The Evolution of Genetic Systems (1958)
C. D. Darlington


i. Male Breakdown

So far we have considered evidence of genetic particles that was mostly available twenty years ago and the conclusions that could be derived from this evidence.1 We may now consider the new enquiries and new interpretations which have sprung from the notion of the plasmagene: a genetic particle which is versatile in its behaviour like the larger organellar genes, but with even more remarkable consequences of this versatility owing to its free situation.

These enquiries and interpretations have confirmed the expectation that there are physiologically and genetically many kinds of ultra-microscopic plasmagenes. Four of them may be usefully described.

The first class of plasmagene includes those responsible for male-sterility in plants. It is stable and purely maternal and for this reason it was the first kind to be identified. Also, of course, the anther is the first organ the plant produces which can be upset without upsetting the general vegetative life. This action in upsetting the development of the anthers always depends on the failure of the cytoplasm to react with a particular nucleus. Each species or race therefore has a nucleus and a cytoplasm each capable of varying in its effect on the development of the anther and with the two therefore nearly always adjusted in nature.

The co-adaptation, as we may call it, of nucleus and cytoplasm is best shown in Zea mays. Here numerous local and varietal differences in nucleus and cytoplasm are found to go together. Crossing therefore readily produces male-sterile forms. Just as cytoplasmically male-sterile forms are of value in the breeding systems of species in nature, so these are of value in the raising of hybrid seed in cultivation.

How do these cytoplasmic differences arise? They arise from mutation which can be induced, or rather vastly increased, by the choice of special nuclear genotypes. In many species of plants particular gene-differences render the plastids unstable. They thus cause irregular variegation. And, since the plastids are maternally inherited, the progeny include white seedlings. In maize a recessive gene, iojap, has this effect. It also causes patches of male-sterility in plants where it is homozygous. This male-sterility is also maternally inherited. The patches of aborted anthers do not however correspond with the stripes of white leaf. Thus it seems that the iojap gene has the effect of rendering unstable some cytoplasmic element or process which reacts independently with the two kinds of genetic particle.

This evidence of a genotypic control, or more strictly a nuclear control, of the stability or mutability of plasmagenes is worth connecting with the corresponding evidence for the control of nuclear genes. For this kind of control is a key property in the evolution of genetic systems. It was in maize also that a particular gene Dt had the property of causing another gene a1 to mutate to A1 and so give a purple flaking on green leaves. Or we may say the allele of Dt has the property keeping a1 stable. This gene difference has no other known effect. Evidently this is the same relation as that found between nuclear genes and plasmagenes. For the stability or instability of each element in the cell-system all the elements are bound up together; they are mutually adjusted, or selectively balanced.2

Male sterility plasmagenes are now known in many plants to be strictly matrilinear and stable in inheritance. Only those which happen to have these qualities have any possibility of being discovered and identified. Many more that are less predictable may therefore also exist.

Floating in several populations of Drosophila bifasciata is a plasmagene of a similar kind. It kills nearly all the XY embryos, that is the males, in the progeny, before hatching.3 This reaction of nucleus and cytoplasm thus produces much the same effect on the population and on its breeding system as the sex-ratio gene complex in other Drosophila species: it raises the proportion of females in the whole population. In its effect on the individual, that is physiologically, the reaction is much the same as those producing male-sterility in plants.

ii. Age Breakdowns

A second class of plasmagene is that which arises frequently as a mutation during development and seems indeed to be a symptom of development. It suggests a quantitative rather than a qualitative change in the effective particles. The classical example is Bateson's rogue pea.

Rogues in peas arise in certain varieties (that is in certain nuclear and cytoplasmic genotypes). They have narrower leaves and fewer flowers than the normal type. The rogue pea passes on its character to all its progeny. Selfed it always breeds true. When crossed with the type either as male or female it yields progeny which become rogue gradually or suddenly during development. On these changing plants the selfed progeny of successive flowers also show a changing proportion of rogues. The heredity goes with the development.

Here we have cytoplasmic inheritance since there is no segregation. But it is biparental. Indeed the crosses become rogue more quickly if the rogue parent is the male than if it is the female. Thus once again the plasmagene contradicts what was formerly the critical test of cytoplasmic inheritance, the predominance of the maternal influence. But from the genetic point of view influences are of no more use than essences. We have to think in terms of the propagation of particles. We must then say that, like the plastid-determinants in Pelargonium, but not of other plants, this type of plasmagene seems to multiply or to be localised in one tissue more than in another, in the pollen rather than in the embryo-sac.4

Similar to the rogue in peas is the breakdown in strawberries which seems to make their plastids unstable in the strongest light and is known as June Yellows. Like the rogue, this breakdown is an irreversible and hereditary change which has been attributed to plasmagene mutation.5 It takes place in clones after a certain number of years of vegetative propagation. All the plants in different places behave in a roughly similar and therefore predictable way. And like the rogue in peas, the yellows in strawberries is transmitted more strongly and at an earlier stage of life through the pollen than through the embryo-sac.6

A third type of plasmagene is that responsible for several other kinds of mutations in plants. The most notable is the rogue tomato with poorer fruiting than the otherwise purebreeding varieties in which it occurs. Among tomatoes grown in glasshouses in England the seed is germinated at a higher temperature than when the crop is grown in the open in warmer climates. Also seed is used from fruits of the later, the higher, infiorescences. These two circumstances to which the species is not adapted combine to induce a proportion of seedlings, up to one half, of the abnormal type or rogue.

How do these rogue tomato seedlings breed? Unlike the rogue peas they have no more rogues in their progeny than their normal sisters. Hence we might argue that this is not a genetic or hereditary problem in the ordinary sense at all but a developmental one. A high temperature merely shifts the mode of development of the young seed from normal to rogue without having any effect on the next generation. But this is not the whole story. Seed from the top inflorescence of any plant gives a higher proportion of rogues than seed from the bottom inflorescence; the two lots are grown of course at the same temperature. Thus seedlings of the variety Ailsa Craig grown at 14°C give 5 per cent of rogues from the two bottom, 10 per cent from the two top inflorescences. At 30° they give 15 per cent of rogues from the two bottom, 22 per cent from the two top inflorescences.

How are we to describe this? The seeds in different parts of the tomato plant differ in their inherent inborn capacity. And we cannot fail to ascribe this difference to variation in a genetic particle. The plant changes in its genetic capacity during development in the same way as the normal-rogue crosses do in peas. But the genetic difference which in peas can be a difference between whole rogue plants and whole normal plants is never more than a difference between the parts of a rogue plant or the parts of a normal plant in tomatoes.7

The importance of this genetic aberration is twofold. On one hand it shows a relation with development like that of the rogue pea combined with a kind of physiological response like that of the Euglena plastids. In these respects it has the character of a genetic particle. On the other hand it shows where one of the fundamental antitheses of genetics seems to break down, that between genotype and phenotype. Lewis overcomes this diffi­culty by saying that 'the difference between rogue and normal is phenotypic and not genetic'. To this however we must add four peculiar circumstances:

(i) Its determination arises or begins in the preceding generation. (ii) It varies in development. (iii) It is in part a maternal effect carried over to the progeny. (iv) The maternal effect does not vary in relation to the phenotype of the mother.

In these circumstances we find the distinction between heredity and development breaking down. The breakdown is genuine in the sense that a particle whose variations are evidently playing a part in development can have these variations transmitted through germ cells, that is in heredity. The same kind of genetic particle, which we may call a plasmagene, can be effective both in heredity and development. We may also notice that organisms benefit from separating the two modes of change; and they are usually successful in their natural environments in so organising their genetic materials that the separation is effective. Here the cultivated tomato fails.

iii. Heredity and Development

The genetic properties of rogues in peas, tomatoes, potatoes and many other cultivated plants compel us to admit that genetic particles which vary in heredity may also vary in development. The distinction between heredity and development, which is so general and so useful under natural conditions, breaks down here and there with artificial plants and artificial treatments. We must then refer to variations occurring within vegetative individuals, variations which for different reasons can be transmitted by sexual reproduction, as genetic and due to the action of genetic particles. In the higher animals the important group of such variations concerns cancer. In microorganisms the important group concerns the antigens whose transmission has been studied in Paramecium.8 In plants the important group concern the transformations of development in long-lived trees and shrubs.

Trees, in the course of their development, undergo certain irreversible changes of character. These prevent the propagator using their shoots as cuttings for they deform the tree that is produced. The most striking of such changes are known in Cedrus. In the Cedar of Lebanon side branches may be propagated and they produce a tree bearing no resemblance in habit to the ordinary cedar or indeed to any other conifer. This situation is in marked contrast to that in the small ornamental and fruit trees which are regularly propagated by grafting. It indicates an irreversible change in the normal vegetative life of the cells which stands half-way between the differentiation of the higher animals and the abnormal development of rogue plants.

Three other types of vegetative change are well known in plants under cultivation. One is the witches' broom an irreversible dwarfing mutation spontaneous in hundreds of tree species which is responsible for many of the dwarf trees of miniature gardens.9 A second is fasciation, the potentially permanent flattening of growing points so common in herbaceous garden plants which gives, for example, the ornamental varieties of many succulent plants known as monstrosa or cristata. A third is the climbing sport which has arisen in some hundred varieties of modern garden roses and which is sometimes reversible by bud-propagation. All of these might be attributed to an upset in a cell system responsible for the orientation of mitoses. They might be referred to changes in a constellation of particles from one steady state to another. But there is yet a fourth type of change, that from a juvenile to a mature habit, in which the results of experiment seems to require single and specific genetic particles.

The flowering shoots of ivy have regularly lost the capacity which the juvenile shoots had of creeping and rooting. They may be propagated from cuttings and remain stable indefinitely in their arboreal habit. Indeed nurserymen offer them under the learned name of Hedera helix var. arborescens Loudon. The flowering shoots have thus the character of having undergone an irreversible change in regard to a cytoplasmic particle. Irreversible except in the normal course of seed formation. The flowering shoots may however be changed back to the juvenile state by two kinds of experimental treatment. The first consists in treating them with a low temperature shock or with an X-ray dose.10 The second consists in grafting, or growing together in the same solution, juvenile and flowering shoots. The first of these effects may be due to direct actions on the relevant particles. The second seems to be indirect: a watersoluble substance in minute concentrations is altering the conditions of propagation of the particle so as to shift the genetic character of the flowering shoots back to that of juvenile shoots.

The overlap between the work of genetic particles or plasmagenes in heredity and in development is manifested only under exceptional conditions in the flowering plants. And in animals its existence is still a matter of conjecture. In the fungi however this overlap is not exceptional. It is widespread. There are in the fungi examples of strictly matrilinear inheritance. These are attributed to plasmagenes and the plasmagenes may undergo, like those of protista, mutation by the action of specific reagents like acridine.11 But on the other hand there is also evidence of cytoplasmic variation during development which is transmitted to offspring arising from spores. It seems to be more effectively transmitted from asexual than from sexual spores. And it is most effectively transmitted by the fungal mode of vegetative propagation.12 Evidently in the fungi as in the flowering plants sexual cells are adapted to restrict or regulate or exclude cytoplasmic variation.

In this range of possibilities we see an analogy with the range of behaviour in rogues, whether peas or tomatoes. Later it will no doubt be possible to amplify the analogy and use the studies in one group of organisms to guide the exploration of another group. For the present however it is enough to say that in general the processes of sexual reproduction are adapted to distinguish sharply between two kinds of genetic particles in the cytoplasm. There are those whose variation in quality or quantity, in character or concentration, is sufficiently independent of the nucleus to be used in sexual reproduction. And there are those whose variation is sufficiently dependent on the nucleus to be used in development and differentiation. But there is no intrinsic difference, no difference in chemistry and physiology, which distinguishes between plasmagenes of heredity, the field in which they were first postulated, and plasmagenes of development, the field to which they later had to be applied.13

We now have to carry the same enquiry into a third field, that of infection. Chapter 23

  1. that is, in the first edition of this book.
  2. Rhoades, 1938, 1950.
  3. Magni, 1953.
  4. Darlington, 1949c.
  5. Darlington and Mather, 1949.
  6. H. Williams, 1955.
  7. D. Lewis, 1953.
  8. Beale, 1954.
  9. Hornibrook, 1923.
  10. Frank and Renner, 1955; cf. Darlington, 1949b.
  11. Ephrussi, 1953.       
  12. Jinks, 1956, 1957.
  13. Darlington, 1944; Sewall Wright, 1945.