Eine wichtige Anmerkungung zuvor: Der nachstehende Artikel ist 1987 verfasst worden. Damals war der Unterschied zwischen weiss-schwarzem Neufundländer und Landseer noch nicht so klar umrissen wie zur heutigen Zeit. Nichtsdestotrotz können Sie sich bei der Erwähnung der Rasse Landseer ebenso den weiss-schwarzen Neufundländer hineindenken, denn der genetische Ablauf ist derselbe.

---

Genetica 80: 115-128, 1990.
© 1990 Kluwer Academic Publishers. Printed in Belgium.

The inheritance of the piebald spotting pattern and its variation in Holstein-Friesian cattle and in Landseer-Newfoundland dogs

H. Pape
Institut für Geophysik, Postfach 1253, D-3392 Clausthal-Zellerfeld, FRG

Received 10. 11. 1987 Accepted in revised form 23.9.1989

Abstract

The black and white spotting patterns of Landseer dogs are divided into qualitatively recognizable phenotypic classes. Breeding data were obtained from the Swiss Dog Stud Book (SHSB) and from breeders' recent records. A plausible interpretation assumed qualitative inheritance of the generally accepted piebald spotting gene s^p₁ with at least two modifiers, s₂ and s₃. The modifier genes are regarded as minor spotting genes and may be responsible for white markings in the related Newfoundland breed which has been cross-bred with Landseers.
The proposed scheme of polygenic inheritance can also be applied to the piebald spotting pattern of Holstein-Friesian cattle, using breeding data from literature.

Introduction

White spotting is a common observed trait in birds and mammals, especially in domesticated species. lf spotting is present, the propagation, migration and differentiation of melanocytes has been disturbed so that the white areas lack melanocytes. Whether melanoblasts, the melanosome-free precursers of melanocytes, are present depends on special circumstances. The table of genes in the laboratory mouse of Green (1975) contains 18 loci with 36 genes causing white spots, white hairs or complete white fur from deficient melanocytes. It seems hopeless to guess homologies of spotting loci between different species purely on the basis of the black and white ratio of phenotypic expression and the dominance behaviour with respect to the normal gene.

Little (1957, 1958) explained recessive spotting in dogs with an allelic series:

S	self coloured
si	Irish spotting (white head blaze, breast, paws and tail tip)
sp	piebald spotting (coloured plates, separated or confluent)
sw	extreme-white piebald spotting (colourplates reduced in size and number or even absent)

This series has been established mainly on the assumption that the well-investigated hooded series of the Norway rat should be present with homologous genes in dogs. As the phenotypic variation of the four alleles will be extremely wide due to the action of unspecified modifiers, Little gave an explanation for each result of his reported matings, but there is no evidence that, if we accept s^padditional factors sⁱand s^w existin the same series.
For the present, one should be cautious and confine oneself to the postulation of a single mutation s^p₁for piebald spotting at the locus S₁. As genetic analysis, including additional genes, is more the objective of studies of qualitative than of quantitative inheritance, the qualitative traits of spotting should be recognized. The key for this is the characteristic pattern of each type of spotting. From the point of view of pattern comparison, piebald spotting in dogs with its typical distribution of coloured spots poorly fits the pattern of the hooded rat, whereas in other cases there will exist comparable patterns in different species.

The pattern of piebald spotting in dogs and cattle

Fig. 1. Examples of piebald spotted phenotypes in Landseer dogs: above - dark (D) or 'Mantel' class, middle - medium (M) or 'Iarge spots' class, below - light (L) or 'restricted spots' class.

Fig. 2. Examples of piebald spotted phenotypes in Holstein Friesian cattle: above - class D, middle - class M, below - class L.

The piebald pattern is composed of solid spots which may join one another. In every case where they form a border line against white, the boundary is composed of convex arcs (Figs. 1 and 2). From this shape it can be concluded that each spot has grown radially. In piebald dogs, a postnatal progression of the solid areas can be observed. This general property is common to a group of spotting patterns in contrast to another group, to which belongs spotting in Nubian goats, where the boundaries are convex with respect to the white areas, which result from a degeneration of melanocytes. Consequently the amount of white area may increase with age in such cases.
The formation of isolated pigment centers represents a normal developmental stage during embryonic ontogenesis, which leads to the selfcoloured condition before birth. In the mouse fetus, Shaible (1963) determined two medial and six pairs of lateral regions where melanoblasts converge. This pattern resembles a so-called demixing pattern which appears in all fields of nature, where a homogeneous phase separates into two different phases, for instance when a solid phase crystallizes from a solution. In such a process the first step is the formation of germs. That is important for a clear understanding of piebald patterns, one cause of which is the hindrance of germination, so that several pigment centers are omitted. The distribution of pigment centers reveals a certain regularity, insofar as they keep a special distance from their neighbours. That is why each germ is surrounded by an aureole, which is infavourable for further germination. For this reason the number of primary pigment centers is limited. In the white areas of a piebald coat, a second generation of germs may occur at a later stage resulting in 'ticking'.
The second step of pigment spot formation is lateral growth. This process can also be hindered with the result that the white areas between the spots remain partially uncovered by the coloured , phase'. A common misunderstanding assumes that the white areas are not invaded by melanoblasts in any case. Although there is a deficiency of melanoblasts, this region may contain several of them which become evident on secondary germ formation.
The germination of pigment centers and the growth of spots influence each other. As germination is hindered near the boundary of a plate, a larger number of germs than normal can be created if growth is retarded. Ort the other hand, rapid growth may be combined with a reduction of germ number. Usually, germination and growth are simultaneously reduced in some way. Generally, the location of pigment centers depends on the idiosyncrasies of each spotting pattern. Therefore Allan's famous centers of retreat (Allan 1914), which describe the most conservative spots for the whole range of variation of piebald patterns, have no general meaning, but solely give an insight into special spotting types.
The piebald pattern of Landseer dogs, which is much the same as that of Holstein-Friesian cattle, will be described in terms of germination and growth of spots. The variation reaches from almost solid coloured phenotypes to nearly completely white phenotypes. Even in the dark types, the lower parts of the legs are white with more extension of white on the hind legs.
For all types of variation the anterior part of the body, especially the head, reveals a relatively larger portion of coloured area than the posterior part. In other words, the white areas increase downwards and backwards. This recognition can be compared with the migration path of melanoblasts, which originate from the neural crest (Rawles, 1940, 1947). The distal parts of the limbs are farthest away from the starting line of the process. It is also necessary to take into account that embryonic development proceeds from the anterior part of the body. Thus, the typical piebald pattern is caused by a normal beginning of the pigmentation process, followed by reduced growing velocity of spots in later stages. The darker and even the medium phenotypes possess all the pigment centers of normal development. In the lighter types, simultaneously with extreme reduction of spot growth velocity, the number of pigment centers is restricted. The most conservative spots are one pair of head spots and a caudal spot.
lf we compare the typical piebald pattern of dogs with the hooded rat pattern, several differences are obvious, though even in the hooded rat the head region is the darkest part of the body. The trunk of the hooded rat lacks the typical plates of the piebald pattern and reveals a dorsal stripe in stead. The domin ance relation ship s differ, as the heterozygous H⁺h is clearly distinct from the homozygous self H⁺H⁺by having white legs, whereas the s^p₁ allele can be regarded as a recessive with respect to S₁. White marks in the progeny of self x piebald matings may occur, but this has to be explained by auxiliary spotting genes. lf one wants to look for comparable spotting patterns, the hooded pattern resembles much more the English-type spotting of rabbits, which is similar with respect to a dorsal stripe and incomplete dominance, than the piebald spotting of dogs. Therefore the hooded series should not be transferred to dog genetics. Another spotting pattern is the Irish spotting (of dogs and not of rats in this context) which is generally present or allowed in Basenjis, the Sennenhund' breeds from Switzerland, Border Collies, Bull Terriers, Boston Terriers and Boxers. It is characterized by the presence of a head blaze, a white breast spot, white paws and a white tail tip. Although the same marks appear in the piebald pattern, there is a principal difference which becomes clear if one compares Irish spotting with piebald spotting of the same grade of darkness. For instance the head blaze is much larger in the Irish type for the same percentage of coloured area. In most cases the white portion is larger on the forelegs than on the hind legs. In contrast to the piebald spotting, the anterior part of the body is most affected by restriction of spots. This indicates that the early part of pigmentation history is especially influenced. In any case, the Irish pattern cannot be regarded as a moderate piebald type pattern, produced by a `weaker' allele of the same series, but acting in the same way. It seems to be more likely that there exists a quite different, non allelic gene sⁱ_n.

Inheritance of the piebald pattern variation

Definition of classes of spotting type

These studies have been made for a dog breed (Landseer) and a cattle breed (Holstein-Friesian), which seem to have been selected for a typical piebald pattern over a long period. The Landseer data were collected by the breeder Mrs. Karin Brönnecke, who evaluated the Swiss Dog Stud Book (SHSB) up to 1979 and breeders' recent litter records, consisting of schematic drawings of the piebald pattern of puppies. The coloration of the ancestors is documented by photographs. The total range of variation of the piebald spotting has been divided into three classes (Figs. 1 and 2) which are more qualitatively than quantitatively defined:

D = Dark = 'Mantel' class

All pigment centers have germinated and grown to such an extent that instead of isolated spots, a continuous overcoat ('Mantel') covers the back and sides. Only small marks and usually most of the legs have been left white.

M = Medium = 'large spots' class

All or nearly all pigment centers are developed, but growth has been reduced in such a way that especially the sides bear several isolated spots.

L = Light = 'restricted spots' class

The number and the size of spots are severely restricted, usually to the most conservative of the Allan's pigment centers, for instance a pair of dark spots in the head region and a spot at the tail base. Instead of ordinary first-order spots, a number of small spots of a second order germination of melanoblast cumulation may be found on the back and sides. These spots should not be counted as pigment centers.
Several investigations of the inheritance of the piebald pattern in Holstein-Friesian cattle have used a similar three class division. The classifications of Lauprecht (1926a, b) and of Wiesch (1929) are based on the development of pigment centers. Dunn, Schneider and Webb (1923) measured the quotient of black and white present in the coat. Their classes are defined as

Dark	(0,70-0,95),
Medium	(0,25-0,699),
Light	(0-0,249),

with per cent black in brackets. Although a quantitative one, this division coincides more or less with the qualitative ones above. This is not the case for the quantitative division of Pfähler (1931), because his medium class covers a very small interval (0,46-0,65 per cent black). It is a pity that the data of his thorough studies in 'Höhenfleckvieh' and Holstein-Friesian cattle could not be included in the present calculations.


Hypothetical genotypes related to classes of piebald spotting types


In dogs there is little doubt that Landseers are homozygous for s^p₁ which is the main spotting gene of the piebald pattern. Heterozygotes S₁⁺s^p ₁ seem to be clearly distinguishable from the darkest piebald types and can be classified as selfcoloured Newfoundlands together with the homozygous S₁⁺ S₁⁺. The offspring of crosses between homozygous Landseers and Newfoundlands are self-coloured with or without white markings. Such markings could be independent of the existence of a single s^p₁gene, as small white spots are common in Newfoundlands even in populations which never produce piebald offspring.
Even in cattle, a clear cut division into selfcoloured breeds and piebald breeds is possible. Therefore the author disagrees with Lauprecht (1926a, b) who supposes that the HolsteinFriesian breed mainly consist of S₁ ⁺s^p₁heterozygous individuals, which present the majority of the desired medium spotted type, and the three genotypes S₁⁺S₁⁺, S₁⁺s^p₁, and s^p₁ s^p₁ overlapping in their phenotypic variation due to external influences or modifier genes.
Wiesch (1929) accepts the genetic explanation of Funquist and Boman (1923) and Funquist (1927), who established a quantitative inheritance of the variation of head patterns of cattle involving three gene pairs. Having shown that the head pattern types can be correlated with the main spotting types, Wiesch takes over the same scheme, which does not discriminate between major and minor spotting genes.
The hypothesis proposed here differs from previous schemes by assuming that the main spotting gene is homozygous through the whole breed.The phenotypic variation is attributed to two pairs of modifier genes. The hypothetical mode of action of the genes can be described more precisely: The homozygous piebald spotting factor s^p₁ does not influence the embryonic germination of pigment centers, if it acts alone. The only visible effect is a reduction of growth in the later stages of spot development.
The phenotypic expression of the minor spotting genes in the absence of homozygous s^p₁will not be discussed here. Now we are interested in the effects of interaction of several spotting genes. From the golden hamster (Robinson, 1975), the rabbit (Robinson, 1958) and the pigeon (Hollander, 1983), we can learn that the white spotting genes characteristically interact synergistically to produce greater areas of white than would be expected on simple proportional action (Robinson, 1975). The following assumption about the minor spotting genes proved to give the best results for the evaluation of matings.
First, a recessive spotting gene s² is proposed which alters the piebald pattern from the 'Mantel' type to the 'large spots' type, if it is homozygous in addition to s^p₁s^p₁. Then, a second modifier s³is introduced, which seems to be recessive, if it acts alone, but which becomes dominant, when interacting together with s^p₁s^p₁ Even in the heterozygous condition, it is able to change the 'Mantel' type to the 'large spots' type.
The restriction of pigmentation is considerably enhanced, if s^p₁s^p₁is supported by both modifiers, namely s²in homozygous and s³ in heterozygous or homozygous state. Not only is the growth of plates there greatly reduced, but also the germination of pigment centers is hindered, which is typical for the 'restricted spots' spotting type.
In all, we get nine genotypes, which are thought to compose the piebald breeds under consideration. Each phenotypic class of spotting pattern comprises more than one genotype according to the following synopsis:

Genotypes			Phenotypes
sp1sp1	S2+S2+	S3+S3+	'Mantel' type	DARK
sp1sp1	S2+s2	S3+S3+
sp1sp1	s2s2	S3+S3+	'large spots' type	MEDIUM
sp1sp1	S2+S2+	S3+s3
sp1sp1	S2+S2+	s3s3
sp1sp1	S2+s2	S3+s3
sp1sp1	S2+s2	s3s3
sp1sp1	s2s2	S3+s3	'restricted spots' type	LIGHT
sp1sp1	s2s2	s3s3

Results of crosses among classes of piebald spotting type

The breeding data were collected so that parents and offspring were determined with respect to their phenotypic spotting class. From the combination of two parental classes, six phenotypic matings are possible. One data set was obtained for Landseers and the following sets for Holstein-Friesian cattle: three populations, investigated by Lauprecht (herd A, herd B and herd C), data from Dunn, Webb and Schneider and data from Wiesch. Additionally all cattle data were combined and treated as a single set.
On the basis of the hypothesis that attributes the variation of spotting pattern to the action of two pairs of modifiers in the previously described mode, the proportions of phenotypes in offspring can be calculated for the reported matings.As each phenotype consists of two or more genotypes, the results depend on the frequencies of genotypes. Assuming Hardy-Weinberg equilibrium within the populations, the probabilities of genotypes can be easily calculated from the frequencies of alleles of each locus, S₂/s₂and S ₃⁺/s₃ in our case.
For each of the six phenotypic matings the distribution of spotting types in offspring was calculated, the frequencies of the modifier genes increasing from 10 to 90 per cent by steps of 10 per cent. By comparing the calculated tables with the observed values, for each type of mating good agreement could be found within a range of gene frequencies. As the two gefie frequencies will be characteristic values of a closed population, in an ideal case they should be independent of the type of mating. Therefore for each population two favourable modifier frequencies, common for all mating types, were chosen. With these. frequencies the agreement between observed and calculated values was studied by applying the χ² test to each mating type (Tables 1-5). The probability P for a set of matings was calculated with a combined sum and Fisher test (Koziol and Perlman, 1978).

Table 1. Results of crosses among Landseers of various spotting classes.

Matings between spotting classes	Spotting classes of offspring	Observed values	Expected values for S2, S3		Gene frequencies S2, S3	χ2	Probability P
M x M	D	4	3.9		0.5; 0.4	4.54	0.11
	M	31	35.7
	L	10	5.4
M x D	D	5	4.7		0.5; 0.4	0.02	0.89
	M	11	11.0	} 12.3
	L	1	1.3
M x L	D	0	0.6	} 8.0	0.5; 0.4	2.40	0.13
	M	10	7.4
	L	2	4.0
D x L	D	5	6.8		0.5; 0.4	1.36	0.51
	M	24	20.6
	L	7	8.6

Combined Sum and Fisher test: P= 0.23 ´


Table 2. Results of crosses among Holstein-Fresian cattle of various spotting classes. (combined data from Wiesch (1929), Dunn et al. (1923) and Lauprecht 1926)).

Matings between spotting classes	Spotting classes of offspring	Observed values	Expected values for S2, S3	Gene frequencies S2, S3	χ2	Probability P
M x M	D	85	76.0	0.4; 0.5	1.42	0.49
	M	401	404.4
	L	90	95.6
M x D	D	108	97.9	0.4; 0.5	1.69	0.44
	M	160	167.3
	L	20	22.8
M x L	D	0	12.6	0.4; 0.5	3.07	0.22
	M	122	121.2
	L	88	84.2
D x L	D	10	9.2	0.4; 0.5	2.65	0.27
	M	28	23.9
	L	6	11.0

Combined Sum and Fisher test: P= 0.37


Table 3. Results of crosses among Holstein-Fresian cattle of various spotting classes (data from Wiesch, 1929)..

Matings between spotting classes	Spotting classes of offspring	Observed values	Expected values for S2, S3		Gene frequencies S2, S3	χ2	Probability P
M x M	D	8	7.7		0.4; 0.4	0.58	0.75
	M	69	63
	L	7	7.0
M x D	D	4	4.8		0.4; 0.4	0.18	0.68
	M	13	10.7	} 12.2
	L	0	1.5
M x L	D	0	1.3	} 17.4	0.4; 0.4	4.82	0.03
	M	23	16.1
	L	5	10.6

Combined Sum and Fisher test: P= 0.29


Table 4. Results of crosses among Holstein-Fresian cattle of various spotting classes. (data from Dunn et al. (1923).

Matings between spotting classes	Spotting classes of offspring	Observed values	Expected values for S2, S3	Gene frequencies S2, S3	χ2	Probability P
M x M	D	2	2.8	0.4; 0.5	2.22	0.34
	M	13	14.7
	L	6	3.9
M x D	D	20	14.3	0.4; 0.5	5.39	0.07
	M	17	24.4
	L	5	3.3
M x L	D	6	6.0	0.4; 0.5	2.14	0.35
	M	19	15.7
	L	4	7.3
D x L	D	2	2.9	0.4; 0.5	6.21	0.05
	M	12	20.6
	L	23	14.3
D x D	D	4	4.3	0.4; 0.5	0.15	0.70
	M	1	0.7
	L	0	0
L x L	D	0	0	0.4; 0.5	0.07	0.79
	M	2	1.7
	L	13	13.3

Combined Sum and Fisher test: P= 0.11


Table 5. Results of crosses among Holstein-Fresian cattle of various spotting classes. (data from Lauprecht (1926), herd A, B and C.

Matings between spotting classes	Herd	Spotting classes of offspring	Observed values	Expected values for S2, S3	Gene frequencies S2, S3	χ2	Probability P
M x M	A	D	49	37.6	0.4; 0.5	5.27	0.08
		M	187	200.1
		L	49	47.3
M x M	B	D	7	10.1	0.3; 0.4	1.21	0.5
		M	73	72.2
		L	23	20.7
M x M	C	D	19	13.8	0.5; 0.6	5.23	0.08
		M	59	58.3
		L	5	10.5
M x D	A	D	52	47.6	0.4; 0.5	1.90	0.40
		M	81	81.6
		L	7	11.1
M x D	B	D	3	5.3	0.3; 0.4	1.39	0.50
		M	13	10.9
		L	2	1.8
M x D	C	D	29	27.4	0.5; 0.6	1.50	0.47
		M	36	39.8
		L	6	4.0
M x L	A	D	2	5.0	0.4; 0.5	2.50	0.37
		M	53	47.8
		L	31	33.2
M x L	B	D	2	2.4	0.3; 0.4	0.15	0.93
		M	32	30.7
		L	24	24.9
M x L	C	D	2	0.7	0.5; 0.6	0.17	0.18
		M	2	5.2
		L	5	3.1

Combined Sum and Fisher test:
Herd A P= 0.19
Herd B P= 0.84
Herd C P= 0.24

P values for the separate and combined test matings for all populations are high. That means the probability is high that the deviation of observed from calculated values is caused by random variation and not by a false hypothesis.
From the distributions of offspring phenotypes (which are not presented), calculated for the whole range of modifier frequencies, it appears that moderate variation of gene frequencies does not greatly affect the distribution values, whereas several attempts with other hypotheses led to quite different distributions, which could not be brought into an agreement with the observations.
In Landseers and in Holstein-Friesian cattle, the frequencies of the modifier genes s² and s³ are each near to 50 per cent. Obviously, the piebald breeds with a desired 'large spots' pattern remain heterozygous with respect to minor spotting genes, although homozygous with respect to the piebald gene s^p₁. A comparison of the three herds A, B, C according to Lauprecht suggests the following: Herd A, with abundant 'large spots' phenotypes, has the modifier frequencies 0,6 s₂ and 0,5 s₃. These values are shifted to higher frequencies (0,7 s₂ and 0,6 s₃) for the herd B, which has been selected for light phenotypes. The opposite is true for herd C, which is rich in dark phenotypes, with a consequent decrease of modifiers (0,5 s₂ and 0,4 s₃). Nevertheless, this short term time selection in one or the other direction has not altered the mainly heterozygous condition of modifier genes.
According to the hypothesis, that matings M x M, D x M, M x L and D x L produce offspring distributed among the three clases. If the 'lage spots' pattern is desired, mating within this class and the compensating cross D x L give the best results. The matings D x D and L x L theoretically give offspring of the parental phenotype and of the medium class, but not of the opposite extreme class. This is in agreement with the observed data, as stated by Wiesch (1929). As the dark and light class contain one double homozygous genotype each, and the medium class two, it should be possible to establish pure breeding strains of each pattern. This would be a trouble-some task, involving a lot of test matings. If the two homozygous Iarge spots' pattern genotypes could be separated all efforts would have been for nothing when the two strains were mated together. Simple selection for the Iarge spots' pattern will achieve the heterozygous condition of minor spotting genes. Thus the present hypothesis offers an explanation for the vast variation in patterns, a peculiarity typical of piebald spotted breeds.

Inheritance of small white marks in dogs

Distinction between phenotypic classes in Newfoundlands
The breed closest related to the Landseer is the Newfoundland. The origin of both breeds has been assumed to lie in the Newfoundland region, where the dogs have been used for water work by fishermen. When these dogs became famous, puppies were raised for export to Europe, where pure breeding started. Study of stud books shows that Landseers have been introduced into the Newfoundland breed and vice versa. Nowadays breeders question whether this practice should be allowed or not. One aspect is the fear of mismarking. Therefore the following study on white marks in Newfoundlands has been carried out, assuming that the same spotting genes are acting as in Landseers.
As a first approximation, the common Newfoundland shall be regarded as the selfcoloured and the Landseer as the piebald spotted variety of the same breed. lf one looks very carefully at the coat of dark Newfoundlands, sometimes one will find white marks. Thus, even Newfoundlands cover a range of spotting grades including the absolutely self-coloured type. Fortunately, there is a gap between the range of Landseer piebald patterns and the range covered by Newfoundland phenotypes. In other breeds this gap is occupied by the Irish type of spotting. Consequently, it may be assumed that the respective gene sⁱ₄ is absent from the Newfoundland/Landseer breed. In the dark Newfoundlands, the minor spotting genes s₂and s₃, established for the Landseers, may be present. In order to investigate the effects and the inheritance of these minor spotting genes independent from a major spotting gene, the white marks of Newfoundlands are of interest. Therefore Mrs. Karin Brönnecke, a breeder of black and piebald Newfoundlands, evaluated the Swiss Dog Stud Book (SHSB) up to 1979 for matings between phenotypic grades of dark Newfoundlands. Primarily she distinguished between cases of different distributions of white marks. After putting together related patterns, the following main classes were established:

class I	self-coloured without any white marks
class II	small white marks are present, restricted to the breast and paws
class III	'pseudo-Irish' type of spotting: marks which are found at the same places of the body as in class II, are enlarged remarkably or additional marks appear, especially at the head, belly and the tail tip.

The most common phenotype of class II has only a white breast spot. A very small number of animals bear marks restricted to one or more paws. Both types are included in subclass IIa. The rest, forming subclass IIb, are marked on the breast as well as on one or more paws. Thus the phenotype of IIb represents a stronger spotting type than IIa from a qualitative point of view.

Hypothetical genotypes related to classes of self-coloured and white marked phenotypes

With respect to the piebald spotting locus S, the S₁⁺S₁⁺ and S₁⁺ s^p₁ genotypes are possible within the self coloured Newfoundlands. As piebald offspring are very rare within this breed, the gene frequency of s^p₁seems to be extremely low. Therefore the piebald spotting gene shall be neglected in the following. We assume that a single s^p₁gene has no influence on the intensity of white marking, but strictly this may not be true. In any case, completely self-coloured offspring have been reported from Newfoundland x Landseer crosses. Thus we propose to analyse the phenotypic effects of the minor spotting genes s₂and s₃ on the base of a 'self-coloured' background with respect to major spotting genes.
The main assumption is that s₂and s₃ act as incomplete recessives. No marking effect will be generated by the heterozygous genotypes S₂s₂S₃S₃ , S₂S₂S₃s₃ and S₂s ₂S₃s₃. Thus s₃, which acts as a dominant in combination with s^p₁s^p₁ , should have a small letter, even in the previous section (p. 000).
Class II contains all genotypes with one homozygous minor spotting gene. In this condition, a weak effect is proposed if additionally a single copy of the other minor spotting gene is present. Then, a stronger expression of marking should appear. The effect is so small, that it is not always visible. This can be concluded from the observation that one leg, left or right, may be spotted, while its opposite is not marked. Thus, the left and right leg may both be self-coloured, although genotypically marks could appear. It was not possible to establish two genotypic subclasses IIa and IIb, but one can assume that the genotypes containing three spotting genes are enriched in phenotypic subclass IIb and reduced in phenotypic subclass IIa.
The double homozygous s₂s₂s₃s₃ will be put in a class of its own as a remarkable increase of white spotting has to be expected from the general property of spotting genes of mutual enhancement. The result could be a pseudo-Irish type of spotting.
This is the complete list, relating genotype to phenotypic class:

Genotypes			Phenotypes
Sp1	S2+S2+	S3+S3+	class I
Sp1	S2+s2	S3+S3+
Sp1	S2+S2+	S3+s3
Sp1	S2+s2	S3+s3
Sp1	s2s2	S3+S3+	class II
Sp1	S2+S2+	s3s3
Sp1	s2s2	S3+s3
Sp1	S2+s2	s3s3
Sp1	s2s2	s3s3	class II

Results of crosses amongphenotypic classes of dark Newfoundlands

Table 6. Results of crosses among Newfoundlands of various spotting classes regarding white marks.

Matings between spotting classes	Spotting classes of offspring	Observed values		Expected values for S2, S3		Gene frequencies S2, S3	χ2	Probability P
I x I	I	1765		1740		0.65; 0.65	2.62	0.11
	II	234	} 234	250	} 259
	III	0		9
I x II	I	627		641		0.65; 0.65	0.94	0.34
	II	320	} 323	293	} 309
	III	3		16
IIa x IIb	I	12		15.4		0.65; 0.65	1.13	0.58
	II	37		36.9
	III	5		3.8
II x II	I	88		88.3		0.65; 0.65 probability of genotypes s2s2S3+s3, S2+s2s3s3 devided by 4	4.96	0.09
	II	127		130.2
	III	6		2.5

Combined Sum and Fisher test: P= 0.16 ´


The evaluated breeding data were collected to give a set of matings (Table 6). The procedure to test the hypothesis is the same as for the piebald spotting studies. As some phenotypic classes contain more than one genotype, the variation of offspring depends on the gene frequencies. The hypothesis presented in the above sector gives a good agreement with the observations in a wide range of gene frequencies, in contrast to the author's calculations for other modes of classing genotypes to phenotypes.
In the matings I x I and I x II, the offspring classes II and III have been put together for statistical calculations, as the reported values of class III are surely lower than the true values. This has to be expected from the practice of breeders to reduce the number of puppies in a litter and to eliminate the disliked, heavily white marked class first of all. Especially in matings between class-II parents, the special condition of dog breeding in contrast to scientific breeding experiments has to be taken in account. Whereas in the mating I x II the white marking of the class-II parent will be compensated by the other parent and all individuals of class II can be used, in the mating II x II, which is very risky, only individuals with very faint white marks will be mated. This selection within class II can be simulated by reducing the probability of the genotypes s₂s₂ S₃⁺s₃ and S₂ ⁺s₂s₃s₃ to one-fourth of the equilibrium value in the population for the mating II x II. On the other hand, the conditions comparable with a planned breeding experiment are maintained, if all matings are collected where one parent is faintly marked and the other has a more heavy expression of white marks, which means II_ax II_b matings.
The statistical tests yield high probability values P for individual matings as well as for the complete set, calculated for the gene frequencies regarded as representing the equilibrium state during the considered breeding period. The best fitting gene frequencies are 0,35 s² and 0,35 s³. Compared with the Landseer breed, the gene frequencies of the minor spotting genes have been slightly reduced.
Furthermore, the inheritance of the white breast spot is better explained by a hypothesis involving two recessive spotting genes than by a single dominant as proposed by Wegner (1979).

Conclusions

In dog and cattle breeds selected for piebald spotting with the 'large spots' pattern as the most desired type, inheritance of spotting can be explained by a recessive major spotting gene s^p₁, working with at least two modifying genes s₂ and s₃ , one of them, s₂, being regarded as a recessive, the other, s₃, as a dominant on the s^p₁ s^p₁background. From this interaction of genes results phenotypic variation reaching from the relatively dark 'Mantel' type to the almost white 'restricted spots' type.
The modifiers act as minor spotting genes in their own right on an S₁⁺ background. As s₃ is a recessive without mutual enhancement with other homozygous spotting genes, it is written in small letters. The variation of phenotypes reaches from pure self-coloured to 'pseudo-Irish' marked.
Whereas in piebald phenotypes, the head region is least affected by the restriction of coloured area, quite the opposite is true in the original 'lrish' type of dog spotting. Therefore a major spotting gene sⁱ₄ non-allelic to s^p₁ has been assumed. Homology between s^p₁of dogs and h of the Norway rat, postulated by Little (1957, 1958), has to be rejected, because the patterns are quite different. lf independent, the major spotting genes sⁱ₄and s^p₁can be combined in a double homozygous individual, which will be restricted in the anterior part as well as in the posterior part of the body with respect to plate development. The result will be 'almost white' in phenotype. In breeds which are homozygous for sⁱ ₄, a bimodal distribution of phenotypes has to be expected with a maximum for the 'lrish' type and another for the 'almost white', separated by a gap. Exactly this happens in Boston Terriers, Bull Terriers and Boxers (Little, 1958).
The factor s_w of Little's hypothesis (1957, 1958) proved to be unnecessary, as the respective phenotype can be explained by homozygous s^p_1, enhanced either by two minor spotting genes (s₂s₂S₃s₃s₂s₂s₂s₃) or by the Irish spotting gene (sⁱ₄sⁱ₄).
These studies were performed to find out what happens if self-coloured and spotted strains of related breeds are cross-bred. Will there, for instance, be an increase of white marked Newfoundlands or of dark Landseers? This is a question of the minor modifiers s₂and s/₃ and not of the piebald spotting gene. As both breeds, each selected for desired types, differ in modifier gene frequencies, an uncontrolled mixture would give unfavourable results. On the other hand, with aid of 'Mantel' type Landseers, the gene frequency of at least s₃ could be reduced in Newfoundlands. lf interested in Iarge spots' type Landseers out of S₁⁺ s^p₁ Newfoundlands, one should avoid mating unmarked Newfoundland x 'Mantel' type Landseer and heavily white marked Newfoundland x 'restricted spots' type Landseer.

Zusammenfassung

Die Vererbung der Plattenscheckung und ihrer Variation bei schwarzbuntem Niederungsvieh und Landseer-Neufundländer Hunden

Für Hunde- und Rinderrassen, die nach einem typischen Plattenscheckungsmuster selektiert sind, läßt sich die Vererbung der Scheckung mit Hilfe eines rezessiven Hauptscheckungsfaktors s^p ₁erklären, der unterstützt wird durch wenigstens zwei Modifikatoren s₂ und s₃. Einer davon, s₂, wird als rezessiv angenommen. Der andere, s₃, soll dominant sein, allerdings nur bei genetischem Hintergrund mit s^p₁ s^p₁. Aus dem Zusammenwirken dieser Gene ergibt sich eine phänotypische Variation im Bereich von 'Mantelscheckung' bis zum fast weißen 'reduzierten Plattentyp'.
Die Modifikatoren wirken auf S₁⁺ Hintergrund als selbständige schwache Scheckungsfaktoren. Wenn kein anderer homozygoter Scheckungsfaktor zur gegenseitigen Verstärkung da ist, ist auch s₃ rezessiv und wird deshalb mit kleinem Buchstaben geschrieben. Die Variation der Phänotypen reicht von voll ausgefärbt bis zu 'pseudoirischer' Scheckung.
Während bei typischer Plattenscheckung die Kopfregion am wenigsten von der Einschränkung der Ausfärbung betroffen ist, gilt das Gegenteil für die echte 'irische' Scheckung beim Hund. Deshalb wird der Faktor sⁱ₄ als nicht zu s₁ alleler Hauptscheckungsfaktor für irische Scheckung angenommen. Der Homologisierungsversuch Littles (1957, 1958) mit dem Locus h der Haubenratte kann nicht nachvollzogen werden, da dieses Muster grundsätzliche Unterschiede zur Plattenscheckung des Hundes aufweist. Wenn sⁱ ₄und s^p₁unabhängig mendeln, können sie doppelt homozygot zusammentreffen. Als physiologische Wirkung ist die Einschränkung der Ausfärbung sowohl im vorderen als auch im hinteren Körperbereich zu erwarten, so daß ein fast weißer Phänotyp resultiert. Dieser Typ sollte bei Rassen, die homozygot in sⁱ₄ sind und außerdem s^p₁enthalten, auftreten und ein zweites Maximum in der Verteilungskurve der Phänotypen bilden. Das ist tatsächlich bei den Rassen Boston Terrier, Bull Terrier und Boxer der Fall (Little, 1958).
Littles Faktor s^werweist sich als überflüssig, da der betreffende Phänotyp bereits durch homozygotes s^p₁ erzeugt wird, wenn die beiden Modifikatoren in den Kombinationen (s₂s₂ S₃s₃) und (s₂s₂ s₃s₃) oder homozygotes sⁱ₄ vorhanden sind.
Diese Untersuchungen wurden durchgeführt, um die Auswirkungen der Verpaarung eines ausgefärbten und eines plattengescheckten Farbenschlages einer Großrasse zu verstehen. Treten eventuell gehäuft unerwünschte Zeichnungen auf wie weiße Abzeichen beim ausgefärbten Teil und zu dunkle Scheckung beim gescheckten Teil der Nachzucht? Dies ist allein eine Frage der Modifikatoren s₂ und s₃. Die beiden verwandten Rassen Neufundländer und Landseer unterscheiden sich in den Genfrequenzen in der berechneten Weise. Eine unkontrollierte Vermischung würde tatsächlich unerwünschte Ergebnisse bringen. Andererseits kann durch Einkreuzung von Landseern des 'Manteltyps' wenigstens die Genfrequenz von s₃in Neufundländern erniedrigt werden. Wenn das Hauptinteresse an typisch plattengescheckten Landseern mittleren Grades besteht, sind die Paarungen Neufundländer ohne Abzeichen x Landseer mit Mantelscheckung und stark gezeichneter Neufundländer x sehr heller Landseer zu vermeiden.

References

Allan, G. M., 1914. Pattern development in mammals and birds.
Castle, W. E., 1951. Variation in the hooded pattern of rats and a new allele of hooded. Genetics 36: 254-266.
Curtis, M. R. and Dunning, W. F., 1937. Two independent mutations of the hooded or piebald gene of the rat. J. Hered. 28, 382-390.
Comberg, G., Sponer, G., Feder, H., Aschermann, G., Plischke, R. and Wegner, W., 1972. Einige qualitative und quantitative Eigenschaften und ihre Beziehungen zueinander in Populationen Deutscher Schwarzbunter Rinder. Z. Tierzüchtg. Züchtbiol. 89: 109-122.
Dunn, L. C. and Charles, D. R., 1937. Studies on spotting patterns, 1. Analysis of quantitative variations in the pied spotting of the House Mouse. Genetics 22: 14-42.
Dunn, L. C., Webb, H. F. and Schneider, M., 1923. The inheritance of degrees of spotting in Holstein Cattle. J. Hered. 14: 229-240.
Funkquist, H., 1927. Vererbung 'weißer Abzeichen' am Kopf bei schwarzbuntem schwedischem Niederungsvieh. Hereditas 9: 289-302.
Funkquist, H. and Boman, N., 1923. Vererbung 'weißer Abzeichen' bei Rindern. Hereditas 4: 65-80.
Green, M. C., 1975. The laboratory Mouse, Mus musculus. In: Handbook of Genetics, R. C. King (Ed), Bd. 4, 203-241, New York.
Hollander, W. F., 1983. Origins and Excursions in Pigeon Genetics. The Ink Spot, Burrton, Kansas.
Koziol, J. A. and Perlman, M. D., 1978. Combining independent chi-squared tests. J. Amer. Statist. Ass. 73: 753-763.
Lauprecht, E., 1926a. Über die Scheckung des schwarzbunten Niederungsrindes und ihre Vererbung. Z. indukt. Abstamm.- Vererb.Lehre 40: 139-196.
Lauprecht, E., 1926b. Über die Zeichnung des schwarzbunten Niederungsrindes. Züchtungskunde 1: 309-331.
Lauvergne, J. J., 1966. Genetique de la couleur du pelage des bovins domestiques (Bos taurus, Linné). Bibliogr. genet. 20: 1-68.
Little, C. C., 1957. The Inheritance of Coat Color in Dogs. - New York.
Little, C. C., 1958. Coat color genes in rodents and carnivores. Rev. Biol. 33: 103-137.
Little, C. C., 1976. The Inheritance of Coat Color in Dogs. 6th Ed. Howell, New York.
Pape, H., 1983. Revision des Erbfaktorenschemas für die aus Loh und Schwarz zusammengesetzte Grundfärbung bei Hunden sowie Aufdeckung paralleler Verhältnisse bei Kaninchen. Z. Tierzüchtg. Züchtbiol. 100: 252-265.
Pflähler, R., 1931. Untersuchungen über die Vererbung der Scheckung und Farbe beim Höhenfleckvieh. Z. indukt. Abstamm. Vererb.Lehre 58: 177-221.
Rawles, M. E., 1940. The development of melanophores from embryonic mouse tissues grown in the coelom of chick embryos. Proc. natn. Acad. Sci. U.S.A. 26.693.
Rawles, M. E., 1947. Origin of pigment cells from the neural crest in the mouse embryo. Physiol. Zoo,1, 20, 248.
Robinson, R., 1975. The Golden Hamster, Mesocricetus auratus. In: Handbook of Genetics, R. C. King (Ed.), Bd. 4, 261-274, New York.
Robinson, R., 1958. Genetic studies on the rabbit. Bibliogr. genet. 17: 229-558.
Shaible, R. H., 1963. Development Genetics of Spotting Patterns in the Mouse. Ph.D. Thesis, lowa State Univ., Ames, lowa.
Wegner, W., 1979. Kleine Kynologie. - Konstanz.
Wiesch, A., 1929. Vererbung der Scheckzeichnung und korrelative Beziehung der Färbung bei Rindern. Kühn Archiv 22:347-397.
Wright, S. and Chase, H. B., 1936. Ort the geneties of the spotted pattern of the guinea pig. Genetics 21: 758-787.

© Karin Butenhoff-Taufertshöfer - All rights reserved - mailto:webmaster@newfoundlanddog.biz