used by the Germans goes along with a gene pool and breed philosophy quite distinct from that associated to the field trial competition approach. While the details may differ from breed to breed, I can put forth the ideas that guide Mr. Erwin Wallmann, the chairman of the breed committee for KlM e.V.. Although I cannot go into the statistical details underlying the German approach, I think I may be able to give the reader an overall grasp of the philosophy, and point him/her toward some of the important consequences this approach has with regard to breed development.
The central goal of the German breeding philosophy is to maintain a broad gene pool with as many dogs as possible in which all dogs are carefully tested and measured for as many characteristics as the Germans can develop tests. All breedable dogs are then sorted on all these characteristics according to statistical techniques similar to those used in livestock breeding such as BLUP (Best Linear Unbiased Predictor). This is not just a one-generation process; rather the Germans factor in the results of their methodical record keeping over about five generations. The goal is to use “mix and match” breedings based on various characteristics in an attempt to improve over time the average value for all tested characteristics and to limit as much as possible the negative variations in desired characteristics. They are not just interested in the "one in a million" exceptional champion; their focus is on the entire breed.
Let’s begin with the use of tests rather than competitions. The goal of a test is to develop a protocol from which one can assign a fixed number to the outcome. Wherever possible, variations introduced by human judgment are to be minimized. When this seems impractical, the Germans have set about to develop a corps of judges who work together to produce a numerical result that is as reproducible as possible.
The standardized testing procedures used to support the German performance-based approach to breeding hunting dogs are under the jurisdiction of the Jagdgebrauchshund Verband, or JGHV. Most of the hunting dogs commonly referred to here in the U.S. as the Continental breeds as well as the American and English hunting dog breeds (that have a breed club in Germany) are all under this German umbrella organization. As to the standard testing procedures of the JGHV, it is my understanding that there are a variety of standard tests set forth which are “breed independent”, in that the score of any of these tests at JGHV sponsored tests will be accepted by the appropriate breed club. However, while there is considerable overlap among the tests required by the various Continental breed clubs, each club has its own set of required tests, and some breeds, particularly the Deutsche Kurzhaars, have quite a few breed specific tests administered outside the purview of the JGHV (see, for example, http://www.dk-weser-bremen.de/gsp_in_english.htm ). Thus, for instance, the tracking of the hare test—considered the “best” upland game recovery test—is required for Münsterländers and Deutsche Drahthaars, but not for Deutsche Kurzhaars (a slide of Mr. Wallmann starting out on such a hare tracking test with two hare for two dogs appears on the slide show). I’ll return to discuss more about the JGHV later.
The tests administered in Germany generally are performed at three stages during the development of a young hunting dog. The first test, Verbandsjugendprüfung (VJP), is performed during the first spring after the puppy has whelped (if the puppy whelps after October 1, the test process begins the following year). The second test, Verbandsherbstzuchtprüfung (HZP), is performed the following autumn, while the final test, Verbandsgebrauchesprüfung (VGP) has no special time limits. There is no retesting allowed outside these time frames, although testing a dog twice is permitted within the given time frame. Only dogs that have taken both VJP and HZP tests may be used for breeding. A passing score (Prize 1,2,3) in a VGP test qualifies a dog for registration in the German Register of Utility Dogs (DGStB). There are other tests as well, such as participation in the National HZP and VGP tests as well as various specialty tests, most of which are related to the hunting of animals rather than birds.
The use of multiple tests with numerical values to characterize the performance of a dog is essentially different from a non-numerical competitive approach to evaluating a dog. While I’m not aware of any German tests in which braces are used to provide a score (as is done in the NAVHDA Invitational with regard to honoring), there is nothing to preclude that type of test. It seems, however, that the German approach to breeding hunting dogs does not even consider statements such as “my dog” can find more game than “your dog”, or “my dog” has better style than “your dog”. As an example, here is an edited excerpt from the Altmoor Drahthaar website:
“Q - I'm a very
competitive person and would like to be involved in many different types of
competitions where my dog and I can show that we're winners. Is the DD for me?
A - Probably not. The hunting tests upon which the breeding system is based are
evaluations, rather than competitions. While we're all proud of a high score
when it's earned, the dogs are there to be rated against an objective standard
in a number of areas. No attempt is made to rate one dog in comparison to
another. In most cases a dog can be evaluated in a given type of test no more
than twice - what the Germans refer to as "sport handling" is discouraged;
competition is not really what the DD is about. The bottom line is that we are
hunters and breed the DD for other hunters, and we consider these other types of
activities to be simply a prelude to, or a means of training for, the real
thing.”
Now, returning to the philosophy of the German breeding program, once you’ve gotten by the front end task of establishing tests for measuring a dog, you end up with a set of numbers for each dog. For some breeds today, this is already 12 or more numbers per dog. This is much too complicated a situation to illustrate some of the basic principles involved, so let’s use an example with only two test numbers per dog:
In this fictitious sample made up in Microsoft Excel, we have the points for 60 dogs, 30 males and 30 females that have been tested for “Nose” and “Style” with values ranging from 0 to 10, with 10 being the highest possible score achievable. The average value for all the 60 dogs is about 5.245 for nose and 4.75 for style, i.e. the dogs are presumably overall a little better at sniffing than style. However, if you look just at the male dogs, the average is 4.99 for nose and 5.01 for style, while for females the average is 5.5 for nose and 4.49 for style. Now, suppose you’re most interested in style. Then, for example, you might want to only consider dogs with a score of 5 or better in style; in that case, the average for nose is only 4.02 for males and 4.91 for females, while if you focus on dogs with a score in nose of 5 or better, the style average for males is 3.78 and for females it is 3.9. In either case, if you concentrate on one quality, you tend to lose something on the other.
The potential breeder can play with several different breeding strategies in this make-believe example. However, overall, it’s reasonable to expect that despite differences in breeding strategies, the averages for both nose and style will increase with time as breeders “mix and match” different combinations while only choosing to breed puppies that conform to this breeding philosophy, which coincides with that of their "market"--subject to other conformation market forces, such as no white on the face for Small Munsterlanders. In addition to the upward drift of the average, the spread of the scores about the average value for each test (depending on the statistics, the spread is related to the standard deviation of the scores about the average) will tend to diminish, i.e. the scores will tend to cluster more tightly around the average value. The amount of the spread is influenced not only to the connection between the canine genome and the test protocol (as well as conformation protocols), but also to the number of dogs in the gene pool. However, all of these factors are subject to a statistical phenomenon similar to the "uncertainty principal" from quantum mechanics; if you multiply together the deviations associated to the qualities for which you are testing, the resulting number can never get less than a certain amount. That means that if you focus too much on style (in this example), the variation in the nose quality will necessarily become large.
In the example I have given, there are only two types of test values, but as the number of test types increases, the theoretical number of dogs required to achieve the same degree of clustering increases exponentially, so that for real life situations, many compromises as well as statistical techniques need to be applied to speed up the improvement process. One corollary of this fact is that popular breeds with large numbers of dogs, such as German Shorthairs and other AKC recognized breeds, if subjected to the German breeding program would produce dogs of more uniform quality.
This is a brief summary of the ideas behind the German breeding program. To make this breeding program work: aside from good data, you need a large and diversified gene pool in which you can use dogs superior in some traits to try to improve others. In this approach it’s important to recognize that the spread of values is required in order to maintain the health and vigor of a breed; indeed, the German approach puts a methodology to the phrase put forth by dog geneticists that “you should breed for the whole dog”.
