BASICS OF GENETICS           Fred Lanting

It's a scary word to most people - akin to "statistics", "calculus", "tax return forms", and "down sizing". And "genetics" has its own subset of words and terms that tend to make the reader/listener stand further away from the writer or speaker as if the latter had lice or bad breath: "heritability index", "heterozygous", "penetrance", and "alleles", to throw out a few. Perhaps the best approach is to read through such an article in a somewhat cursory manner the first time, then go back to it a while later and try to absorb a little more each time. No one expects you to understand the biochemical nature of DNA if you have never taken chemistry in school (or taken it seriously). But if you are relaxed, the introduction can be much more pleasant. The chemical basis of genes and chromosomes is key to a full understanding, but you can skip over the parts that are "Greek" to you, and still get a lot out of your readings.

Even if you have never taken a course in chemistry, you probably know that a certain configuration of hydrogen and oxygen atoms in a molecule results in the compound known as water (H20), while another combination gives hydrogen peroxide (H202). When three or more types of atoms are joined together into molecules in all their different ratios, positions, etc., a more complex structure results. Carbon, hydrogen, and oxygen atoms, for example, are arranged in innumerable ways to form a vast variety of carbohydrates, fats, and, with the addition of nitrogen atoms, proteins.

One classification of carbohydrates is the sugar structure. Although sugars include many compounds, they are all somewhat similar in their C-H-O linkage. Deoxyribose sugar makes up part of deoxyribonucleic acid (DNA) along with other molecular structures containing nitrogen and phosphorus. The DNA molecule looks like a spirally twisted rubber ladder with the rungs alternatingly representing phosphate groups, purines, and other collections. DNA is of interest here since it is the substance of the little unit called the gene. Even smaller units called mutational sites are, in essence, mini-genes, and there may be up to several hundred of them in one location (in one gene). These mini-genes may explain why no two noseprints are alike, even though the general physical shape of the nose may be controlled by one or a few genes.

The coiled genes lie on or compose the larger structures called chromosomes, strands of protein-like material found in every cell. When cells divide, chromosome pairs may become visible (though often with great difficulty are they recognized), most of them resembling a box of irregular staples spilled on the floor. One of these chromosomes is shaped like the letter X and obviously is known as the X chromosome. If there is a pair of X-chromosomes, the animal is female, but if there is only one and the other is a Y chromosome, the mammal is a male. Each chromosome has a partner, being twins in the respect that they are shaped about the same (except for the X and Y pair), yet these partners may have slightly dissimilar genes on them. Since chromosomes exist in pairs, so do the genes, which make them up. This is an important fact that you should fix in your mind so as to understand later comments.

It would be an oversimplification, and erroneous, to say that one DNA strand (gene) determines one characteristic, such as length of ears. Although I will use such simplification for convenience's sake, remember that actually all the genes in an individual may have some effect on each of the others, even though their influence sometimes is remote. Differences between dogs may be caused by changes in the sequence of sugars, phosphates, etc. in the DNA sets of bonded chemical molecules, or even half a molecule being lost. These chemical forces are felt far away in other parts of that gene, in the other genes that that chromosome contains, and perhaps even in separate chromosomes. This is known as the "total genotype" theory or concept.

Another school of thought holds that each chromosome is completely independent, and that genes on one do not exert known influence on genes in the others. The more I study dogs, the less I am attracted to that alternative theory. However, it does appear to be true that the further apart two genes lie, the less influence they have on each other. Conversely, the nearest-neighbour concept holds that the greatest mutual effect may be between two adjacent genes.

Location and Strength of Genes
We envision chromosomes as collections of genes up and down their lengths. Each gene has a specific location (locus) on its particular chromosome, and these loci are given arbitrary letter designations, with the letter sometimes standing for a key word in the condition's name or description. In referring to dominance/recessiveness, the letter chosen is often the initial of the dominant trait. Often, the capital is used for the dominant variety and the subcase for the recessive (weaker) variety of the gene occupying that locus. A characteristic governed entirely by one gene pair is called a simple Mendelian trait, and those traits determined by a larger number of genes are known as multifactorial or polygenic. Multifactorial is usually used when stressing the input of non-genetic environmental factors on the various genes for a trait, and multifactorial or polygenic traits are usually seen in a great continuous variation.

Two or more gene pairs may work on a specific characteristic, and they may be positioned on different parts of a chromosome, or a single pair may be involved, with one gene at a certain locus on one chromosome, the other gene (perhaps identical, perhaps opposite in effect) occupying the corresponding locus on the other chromosome in that pair. In the case where they are different yet occupy corresponding loci (same position) on the two chromosomes, one gene may exert a domineering effect over the other. The condition is known as being heterozygous if the genes are different, homozygous if they are identical or call for basically the same trait. The dominant allele (gene version) tends to determine the trait (hair colour, for example), although sometimes a hint of the recessive allele's "desires" will show through. Example: a Black-and-Tan (saddle-marked) German Shepherd who is carrying a recessive for solid black may have considerably greater amount of black, such as larger saddle, than littermates who do not carry the "black factor". By the way, the reasons most beginning lectures on canine genetics are weighted so heavily toward coat colour are that it is a simple ("Mendelian") trait, and it is easier to understand.

We say that genes on one chromosome often stay together, because we have observed that certain characteristics seem to be inherited in bunches. Such is the case when one pup in the litter looks and acts just like one parent and another pup is just like the other parent. A dog that resembles one parent or ancestor has many genes in common with that individual. Often, however, genes formerly found on one chromosome may become attached to another when the chromosomes split and divide, or may end up on an entirely different end of the same chromosome. These things happen through processes known as translocation and crossing over.

PIGMENT
Most black, brown, and yellow colorations are due to the presence of melanin-class pigments. The word melanin is derived from the Greek word for black, and is commonly used to refer to the two or three known chemicals similar to each other which produce the above colour families. It can be said that there are only two colours in the haircoat of the canine: black and yellow, all others being variations of those. Eumelanin gives the black and dark brown colours; phaeomelanin gives the light brown and yellow ones. Certain genes affect the expression of these two so that the dark melanin may appear as black or be diluted to blue or changed to a chocolate brown, and the various shades of gelbe may exist as yellow, tan, light to darker brown, cream, etc. These genes can be referred to as modifier genes. That red-brown-tipped "burnt" appearance of the coat that we see sometimes, especially noticeable in black dogs, can be a result of sun bleaching the melanin in the hair to some extent, more often is caused by too high a protein level in the food, but in many cases it indicates a partially otherwise hidden recessive modifier gene for dilution or lighter-than-normal pigment.

One of those modifiers of black pigment can make a dog "liver" or "blue", or even a combination with various names such as "Isabella" (or "fawn" in Dobes". Now, if the gene we call B is present, eumelanin will show up as black, even when paired with b, but if B is absent and only the double b is present, that "same" melanin will be expressed as brown (a.k.a. liver, seal, chocolate). That is, the saddle and other areas that normally may be black in the German Shepherd (or several other breeds) will be liver-brown instead. The causative gene will have no noticeable effect on the lower tan parts of the dog; a different gene affects the phaeomelanin which controls these tan parts while the eumelanin affects the nose, saddle, etc.

At the follicle, a hair bulb can manufacture more than one type of melanin, and can alternate production in such a way that some hairs, such as most of those on a sable dog, are dark-tipped, followed by a lighter midpiece and an even lighter base (or perhaps a dark base). There may be two shades of yellow: one reddish and the other cream, on one shaft. Sometimes the phaeomelanin is concentrated in the tip and the eumelanin in the base, though not often. The banding or alternating between dark and light sections results in a beautiful variety of colorations in many breeds, especially around the neck, withers, and shoulders.

Capital letters are used when a trait is believed to be dominant or "stronger" than the other genes which could occupy that same locus. The world allele is used to refer to a form of a specific gene. For example, B and b are the two possible alleles that could exist at the B locus in the color chromosome. One calls for black pigment in the saddle, the other calls for brown in the same place. If both members of the pair are B, the saddle (and/or other dark areas) would be black; if both were b, it would be brown or liver-coloured. If the cell had one of each (Bb), the dominant one (B is higher on the map than b) would prevail, and the colour there would be black. Different letters tell the reader that these are different types of traits and that the genes responsible are located on different portions of the chromosomes, but the initial system doesn't give any clue as to whether the genes discussed share the same chromosome or are on different ones.

The concept that genes operate in pairs is true in all cases but one. The exception to this paired existence occurs in gametes, or reproductive (sex) cells: sperm in the male, ova in the female. Each of these cells has only half the usual number of chromosomes, 39 instead of 78 in the case of the canine. When these two types of cells meet in conception, the result is again 78 for all the new individual's cells except those sex cells formed approximately at the same time the fertilized egg divides for the sixth time.

Long versus normal length hair is said to be a simple Mendelian trait, with modifiers affecting differences. Other traits are not so simple, since genes frequently act in concert in what we call polygenic traits. Even those we call simple are not entirely so, but enough so the result can be predicted with high accuracy. In these simpler traits especially, the amount of masking or dominance one gene has over the other may be incomplete; you may be able to look at the phenotype (individual's appearance) and fairly accurately guess as to his genotype (genetic constitution, what he carries as recessive).

Both the forces of nature and the whims of man can cause relationships between genes to change, and when the genes themselves seem to have been changed, the results are called mutations. Most mutants die, are sterile, or have no dominance and thus are swallowed up or lost in the normal population. The greater the mutation change, the less likely it is to survive. I have a hypothesis that the first "dog" (ancestor of wolf, modern dog, jackal, etc.) started with more alleles than we now see in any of them, and many dominant ones were lost in the dog to enable him to survive in a changing world. This played a greater part in diversification than did mutations, since different genes were lost in different parts of the globe over which the canine was dispersed. The wide variety of types we see today is due to the multitude of combinations that are still available and, to a lesser degree, the "plasticity" of the canid genotype - the tendency to vary over great time periods through minor mutations. Crossing over is the major mechanism leading to diversity and variety, but segments that are lost entirely simply do not reappear.

INHERITANCE OF OTHER CHARACTERISTICS
While not as extensively researched as colour genetics, other inherited traits are of equal or greater importance to the breeder. Polygenic traits can be broadly classified as more or less dominant or recessive, but it must be remembered that such traits are exhibited along a wide range of expression, and the dominance is seldom clear-cut. Remember that the more genes that are involved in a trait, the less effect any single gene has on the ultimate phenotype.

As with colour inheritance theories, many expressed here are based on valid and repeated observations, and the generalizations and hypotheses which follow are not based on isolated examples, yet in some cases not enough research has been done to warrant a dogmatic, near-100 percent positive statement. Excitability and auditory sensitivity often seem to be interrelated and also "connected" to other traits, and it may be that certain modifier genes affect more than one "main" gene, or that they are so close together on the chromosomes that it is not likely they would be separated as easily as others.

Behaviour
A trait which is difficult to attribute to one or a few specific genes is the dog's reaction to other dogs. Whitney generalized that German Shepherds seem to avoid a fight whenever possible. However, early selection for herding ability has stamped the German Shepherd Dog with a very strong protective or at least watchful nature concerning whatever he regards as his or his master's property or territory. A similar expression of behaviour is what the Russian investigator, Krushinsky, called "passive defence reaction" which others may call cowardice, as opposed to "active defence reaction" found in the feisty terrier breeds, Doberman Pinschers, hornets, and a few others that react in anger instead of withdrawal. This may be due to a different set of interacting genes, although insufficient research prevents a firm conclusion. Some breeds are notorious fighters (though there are always individual exceptions and matters of training): Malamutes, Akitas, Shibas, and several others have inherited this tendency and perhaps not so coincidentally are similar to each other in other ways. On the other hand, hounds are typically congenial, whether or not they have been raised and trained in packs.

Another trait that may have a separate genetic cause is shyness - a behaviour found in wild animals. When Krushinsky combined the shyness of wolves with the excitability of German Shepherds and other breeds, the result in the offspring was extreme shyness. When the Shepherds were bred to a native, less excitable Russian breed, the progeny were usually shy. Undoubtedly some "wolf-shyness" or wild shyness still exists in the Shepherd breed as a somewhat dominant trait, and breeders should continue to select against it despite the need to preserve the useful excitability nature. The shyness/boldness factor may be highly complex and multifactorial.

One hypothesis is that shyness is an incompletely dominant tendency and therefore can be bred out to some extent by breeding those dogs that exhibit the more recessive, non-shy characteristics. However, for some working breeds, care must be taken to prevent the development of a strain of deadheads. Although that's not likely to happen, a general rule to follow in such breeds is: Don't double up on shyness, or "soft" character. Breed an otherwise good specimen to a mate with unquestionable courage and character (even if it is mostly a result of training), and cull the less confident progeny out of your breeding program. It may take a couple of generations to correct soft temperament, and regardless of how long you try, you may continue to get an occasional "flaky" dog.

An apparently different characteristic is gunshyness or, as E.S. Humphrey and L. Warner, authors of Working Dogs, (Johns Hopkins publisher), referred to it, "auditory over sensitiveness." They suggest a polygenic cause, and indicate it to be not dominant in nature. Dominant traits are relatively easy to eradicate, but the very nature of recessive traits (being hidden from view) means that they may lie undiscovered for many generations, until a partner contributes another such allele and produces a homozygous individual. This means that such "nervous reactions ... to strange sounds and sights," as deplored in the GSD's AKC Standard, may be hard to breed out to a low enough level to be satisfactory.

The Head
Ears
Humphrey and Warner, who extensively studied Shepherds in the 1930s, found erect ear carriage to be dominant over faulty carriage, but the extent of dominance seems to be influenced by other genetic factors. In fact, several genes are probably involved, as ear carriage variations are numerous, even within one litter raised under the same environmental conditions. The veterinarian, breeder, and practical geneticist, Dr. Leon Whitney, found in crossing various breeds that lop-ear carriage was dominant over the erect ear. But remember, he was looking at crossbreeds, not strains within the German Shepherd as did Humphrey and Warner. I have seen many Shepherd-Dobe crossbreeds with hanging ears but I have also seen a few with one or two ears nearly erect, or up however weakly, so I would postulate the lop-ears (natural carriage in Dobes) is incompletely dominant, that it has partial penetrance in crossbreeds. Breeds with greater neoteny (more juvenile, "friendly" appearance) and further from the ancestral wolf type usually have much more soft-ear characteristics (Saints, scenthounds, spaniels, retrievers, etc.), and their ears are generally larger in proportion than are those of the more primitive breeds.

The genes for ear carriage
The genes for ear carriage in the breeds with erect ears may be different from those in other breeds. I have seen Shepherds with semi-erect (weak) ears produce whole litters of puppies with completely normal ear sets. This might indicate weak ears being recessive to fully erect ears without negating the possibility of erect ears being recessive to the hanging ears found in other breeds. Marchlewski also found that in purebred German Shepherds, the correct ear carriage was dominant over poor ear carriage. This means GSD fanciers have been breeding too many recessives to each other, as there have been a larger number of weak ears in recent generations. Spitz breeds, on the whole, may be slightly closer to the ancestral type than is even the German Shepherd, and in those we find smaller ears, which fact contributes further to firmness of carriage.

Eyes
As dark an eye as possible is recommended by many breed standards, and the dark brown eye admittedly contributes to a pleasing expression and therefore a favourable impression. In some breeds, it is purely a matter of convention, such as the desire to have Shepherds not look like their light eyed wolf cousins. Two or three shades are common to the breed; they range from the preferred and dominant dark brown eye to the yellow (actually light brown) eye. Since the light eye appears to be recessive, it will probably be in this breed longer than its devotees will. However, it appears the majority of dogs being shown today have either the dark or the intermediate shade, which latter might suggest many with a Yy constitution, with Y representing the dark eye and y the light eye. The recessive y allele in such a heterozygous dog may give the same kind of partially hidden, partially revealed trait as the B&T pattern often does in a heterozygous sable GSD. In many breeds light eyes have been selected for (by eliminating the dominant dark eyed individuals), and in some the light eye is unavoidable because eumelanin expression has been totally bred out: Weimaraners and liver-pigmented dogs, for examples.

One thing that makes it difficult to guess at the genetics by looking into a dog's eyes is the possibility of modifiers acting on the Y and y genes, but it seems more probable that the background colouring around the eye throws one off; a solid black dog's eyes might not look as dark as the same shade would on a lighter-colored face. Even if the two dogs are placed side-by-side, you may still have an optical illusion. Modifier genes may have much to do with a gradual transition through several hues, and in some breeds, such as Rottweilers, breed survey managers can use commercially produced eye colour guides like marbles set in a box that can be held up to a dog's face to determine the degree of darkness.

Whitney discovered that Bloodhounds with yellow eyes were also always liver-and-tan coloured dogs. Other than a possible linkage with the liver (bb) or blue (dd) genes, there doesn't seem to be any connection with coat colour. The amount of black is not a factor, just the very presence is. I have observed many dogs with relatively little black in the coat (light sables and small-saddle B&Ts) having the darker eyes, but I've also found the obverse to be as true. Possibly the high percentage of dark dogs with dark eyes found in the show ring is due to both characteristics being selected for separately by the breeder, rather than eye colour being linked with coat colour. Or a less-likely explanation might be that the two traits are governed by genes on opposite ends of the same chromosome and that there is some crossing-over or translocation involved. (Crossing-over involves genes on the same chromosome.) This translocation refers to the chromosomes sticking to each other in such a manner during division of the sex cells that the bottom half of chromosome #1 gets attached to the top half of #2, and the bottom of #2 ends up with the top of #1. Light eye is not a problem of any significance in any breed other than aesthetic consideration, but knowing it to be recessive will enable breeders to handle it in their own breeding programs.

Teeth
Standards most often call for complete, normal dentition, but there are many exceptions, the abnormal almost always being recessive in nature. This is why you may occasionally find Shepherds with undershot mouths, but not Boxers with overshot or scissors bites. Indeed, Humphrey and Warner found full dentition to be dominant over missing teeth, and theorized the involvement of more than one gene. Yet the doubling up of recessives has produced many a German Shepherd Dog with missing teeth. That this is not a Mendelian trait is shown by the fact that if you breed a dog with one missing premolar to another with the same tooth missing, sooner or later that pair's progeny or line will have several missing teeth, not just that one. This is slightly similar (and perhaps the mechanism is, too) to the phenomenon of breeding slightly dysplastic or "fair-normal" dogs and coming up with more dysplasia and greater severity than in the parents, as well.

An overshot mouth
An overshot mouth is definitely inherited. While there may not be enough research as yet to determine the mode of inheritance, or whether it is due to the upper jaw being too long or the lower jaw too short, as a long-time breeder I feel the latter is the case. In my own experience and in observed breedings of others, I have seen evidence that the defect occurs because the lower jaw, which often grows at a different rate, fails to catch up with the upper jaw. Most investigators agree that overbite is a recessive trait, probably with modifying factors, which is to say it is polygenic. In fewer cases, the arrested growth of the upper jaw allows the lower teeth to exceed the point of a scissors bite, thereby producing a level bite. It sometimes occurs in litters whelped by bitches that have produced overbites. It has been my untested opinion, supported by other breeder/geneticists, that a level bite is caused by at least some of the same genes that cause overbite and possibly even an undershot mouth; namely, a disruption in the normal growth rate of the lower jaw.

In a true level bite, the four centre incisors on the top meet, edge-to-edge, with their four opposing ones on the bottom. The incisors next to the canine teeth are seldom involved. Often, the two centre bottom incisors are level with the top ones, but the next two are normal (scissors). I make it a practice not to consider this a truly level bite when judging in the ring, though I do take notice of the condition as being less than ideal. I believe many such bites are caused by other genetic factors calling for an overcrowding of an already narrow jaw. Such a jaw is not as wide as should be, and in the Shepherd it is derogatorily called a "Collie mouth". Usually a decrease in underjaw depth accompanies this.

The Neck
The condition of loose folds of skin in the throat area is known as dewlap. Marchlewski in 1930 reported dewlap to be clearly dominant, but the type of dewlap seen in Bloodhounds and other breeds probably is genetically different from what is seen in the German Shepherd Dog. I believe that breed's minor or occasional throatiness in the big-boned, full-coated, loose-ligamented dogs with probable upland heritage represents a polygenic trait.

The Back
At first glance into many breed standards' wordings, there seems to be two opposite demands pulling at the breeder and the length of the dog. A relatively long neck is often called for, as is a fairly long body in general. The preference for a short loin on a long dog may seem to be at odds with the genetics producing length, but is not impossible to achieve. The long appearance can be achieved by shorter legs, deeper chest, longer vertebrae, or a combination of these. But if we breed for a long dog by selecting for longer vertebrae, we should also be sure to breed for good shoulder layback and broad thighs under a long croup, or the length will all be in the true back (that portion between the withers and the croup), which is usually undesirable. The length of the pelvis seems genetically less related to the lengths of the vertebrae, which probably means that the genes controlling each are not right next to each other, and can relatively easily be selected separately. A less often seen reason for a short loin is an abruptly shortened ribcage, the opposite being the norm, one that is carried well back and tapering more gradually. If the exhibitor and breeder want the impression of a short loin on a dog with a long back, it would best be accomplished by plenty of roadwork and adequate feeding after the dog reaches maturity. Unless you fatten the dog excessively and get rid of most of the waist, you can accomplish much by building up the muscles and the coat there and still keep the dog in good health.

Shoulder
The shoulder layback, or slope of the scapula viewed from the side, may be related to the lengths of the individual vertebrae, by the nature of attachment of the muscles. The scapula is always attached via those muscles and ligaments to the same set of cervical and thoracic vertebrae. Genetic research indicates that layback appears to be a polygenic trait. I believe the differences in layback between one dog and another in the same breed are due to growth rates of muscle fibres and bundles, some of which exert a steady pull on the developing bones during maturation (even in the embryo), a pull that is different in intensity or location from dog to dog.

A much greater consideration in upright forequarters is the humerus (upper arm). There is far more variance in lengths and angles of upper arms than in shoulder blade angle from dog to dog. And between the two, bone length determination is far more of a factor than is the upper arm's angle from the vertical or from the scapula. The many minute variations from dog to dog give ample evidence of the polygenic nature of such traits.

If it is true that most "correct" traits are at least slightly dominant, then breeders must be selecting for upright fronts, judging from the large proportion of faulty assemblies seen today in so many breeds. One explanation is that these breeders think they're only selecting for high withers, not realizing that the highest "withers" are obtained by locating the shoulder blades' tops right behind the ears, this being accomplished by "selecting by neglect" for upright forequarters. The other explanation is more forgiving, acknowledging that it is a difficult goal to reach, that of getting and keeping good front angulation. The reason should be apparent to anyone who understands the principle of the "bell curve" in which most characteristics or statistics are found in the middle section of that curve. Such a distribution is seen in rear angulation, but is not the case with the front angles. With the rear, the ideal is in the middle of the spectrum that runs from very straight stifle to very bent; but with the front, the ideal is on one end of the spectrum. When you randomly breed two dogs, the tendency is for most of the offspring to have a phenotype closer to the middle than the parents' type or characteristics. That is why we have no trouble getting moderate rear angulation in GSD’s, but the middle of the curve for the front is not the most desirable structure. Many American breeders have long consciously selected for the extreme rear with its longer lower leg, but not as diligently for the "extreme" front with its maximum humerus length. Most Rottweilers in the quarantine-shackled British Isles have notoriously bad (upright, east-west) front assemblies because of decreased genetic diversity and negligence in attacking the problem.

Rear Angulation
In most breeds there is appropriate emphasis in Standards and efforts on selecting for a good angle in the shoulder assembly, with the need more understandable in the trotting breeds than in the coursing or digging or fighting heritage ones. The inheritance of rear angulation is much more complicated. Before going further, let me redefine what is meant by angulation: it does not refer to the slope of the topline. It refers to the sweep or bend of the stifle area (knee), the angle made by the departure of the line or direction of the lower thigh from that of the upper leg, the tibia from the femur. Very little difference is seen if this area is straight, as in a Chow Chow; much greater in the "American Shepherd". This angulation is created mostly by a lengthening of the tibia and fibula, out of normal proportion. That is to say, too far from the ancestral type that is ideal for endurance and efficiency. Again, the unbroken spectrum of different lengths points to the fact that this too is a polygenic trait.

I hold to the hypothesis that correct structure tends to be dominant, and that with the least amount of human selection, this correctness will be seen to approximate that seen in the wild. Most wolves, jackals, and related wild dogs have much less variation in rear angles than do their German Shepherd cousins. There is another principle hinted at earlier, which is that in random breeding, characteristics "tend toward the norm". This is as true of human I.Q as it is of canine rear angulation. That means that unless selection pressures continue to favour the extreme rear angulation, succeeding generations will have the centre of the bell curve shift toward that found in the wild. Since the German Shepherd Dog is closer to the wild form than is, for example, the Chihuahua or Boxer, American breeders must have been purposely selecting characteristics further and further from the wolf/ wild dog type by concentrating on recessives and mutations. The theory of general dominance of correct structure presupposes that moderate angulation is more correct than either the straight stifle or extreme rears, and general crossbreeding (or outcross breeding if within one breed) would tend to bring the particular dog population toward the centre. If two dogs described as too straight in the hindquarters are bred, extreme puppies are possible but not probable; a greater number of extremely angulated offspring are produced by dogs of moderate angulation because they carry more genes for that end of the range than do the ones lacking rear angulation.

Gait Inheritance
In the typical show ring, too much emphasis is placed on movement, as if judges are commissioned to make all breeds fit one trotter-dog mold. Instead, we should be concentrating on the historical function and utility of each breed separately. In those once required to earn their keep by trotting or some other type of movement, inheritance of tightness and firmness is important; the dog should be "fest und trocken" (firm and dry). Loose ligaments, sloppy hock action, and an overall imbalance due to less angulation in front than in rear are genetic failings that the judge or performance requirements should be weeding out continuously. The dog bred for extreme rear angulation may easily be a poor mover both coming and going. Gait is a matter of inheriting the right (or wrong) polygenic traits on both ends of the dog. Gait is also affected by the length of the spinal column, especially between the withers and the pelvis, as well as by muscle tone, ligaments, length of limbs, other angulations, balance, and orthopedic disorders, all polygenic. All are inheritable, but the mode of inheritance of each is not always fully understood. Another hazard exists in the dog bred for a long back: a loose middle which at certain speeds bounces up and down like a balloon filled with water. While not the only cause of poor backs, an inherited propensity for loose musculature and loose ligaments combined with a long back can have a devastating effect on side-gait.

The condition of gait known as "out at the elbows" is marked by the dog's front legs not moving in a straight line. The elbows move in and out and the forelegs swing forward in a side-winding arc. One cause can be one of the elbow dysplasias, but it is not the only culprit. The pinched-at-the-elbows look, which usually is accompanied by "east-west" feet, seems to be recessive to more normal chest width and depth, but remember the variation that can exist with polygenic traits and the principle of "selecting by neglecting".

Feet
It appears that a compact foot form (but not necessarily or only a cat-foot) is dominant over the hare foot and the paper foot, and moderately short toes (the final phalanges) dominant over long toes. It is possible that each of the characteristics that go into the appearance of feet is controlled by one pair of genes, and that the combination of traits plus environment gives a variation in phenotype that reveals polygenic conditions in complexity. If, for example, a dog were simultaneously heterozygous for short toes and heterozygous for weak foot ligaments, his feet may have a different appearance than those of another dog with the same gene combination because of other genes' interaction or a difference in their environments.

Let us consider for the sake of illustration that a splayed, paper-foot condition is the result of recessives. If the symbol S is used for dominant short toes, and s for recessive long toes, and if further we give compact feet the symbol C and splay feet the designation c, let's see what might happen if we put them together in different ways. A dog that is dihybrid may have SsCc and have a moderate paper foot. Another dog that is homozygous for normal short toes but is also homozygous for weak foot ligaments may have the genotype SScc. Yet these dogs may have the same apparent phenotype to the average observer. If you breed together two dogs that are Ss, you can expect about 25 percent long-toed dogs, 50 percent close to normal, and 25 percent short-toed. It may not be too risky to breed a dog with less than ideal toe length (ss or Ss) to one with slightly splayed feet (possibly Cc), but I personally wouldn't want to. Each condition or genotype is undesirable in a different way, and the resulting litter may not be large enough to yield any pups with the combination of good qualities from each gene locus. I would think long and hard before using a dog with poor feet in my breeding program, yet would rather have a dog with long but compact toes than one with splayed toes. In some breeds such as Toys, this may not be as important as in the working, herding and other breeds.

Complicating the analysis is the possibility that splay foot may be a polygenic trait, as most loose ligament conditions seem to be. But, in my opinion, the number of such genes is probably very low if not actually Mendelian. This conclusion is based on simple observation of many, many cases of linebred dogs I have owned or handled. The front feet show the genotype possibilities and variation much more than do the rear feet, so in choosing breeding partners, look closely at the front feet of the dogs in question as well as those of their parents, if possible. If ligaments are weak in the feet, be sure to check hocks and back, too, as it appears there may be some linkage or commonality of genes controlling ligament condition over the whole body.

Pasterns
I suspect that the correct pastern is both polygenic and generally dominant over a greater (weaker) angle, and that a relatively small number of genes are involved. It may be that the genes for a stiffer, terrier-type upright pastern have for the most part been lost from the herding dog's gene pool by early selection for endurance in sheep-tending dogs. In a few breeds, most infamously the German Shepherd, the weak 35- and 40-degree or worse pasterns in abundance today are probably the result of linebreeding on individuals of that phenotype or carrying the recessive polygenic factors. Besides heredity there are many other minor but noticeable contributors to poor pasterns: nutrition, exercise, and footing; one or two of these may affect the feet at well. Another one, perhaps rare, is a specific disease which shows up in completely flat pasterns, the entire forefoot (equivalent to your hand) being flat on the ground. I've known of this in the GSD and the Malamute, and have heard of it in other breeds. Although there is a tendency for many people to blame environment, I've found that almost all cases of poor pasterns have been traceable to heredity, not to poor footing (the type of surface on which the dog is exercised or kenneled).

Dewclaws and Extra Toes
Dewclaw is a term rightly applied only to the nail and toe of the first digit of the hind paw, not the front paw where such a digit is always present at birth, and considered normal and useful. The true dewclaw is carried so high on the hock that it is really a metatarsal digit rather than a phalange or toe. In most breeds, dewclaws are not appreciated, but also not severally criticized. Dewclaws are occasionally seen in many breeds where they are considered abnormal. The condition could be caused by two or more types of genes, on two different loci, perhaps even two separate chromosome pairs. A number of studies on the inheritance of dewclaws give conflicting results. Whitney explained that in some dogs, the dewclaw is present beneath the skin and other tissues and does not protrude enough to be seen. There could be a gene pair for dewclaws and another pair for hiding/not hiding. Dewclaws are "required" (by fancy, not function) in the Briard, Beauceron, Great Pyrenees, and others.

Polydactyly, the condition of having extra toes, is sometimes applied to the presence of dewclaws, though I think it should be recognized as a separate entity; it is more correctly used when extra toes are found in the normal footpad location. It has been reported that polydactyly is a recessive trait in Collies, but is expressed in other breeds and in humans via dominant genes. Different genes may be involved in some breeds than in others. I've seen many cases in cats and other mammals, but none in most of the breeds I've studied to date. One breed that I judge, the Norwegian Lundehund, or "Puffin Dog" is required by its breed standard to have those extra toes.

Tail
A long (and often bushy) tail, dubiously claimed by some to be useful in maintaining balance when swimming, herding sheep or turning at a gallop, but certainly helpful in presenting an attractive picture in the show ring, is required by many Standards to at least reach the hock. I have found that palpating the tail to count the number of vertebrae is highly inaccurate, but I know that most Shepherd Dogs have twenty to twenty-two tail bones. The length of each vertebra and the amount of legginess the dog has will also contribute to the appearance of tail length or shortness. In the wild, the wolf, coyote, Arctic dogs, and other canids usually have a tail that barely extends to the hock. There are conflicting data on tail length inheritance within many other breeds and breed crosses. The real reason for a bushy tail is to provide protection for the nose from snow, rain, frostbite, dust, and insects while sleeping. It follows that long-coated dogs have thicker, even bushier tails. The wolf or coyote tail is also carried straight out or hanging, while in many dogs a sign of domestication is the higher carriage, easily wagging tail. In many "very" domesticated as well as some primitive breeds, the tail curls tightly over or lies flat on the back.

SUMMARY Several genetic factors have been reviewed in the preceding paragraphs. In some cases, little or no scientific study has been applied to the phenomena discussed, and in these instances I have offered my opinions based on many years of observation. Remember that statements on inheritance may sometimes be based on limited data, and the more reliable the information the better the chance of accuracy. Contributions from readers are always appreciated, as breeders are as rich a resource as veterinary schools, sometimes more so. More information on the nature and specifics of inheritance will someday be available through the processes known as DNA typing, and identification of "marker genes" that give a clue to the physical presence of undesirable (usually but not always recessive) genes. But that day is many years away for most hereditary disorders, a little closer for others. For a discussion of heritability, which term is much misunderstood by the general dog fancier population, and which does not mean the same thing as "inherited", see my 1998 book on orthopedic disorders with emphasis on hip dysplasia (in process of completion at this writing) or watch for a follow-up article in these pages.