selection in the opposite direction gave relatively longer tails. However, 417

Transcription

1 GENOTYPE AND ENVRONMENT N TAL LENGTH N MCE. By S. A. BARNETT. From the Department of Zoology, Glasgow University. (Received for publication 24th February 1965) Mice (Mus musculu8 L.) of (i) inbred strains A/Tb, A2G/Tb and C57BL/Tb, (ii) five F1 types produced by crossing the inbred strains, and (iii) a mixed stock derived from the three inbred strains and GFF, have been reared in environments kept at (a) 210 C. and (b) - 3 C. Body and tail lengths were recorded at 16 weeks and, for A/Tb and A2G/Tb, also at 3 weeks. Strain A2G/Tb had a typical mouse tail, near the length of the body, with about thirty-four vertebrae (including sacral); C57BL/Tb had a slightly shorter tail, with about thirty-two vertebrae, and A/Tb a much shorter tail, with a blunt tip, and about twenty-four vertebrae. Evidence is given that the condition of the tail in strain A/Tb is recessive. The mixed stock and all F1 groups, including those produced by crossing A/Tb with C57BL/Tb, had typical long tails with about thirty-three vertebra. The mixed stock had the longest tails. n all classes of mice absolute tail lengths were lower at - 3 C. than at 210 C. This was the case despite the fact that, in some classes, body length was the same in the two temperatures. Hence the effect of cold on tail length is independent of its effect,on body length. All classes of mice, except the mixed stock, had fewer caudal vertebrae in the cold. The size of the caudal vertebrae was also reduced in the cold. The mixed stock, selected for fertility at - 3 C., developed progressively longer tails over eighteen generations. There was no consistent influence of cold, or of heterozygosis, on variance in tail length. But variance was higher in the reducedtail strain A/Tb than in the other two inbred strains. These observations do not support the hypothesis that the tails of mice have a thermoregulatory function in a cold environment. THE almost hairless tails of the Muridae provide excellent material for studying the interactions of nature and nurture in mammalian development. This applies especially to the tail of the laboratory mouse, Mus musculus L., with its many inbred strains. This type of tail is commonly thought to have a thermoregulatory function. t is shown below that inbred, F1 and heterogeneous mice, with tails of typical length, grow shorter tails in an environment at - 3 C. than at 210 C.; and that this is also observed in a strain with reduced tail length. Variance in tail length is not consistently influenced either by low temperature or by heterozygosis. These observations have some bearing on the question whether the mouse tail is important for the regulation of temperature. Previous Work on Genetical Effects. - Polygenically controlled variation in tail length in mice has been studied by MacArthur and Chiasson [1945]. They, and later Cockrem [1959], found that selection for increased body weight produced mice whose tails were shorter relative to body weight, while selection in the opposite direction gave relatively longer tails. However, 417

2 418 Barnett Falconer [1954] selected another strain for increased body weight and observed an accompanying increase in relative tail length. Effects of single loci on tail length in mice have been reviewed by Gruneberg [1952, 1963]. He describes short-tail (T), a dominant condition which is lethal in the homozygous state; in heterozygotes deformities, such as ankylosing of caudal vertebrae, are common. Another abnormality, tail-short (Ts), investigated by Deol [1961], is a semi-dominant condition in which the tail is not only short, but also kinky; there are other skeletal effects throughout the body. Other abnormalities of the tail are irrelevant for the present study. Previous Work on Environmental Effects. - Przibram [1923] wrote of the Schwanzthermometer of rats and mice. He reviews early work, such as that. of Sumner [1915] on mice, and Bierens de Haan and Przibram [1922] on rats, in which the effects of high and low temperatures were studied. Tail length increases with the temperature at which rats and mice are reared. Sundstroem [1930] confirmed the effect of heat on the tail of laboratory rats; Sakharov [1949] has confirmed an effect of cold; and Chevillard et al. [19631 have reported in more detail effects of both cold and heat. Ashoub [1958], Harrison et al. [1959] and Harrison [1963] have studied the lengthening effect of heat on the mouse tail. Harrison and his colleagues have given a particularly detailed analysis, including observations on several inbred strains and the F1 mice produced by crossing them. MCE AND METHODS The Mice. - (i) Three highly inbred strains were used: A/Tb (a sub-strain with a short tail), A2G/Tb and C57BL/Tb. (ii) F1 mice were produced by crossing members of the inbred strains. Table shows the crosses used; the designation of the male is given first. (iii) A mixed stock was produced by crossing the three inbred strains mentioned and GFF mice. This stock was random-bred at 210 C., but was selected for reproductive success at -3 C. [Barnett, 1961]. Methods. - The mice were studied in two environments, kept at about 210 C. and -3 C., respectively. With minor exceptions, mentioned with the results, all belonged to breeding colonies kept permanently in one of these temperatures: consequently they were conceived, born and reared in one of the two environments and so were their parents. All had cotton wool for nesting. Further information on breeding is given by Barnett [1961]. Most of the mice were killed at the age of 16 weeks. They were then weighed, and the body-length (nose tip to middle of anus) and tail-length (middle of anus to tail tip excluding hairs) recorded to the nearest 1 mm. Some were similarly killed, weighed and measured at the age of 3 weeks. Observations on body weights have been reported separately [Barnett and Scott, 1963]. Some mice were skinned, put in fixative and later X-rayed to give photographs showing the lumbar, sacral and caudal vertebrae. To do this the bodies were flattened so that the whole of the vertebral column was in one plane. The number of caudal vertebrae was counted from the first vertebra after the last lumbar vertebra, and so usually contains three sacral vertebrae. Some measurements were made of tenth and fifteenth vertebrae counted in the same way. They were made to the nearest 041 mm. with a vernier microscope; lengths were taken in the midline of each vertebra and breadths at the: narrowest point.

3 Tail Length in Mice 419 TABLE 1. MEAN BODY AND TAL LENGTHS, WTH STANDARD ERRORS AND PERCENTAGE COEFFCENTS OF VARATON (V). Strain A2G { 21 y - 30 C3 {210 { C57BL { 23 A { { -3 A { 210 A d 210 C57BL C57BL 10 { S? { A2G {- 3 x d{ - 3 Random- 210 bred - 30 C57B Relative length tail length x 100. body length Bodylength Tail-lengths mrm. 97 ±0-6 92±1-4 94±0-6 90±1-1 No. of mice ±2-0 92±1-2 90± ± ± ± ±0-8 85± ± ± ±0-3 94± ± ± ± ± ±0-9 92± ± ± ±0-4 92± ±0-7 87±1-5 96±1-0 89±1-9 93± ± ± ± ±0-6 mm. V relative V 93 ± ± ± ±1-Ot 5 89± ± ± ±1-5t 7 83± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±2.5* 12 91± ±0-6t 3 69± ± ± ± ± ±0-9t 5 96± ±2-9t 11 90± ± ± ±0.7* 4 96± ±1-2t 8 97± ±0-7t 4 90± ± ± ± ± ± ± ± ± ±0-9t ±1-1 5 * Significantly different from 210 C. (P < 0-01). t Significantly different from 210 C. (P < 0-001). Temperature differences of all absolute tail lengths are highly significant (P < 0-001).

4 420 Barnett RESULTS Genetics of Tail Length at 210 C. - Each of the three inbred strains had its own distinctive pattern of tail growth at 210 C. (figs. 1 and 2). Strain A2G had tails around 90 mm. long, i.e. about 4 per cent less than the length of the body (Table ). The modal number of vertebrw was thirty-four (Table, fig. 3). The tail tapered to a fine point in the manner usual in this species. n this strain, as in the other inbred varieties, there was a tendency for taillength to be less in the females than in the males; but there was no evidence TABLE. NUMBERS OF VERTEBRJE Class of No. of Standard mice C. mice Mode Mean error A2G J C57BL C5B A C57BL x A2G A2G xc57bl J * A xa2g J A2G xa x 'Random-bred' of an effect of sex on vertebra number in any mice, inbred or not; hence the figures of vertebra number in Table and fig. 3 are those for males and females combined. Strain C57BL had tails substantially shorter than those of A2G: at 210 C. they were just over 80 mm. long and about 10 per cent less than the length of the body. Their shape, however, was similar. The modal number of vertebrae was thirty-two. These differences from A2G were statistically all highly significant. Strain A was atypical. Tail length was about 65 mm. in males, less in females. n both sexes it was only about 70 per cent of the length of the body; this was not a consequence of greater body length, since the A mice had slightly shorter bodies than the A2G. The modal number of vertebrae was twenty-four. This represents a difference in vertebra number proportional to the difference in tail length. There were no other regularly occurring structural abnormalities, such as deformities of the vertebrae; but occasionally there was a tail 'kink', such as Gruneberg [1952]

5 A A2G C57BL FG. 1. Tle three types of tail foumii( in iiibre(l iii'ee. Males reare(l at 21 C. 1To face p)ye 420

6 A2G e 210C A C57BL g210c Tail Length in Mice A 210C A2G c9-30c aw C57BL d'-30c A d -30C 421 Fia. 2. Typical silhouettes, drawn from X-ray photographs, illustrating the vertebrae of the three types of tail of the inbred mice C 60 - A z w U < >24 A2G ;l C57BL a. LLL. -3 0C W , 21C A2G x x A C57BL mxa2g C57BL -3 C 0< >34 < < >34 < >34 NUMBERS OF VERTEBRAE FG. 3. Numbers of tail vertebre (including sacral vertebre: see text). The two upper rows of histograms compare the three inbred strains at the two temperatures; the two bottom rows, four classes of F1 mice. n every instance there is a shift to the left at - 30 C.

7 422 Barnett describes. The strain A mice already showed their reduced-tail pattern at the age of 3 weeks (Table ). All the F1 mice at 210 C., including A x C57BL, had mean tail lengths of 88 mm. or over; and in all, the modal number of vertebrae was thirty-three. The random-bred mice, derived from four inbred strains, had the longest tails; but in them too the modal number of vertebrae was thirty-three. The genetics of the reduced-tail of the mice of strain A was formally studied at 210 C. by crossing A with A2G; some of the F1 generation were backcrossed to A, others were bred to produce a F2 generation; F2 mice with short tails were also back-crossed to A. All were measured at 16 weeks. Fig. 4 and Table V summarize the main results. Reciprocal crosses gave no evidence TABLE. TAL LENGTHS AT 3 WEEKS: MEANS WTH STANDARD ERRORS AND PER- CENTACE COEFFCENTS OF VARATON. 210 C C. T-;diltlegth Tail x100-',. A length x100 N X S.E. V X S.E. V N i S.E. V X S.E. V A2G d A A of a maternal effect on tail-length; hence the results of such crosses were pooled. Body-length was unaffected in the various crosses. All tail lengths below 77 mm. were counted as reduced (see fig. 4). n both back-crossed and F2 mice there was evidence of segregation into normal and reduced-tail conditions. n the back-crosses the percentage of reduced-tail mice was 50; in the F2 group it was 33. All the offspring of the cross between reduced-tail F2 mice and the A strain had short tails: in the males the mean length was 69 ± 1.0 mm. (N = 30), and in the females, 66 i 1.6 mm. (N = 19). Effects of Cold. - Table shows that all the kinds of mice studied had shorter tails in the cold than in the warm environment. The tail-length of cold-reared male A2G mice was 84 per cent of that in the warm, and of females, 86 per cent. n the C57BL males the figure was 87 per cent, and in the females, 90 per cent. n these strains there were comparable differences when tail-length was expressed as a percentage of body length. n strain A the absolute tail lengths in the cold were 89 and 87 per cent, respectively, of those in the warm; but the differences in relative tail lengths between the temperatures were not statistically significant. The 'random-bred' males in the cold had a mean tail-length 92 per cent of that of controls at 210 C. Of the F1 mice, the crosses C57BL x A2G and A x C57BL displayed a response to cold similar to that of A2G and C57BL. But mice of the other three crosses resembled strain A. The preceding statements concern the tails of mice aged 16 weeks.

8 ..L.n.1.*A?G>L4 6 A do' A2G d' L2 2Ll Tail Length in Mice 423 LU (u F d' LL. 0 coe U = F19 m m F2 d' z 2 r M. n L 7T3 2 BACK CROSS K1u 4 BACK CROSS L TAL LENGTH, mm FG. 4. Genetics of reduced tail: distribution histograms of tail length. TABLE V. GENETCS OF REDUCED TAL. MALES AND FEMALES POOLED N EACH CLASS. Reduced tail No. of - - 'Expected' mice No. percentage percentage X2 A2G A * F F Back cross (There was no significant divergence from expectation.)

9 424 Barnett Table gives the tail-lengths of A and A2G mice at 3 weeks. The effect of cold was already evident at that age. There were also effects of cold on the number of caudal vertebrae. Apart from the mixed stock, all kinds of mice showed a clear, though slight, shift downwards in the cold (Table, fig. 3). The mixed stock males in both temperatures had a modal number of thirty-three; the mean numbers of their vertebrae were at 210 C. (N = 16) and at - 30C. (N = 14). No observations were made on mixed stock females. A loss of one or two, presumably terminal, vertebrae out of about thirty TABLE V. DMENSONS OF CAUDAL VERTEBRE N MALES OF SOME OF THE GROUPS: MEANS (mm.) WTH STANDARD ERRORS. No. of mice A Length 0-43 ± ± th vertebra Breadth ±0O ±0O0074 C57BL B ±0O ±0O *37 ±0O ±O0043 A2G f A ±0± ±0* *41 ± ±0O0508 A2G x C57BL f ±0± ± ±O0014 0*098 ±00043 A2G xa { ±0O ± ± ±0*0058 'Random- 210 bred' *45 ± *111 ±0O ±0O ± th vertebra Length Breadth 0*41 ± *32 ± *071 ±0O ±0O ±0O ±0O ±0O ±0O ±0O ±0O ±0O ± ±0O ±0O ±0O0041 0*45 ±0O ± ± ± ±0O ±0O ±0O ±0O0031 was not, however, the main source of the decrement in tail length. Table V shows that individual vertebra were consistently shorter in the cold than in the warm. They were also narrower: hence there was no substantial effect on shape, but there was a uniform reduction in total growth in the cold. One further effect of cold remains to be described, namely, the results of long-continued breeding in the cold environment. This has little or no effect in the highly inbred strain A2G [Barnett and Widdowson, 1965]. By contrast, there was a marked effect of selecting a genetically heterogeneous stock for ability to produce large litters in the cold environment. The fertility of these mice rose during the twelve generations of the main experiment [Barnett, 1961], and so did their body weight at 16 weeks [Barnett anid Scott, 1963]. The stocks were maintained in the two temperatures for eighteen generations in all, and the body and tail lengths of twelve to twenty-four males, aged 10 weeks, were recorded in each generation except the first two. There was a steady rise in absolute tail length over the generations. Since there was increased total growth, the change in relative tail length was not so striking. Nevertheless, as shown in fig. 5, relative tail length increased

10 Tail Length in Mice from the seventh to the fifteenth generation in the warm and from the sixth to the seventeenth in the cold. By the time the rise had (apparently) ceased, relative tail length in the mice in the cold was close to that of the controls. The increase in the cold, from a mean of mm. in generations three to - five, to mm. in generations sixteen to eighteen, was highly significant (P < 0.001). The smaller rise in the warm over the same number of generations, from 95 i: 0.7 to 106 1: 2.0 mm. was also highly significant (P < 0.001). t may have been due to inadvertent selection, resulting from infertility or losses in utero or in early life. Some of the cold-bred mice of the thirteenth generation were transferred to 210 C. at 5 weeks and there mated. Their offspring were heavier than the controls (with ancestry in the warm) to",.10-.-\ --.. o * - -. z w GENERATONS FG. 5. ncrease in relative tail-length of random-bred mice over eighteen generations. Upper curve, 210 C.; lower curve, - 30 C. 425 at 3 to 16 weeks [Barnett and Scott, 1963]; and the tails of these mice, as shown in the bottom line of Table, were longer than those of the controls. Variance. - Variance in absolute or relative tail length could be influenced by (i) the presence of a mutant gene which influences the length of the tail, as in strain A: a mutant gene in the homozygous state can lead to developmental instability, reflected in increased phenotypic variation [Waddington, 1942]; (ii) heterozygosis; (iii) an adverse environment, such as a cold room. The coefficients of variation in Table show that variance in tail length was never high, and suggest little effect of heterozygosis or cold. Since the distributions of tail length were negatively skewed, expecially in strain A, this was further tested by taking the logarithms of the tail lengths [Falconer, 1960, p. 292 et seq.], and calculating variance ratios. The variance, expressed in this way, in the tail length of strain A, at 3 weeks, was higher than that of A2G, in both sexes, at both temperatures (P < 0 05); the same applies to relative tail lengths except in males at 21 C. The same trend was evident at 16 weeks, but at that age there were fewer mice in each class, and the differences were not statistically significant. For the comparison between A and C57BL there are figures only for 16 weeks; the females of strain A varied more than those of C57BL at both temperatures (P < 0-05 at - 30 C.), but the males did not. This is perhaps an example of developmental instability resulting from the presence of a mutant gene. There was no consistent effect of the cold environment on variance in tail VOL. L, NO

11 426 Barnett length; but the coefficients of variation in Table suggests that, in both crosses between strains A2G and A, the low temperature increased variance. The logarithmically transformed figures did not, however, give statistically significant variance ratios. The only variance ratios which showed a significant effect of cold were those of the mixed stock at - 3 C., compared with both random-bred classes at 210 C.: the mice at - 3 C. varied less than those in the warm (P < 0.01). Finally, there is the question of the effect of heterozygosis. The coefficients of variation indicate no consistent effect of heterozygosis on relative tail length in either temperature. Nor did the variance ratios of the untransformed figures. This was confirmed by calculation of variance ratios from the logarithmically transformed figures. n seventeen out of forty comparisons, the F1 variance was higher than the inbred; but the difference was statistically significant in none. Heterozygosis and Luxuriance. - 'Growth heterosis', or luxuriance in tail length, was displayed in only one cross, namely, A x C57BL. The two inbred strains, A and C57BL, both with shorter tails than the typical (which is represented by A2G or the random-bred stock), produced F1 young with tail lengths very close to those of the A2G mice in both temperatures (Table ). DSCUSSON These observations confirm that a low environmental temperature diminishes the growth of the mouse tail in two highly inbred strains, and also in F1 and random-bred types derived from them. One inbred strain had a reduced-tail with about two-thirds the usual length and number of vertebrae; yet these mice too displayed an effect of cold on their tails. Shorter tails in a cold environment could in principle be due to an effect on the growth of the whole body: the body weight of some, but not all, of the classes of mice used, is lower at - 3 C. [Barnett and Scott, 1963]. Body lengths correspond (Table ). f altered length in the cold were merely a by-product of a general decline in growth, relative tail length should be higher in the cold in strains with reduced growth; but, in fact, it is not. The independence of the effect of cold on the tail, from that on body weight, has been further illustrated by Barnett and Widdowson [1965]. They studied two groups of A2G mice at - 3 C., one of recent ancestry in the cold and lighter than controls at 210 C., the other of long ancestry in the cold and similar in weight to the controls: the tail-lengths of the two groups were similar. Hence cold probably shortens the mouse tail through its direct action on the tail itself. A curious feature, previously (it seems) unreported, is the smaller number of vertebrae in the cold, found in all the classes of mice studied. A separate problem is whether possession of a shorter tail in a cold environment confers any advantage. The functions of the tail in mammals have been reviewed by Bopp [1954]. The tail in the Rodentia usually has few hairs; in this state its most likely functions are (i) balance and (ii)

12 Tail Length in Mice 427 temperature regulation. But not all rodents have naked tails: the members of the families Sciuridae and Ctenodactylidse, most of the sub-family Gerbillinae, and isolated genera such as Glis and Crateromys, have fully covered tails [Ellerman, 1941]. The functional significance of this (apart from their use in communication in some species) is not known. gnoring these exceptions, the hairless type of tail might be supposed to allow an adjustment of heat loss by variation of the blood supply. Hence, in a cold environment, in which heat loss needs to be minimized, shortening might be advantageous. Similarly, longer tails in a hot environment (mentioned in the ntroduction) could also be beneficial. Harrison [1958] gives evidence on this: mice whose tails had been cut off had lowered heat tolerance. By contrast, Wilber and Robinson [1961] found no effect of removing the tail on the thermoregulation of rats. The mice and rats of these two studies were, however, from laboratory stocks; there must always be some doubt whether such long-domesticated types adequately represent their species in their normal, that is, wild forms, as more fully discussed by Barnett [1965]. The mice of strain A/Tb illustrate the way in which a domestic variety can depart from the (probable) optimum: their tails at 210 C. were already shorter than those of most mice at - 30 C.; yet at - 30 C. they were shorter still, and had fewer vertebroe. Unfortunately, we do not know the physiological effects of an abbreviated tail, in any environment. Johansen [1962] has, however, given convincing evidence of increased heat loss from the tail of the muskrat, Ondatra zibethicus, during exercise. The best-known generalization about the appendages of wild animals in cold environments is Allen's 'rule', according to which, if a species has a variety living in an unusually cold habitat, that variety will have shorter limbs and tail. The difference is assumed to/ be genetically determined, unlike the short tails of cold-reared inbred mice. The mixed stock selected for resistance to cold did not conform to Allen's rule: their tails became longer, both absolutely and relatively, during seventeen generations in the cold, and nearly reached the relative length of the controls. The relative tail-length of these mice, after transfer to 210 C. from - 30 C., was the same as, or greater than, that of the controls. The secular change in tail-length in these mice accompanied an increase in body-weight, as well as a progressive improvement in reproductive performance [Barnett, 1961]. Hence, in this case at least, superior adaptation to cold, due at least partly to selection, was not accompanied by the expected change in the tails. Another approach to function has been proposed by Michie and McLaren [1955]: they suggest that, if the length of the mouse tail is adaptively important (i) it should be more uniform in F1 mice than in their inbred parents and (ii) the effect of cold should be more marked in F1 than in inbred mice. The second suggestion does not, however, take into account the superior coldresistance of F1 mice [Barnett, 1964]. n any case, the observations reported above do not support either prediction; hence, once again, the evidence leaves unresolved the problem of t.he thermal function of the mouse tail.

13 428 Barnett ADDENDUM Two recently published papers have thrown light on the role of the tail in thermoregulation. Rand et al. [1965] have recorded vasodilation in the tails of laboratory rats at environmental temperatures of C:; the tail can then lose up to 20 per cent of the animal's heat production. Coldadaptation (to C.) lowers the critical temperature for vasodilation. Thompson and Stevenson [1965] find that in forced exercise vasodilation in the tail and feet of laboratory rats ensures that the colon temperature does not rise above C.; vasodilation evidently depends on colon temperature. f these observations apply to rodents generally, the tail is probably important in thermoregulation only as a means of ensuring heat loss when a certain temperature is exceeded. Except at high temperature, blood flow would then always be at a minimum. RAND, R. P., BURTON, A. C. and NG, T. (1965). 'The tail of the rat, in temperature regulation and acclimatization', Can. J. Physiol. Pharmacol. 43, THOMPSON, G. E. and STEVENSON, J. A. F. (1965). 'The temperature response of the male rat to treadmill exercise, and the effect of anterior hypothalamic lesions', Can. J. Physiol. Pharmacol. 43, ACKNOWLEDGMENTS This work w%vas aided by generous grants from the Medical Research Council and the Wellcome Trust. t depended also on consistent technical help, especially by Kirsteen Borland, Jack Keys, Sandra Neil and Anne Tannoch. Grateful thanks are due to Dr. M. J. Little and Sybil Scott for statistical help. REFERENCES ASHOUB, M. A. EL-R. (1958). 'Effect of two extreme temperatures on growth and taillength of mice', Nature, 181, 284. BARNETT, S. A. (1961). 'Some effects of breeding mice for many generations in a cold environment', Proc. Roy. Soc. B. 155, BARNETT, S. A. (1964). 'Heterozygosis and the survival of young mice in two temperatures', Quart. J. exp. Physiol. 49, BARNETT, S. A. (1965). 'Adaptation of mice to cold', Biol. Rev. 40, BARNETT, S. A. and SCOTT, S. G. (1963). 'Some effects of cold and of hybridity on the growth of mice', J. Embryol. exp. Morph. 11, BARNETT, S. A. and WDDOWSON, E. M. (1965). 'Organ-weights and body-composition in mice bred for many generations at - 3 C.', Proc. Roy. Soc. B, 162, BERENS DE HAAN, J. A. and PRZBRAM, H. (1922). 'Erniedrigung der K6rpertemperatur junger Wanderratten (Mus decumanus) durch chemische Mittel und ihre Einfluss auf die Schwanzlange', Roux' Arch. 50, Bopp, P. (1954). 'Schwanzfunktionen bei Wirbeltieren', Rev. Suisse Zool. 61, CHEVLLARD, L., PORTET, R. and CADOT, M. (1963). 'Growth rate of rats born and reared at 50 and 300 C.', Fed. Proc. 22, COCKREM, F. (1959). 'Selection for relationships opposite to those predicted by the genetic correlation between two traits in the house mouse (Mus musculus)', Nature, 183,

14 Tail Length in Mice 429 DEOL, M. S. (1961). 'Genetical studies on the skeleton of the mouse. XXV. Tailshort', Proc. Roy. Soc. B. 155, ELLERMAN, J. R. (1941). Families and Genera of Living Rodents, Vol. 2. British Museum, London. FALCONER, D. S. (1954). 'Validity of the theory of genetic correlation', J. Hered. 45, FALCONER, D. S. (1960). ntroduction to Quantitative Genetics. Oliver & Boyd: Edinburgh. GCRUNEBERG, H. (1952). The Genetics of the Mouse. Nijhoff: The Hague. GRUNEBERG, H. (1963). The Pathology of Development. Blackwell: Oxford. HARRSON, G. A. (1958). 'The adaptability of mice to high environmental temperatures', J. exp. Biol. 35, HARRSON, G. A. (1963). 'Temperature adaptation as evidenced by growth of mice', Fed. Proc. 22, HARRSON, G. A., MORTON, R. J. and WENER, J. S. (1959). 'The growth in weight and tail length of inbred and hybrid mice reared at two different temperatures', Phil. Tran8. B. 242, JOHANSEN, K. (1962). 'Buoyancy and insulation in the musk rat', J. Mammal, 43, MACARTHUR, J. W. and CHASSON, L. P. (1945). 'Relative growth in races of mice produced by selection', Growth, 9, MCHE, D. and MCLAREN, A. (1955). 'The importance of being cross-bred', New Biol. 19, PRZBRAM, H. (1923). Temperatur und Temperatoren im Tierreiche. Leipzig: Deutige. SAKHAROV, P. P. (1949). 'The inheritance of altered characters in animals', Zool. Zh. 28, SUMNER, F. B. (1915). 'Some studies of environmental influence, heredity, correlation and growth, in the white mouse', J. exp. Zool. 18, SUNDSTROEM, E. S. (1930). 'Contributions to tropical biochemistry and physiology.. Supplementary experiments on rats adapted to graded levels of reduced cooling power', Univ. Calif. Publ. Phy8iol. 7, WADDNGTON, C. H. (1942). 'Canalization of development and the inheritance of acquired characters', Nature, Lond. 150, WLBER, C. G. and ROBNSON, P. F. (1961). 'Temperature regulation in decaudated rats', U.S. Army Chem. Res. Dev. Lab. Tech. Rep. CRDLR 3063, 1-11.