IV Conference internationale de genetique (1911) pp. 381-393
Variation in First Generation Hybrids (Imperfect Dominance):
Its Possible Explanation Through zygotaxis1
W. T. Swingle

1 Communication faite à la troisiènne séance de la Conférence.

Current theories of heredity and variation give no method of explaining variations in the first generation hybrids between pure bred parents.

It is generally assumed that differences in the characters of organisms are due to corresponding differences in the composition of the gametes that unite to form the individual organism. No doubt that is true in many cases, but it is my purpose to point out cases in which this asumption cannot be sustained.

Some first generation hybrids show a very considerable range of variation, far more than can possibly be accounted for by any qualitative differences in the chromosomes or other bearers of heredity received from the two parents.

Inasmuch as during the whole of the first (or conjugate) generation the chromosomes from the two gametes persist side by side in an unfused state, such variations cannot be due to any quantitative differences in the bearers of heredity possessed by the sister hybrids.

If proof can be given to show that in certain specific cases pairs of gametes of identical hereditary composition give rise to very diverse organisms, the way has been opened for a general reinvestigation of the validity of our modern theories of heredity.

2Kellerman, W. A., and Swingle, W. T., 1899. Crossed Corn the Second Year, in 2nd Annual Report Kan. Agric. Exp. Sta. pp. 334-346, pl. 11. Topeka.

Some years ago when I carried on breeding work on maize in conjunction with my teacher, the late Prof. W. A. Kellerman, I was profoundly impressed by the diversity shown in different plants of first generation hybrids of identical pure parentage. All the ears on a single plant, no matter how many, were of the same general type, but those on other plants descended from the same two races were very different, often very decidedly different. A considerable number of hybrids showed this behavior.2

3Cook, O. F., and Swingle, W. T., 1905. Evolution of Cellular Structures, Bull. 81, B.P.I., U. S. Dept. Agriculture. (Aug. 4)

It must be remembered that such hybrids are made up of cells, the nuclei of which contain side by side the identical chromosomes defived from the two parental gametes. If the parents are pure bred, all the first generation hybrids are derived from identical gametes and have an identical equipment of chromosomes.3

4Cook, O. F., 1907. Transmission Inheritance distinct from Expression Inheritance in Science (N.S. 25: 911-912 n. 649, 7 June).

Any differences observed between such hybrids must be due to differences in the expression of hereditary tendencies and not due to differences in their transmission to the gametes as Cook has pointed out.4

In recent years I have found striking differences between first generation sister hybrids of two very different species of Citrus, which although showing some slight variations, are relatively pure bred in contrast with the wide chasm separating the very different species crossed to obtain the hybrid.

In 1897 I crossed the common sweet orange (Citrus Aurantium, sinensis L.) with the Oriental trifoliate orange (Citrus trifoliata L.), a deciduous species very distinct from the common citrous fruits.

In comparison with the wide chasm that separates these species, the variations they show inside their respective limits are insignificant. In other words, for purposes of genetic study, the germ cells may be considered pure in so far as neither of the parent species show variations that to any perceptible degree bridge the wide gulf separating them. In fact, the older trees of Citrus trifoliata growing in America are all descended from a very few plants imported from Japan some forty hears ago and show few variations and these of minor extent.

Fig. 1 — Colman cotrange, showing the spots due to fungous growths among the fuzzy hairs covering the skin. The oil glands are very small.

The sweet orange was grown from seed for a couple of centuries in Florida before the European varieties were introduced by General Sanford, the Rev. Lyman Phelps, and other progressive men about half a century ago. It is true that a number of varieties of so-called Florida oranges were found among these tens of thousands of seedlings, but they are all of the same general type and cannot for a moment be held to obscure the profound differences separating them, and in fact all known sorts of the common orange, from Citrus trifoliata. Nevertheless, eleven of these sister hybrids grown in 1897 from seeds of a single fruit of Citrus trifoliata crossed with pollen of a single flower of the common orange showed considerable diversity in leaf character and very striking diversity in the fruit characters. One variety of citrange as these new hybrides are called the "Morton" has very large, round, smooth, orange-colored fruits; another, the "Colman", has depressed, globose, yellow, fuzzy-hairy fruits; the "Willits" has a large percentage of fingered fruits; the "Rustic" often has double fruits, a smaller fruit situated on top of the larger one, and the calyx is frequently enlarged. The "Phelps" is very bitter, while the "Saunders" has small fruits with scarcely a trace of bitter in the juice. Every one of these eleven hybrids is decidedly different from all the others in its fruit characters and some, like the depressed, fuzzy "Colman" and the fingered "Willits" are strikingly different.

These citranges are a new type of acid citrous fruits decidedly more resistant to cold than any now grown. They can be used for making acid drinks and for culinary purposes in place of lemons or limes. They are being grown in very many parts of the southern United States where no oranges or lemons could stand the severe winter weather.

Fig 2 — Willits citrange, showing a fingered fruit common in this variety. One fruit from above, one from the side (sectioned).

Great as are these differences in the first generation hybrids between the orange and Citrus trifoliata, still greater differences have come to light in those between the lemon and Citrus trifoliata. All hybrids between the sweet orange and Citrus trifoliata (citranges, as they are called) have trifoliate leaves and germinate in a normal manner. Hundreds of new citranges have been produced within the last few years, so it is possible to speak with assurance on this point.

Fig 3 — Rustic citange, in cross section and side view showing a double fruit, not infrequent with this variety.
Fig 4 — Saunders citrange, showing the skin roughened by the very large and prominent oil glands.

The citremons, or crosses between the lemon and Citrus trifoliata, on the contrary, show nearly 20 per cent of seedlings with an exaggerated development of hypophylls, and in the majority of cases never produce any normal foliage leaves at all, dying from starvation shortly after the reserve food material of the seed is used up. One parent, Citrus trifoliata, has a few hypophylls along the base of the stem of the young seedling, while the lemon, like the orange and all other common citrous fruits, shows a pair of rather large, rounded, sessile, opposite leaves. Some show five leaflets (this occasionally occurs in citranges also), others show unifoliate leaves or leaves with very much reduced side leaflets (which is rarely or never seen in citranges). Most of the citremons have trifoliate leaves with large lateral leaflets, which is the type characteristic of citranges.

There are several hundred of these hybrids under study. Three varieties of lemon, the Eureka, Villa Franca and Lisbon, were used in crossing with Citrus trifoliata and all three yielded a considerable proportion of seedlings (usually about one-fifth) with pronounced hypophylls. As yet none of the citremons have borne fruit, so it is not possible to compare the variations in the fruits with those of the leaves. Doubtless the striking fruit variations shown by the citranges will be equaled or more probably exceeded by the citremons, since the latter show a decidedly greater range of foliar variation.

Fig. 5 — Citremon, No. 46750 (Lisbon lemon crossed with Citrus trifoliata). Leaves mostly in form of hypophylls. Citremons show nearly 20 per 100 of seedlings with an exaggerated development of hypophylls. Fig. 6 — Citrus trifoliata, No 09MT401. A seedling plant, showing the hypophylls along the lower part of the stem. This species crossed with the orange produces the citrange; crossed with the lemon, the citremon. Fig. 7 — Citrange, No. 45095 (Citrus trifoliata crossed with the Thompson Navel orange); shows many 5 foliate leaves.

Enough has been said to show that there are striking variations among first generation hybrids of citrous fruits.

Messrs. Collins and Kempton, in an article presented to this congress, have given conclusive evidence of the varying hereditary nature of different individual plants of first generation plants of maise, although these differences were indicated by the nature of the endosperm itself representing F2, or the perjugate generation. A Chinese maise with a waxy endosperm was crossed with Mexican maize with a horny endosperm of the usual type. The horny endosperm was dominant, so all the kernels on the crossed ears showed the horny character. These kernels sowed separately gave rise to plants of the first generation (F1) which were self-pollinated. the resulting ears showed varying percentages of kernels with a waxy endosperm from 15.7 to 55.5, approaching nearly but not quite to the expected Mendelian ratio 1-3 (being 23.1 instead of 25 per 100).

In as much as the ears contained a large number of kernels there can be no doubt but that their varying percentages represented real differences in the hereditary composition of the first generation plants. It woulld be hard to find a more conclusive case since there could be no doubt as to the purity of the parents and what is more rare no possible doubt as to whether a given kernel had a waxy or a horny endosperm.

Such variations among individuals of first generation hybrids are not unknown to students of genetics and these phenomena have been referred to variable dominance by those who use the Mendelian terminology. It has not, however, been realized how fatal this phenomenon is to some of the chief tenets of modern theories of heredity. Already in 1897 at the Ithaca meeting of the American Association for the Advancement of Science, I called attention to this phenomenon and, to account for it, proposed a provisional hypothesis of which a summary was given in the following words:

"It was pointed out that Weismann's theory of reduction of chromosomes, though giving a plausible explanation of the differences observed between the first (uniform) and second (polymorphic) generations of most hybrids, is not only in disaccord with the observed phenomena of spore and pollen formation in higher plants, but fails to account for the extreme polymorphism often observed in the first generation of hybrids between races of cultivated plants, or between closely related species, as for example some racial hybrids of maize and some specific hybrids of Lychnis and Digitalis. Mr. Swingle considered it necessary to assume in some such cases, at least, a predetermination of the characters of the hybrid at the time of fusion of the male and female nuclei. The male and female chromosomes probably persist side by side unchanged in number, and possibly unchanged in quality during the whole of the ontogeny of the hybrid, reductions not occurring until the close of the first generation. It is therefore necessary to assume, in order to explain the observed fact of divergence of character in the first generation of some hybrids, that the influence exerted during ontogeny of the hybrid by the material bearers of heredity, is at least in some cases, a function of their relative positions; and further, that in most cases the relative positions of these bearers of heredity, as determined at the moment of fusion of the male and female nuclei, persist unchanged throughout ontogeny of the offspring. Some phenomena, such as reversions to the one or the other parent form by a larger or smaller portion of the hybrid, would be explained by assuming some change in the disposition of the hereditary substance, whereby they assumed a new position of partial or complete stability. The suggestion was made that possibly the difference between uniform and polymorphic hybrids of the first generation is due to a more complete intermingling of the hereditary particles in case of polymorpic hybrids (offspring of closely related organisms), whereby many differing combinations would be possible, and in case of uniform hybrids (mostly offspring of distinct species or very different races of the same species), to greater or less aversion to commingling between the two more diverse sorts of particles, whereby they would remain in two separate groups and affect ontogeny uniformly and equally.

Since this announcement was made some evidence has been published to show not only that individual chromosomes persist through the so-called resting stages of the nucleus but also that the particular configuration is preserved from one cell generation to another with some accuracy.

It seems not improbable that this theory of a positional or vectorial influence of chromosomes may be further amended by assuming that the chromosomes which happen to lie nearest the nuclear wall and, ex hypothesis, thereby exert the greatest influence on the character of the cells containing them, are likewise better nourished than the chromosomes located near the center of the nucleus which are therey prevented both from exerteing their full measure of influence on the developing cell and from receiving the most abundant nourishment. It might be supposed that the well nourished chromosomes which from their superficial position on the nucleus had been preponderant in determining the character of the individual would also dominate during synapsis and tend to give as a result gametes similar in their hereditary character to the first generation hybrids whose character was determined by the configuration of the chromosomes which accidentally came about at the moment of fusion between the nuclei of the syngamete.

Three types of nuclear configuration might be assumed to occur in higher organisms. If a violent cross be made, very unlike species being hybridized, the paternal and maternal chromosomes might repel each other and occupy opposite sides of the syngamete nucleus, exerting equal influence upon the developing organism and as a consequence the first generation hybrids between widely distinct species are often but very slightly variable and are almost strictly intermediate between the two parents. Such hybrids are usually sterile, that is to say, the chromosomes are so unlike that synapsis (mitapsis) cannot occur. The mule is a striking example of this class of hybrids, which are also common among plants.

The other extreme is seen when abnormally inbred races of domesticated animals and plants are crossed. Such forms are likewise usually intermediate in the first generation and are fertile, the nuclei being able to pass through synapsis and form fertile gametes. The synapsis, however, is more or less imperfect and the paternal and maternal idioplasmic particles may not, properly speaking, fuse or blend at all but go over more or less unmixed to the resulting gametes. Possibly because of long continued narrow breeding and resulting great affinity for a non-related mate, the pairs of chromosomes might remain juxtaposed in the first generation, preventing much vectorial influence during the first generation. The dominance of certain characters which is regularly observed in Mendelian hybrids is due rather to the nature of the chromosomes from the different parents than to their relative position in the syngamic nuciei.

1The law of filial regression discovered by Francis Galton (Natural Inheritance, Macmillan Co., 1889, pp. 95 et. seq.) may in part be explainable as representing the measure of the extent to which these accidental configurations are wiped out during the succeeding synapsis. If such configurations were completely wiped out, filial regression would be normally 100% instead of something like 55% as found by Galton.
The third condition, probably normal in wild species, is that in which the chromosomes freely intermingle at the moment of fecundation, the configuration they then take on accidentally being perpetuated throughout the life of the individual. The different chromosomes exert varying degrees of influence on the cell, depending on whether they are located near the periphery or near the center of the nucleus.

The following synopsis exhibits the three types of fecundation.

  1. Interspecific hybrids: Usually sterile and intermediate.
    Violent crossbreeding. Chromatic material tend to remain separate:
  2. Mendelian crosses: Usually intermediate, dialytic at synapsis.
    Mendelism, crossing of inbred hybrids with imperfect or false synapsis.
  3. Normal cross-bred species: Usually vigorous, fertile and variable.
    Normal crossbreeding. Chromosomes freely intermingle at moment of fecundation and the configuration there taken persists through the ontogeny of the individual, giving great variety of forms. Synapsis normal.

As to the fact of considerable variations in first generation hybrids of relatively stable species or races there can be no doubt.

In the light of what has been said above, it cannot be doubted that we have in such variations a vital and important principle of heredity and one which goes far to explain many obscure phenomena of the everyday experiences of animal and plant breeders.

The suggestion, first made in 1897 and here reiterated, that such variation may, in part at least, be due to positional differences in the material bearers of hereditary characters is of course merely a provisional hypothesis. As such it may prove helpful in getting a clear idea of the phenomena to be explained efen if it does not aid us materially in such an explanation.

2Cook, O. F. 1910. Mutative reversions in cotton. Circular no. 53. U.S. Dept. Agric. p. 13. Mar. 21.

Cook2 has pointed out that this theory, of positional relations of the idioplasmic materials of the cell, puts a new resource in the hands of the evolutionist in the study of the bafflingly complex expression relations of hereditary characters.

It is easy to see that if we accept the hypothesis of the positional relation of chromosomes, a very large number of different forms could be expected to occur among hybrids of the same relatively pure-bred parents. In this way the practice of almost all succesful plant breeders in selecting the best individuals from among a large number of hybrids receives a new justification.

In order to have a name for this supposed positional influence of the parental idioplasm in the cells of cross-bred individuals, I propose the term zygotaxis.

From zygos, yoke (for zygote), and taxis, arrangement, with reference to the fact that such an arrangement of the chromatin would be most likely to occur at the moment of fusion of the parental gametes to form the zygote. The fundamental idea underlying the term zygotaxis is that the architecture of the zygote with reference to its idioplasmic particles, as well as its mechanisms for transmitting hereditary tendencies into expression, is determined to some extent at the moment of fusion of the parental gametes and that this particular arrangement of parts is transmitted to the cells of the organism to which the zygote gives rise.

By zygotaxis3 we mean the arrangement in the syngamete (zygote) of the chromatin and other hereditary substances derived from the parental gametes and the persistence of this arrangement in the cells produced by the subdivision of the syngamete. It is assumed that the particular zygotactic arrangement taken up by the chromosomes of the parental gemetes usually persists with little or no change throughout the life of the organism. It is easy to see, however, that many bud variations, sports, and other mutations could be explained by zygotactic changes occurring during ontogeny.

It must be remembered that the effects produced by zygotaxis are usually obscured by variations due to the varying hereditary composition of the gametes and only rarely when very constant races or widely different species are crossed can we see variations due to zygotaxis alone.

This does not mean that zygotactic effects are rare, on the contrary they probably occur as contributary causes of variation in much if not all of the material now being investigated by students of heredity.

Fig. 8 — Citremon, No. 46079 (Lisbon lemon crossed with Citrus trifoliata), showing leaves with very small side leaflets; an unusual type, seldom found in citranges. Fig. 9 — Citremon, No. 46554 (Lisbon lemon crossed with Citrus trifoliata), showing trifoliate leaves with very lage lateral leaflets as in citranges. Fig. 10 — Lisbon Lemon. The variety used in making the hybrids (Citremons) illustrated in Figs. 5, 8 and 9.