ABSTRACT
All six possible reciprocal crosses were made between C57BL, C3H/He, and C3H/Bi mice. The last group differs from the other two in having predominantly 5 lumbar vertebrae.
The offspring of the C3H/He ♀ ♀ × C57BL ♂ ♂ cross had a lower average number of vertebrae than either parental population. The other crosses gave offspring falling within the range between the parental means.
There was a significant difference between reciprocals in the two pairs of inter-strain crosses. The F1 offspring tended towards the vertebral type of the maternal strain. No significant effect was apparent in the inter-substrain cross.
The evidence from the reciprocal crosses suggests that, although the C3H/He substrain has since 1930 diverged from the C3H/Bi substrain in the direction of a higher number of lumbar vertebrae, the two substrains continue to show similar maternal effects when crossed with the C57BL strain. Hence the genetic control of maternal influence upon vertebral type can operate independently of the genetic control of vertebral type itself.
INTRODUCTION
It has long been realized that the biological properties and potentialities of the new-born animal are not conditioned solely by the chromosomal endowment received from its parents. Other means whereby a parental imprint may be stamped upon pre-natal development include the extra-chromosomal constituents of the gametes and, in species characterized by internal fertilization, the biological environment acting during post-zygotic existence within the mother’s body.
In mammals the post-zygotic connexion between the mother and her unborn young is intimate and prolonged. Yet the number of recorded instances of ‘maternal effects’ in mammals is not large, while cases which have received detailed study are very few.
We are concerned to analyse maternal effects upon a skeletal character—the number of lumbar vertebrae—for which natural variation exists both in laboratory (Green, 1941) and in wild (Weber, 1950) populations of mice.
The facts which provided the basis for our investigation are the following. Whereas over 95 per cent, of mice of the C57BL inbred strain have 6 lumbar vertebrae, there are substrains of the C3H strain which have predominantly 5 lumbar vertebrae (Green, 1941). Russell & Green (1943) and Green & Russell (1951) made reciprocal crosses between mice from C3H substrains of this type and C57BL mice. They found that when the maternal parent was from the C57BL strain only 29 per cent, of the F1 offspring had 5 lumbar vertebrae, as opposed to 57 per cent, in the reciprocal cross. The difference was manifest in the female as well as in the male offspring, and hence could not be attributed to sexlinkage. In the light of this result we resolved to investigate in detail the causes of vertebral variation within and between the two strains, with particular reference to maternal effects.
Our first step was to make an extensive survey of extant colonies of the three main substrains of the C3H strain, with results which were reported in the first paper of this series (McLaren & Michie, 1954). We confirmed Green’s (1953) finding that although the C3H/St and C3H/Bi substrains of the C3H strain have predominantly 5 lumbar vertebrae, the other main substrain, C3H/He, has predominantly 6. We also established that this difference has a genetic basis.
In this paper we report the results of reciprocal crosses of the C3H/Bi and C3H/He substrains both with the C57BL strain and with each other.
MATERIAL AND METHODS
Our colony of C57BL mice belongs to the C57BL/How substrain and is derived from foundation stock obtained in 1950 from Dr. Alma Howard. The foundation stock of our C3H/He colony came from the same source, but that of our C3H/Bi colony was obtained from Dr. J. Craigie. The C3H/He and C3H/Bi substrains have been separated since 1930. For a full genealogy of C3H colonies and substrains see McLaren & Michie (1954).
In all our work on vertebral type the mice have been maintained in the temperature range 65°–70° F., and fed on diet No. 86 of the Rowett Research Institute supplemented with mixed grain. Classification of vertebral type is done exclusively by X-ray photography.
TERMINOLOGY
As in the previous papers, vertebral types will be denoted by the following symbols:
5/5 = five lumbar vertebrae on each side.
6/5 = six lumbar vertebrae on the left side and five on the right.
5/6 = five lumbar vertebrae on the left side and six on the right.
6/6 = six lumbar vertebrae on each side.
When we wish to characterize a particular strain or substrain as having predominantly 5 or predominantly 6 lumbar vertebrae, we shall add the number 5 or 6 in brackets, thus—C3H/Bi(5), C3H/He(6), C57BL(6).
RESULTS
We performed all six possible types of cross between C57BL, C3H/Bi, and C3H/He mice. Data on vertebral type in the three parental populations and the six F1 progenies are set out in Table 1.
STATISTICAL ANALYSIS
We have tested the significance of the difference between reciprocal Fi hybrids by the χ2 test, as shown in Tables 2, 3, and 4. There is evidence of heterogeneity between litters, since the corresponding χ2 value is in each case greater than the number of its degrees of freedom, significantly so in the data of Tables 2 and 4. Highly significant evidence of this phenomenon has been described in other material (McLaren & Michie, 1954, 1955). We have therefore assessed the level of statistical significance, not directly from a table of χ2 which might overestimate it, but from the mean squares obtained by dividing each χ2 by the number of its degrees of freedom. The ratio of the mean squares can be assessed from a table of the variance ratio (Fisher & Yates, 1938–53, Table V).
DISCUSSION
Russell and Green found a significant difference between the offspring of reciprocal crosses between C57BL(6) and C3H(5) mice. The first two Fi items of Table 1, as shown by the analysis in Table 2, strongly confirm their finding. Table 3 shows that the difference between the offspring of the other pair of reciprocal inter-strain crosses is also statistically highly significant. Table 4 shows that there is no significant difference between the offspring of reciprocal crosses between the two C3H substrains.
Two major points of interest emerge from these data:
Whereas in the crosses between C3H/Bi and C57BL the F1 offspring are intermediate between the two parental strains, this is not so of the cross between C3H/He females and C57BL males. Expressed as the percentage showing the 6/6 type, the parental values are respectively 80 ·4 per cent, and 98 ·9 per cent., while only 64 ·6 per cent, of the offspring are of this type.
Although the C3H/He substrain is predominantly of the 6/6 type, the females exert a maternal effect in the 5 / 5 direction when crossed to C57BL, just as do females of the C3H/Bi substrain which is predominantly 5/5. Judged on the crude percentage scale the C3H/He maternal effect might appear less strong than that of C3H/Bi, but on a normalized scale (McLaren & Michie, 1955) the two effects have approximately the same magnitude; the difference between reciprocals is in each case about
standard deviations.
It follows that the genetic changes which have differentiated the C3H/He and C3H/Bi substrains since their separation in 1930 have had effects on phenotype independent of those which they have had on the maternal influence. As skeletons, C3H/He mice differ from C3H/Bi mice in the 6/6 direction; as mothers, the two substrains exert maternal effects of approximately equal magnitude.
In order to determine whether the maternal effects described in this paper are exerted through the cytoplasm of the egg or through the uterine environment, we have embarked upon a programme of egg transfer from C3H donors into C57BL uterine foster-mothers.
ACKNOWLEDGEMENT
We wish gratefully to acknowledge the support of the Agricultural Research Council.