Morphogenetic studies of the rabbit XXXVI. Effect of gene and genome interaction on homeotic variationкод для вставкиСкачать
Morphogenetic Studies of the Rabbit XXXVI. EFFECT OF GENE AND GENOME INTERACTION ON HOMEOTIC VARIATION P. B. SAWIN AND MARYANN GOW The Jackson Laboratory, Bur Harbor, Maine ABSTRACT Study of the homeotic shifts of vertebral borders as affected by the D a gene when on two different genetic backgrounds shows that the effects of both gene and genome are distinguishable. The Du in either one or two doses shifts both thoracolumbar and lumbosacral borders forward, particularly the former. The effect of the DA genome is in the same direction, but significantly greater, and the two combined are additive. The effect of the IIID, genome, by contrast, is in the posterior direction and epistatic. It suppresses and tends to shift the localization of D a effect posteriorly. Significant differences between borders when considered separately and in relation to each other indicate differences of interaction between gene and genome. The fact that the only signacant difference in the three way comparisons of border, genotype, and genome is between the Du/+ genotypes of IIIoa and DA as a manifestation of overdominance can be discounted because of the differences i n border interaction. As reference points defining the relative size and position of the thoracic and lumbar regions, these borders reveal the activity of gene and genome to be a n alteration of the relative size and position of the growth gradients of these two regions rather than a direct gene specific morphology. The importance of genic interaction in morphological genetics has been demonstrated (Wright, '63). One of its most signiiicant features is the dmculty of isolating specific genes and defining the limits of their action, due to the fact that morphology is by nature a differential growth phenomenon. Mammalian growth as revealed by study of the rabbit was demonstrated to have a polygenic basis subject also to extrinsic and other intangible factors (Castle, '31, '41) and control in time by developmental rate (Castle and Gregory, '20). Studies of many of the effects of major mutant genes in mice have also shown the influence of numerous minor so-called modifying genes affecting both external (Wright and Chase, '36; Dunn and Charles, '37; Dunn, '42) and internal morphological variations (Dunn and Gluecksohn - Schoenheimer, '45 for short tail; Green and Green, '46, McNutt, '54, and Green, '57 for the short ear gene) indicating that these mutations are also under polygenic influence. Viewed in this way, analysis of morphological variation and malformation requires an approach formulated in terms of the possible interaction of a number of genes acting by way ANAT. REC., 157: 425-436. of growth processes rather than specificity of gene per discrete morphological unit. The importance of gene number, if not the entire genome (gene milieu), was recognized by Landauer ('58). The importance of developmental processes at particular times and places was emphasized by Willier ('54) in discussion of morphogenesis in the chick and by Gruneberg ('57) in the mouse, both of whom have shown the importance of the continuum or sequence of events, a characteristic also beginning to be recognized by the teratologists (Lash, '64). Few investigators, however, have associated these facts with the species pattern of gradient growth as outlined by Child ('25). The rabbit species pattern of gradient growth of the axial skeleton as illustrated by strain I11 was described qualitatively and quantitatively by Sawin and Crary ('64) and discussed with reference to malformations. The manner in which the pleiotropic effects of the Da (chondrodystrophy) gene arise through retarded growth and are expressed by deficiency in number of skeletal units, compensatory induction of accessory centers of ossification, and overgrowth of adjacent normal units 425 426 P. B. SAWIN AND MARYANN G O W was described by Sawin and Trask ('65). Collectively, these lead to homeotic shifts in position of regional borders (Bateson, 1894) and to crowding or fusion of units. These observations, together with those of Sawin and Crary ('57), Lamb and Sawin ('63), and Crary ('64), show that the Da genotypes Da/+ and Da/Da may be distinguished from the normal and make it possible to study the relative effects of gene and genome upon the rabbit gradient growth pattern. This communication is concerned with the effects of the Da gene upon the thoracolumbar (ThL) and lumbosacral (LS) borders and the extent that they are shifted by association of Da with two unrelated genomes. The borders serve as cartographic reference points, determining the limits of thoracic and lumbar regions and of the relative anteroposterior position of the lumbar growth gradient. MATERIALS AND METHODS Skeletons of 1555 rabbits were studied. They were obtained, with seven exceptions, during the first five postnatal days and were examined either roentgenographically or as alizarin-stained preparations (Crary, '62). Seven populations were involved (fig. 2). The two unrelated parental strains were of New Zealand White origin, DA,' which was the origin of the Da mutant, and 111' which does not transmit Da but which has been partially inbred and selected for the 7C/13Th/7L/3s/18cd vertebral formula (VF). The IIIDa (transfer) strain was synthesized by outcrossing I11 to DA with subsequent continued backcrossing of the D a / + genotype to race I11 for at least five generations. Both DA and IIIo, were maintained under forced heterozygosity for the Da gene by selective breeding of tested heterozygotes ( D a / + ). Within each of these two strains, the three genotypes (+/+, D a / S , and D a / D a ) are readily identified by discrete variation at the base of the external ear (Sawin and Crary, '57; Lamb and Sawin, '63; Crary, '64). This permits evaluation of 0, 1, or 2 doses of the gene upon each of the pleiotropic characters induced by the gene. These six populations and parental strain I11 provided tests of the differences between Da genotypes and normal genomes. The IIID, populations studied were derived by intercrosses among the 5-7 backcross generation offspring. Many of the young of populations on which the tabulations were based were sacrificed at 33 days post coitus to provide satisfactory information on initial appearance of the vertebral epiphyses, to be reported later. AU potential breeders (females and males) necessary for the succeeding generations were typed roentgenographically for vertebral formula. Chisquare tests of the variation in VF between the five age groups (30 to 34 days) and generations revealed no significant differences, hence the data could be pooled irrespective of age and generation. As described by Sawin ('45) the VF of both alizarin preparations and roentgenograms were read with reference to the relative thirteenth rib expressivity on each side. A scale of 0-4 with relation to the size of the twelfth rib was used. Similarly the expressivity of sacro-pelvic attachment was recorded on a scale of 1-6. Type I involved only the twenty-seventh vertebra, type 5 the twenty-eighth only, type 3 both 27 and 28 equally, and 2 and 4 were intermediate classes (fig. 1 ) . A sixth type, similar to 5, involved a very small part of the twentyninth in one case. 1 The DA strain of the Da gene was obtained from California i n 1950 (Crary and Sawin, '72; Sawm and Crary '57. Sawm Ranlett and Craw, 59, '62; Lamb and dawid, '63; &win and Trask, '65). 2 Strain I11 was first descrlbed by Sawin ('37) with reference to the homeotic variations of the axial skeleton. It was also studied in connection with malformation of the vena cava (McNutt and Sawin, '43), vanations of the ventral spinous processes of the lumbar vertebrae (Sawin and Hull '46) aortic arch ( Sawin, and Edmonds, '49), gall blhdder' ( Sawin and Crary, 51), growth of vertebra! centra (Tanner and Sawin, '53; Crary and Sawin 57). growth and maturation (Crary and Sawin, '6'0).adult body size and size of prgans and skeletal parts (Latxmer and Sawm, '571 h 58 '59a.b. '60a.b. '61. '62. '63a.b: Sawin and &a-:' '56.' CraG'and Bawi?' '57) growth and body w e i g k (drary and Sawin 60), h d malformations (Sawin and Crary, '64). ks reported by .Sawin ('37). strain I11 tended to produce a much hlgher proportinn of Droeenv of the vertebral formula 7C/13Th/7L ( 1 3 pa& 6f fibs and 27 presacral vertebrae, PSV) than any other race. As close inbreeding as possible consonant with maintenance of reproductive capaclty plus select!on has materially increased the proportion unul in two sublmes the penetrance appears to be 100%. One other subline, was much more variable in vertebral formula but was assigned for inbreeding and maintenance. of the ep gene (audiogenic seizures) without selection for vertebral formula. Since the Da gene had already been introduced without reference to sublme differenbatlon the experiment was conducted with due attentlon only to secure sufficient progeny from each subline to permit adequate test of the genetics of possible subline differences. 427 GENE AND GENOME INTERACTION 1 NORMAL THORAX 12 17 2 NORMAL LUMBAR (7) 13/6L REDUCED LUMBAR (61 3 13/6H ENLARGED 4 THORAX 4 13/7H NORMAL LUMBAR (71 SHIFTED POSTERIOR 5 13/7L Fig. 1 For the purposes of this paper, expressivity of rib (H and L ) and presacral vertebrae were ignored and classification was confined to an all or none basis, presence or absence of extra ribs and grades 1-3 considered as 26, and 4, 5, and 6 as 27 presacral vertebrae. 428 P. B. SAWIN AND MARYANN G O W OBSERVATIONS 1 . Effect of genome differences on ver- tebral formulae. In using the skeletal pattern of growth and development to portray the true effect of any one gene it is essential first to consider possible genetic differences in that background. Figure 2 shows the difference between the normal genome effects on VF of race 111, which segrelacks the Da gene, and the gates of the DA race. Race I11 has over twice as many individuals of type 7/13/7 as does the DA and relatively few of the other three types. The DA race has significantly larger proportions of 7/12/7 and 7/13/6 classes (P < 0.005). The segregating genotypes of strain I&, however, do not differ from race I11 (P > 0.05), indicating that the genetic background is essentially similar to that of race 111. The race I11 genome tends to shift the vertebral border in such manner as to increase the proportion of individuals in the two classes to the right (posteriorly) and the DA to decrease them or shift the border to the left. When the Da gene is present in either heterozygous or homozygous state (fig. 2), the shifts are complementary and the genome differences, DA vs I&., likewise are highly significant ( P < 0.005, see table 2). +/+ +/+ Genotype 2. Gene effects on vertebral formulae. The effects of one and two doses of the Da gene may be compared similarly as shown in figure 2. The differences between the DA race normals and both heterozygous and homozygous dachs (Da/+ and D a / D a ) are signilkant (P < 0.01 and < 0.005 respectively). The shift effected by one dose of the Da gene appears to be very little different from that of two doses as indicated by the fact that the difference between the two Da genotypes is not statistically significant (P > 0.05). The same comparisons within the III,, race (fig. 2 ) show that the X2 between normal and heterozygote is significant at the 0.02-0.01 level, but not between normal and dachs or between heterozygote and dachs. In general, the Da effect is manifest in both DA and III,, as a reduction of penetrance of extra ribs and PSV resulting in a n anterior shift of the borders. The effect, however, is greatest in the DA and expressed more anteriorly than in IIIo,. Since these borders determine the relative size of the thoracic and lumbar regions, reduction of rib number represents a decrease in size of the thoracic area in the DA, whereas in the I11 and IIID. the greater proportion of individuals with 13 ribs indicates that these strains actually have a relatively Race x2 df P Total .3-.2 132 DaDa DA 4.0 3 Da+ DA 18.4 3 .005 241 +I- DA 12.1 3 .01 116 DaDa IIIDa 2.0 3 .7-.5 131 Da+ IIIDa 7.2 4 .2-.1 190 $+ IIIDa i+ 111 621 /% 20 40 60 80 100 1555 Fig. 2 Bar graphs show the distribution of individuals according to vertebral type in each of the three Da genotypes for the DA, 111, and IIIoa races, with significance and number of progeny at the right. Blank bars are normal (12 thoracic/7 lumbar) vertebral formula, diagonal lined 13/6, black 13/7, and stippled 12/8. Significance of intrastrain differences are shown by brackets. Interstrain differences of (+/+) phenotypes are significant between I11 and DA ( p 0.005), but not between 111 and IIID,. Actual numbers of each vertebral type are shown in table 5 by strain and genotype. < GENE AND GENOME INTERACTION W 4 2 iij B 4 Y ;i W s Fr; al PI ,x d d 429 larger thoracic region. The question arises as to whether the Da gene and DA genome affect both borders and both regions in the same way, or whether regional differences of effect exist and, if so, how they are localized. 3. Gene and genome effects upon the separate vertebral borders. To test more precisely the localization of gene and genome effect on the two regions the racial differences at the two vertebral borders have been compared separately. As shown in table 1, the differences in both borders between I11 and DA in the absence of the Da gene are significant, race I11 having a greater proportion of individuals with both 13 ribs and 27 PSV than does the DA (P<O.O05). The same is true of DA vs I I L , although the X2 is not as great. Although race I11 and IIL, are not significantly different from each other in rib number (P>O.O5), a 6% reduction in PSV suggests the possible existence of some minor genetic differences, reducing the proportion of PSV in the transfer strain . The effects of the genome on distribution of position of the individual borders within the two Da genotypes are shown in table 2. Considered independently, all are highly significant at both borders, showing again that there is a difference between the two regardless of the genotype. The largest between-strain chi-squares are found in the proportion of PSV of the two Da/Da populations and in the number of ribs of the two Da/+ populations. Further evidence of this is found in comparison of the separate borders in the Da genotypes as found in the DA (table 3) and IIID, (table 4 ) genomes. 4 . Gene effects on the separate borders. In the DA race (table 3) the chi-squares for the thoracolumbar border (ThL) are significant between both Da genotypes and normal (+/+) (P<O.O05), and not between the Da/Da and D a / + (P=O.30.2). Those for PSV also are significant though at lower levels (P < 0.05), thus showing some, though perhaps less, effect on the lumbar border. In the IIIDa (table 4), however, the effects of the Da gene separately on the borders are not significant at the thoracolumbar in any case, but are significant 430 P. B. SAWIN AND MARYANN GOW at the lumbosacral ( P < 0.01) between Da/+ and + / C and (P<O.O5) between Da/Da and 5 . Relation of thmacolumbar and lumbosucral borders. To test the possible relationship of the two borders, the data for +/+. the seven populations were assembled as shown in table 5 and tested by Xa. All were significant ( P < 0.005), indicating a high degree of dependence or interaction of the borders within each genotype of both strains. The most obvious differences between the two strains was the relative proportion of the minority types, 12/8 and 13/6, and that which is the modal (12/7 in DA vs 13/7 in I I L ) , indicating some degree of interaction between the Da gene and its associated genome. To test this, the method devised by Bartlett (described by Snedecor, '46) for the combined interaction of three attributes, in this case regional borders, strain, and genotype, was applied. As shown in table 5, of the groups tested the only ones showing significant differences are the heterozygous groups (Da/+) of the DA and 111,. races (p = 0.05-0.02). Since no significant differences between the genotypes were found, those of both DA and IIID, were pooled, and similar 3-way tests of the three genomes (DA, 111, and 1110,) were made. Again no significant differences were found. Therefore, the only significant difference in the dependent relationship of the borders is that arising in the two Da/+ groups. Hence, we may conclude that it is interaction of the Da gene with the two genomes which creates the difference. It is supported by the fact that the effect of Da is greater on the posterior border when associated with the I11 genome than when with DA (tables 3 and 4). DISCUSSION The results of this study clearly show that the homeotic shifts of the ThL and LS borders which were first regarded as pleiotropic effects of the Da gene are not in any sense exclusive to specific skeletal units. Instead, as parts of the metameric organization of the species, addition or deletion of ribs or presacral vertebrae can result from the cumulative effects of some greater or lesser portion, if not the entire genome, in much the same sense that 431 GENE AND GENOME INTERACTION TABLE 3 Significant effects of Da on the thoracolumbar and lumbosacral borders of the DA race considered inddpendently Thoracolumbar Genoqpe companson + + vs. DA+ + + vs. DaDa Da+ vs. DaDa Lumbosacral XZ df P xa df P 12.1 16.2 1.2 1 1 1 0.005 0.005 0.3-0.2 4.5 5.6 0.3 1 1 0.05-0.02 0.02-0.01 0.7-0.5 1 TABLE 4 Significant effects of Da on the thoracolumbar and lumbosacral borders of the HI,, race considered independently Thoracolumbar Genotype cornpanson + + vs. Da+ + + vs. DaDa Da+ vs. DaDa Lumbosacral XZ df P X2 df P 1.73 1.06 0.05 2 2 0.5-0.3 0.7-0.5 0.9-0.8 7.6 5.1 0.1 1 1 1 0.01 0.05-0.02 0.8-0.7 1 TABLE 5 Shows the populations (portrayed by bar graphs) of figure 1 arranged with Telationship o f anterior (ribs) to posterior (PSV) borders for tests of significance between borders and f o r interaction of the three attributes. Significances of all tests are based on 1 degree of freedom Strain DA Strain 111, PSV PSV Ribs 26 12 13 71(54) 21(16) 92 Ribs T 27 ~~ 14(11) 26(20) 40 x2 = 21.8 p= 0.005 85 47 132 15(6) 65(27) 80 77.9 141 100 241 5(4) 47(41) 52 33.9 0.005 45 71 116 DaDa 12 13 12 13 126(52) 35(15) 161 x2 p 12 13 = x2 p Da + = =< X= 21 vs. 24 22 vs. 25 23 vs. 26 21 vs. 22 21 vs. 23 22 vs. 23 24 vs. 25 24 vs. 26 25 vs. 26 23 vs. 27 26 vs. 27 1.10 5.167 0.236 0.09 1.08 0.097 2.58 0.056 2.58 0.342 0.479 22 109 131 16(8) 121(64) 137 14.7 0.005 35 155 190 9(7) 96(77) 1 18 106 x2 = 31.4 P = 0.005 20 103 1 124 =< 19(10) 34(18) 53 x2= p=< +f 12 13 14 11(9) 7(6) < Significance of tests between 3 attributes Groups 4 ~~~~~ 7(5) 90(69) 97 24.6 0.005 15(11) 19(15) 34 p 12 13 = < 0.005 40(34) 24(21) 64 ~~ x2= < T 27 26 P 0.3 -0.2 0.05-0.02 0.7 -0.5 0.8 0.30 0.8 -0.7 0.3 -0.2 0.9 -0.8 0.3 -0.2 0.7 -0.5 0.5 -0.3 12 13 Strain I11 32(5) 52(8) 24(4) 513(83) 56 565 x2= 100.10 84 537 621 P = < 0.005 *.Groups of genotypes for DA and IIID. are numbered 21-23 and 24-26 respectively, from top to bottom and III is 27. 432 P. B. SAWIN AND MARYANN GOW growth when measured by body size is polygenic. The fact that both genotypes and genomes appear capable of producing at least some proportion of all four vertebral types, but in different proportions, indicates that the skeletal units are affected by more than one genetic factor. Although the Da gene does have an effect upon the homeotic shifts, those shifts are varied by other elements in the genome, and furthermore the apparent effect of Da in the one genome can be altered or suppressed by interaction with a different genome. Furthermore, it is clear that in spite of this interaction the effects of both the Da gene and the DA genome can be distinguished from each other and can be within rough limits defined. The significant differences in proportion of individuals manifesting one or the other or both border shifts show that Da in either one or two doses tends to shift both borders anteriorly, but that the effect on the ThL border is the greater. The DA genome has a similar directional effect but of greater magnitude and the effects of Da and DA when combined are complementary. The effect of 111,. upon the borders is in the posterior direction. It tends to be epistatic, suppressing or masking the effects of Da, and what is more significant, to alter its localization. The significant effects of Da in the III,, genome are more apparent at the LS border. In general, between genome and between genotype, the significant differences in VF and in the two borders considered separately indicate differences of interaction between several genetic components. This is supported by the highly significant Xz tests for independence of the two borders within genotypes and strains and by significant evidence of strain difference between the interaction of the two borders when the Da heterozygote of the two strains is compared. Accordingly, these variations are in harmony with the concept of a major gene (Da) plus modifiers as the most likely and presently accepted interpretation for the range of expression and penetrance which is usual in morphological variations. When the overall pattern of significant differences in homeotic border shifts is considered, it is apparent that the actual effect is either (1) to alter the size of thoracic and lumbar areas in relation to each other by some fraction of one vertebra, or (2) in those cases where both borders are shifted in the same direction to shift the position of the entire lumbar region by that amount. These two effects, size and position, are essentially integrated in determining the borders and together are determined by genome entities acting on basic growth processes. Previous measurement of the lengths of centra of race I11 has shown that ( 1 ) the thoracolumbar border is almost identical with the peak of gradients 3 and 4 and that the lumbosacral border coincides with the valley between gradients 4 and 5, and (2) these reference points are relatively constant from beginning ossification to well after birth (Sawin and Crary, '64) at which time they are undoubtedly permanently established. In the same communication, certain significant relationships between skeletal malformation and the species gradient pattern were shown. When considered in relation to that pattern, these border shifts are understandable as reference points portraying the interactions of adjacent gradients in response to genetic changes in relative growth capacity. Because of the broad distribution of effects of the dachs gene, Gruneberg ('63) has classified Da under the systemic disorders of the cartilaginous skeleton and the study of these border effects of the axial skeleton of Da constitutes further substantiating evidence. Time is being widely recognized as a critical factor in determining induction of malformation both experimentally and genetically. The effect of Da being greatest at the anterior border in DA, its reduction of ribs (or anterior shift) may indicate an earlier genetic growth retardation in the cartilaginous skeleton, possibly even in the late membranous stage, and hence on the primary anteroposterior gradient. The effects of either genome, however, are more extensive than are Da's and particularly that of I&, which is exerted more posteriorly. This may indicate a slightly later origin in the metameric process, actually perhaps effects of the differentiating secondary (thoracic and lumbar) gradients. However, the fact that the Dd anterior border shift (as observed in the GENE AND GENOME INTERACTION DA) when transferred to the III,, appears more posteriorly suggests that it is not exclusively a “spuriously” pleiotropic effect of Da (Gruneberg, ’63), nor solely an expression of its generalized retardation (Sawin and Crary, ’57), in which case it would be a “genuine” pleiotropic effect, but is a manifestation of the interaction of several gene-induced growth processes best described at present with reference to the primary and secondary gradients. Although of all the effects, those of genome appear much greater than those of the Da genotypes, it seems reasonable to assume that this is due to the number of genes involved which have been concentrated by selection and inbreeding in race I11 and transferred to IIID,.As a single gene in the DA background Da’s effect is prominent in both one and two doses and although less so in IIL, the shift in localization of effect is highly indicative of the mechanism involved, i.e., the gradient being a result of the cumulative contribution of both a mutant gene and multiple units of the genome. Such difference in timing and localization could account for the irregularity revealed graphically in figure 2 and statistically in the heterozygous interstrain difference (table 5), where the effects of the gene ( D a ) and genomes are both most clearly expressed and effective in alteration of the vertebral type. Logically, it may well be questioned whether specific units of the modifier genome can ever be identified with more precision than those determining body size. However, the fact that some degree of identification has been achieved here by use of these border units suggests that closer examination of adjacent skeletal units may in a similar way reveal the contour of others of the localized gene determined processes. In this paper, no consideration has been given to either the parental phenotype or genotype insofar as the VF or separate borders are concerned. Furthermore, thus far no attention has been given to other reference points either between these borders or in either direction beyond them which would help to define the gradient limits. Nor has attention been given to precise chronological aspects of the gradient growth processes. Each of these could be closely associated 433 in the gradient pattern and thus contribute additional knowledge of the gradient mechanism. Although the possibility of overdominance or one gene heterosis as an explanation for the interstrain difference in heterozygous effect cannot be ruled out, the suggestion that “overdominance may be a function of the background genotype rather than a characteristic of a given locus,” (Wallace and Madden, ’65) at present seems the more applicable and tenable for this variation. It can be explored by breeding tests of the interaction of these racial gradients using ( 1 ) matings of specific vertebral phenotype and genotype, and ( 2 ) other reference points in the skeleton to portray the limits of gradient activity and points of interaction. The latter would portray in more detail the magnitude and characteristics of the gradients. Such projects are in progress, some of which are approaching completion. As indirect interaction products effected by way of the species growth gradients rather than direct gene induced malformations, these border shifts and other spontaneous differences arising in relation to gradient borders merit further intensive genetic investigation. They have a bearing upon several major clinical problems, particularly those of spontaneous malformation of the teratologist (Wilson, ’65; Sawin, Crary, Fox and Wuest, ’65), and the quantification of morphological differences of the anthropologist (Haronian and Sugarman, ’65). Of most value to the first is the knowledge of specificity of effect of one gene Da (as one experimental insult) and how its effects may be modified in the society of two different genomes or as Leck (’65) has suggested, by “fortuitous mismatching of quantitative factors that in other combinations would contribute to normal development.” To the second, is the cartographic use of the skeleton as a substitute or supplement to present methods of quantifying morphological differences. Although these specific homeotic variations are usually of minor clinical significance and therefore this information may seem irrelevant to the malformation problems generally, the same homeotic variation when asymmetrical becomes scoliosis. When it involves an articulation 434 P. B. SAWIN AND MARYANN GOW such as the occipitovertebral border it becomes functionally handicapping. It is anticipated that this knowledge of gradient interaction gained from relatively simple homeotic shifts will be helpful in analysis of mechanisms involved in more complex malformations and syndromes of malformation (see also Sawin and Trask, ' 6 5 ) . CONCLUSIONS Study of the homeotic shifts of vertebral borders as affected by the Da gene when on two different genetic backgrounds shows that the effects of both gene and genome are distinguishable. The Da in either one or two doses shifts both thoracolumbar and lumbosacral borders forward, particularly the former. The effect of the DA genome is in the same direction, but significantly greater, and the two combined are additive. The effect of the III,, genome, by contrast, is in the posterior direction and epistatic. It suppresses and tends to shift the localization of Da effect posteriorly. Significant differences between borders when considered separately and in relation to each other indicate differences of interaction between gene and genome. The fact that the only significant difference in the three way comparisons of border, genotype, and genome is between the Da/+ genotypes of IIIn. and DA as a manifestation of overdominance can be discounted because of the differences in border interaction. As reference points defining the relative size and position of the thoracic and lumbar regions, these borders reveal the activity of gene and genome to be an alteration of the relative size and position of the growth gradients of these two regions rather than a direct gene specific morphology. ACKNOWLDGMENTS This study was supported in part by Public Health Service grant C281 from the National Cancer Institute and in part by grant E40 from the American Cancer Society. Our thanks are due to Dr. R. R. Fox, Dorcas Crary, Adelaide Cousens, Esther Clark, Eugene Farrin, Bruce Plummer and Marjorie Muehlke who assisted in various ways in analysis and collection and preparation of specimens or roentgenograms. LITERATURE CITED Bateson, William 1894 Materials for the study of variation. Macmillan, London. Castle, W. E. 1931 Size inheritance in rabbits; the backcross to the large parent race. J. Exp. Zool., 60: 325-338. 1941 Size inheritance. Amer. Naturalist, 75: 488-498. Castle, W. E., and P. W. Gregory 1929 The embryological basis of size inheritance in the rabbit. J. Morphol. Physiol., 48: 81-103. Chai, C. K., and K.-H. Degenhardt 1962 Development of anomalies in inbred rabbits. J. Hered., 53: 174-182. Charles, D. R. 1938 Studies on spotting patterns. IV. Pattern variation and its developmental significance. Genetics, 23: 523-547. Child, C. M. 1925 The physiological significance of the ccphalocaudal differential in vertebrate development. Anat. Rec., 31: 369-383. Crary, Dorcas D. 1962 Modified benzyl alcohol clearing of alizarin stained specimens without loss of flexibility. Stain Technol., 37: 124-125. 1964 Development of the external ear in the Dachs rabbit. Anat. Rec., 150: 441-448. Crary, Dorcas D., and P. B. Sawin 1952 A second recessive achondroplasia in the domestic rabbit. J. Hered., 43: 255-259. 1958 Morphogenetic studies of the rabbit. XVIII. Growth of ossification centers of t h e vertebral centra during the twenty-first day. Anat. Rec., 127: 131-150. Crary, D. D., P. B. Sawin and N. Atkinson 1958 Morphogenetic studies of the rabbit. XXI. The nature of disproportionate dwarfism induced by the Da gene revealed by the early fetal ossification pattern. Am. J. Anat., 103: 69-98. Crary, D. D., and P. B. Sawin 1960 Genetic differences in growth rate and maturation of rabbits. Growth, 24: 111-130. DUM, L. C. 1942 Studies of spotting patterns. V. Further analysis of minor spotting genes in t h e house mouse. Genetics, 27: 258-267. Dunn, L. C., and D. R. Charles 1937 Studies on spotting patterns. I. Analysis of quantitative variations in the pied spotting of the house mouse. Genetics, 22: 14-42. Dunn, L. C., and S. Gluecksohn-Schoenheimer 1945 Dominance modification and physiological effects of genes. Proc. Nat. Acad. Sci., 31: 82-84. Green, E. L., and Margaret Green 1946 Effect of the short ear gene on number of ribs and presacral vertebrae in the house mouse. Amer. Nat., 80: 619-625. Green, M. C. 1957 Modifiers of the pleiotropic effects of the short ear gene in the mouse. J. Hered., 48: 205-212. Griineberg, Hans 1957 Genetical studies on the skeleton of the mouse. XIX. Vestigial tail. J. Genet., 55: 181-194. 1963 The pathology of development. Wiley, New York. Haronian, F., and A. A. Sugarman 1965 A comparison of Sheldon's and Parnell's methods for quantifying morphological differences. Am. J. Phys. Anthrop., 23: 135-142. - GENE AND GENOME INTERACTION Lamb, N. P., and P. B. Sawin 1963 Morphogenetic studies of the rabbit. XXXIII. Cartilages and muscles of the external ear as affected by the Dachs gene (Da). Am. J. Anat., 113: 365388. Landauer, Walter 1958 On phenocopies, their developmental physiology and genetic meaning. Am. Nat., 92: 201-213. Lash, James W. 1964 Normal embryology and teratogenesis. Am. J. Obstet. Gynecol., 90: 1193. Latimer, H. B., and P. B. Sawin 1957a Morphogenetic studies of the rabbit. XV. Measurements of the digestive tube and of its parts in normal and dwarf rabbits of race X. Anat. Rec., 129: 1-16. 1957b Morphogenetic studies of the rabbit. XIX. Organ size in relation to body size in large race I11 and in small race X. Anat. Rec., 129: 457-472. 1958 Morphogenetic studies of the rabbit. XX. A comparison of the weights of the brain and of its parts in a large and in a small sized race of rabbits. Anat. Rec., 132: 619-631. 1959a Morphogenetic studies of the rabbit. XXII. Linear measurements of large race I11 and small race X. Anat. Rec., 134: 6986. 195913 Morphogenetic studies of the rabbit. XXIV. The weight and thickness of the ventricular walls in the rabbit heart. Anat. Rec., 135: 141-147. 1960 Morphogenetic studies of the rabbit. XXVI. Correlation coefficients of external measurements of two inbred races of rabbits. Anat. Rec., 136: 417-425. 1960b Morphogenetic studies of the rabbit. XXXII. The postnatal growth in weight and thickness of the ventricular walls of the heart. Growth, 24: 59-68. 1961 Morphogenetic studies of the rabbit. XXVIII. Variations in shape and position of some of the viscera. Anat. Rec., 139: 85-88. 1962 Morphogenetic studies of the rabbit. XXXI. Weights and linear measurements of some of the bones of 65 race I11 rabbits. Am. J. Anat., 11 0: 259-268. 1963a Morphogenetic studies of the rabbit. XXXN. Weights and linear measurements of the bones of small race X rabbits compared with large race 111. Am. J. Anat., 113: 235-243. 1963b Morphogenetic studies of the rabbit. XXXV. Comparison of the weights and linear measurements in normal and heterozygous dwarf rabbits of race X. Anat. Rec., 146: 85-92. Leck, I. 1965 Etiology of malformations. New Eng. J. Med., 273: 864-866. McNutt, C. W., and P. B. Sawin 1943 Hereditary variations in the vena cava inferior of the rabbit. Am. J. Anat., 72: 259-289. McNutt, C. W. 1954 Gene expression as influenced by genetic background. J. Hered., 45: 235-240. Sawin, P. B. 1937 Preliminary studies of hereditary variation in the axial skeleton of the rabbit. Anat. Rec., 69: 407-428. - 435 1945 Morphogenetic studies of the rabbit. I. Regional specificity of hereditary factors affecting homoeotic variations in the axial skeleton. J. Exp. Zool., 100: 301329. Sawin, P. B., and Dorcas D. Crary 1951 Morphogenetic studies of the rabbit. X. Racial variations in the gall bladder. Anat. Rec., 110: 573590. 1956 Morphogenetic studies of the rabbit. XVI. Quantitative racial differences in ossification pattern of the vertebrae of embryos as an approach to basic principles of mammalian growth. Am. J. Phys. Anthrop., 14: 625-648. 1957 Morphogenetic studies of the rabbit. XVII. Disproportionate adult size induced by the Da gene. Genetics, 42: 72-91. 1964 Genetics of skeletal deformities in the domestic rabbit (Oryctolagus cuniculus). Clin. Orthopaedics, 33: 71-90. Sawin, P. B., Dorcas Crary, R. R. Fox and H. M. Wuest 1965 Thalidomide malformations and genetic background in the rabbit. Experientia, 21: 672-685. Sawin, P. B., and H. W. Edmonds 1949 Morphogenetic studies of the rabbit. VII. Aortic arch variations in relation to regionally specsc growth differences. Anat. Rec., 105: 377-398. Sawin, P. B., and I. B. Hull 1946 Morphogenetic studies of the rabbit. 11. Evidence of regionally specific hereditary factors influencing the extent of the lumbar region. J. Morph., 78: 1-26. Sawin, P. B., Mary Ranlett and Dorcas D. Crary 1962 Morphogenetic studies of the rabbit. XXIX. Accessory ossification centers at the occipitc-vertebral articulation of the dachs (chondrodystrophy) rabbit. Am. J. Anat., 111: 239257. Sawin, P. B., and M. Trask 1965 Morphogenetic studies of the rabbit. XXXV. Pleiotropic effects of the dachs gene and the gradient growth pattern. J. Morph., 117: 87-114. Snedecor, G. W. 1946 Statistical methods (4th Ed.). Iowa State College Press, Ames, Iowa. Tanner, J. M., and P. B. Sawin 1953 Morphogenetic studies of the rabbit. XI. Genetic differences in the growth of the vertebral column and their relation to growth and development in man. J. Anat., 87: 54-65. Wallace, B., and C. Madden 1965 Studies of inbred strains of Drosophila melanogaster. Amer. Naturalist, 90: 495-510. Willier, B. H. 1954 Phases in embryonic development. J. Cell. and Comp. Physiol., 43: 307317. Wilson, J. G. 1965 Embryological considerations in teratology. Ann. N. Y. Acad. Sci., 123: 219. Wright, S. 1963 Genic interaction. In: W. J. Burdette, ed. Methodology in mammalian genetics, pp. 159-162. Holden Day, San kancisco. Wright. S., and H. B. Chase 1936 On the Benetics of the spotted pattern of the guinea-plg. Genetics, 21 : 758-787.