Evolution of Genetic Systems, pp. 180-184 (1958)
C. D. Darlington

iii. Invisible Particles

Knowing the special properties of extra-nuclear inheritance shown by the visible plastids, we are in a better position to test the evidence for invisible determinants outside the nucleus. We must not expect absolute constancy; we must allow for sorting out. We must allow for differential rates of propagation subject to changing external conditions. We must allow for varying aggregations and localizations. We must not therefore expect absolute matrilinear descent in plants; indeed we may expect a reversal of it; and accordingly we must be prepared for the predominance of either one side or the other in reciprocal crosses.

Reciprocal crosses between many pairs of species in plants differ, for example in Epilobium, Streptocarpus21 and Geranium. They differ, not in resembling the mother plant more closely, but in new and precise modifications. Reciprocal crosses between Funaria mediterranea and F. hygrometrica, for example, differ in sporocarp shape. By regeneration meiosis can be avoided and diploid gametophytes produced vegetatively. These again differ in leaf shape. They still differ if maintained vegetatively for 13 years. And further, if haploid gametophytes are backcrossed to the male parent eight times, so that the product has the nucleus of the male parent and the cytoplasm of its matrilinear ancestor, it still does not agree in form with its male parent.22

Such tests show the action of a permanent cytoplasmic determinant. Or, should we say, system of determinants? The answer depends on the nature of the crosses. In a particular difference between reciprocal crosses of individuals with a single gene difference, a specific cytoplasmic determinant may be assumed. In Linum a special gene from the tall flax aborts the anthers in a quarter of the F2 progeny of its cross with a procumbent species when this species is used as the female parent. A gene mutation has been accompanied by a corresponding cytoplasmic change.23 In these circumstances it is not surprising that the plant populations have been able, in the ways we have seen, to seize on cytoplasm-controlled male-sterility as an instrument in correcting their breeding systems.

Where generalized or multiple effects are concerned in a number of different determinants must be supposed to take part. Thus in Vicia Faba a number of plasmatic differences are to be inferred from reciprocal crosses between the subspecies major and minor. Amongst other effects the minor cytoplasm seems to eliminate the F2 segregates which are pure for certain genes from the stock. There are six genes concerned and they are linked in one complex.24 When we recall the importance of such complexes in developing the discontinuity between species it begins to appear that this action on linked genes is not a coincidence but a consequence of special adaptation of the cytoplasm to the nucleus during the differentiation of the two races  It implies the existence of a number of different and specific determinants in the cytoplasm.

The variety of changes in plasma types similarly argues a specificity in these changes. In Epilobium hirsutum half a dozen stocks from different localities differ in their plasmatic properties, as shown in reciprocal crosses with one another and with E. luteum. They differ not only quantitatively but qualitatively and give in the extreme cases a variety of probably unrelated effects.25

The cross luteum-hirsutum has been back-crossed to the male parent thirteen times so as to give a hirsutum nucleus in an approximately luteum cytoplasm. This product may be represented as Lhn. We then find that Lh14 x luteum resembles luteum x hirsutum and not the reciprocal cross. For example, it is male sterile. The cytoplasm of Lh14 remains after fourteen generations essentially luteum.

This is not the whole story. When a white-over-green hirsutum is used as the male parent of Lh14 one in 400 plants shows white flecks. Evidently this is due to the sorting out of pollen plastids from egg plastids. Examination of this generation shows that about one in 400 also has shoots with fertile anthers. Further, the male-fertile flowers selfed give fewer sterile progeny than the male-sterile flowers of the same plant (using presumably the pollen of the fertile flowers). The plasmatic determinants therefore just as much as the plastids are sorted out in development. In fact plasmatic inheritance of all kinds, like Mendelian inheritance, is particulate.

Particulate inheritance of the plasma does not mean that a single type of particle distinguishes luteum from hirsutum. As we saw, the several effects in Vicia and the several races in Epilobium indicate several kinds of particle. This is brought home to us in a special way by the same mutating plants of Lh14. When the fertile flowers of Lh14 are crossed with luteum pollen they give less fertile progeny than do the sterile flowers in the same cross. The sorting out of a determinant affecting the fertility of pure hirsutum nuclear type does not mean the sorting out of a determinant affecting the fertility of a hirsutum-luteum hybrid. Different and independently assorting determinants are at work.26

Taking these experiments together we see that genotype must be supposed to embrace three elements, the nucleus, the determinants of plastids and organellae, the pigment and fibre producers, and the submicroscopic particles which we cannot yet identify because they have no products attached to them.

All of them depend on self-propagating or genetic particles. The nucleus, to be sure, consists of particles of different sizes and is subject to variation of different kinds: for the other genetic particles we are still in the dark about these things. Yet if we give the name of genes to the particles of the nucleus, we have to use the same kind of name for the free particles in the cytoplasm. Those in the plastids and the cytoplasm may be treated more rigorously if we also think of them also as genes—plastogenes and plasmagenes. To be understood, and even sometimes to be discovered, we need to know them in relation to specific differences in nuclear genes. But this in no way robs the free genes of the specificity or integrity. It merely shows that we cannot so exactly control them.

In order that we should find out something more about these free genes we must try to form a more precise picture of how they live, move and multiply. It seems likely that they are nucleo-protein molecules or aggregates. Evidently they are such that, like trye genes, they can rise only from nucleo-proteins of the same kind—apart from mutation. Unlike true genes however their reproduction is not controlled by a mechanical equilibrium but will be subject rather to conditions of equilibrium genotypically controlled but specific for each type of gene. Where the particles are as large and complex as plastid determinants we may describe their equilibrium as physiological and we have seen how this equilibrium may be physiologically upset by light and nutrition. Where the particles are smaller as they doubtless are for most plasmagenes we may imagine that their equilibrium approaches nearer to a condition of chemical equilibrium. The conditions of cytoplasmic heredity are therefore likely to show a wide variation according to the type of equilibrium to which the plasmagenes concerned are subject. Probably also the varying kinds of differentiation found in animals, plants and micro-organisms will favour the development of varying types of plasmagenes.

The possible nature of the plasmagene is indicated by what we know of virus diseases. The virus is a protein molecule which, introduced into one organism, disappears; into another, multiplies to give a neutral equilibrium; into yet another, multiplies without limit and has deleterious or fatal results. If we look upon the virus as a protein taken out of one organism (usually by a parasite) and injected into another to which it is not properly adapted, we see that it has the properties often shown by a plasmagene in crosses between species. This is to suppose that a virus is not a primitive enemy of nature but just a protein out of place. Such is the position of the cytoplasm of one species which harbours the nucleus of another, and the varying behaviour of the one indicates the varying possibilities of the other.