Breeding data from Eire and Great Britain conf'mn the hypothesis of Sponenberg (1985) that the harlequin colour of the Great Dane breed of dog is due the ...
Genetica 78: 215-218, 1989. © 1989KluwerAcademic Publishers. Printedin Belgium.
215
Harlequin colour in the Great Dane dog N. O'Sullivan & R. Robinson 1Department of Agricultural Zoology and Genetics, University College, Dublin, Dublin 4, Ireland 2St. Stephens Road Nursery, London W13 8HB, England Received 18.4.1988
Acceptedin revisedform 23.3.1989
Abstract
Breeding data from Eire and Great Britain conf'mn the hypothesis of Sponenberg (1985) that the harlequin colour of the Great Dane breed of dog is due the combined action of a dominant gene H with the merle gene M in the genotype H + M +. The typical bluish coloration induced by M is modified to white by the action of H. The H gene is a prenatal lethal when homozygous H H and this study offers clear indication that the heterozygous H + interacts with M to reduce the viability of white merle homozygotes H + MM. Introduction
The Great Dane breed of dog is bred in a number of coat colour varieties: namely, black, blue, brindle, fawn, merle and harlequin. The last two colours are of particular interest. The colour of these varieties is a characteristic mosaic of ragged edged black patches of varying sizes against a bluish (merle) or white (harlequin) background. The mosaic pattern is inherited as a dominant trait to normal (Little & Jones, 1919). Breeders are only interested in the harlequin pattern as merle is not an acceptable pattern for show purposes. The continued occurrence of merle when it has been selected against for many generations of harlequin breeding is suggestive that the two patterns may be related at the genotypic level. In recognition of the close phenotypic similarity, the two varieties were considered to be produced by the same gene, merle, symbol M (Little, 1957), until Schaible and Brumbaugh's (1976)observations gave rise to the possibility that merle and harlequin are due to two allelic genes (Robinson, 1982). Recently, Sponenberg (1985) has presented evidence that merle and harlequin both possess
the Mgene but that the harlequin has an additional gene (symbol H) which converts the background coloration from bluish to white. The H gene is a prenatal lethal when homozygous and only engenders the harlequin phenotype in the double heterozygote H + M +. The M gene produces an all white phenotype when homozygous M M and the merle phenotype when heterozygous M +. These phenotypes have relevance for the inheritance of H since the gene is not expressed in the white H + M M nor normal coloured H + + + genotypes. There is reason to suppose that the genotype H + M M is of lower viability in comparison with other genotypes. The present report summarises breeding data for Irish and British populations of Great Danes which confirm and extend Sponenberg's findings for the United States population.
Materials and methods
The data consist of compilation of litters which involve the assortment of the merle and harlequin varieties of the Great Dane. Two independent
216 compilations have been completed for Irish and British populations. The viability (v) of the H + M M genotype for the harlequin to harlequin mating may be estimated by maximum likelihood from the expectations: White phenotype
Coloured phenotypes
1 +2v
9
10 + 2v
10 + 2v
whence, v = (9a - b)/2b, with variance = (1 + 2v) (10 + 2v)2/36(a + b), where, a = the number of white dogs and b = number of coloured dogs.
Results and discussion
The frequency distributions of offspring for the various matings between the colour varieties are
summarised by Table 1. The decisive mating for the Sponenberg hypothesis is that of harlequin to harlequin (entries 1 and 7 of Table 1). Because of the hypostatic behaviour of the H gene in genotypes other than H + M +, this is one of the few matings for which the genotypes are precisely known. From the data to hand, all harlequins are capable of producing merle offspring, hence the variety is an obligate double heterozygote. Table 2 presents an analysis of the observed frequencies of the offspring for the mating of harlequin to harlequin. The successive entries, A to D, reading from left to right, are tests of hypotheses to explain the observed distribution of phenotypes for United States, Irish and British populations, and combining all three. Test A is essentially a check for independent assortment of the H and M genes and of the normal viability of the different genotypes. The significant values for three of the populations indicate poor agreement
Table 1. Classification of offspring for various matings between colour varieties of Great Dane dogs for Irish (first six entries) and British (last three entries) populations.
Mating
White
Harlequin
Merle
Black
Total
Hare x Harl. Harl. x Black Harl. x Merle Harl. x White Merle × Black White x Black Harl. x Harl. Harl. x Black White x Black
35 11 5 25 -
123 67 32 6 5 10 49 5 17
52 40 23 1 6 29 20 9 4
74 101 25 14 26 24 -
284 208 91 12 25 39 120 38 21
Harl. = Harlequin. Table 2. Chi-squared tests for agreement between observed and expected frequencies ofcolour classes among the offspring from the mating of harlequin to harlequin upon the assumption of differential viability for specific genotypes as defined below.
Population
A
B
C
D
Viability of H + M M
United States Irish British Combined
16.0 28.4 3.4 32.1
8.8 27.3 3.4 26.8
9.8 4.2 17.5 24.9
0.7 9.3 2.7 3.7
0.380 0.179 0.684 0.364
+ + + +
0.163 0.125 0.266 0.095
Definitions: A = all genotypes viable, B = H H inviable, C = H H and H + M M inviable, D = H H and 50 per cent of H + M M inviable. The United States population date are those of Sponenberg (1985). The Z2 values in the body of the table have 3 df, for which the 5 per cent level of significance is 7.82.
217 between observation and expectation. However, Sponenberg (1985) has proposed that the H gene is lethal when homozygous. Test B examines this possibility and reveals a somewhat closer agreement but the values are still mostly significant. The major discrepancy between observation and expectation in both tests is a deficiency of white dogs. This could imply that the genotype H + M M is inviable. Test C examines this possibility and the values still indicated poor agreement, but the discrepancy is now in an excess of white dogs. Sponenberg (1985) has suggested an inviability interaction between H and M M such that approximately half of the H + M M genotype die. Test D examines this possibility and the values indicate satisfactory agreement between the expected and the observed frequencies for two of the three populations and for the combined data. The last column of Table 2 lists the estimated viability of the genotype H + M M for the three populations and for the combined data. Obvious differences exist between the populations for the degree of inviability of the genotype. Such differences could be explainable on the basis that the inviability is dependent in part on the genetic background of the population. A factor to consider is that the white merle M M might be somewhat inviable. However, a search of the published literature (Anker, 1925; Mitchell, 1935; Sorsby & Davey, 1954) for details of merle to merle matings does not support this possibility. The pooled results are 39 white (MM), 84 merle (M + ) and 49 black (+ + ) while the expected frequencies are 43, 86 and 43, respectively; the figures are in sufficiently close agreement to imply normal viabilify for all three genotypes. Sponenberg (1985) has noted that the mating of harlequin to black will give harlequin, merle and black progeny in either a 2 : 1 : 3 or a 1 : 1 : 2 ratio, depending whether or not the black is heterozygous for H. His data indicated a near perfect ratio of 1 : 1 : 2 and he concluded that it is unlikely that any of the blacks in his sample carried the Hgene. However, the Irish data of this report are in poor accord with the 1 : 1 : 2 ratio (~2 = 7.18, for 2 dr) but in good accord with the
2 : 1 : 3 ratio (Z2 = 0.97, for 2 dO. The implication is that black dogs can be heterozgous for H. The mating of harlequin to merle could provide evidence for the inviability interaction of H and M. The white individuals from this mating will be composed of 50 per cent ofH + MMand a proportion of these will be inviable. A deficiency of white dogs is to be expected and this is borne out by the data (Z 2 = 8.09, for 1 dO. However, the shortage is so great to exceed the expectation even if all of the H + M M genotype is inviable. However, the estimated viability of the genotype for the Irish population is low, as shown by Table 2, and the data would be consistent with a low value. On the other hand, the present result raises the possibility that H + M M i s totally inviable. The analysis C of Table 2 does not support this supposition, however, while the deviation from expectation is significant (Z2= 24.9 for 3 dO, the heterogeniety between populations is not (~2 = 6.6 for 3 dO. The mating of white to black will give harlequin and merle progeny in the ratios of 2 : 1, 1 : 1 or merle only, depending if both, one or neither parent is heterozygous for H. The present data are consistent the 1 : 1 ratio ()r2 = 0.6 for 1 df) but inconsistent with the 2 : 1 ratio (X2= 12.68 for 1 dO. That black individuals can be heterozygous for H is shown by the occurrence of harlequin offspring from the mating of merle to black (see Table 1). It is unfortunate that the mating of white to black cannot discriminate between which partner is heterozygous for H. Independent evidence for the existence o f H + M M individuals would be provided by the occurence of harlequin offspring from the mating of white to merle but this mating does not find favour with breeders and none are reported. The distribution of sizes of litters for the harlequin to harlequin matings versus all other matings is of interest. For the former, the sizes ranged from 1 to 13, with a mean of 5.97 + 0.36 (66 litters); while for the latter, the sizes ranged from 1 to 16, with amean of 7.73 + 0.6 (41 litters). The difference between means of 1.76 + 0.7 is significant and is indicative of the higher rate of foetal loss to be expected for the former mating.
218 Acknowledgement The authors wish to thank those breeders who kindly made available their breeding records.
References Anker, J., 1925. Die Vererbung der Haarfarbe beim Dachshunde. Dansk Viden. Selskab. biol. Medd. B. 4: 1-72. Little, C. C., 1957. The inheritance of coat colour in dogs. Cornell University Press.
Little, C. C. & Jones, E.E., 1919. The inheritance of coat colour in great danes. J. Hered. 10: 309-320. Mitchell, L. M., 1935. Dominant dilution and other colour factors in collie dogs. J. Hered. 26: 424-430. Robinson, R., 1982. Genetics for dog breeders. Pergamon Press. Schaible, R. H. & Brumbaugh, J. A., 1976. Electron microscopy of pigment cells in variegated and non-variegated, piebald spotted dogs. Pigment Cell 3: 191-200. Sorby, A. & Davey, J.B., 1954. Ocular associations of merling in the coat colour of dogs. J. Genet. 52: 425-440. Sponenberg, D. P., 1985. Inheritance of the harlequin colour in great danes. J. Hered. 76: 224-225.
