Popular Science Monthly/Volume 63/July 1903/Evolution, Cytology and Mendel's Laws


By O. F. COOK,


THE debt of science to theory is a truism. Bad theories are only less valuable than good ones, and for some purposes they are even better. We do not arrive where we expected to go, but reach; an undiscovered country which a more direct route would have left unexplored. The recent history of biology furnishes two excellent examples of the fertility of false theories in the development of the related sciences, embryology and cytology The theory of organic recapitulation, to the effect that the phylogeny or evolutionary history of natural groups must be repeated in the ontogeny or development of each individual organism promised the student of embryology an easy wealth of scientific discovery, and within a few years hundreds of razors were paring thin the mysteries of evolution. Libraries of new facts were discovered and published, but as our knowledge of life histories increased there was a corresponding decline in the probability that any particular stage in the growth of the individual is necessarily more ancestral than any other. That no general doctrine of recapitulation could be maintained was perceived by Sir John Lubbock as early as 1873,[1] but vertebrate embryologists did not permit their zeal to be dampened by even the most obvious facts of entomology. Indeed, one of our prominent investigators, finding that recapitulation is elusive by microscopical methods, now proposes to test it by breeding experiments, the results of which may be available in a future geologic epoch.[2]

The organism having been followed back to its unicellular stage without discovering any process or mechanism by which its adult form was predetermined, believers in such a device must needs seek it inside the cell, and thus was opened another highly fertile field of investigation. Instead of mere homogeneous jelly, surprisingly complicated intracellular structures and processes have been discovered and described, and to identify some of these as the long-sought 'hereditary mechanism' is now the dream of the cytologist.

To judge from his recent article on 'Mendel's Principles of Heredity and the Maturation of the Germ-Cells'[3] Professor Wilson, at least, is still very strongly of the belief that the investigator of reproductive cells holds the keys of evolution, and he even finds it remarkable that a general cytological explanation of these 'principles of inheritance' was not suggested before. According to Professor Wilson the facts discovered by Mendel, that in some hybrids, characters of the parents are not permanently combined, are explainable by the 'normal phenomena of maturation,' that is, if we admit that 'individual chromosomes stand in definite relation to transmissible characters,' and that the 'reducing division' by which the reproductive cells are formed 'leads to the separation of paternal and maternal elements and their ultimate isolation as separate germ-cells.' This would be important if true, but the Mendelian facts are unable to accept this proffered support of cytological theory, because they have already demonstrated its falsity.

The commonly accepted view of organic descent may be illustrated by a simple diagram which indicates that a single individual may

PSM V63 D224 Organic descent inheritance from the four grandparents.png

inherit characters from all four of its grandparents. Professor Wilson's explanation of Mendel's law would deny this possibility, and would limit the descent of all individuals to two grandparents, so that, the form of our family tree would be completely altered.

PSM V63 D224 Organic descent inheritance from the two grandparents.png

It would be impossible to have any such compound as ABCD, but we should get instead one of the four character-combinations AC, AD, BC, BD. The inheritance of a single character from one grandparent would certify the inheritance of all, and thus establish an alibi for the other ancestor of the same side of the house. What a resource for genealogists to be able to prove that a man was no relative of his grandfather, or even that he had no consanguinity with his own brother! Alas that Mendel and other 'experimental evolutionists' have proved that inheritance is by characters and not by chromosomes, if these behave as Professor Wilson indicates. Only the so-called monohybrids, those differing by a single character, would tolerate such an interpretation, and the fallacy of it is obvious as soon as we remember that hybrids may be assembled with reference to two, three or more characters derived from different ancestors. Moreover, Professor Spillman has recently drawn from his experiments with wheat concrete and detailed examples of the fact that definite proportions of such combinations are permanent, since two dominant characters do not antagonize each other.[4] Unless it be in the case where the varieties crossed differ in but a single character we know of a certainty that the germ-cells are not of pure descent with respect to parentage.[5] The most that can be claimed is that they are organized reciprocally with reference to the divergent parental characters, since it seems that the different features may be distributed and recombined quite without reference to the manner in which they were grouped in the parents. In his hybrid wheats Professor Spillman finds all the combinations possible under the mathematical theory of chance. How this could be managed by the chromosomes our cytological friends may be able to conjecture, though from the outside it looks like a rather difficult question.

Hybridization is possible only between groups of common origin, and the characters which show the Mendelian effect are those on which the greatest divergence has taken place. That such characters may be changed about or substituted, and are able to enter freely into all varieties of combination, not only does not prove that the chromosomes are mechanisms of heredity, but it greatly decreases the probability of mechanical theories of evolution, since it shows the facility with which characters may be accumulated in normal interbreeding, before the Mendelian degree of divergence has been reached.

If two plants different in other respects are found to differ also in chromosomes this does not prove that the chromosomes cause the other differences, even though the differences of the chromosomes interfere with the conjugation of the reproductive cells and thus prevent the hybridization of the plants. Species or varieties seldom, if ever, differ by single characters or at one stage merely, and there is no known reason why related species should not diverge in their single-celled condition as well as at any later period. It is rapidly becoming apparent that the internal organs and functions of cells are as diverse as those of embryos and adult organisms, and as much in need of a general evolutionary explanation,[6]

The notion that heredity, variation or other phases of evolution are the functions of special organs or mechanisms of cells, has no ascertained basis of fact, and is but an inference from the traditional evolutionary errors that species are normally constant or stable, and that developmental changes are the results of external influences. To move a stationary organism some sort of 'hereditary mechanism' would be needed to bring about the inheritance of characters 'acquired' from the environment, but if we consider that the individuals of a species are normally diverse, and that the species as a whole is normally in motion, a 'hereditary mechanism' becomes quite superfluous, or may be identified with the organism itself, whether in a unicellular or a polycellular stage.

Heredity is the term under which we allude to the fact that organisms exist in series of similar individuals; we have as yet no warrant for holding that it is special 'force' or agency. Crystals of the same substance are thought of as repeatedly taking the same form because of certain properties of matter, not because of a special crystallizing mechanism. The analogy of crystals is, of course, quite inadequate for biological purposes, but we need not reject it entirely, since for all purposes of expression heredity is a general property of living things, and with these there is even less reason than with crystals to seek a cause in the function of a special organ.

Inorganic elements and compounds are homogeneous and similar in all masses or parts; but diversity is the rule among organisms, no two of which are exact duplicates. The idea of a heredity which maintains identity of structure or form represents no fact in nature. The necessity of continued readjustment is general in life, and is not confined, even in complex organisms, to preliminary stages or to reproductive cells. The individual is not constant nor permanent, but has its own cycle of growth, reproduction and decline, accompanied by continuous changes in all parts of the bodily form and structure. Organisms are not made up merely by the few characters enumerated by the systematist; an infinite number of differing relations of parts might be formulated. Evolutionary divergence is not confined to external adult characters, but may appear in any structure, function or instinct, and at any time in the life history. Species very different as adults may have closely similar young, or larvæ may be much more diverse than the mature insects. Only the inadequacy of our notions of the vital structure and activities has led us to expect that reproductive cells will be found to contain special 'hereditary mechanisms' for the predetermination of the characteristics of adults. The largest and most complex individuals are still groups of cells, and no adequate reason has been shown for believing that particular cells or links of the organic sequence are more hereditary or more determinant than the others. Characters are to be thought of as lines of biological motion, not as structures or entities of reproductive cells. The predetermination of the infinity of structural and morphological characters and positional relations of the millions of cells of the adult by a working model resulting from the conjugation of sexual elements may be dismissed as a crudely anthropomorphic notion of biological processes, as unsupported by facts as it is illogical in conception. Cells have their functions and organs, but evolution is not confined to these; it is also a supercellular or organic process. Cytology is a very interesting branch of descriptive biology, but it enjoys no special evolutionary facilities.

Polycellular organisms grow by the division of cells; but instead of proving that all cells divide in the same way cytologists have found that the same result may be accomplished by a great variety of protoplasmic organs and processes. Unicellular organisms are known to be extremely diverse cytologically, and the cells of compound organisms are, if possible, more so. We know also that the diversity of organisms is not due so much to differences of the individual cells as to differences of number and arrangement in the cell-colonies of which they are constituted.

Heredity is the unknown means by which successive generations of organisms are able to construct themselves in similar, though not identical, forms; it is, in short, an organic memory, and is responsible, not alone for the repetition of the structural type, but also for vast numbers of involuntary functional coordinations and instinctive acts, whether of unicellular or of compound organisms, or of whole colonies of organisms. A colony of social termites is as truly an evolutionary unit as a tree with its many branches, and the cooperative instincts which pervade the individual insects are as truly a hereditary phenomenon as the peculiar arrangement of branches which we term a 'character' of the tree.

To compare heredity with memory explains nothing, of course, since we know as little of the physical basis of the one as of the other, but if the analogy be admitted it will prevent the too confident insistence upon the theory that heredity depends entirely upon positional or other mechanical relations of molecules, or is in some way embodied in particular granules or chromosomes.

But even the most fantastic theories often have some basis of suggestion in fact, and although we can not accept Professor Wilson's cytological explanation of Mendel's laws, nor even share his hope that cytology will elucidate evolution, it is by no means impossible that the normal individual diversity of organisms has a cytological as well as an evolutionary significance. That normal development or growth by cell division is advantaged by cross-fertilization may mean that the cells divide more readily and normally when they contain protoplasmic 'elements' of a proper degree of diversity than when they have only one kind of protoplasm, as would happen in narrow inbreeding, and also when cross-breeding is too wide for the intimate cooperation required for true fertilization. The Mendelian effect would then be explainable on the suggestion of a partial cooperation which has to be abandoned in the formation of new individuals, because, while the organism can follow either of two diverging parental roads with respect to any character, it is, as it were, a stranger to the path that an average would require.

The conjugation of cells may be viewed as a process quite distinct from reproduction, though it is a necessary preliminary to the long series of cell divisions required to build up the complex bodies of the higher animals and plants. As we descend in the organic scale the conjugating cells become more and more similar to each other and to the so-called vegetative or somatic cells of which the body of the organism is composed. Among simple organisms all the cells are alike, including those formed immediately before and after conjugation, and it is not strange that with the diversification of the cells which constitute the various tissues of the plant or animal body the germ cells should become specialized and unlike any of the others. The existence of special reproductive cells among the higher animals and plants is therefore to be looked upon as corresponding to the general complexity of the organism, rather than as an indication of a special mechanism of heredity resident in the germ cells. As founders of new cell-colonies or compound individuals they develop, it appears, on one or the other of divergent parental lines instead of striking out on an untraveled road between.

Notwithstanding their great significance Mendel's laws are negative rather than positive in their bearing upon descent, since we do not learn from them the nature of that process, but one of its limits. Moreover, the probability that these laws are of general application is greatly lessened by the fact that they are demonstrable only in connection with narrowly inbred and much divergent varieties of plants and animals, to which condition of the experiment the phenomena discovered may prove to be due, rather than to any general fact or mechanism of heredity. It seems certain, however, that neither theory nor experiment will make permanent progress in this direction as long as we continue to confuse under the word hybrid several extremely diverse evolutionary conditions, and fail to realize that generalizations based on any one kind or type of hybrids are quite premature and irrational.

Sterile Hybrids.—The original notion of a hybrid, or at least the most popular meaning of the term, is that of a cross between organic types so widely diverse that the progeny are in some way abnormal or defective, especially with reference to reproduction. Among animals sterile hybrids can not be propagated, but in plants they can be grown from cuttings or buds, and are thus preserved as horticultural * varieties. '

Aberrant Hybrids.—The second and succeeding generations of hybrids not completely sterile often show striking deviations from both parental types. As these new characters are analogous to the abrupt variations of close-bred plants described by Darwin as 'sports' and more recently renamed 'mutations' by De Vries, it has been suggested that they may be due to the same causes, that is, they may not be in reality the result of crossing, but rather of an inadequate conjugation or defective fertilization which allows a lapse from the normal form. Both mutations and mutative hybrids are comparatively infertile, so that their suddenly attained new characters should not be looked upon as true contributions to evolutionary progress.

Reciprocal or Mendelian Hybrids.—Mendel and his successors have proved that there is still a third type of less abnormal hybrids, in which there is no permanent combination or averaging of divergent parental characters, although it is not known that vigor and fertility are notably diminished. Mendel's so-called laws are generalized statements of the results of his experiments upon the crossing of different garden varieties of the pea; he himself found that the same was not true among hybrids of Hieracium.[7] A part of the scientific community which a few years ago was properly characterized as more Darwinian than Darwin might now be described as more Mendelian than Mendel, and expects to find in 'Mendel's laws' an explanation of heredity, to say nothing of other things. Crosses between some twenty-six close-bred varieties of plants and animals have been found to 'Mendelize,' as the new expression is, and it may be expected that others will do the same wherever the conditions of the experiment can be met, though no amount of similar facts would justify the general conclusions which some recent writers have so promptly drawn. 'Mendel's laws' have already had many different statements, but the most that can be said with certainty is that after close-bred varieties of a plant or animal have sufficiently separated, their divergent characters do not again blend or reduce to an average, but draw apart into definite proportions of each succeeding generation of offspring. Obviously, this is not a method or law of inheritance, but of non-inheritance or fractional inheritance. The sterile and aberrant hybrids are evidence that too wide crossing is not advantageous and makes no contribution to evolutionary progress. Mendel's experiments afford further evidence of the same fact, in that the organisms themselves are found to have means of dissolving such alliances and thus of holding to the paths on which their varietal divergencies have gone forward. The theory that hybridization assists evolution by encouraging variability is shown to have a distinct limit, since little evolutionary progress would come from mere combination of the stable or divergent characters which are a prerequisite of the Mendelian experiments.

Synthetic or Blended Hybrids.—If the normal flexibility of the organism has not been diminished by narrow segregation or inbreeding, the Mendelian repugnance of divergent characters does not appear; Mendel's law of reciprocal characters gives place to Spillman's law of blended or graded characters.[8] Thus there is no record of a normal straight-haired white child as the offspring of two mulattoes. Inbreeding to an extent far beyond anything usual in nature is the rule among domesticated plants and animals, but if the varieties are not too divergent they cross freely and with obvious advantage, as shown by increase in vigor, though such 'new characters' soon disappear under renewed inbreeding. Characters which would become dominant in the Mendelian hybrids are in the less divergent stages termed prepotent, that is, they are impressed with increased intensity upon increasing numbers of each successive generation. On the other

hand, characters acquired through inbreeding or other debilitating causes may disappear or become recessive as soon as crossing permits a return to a more normal and vigorous ancestral type of organization, as in the historic pigeon experiments of Darwin. The popularization of Mendel's laws should make it more easy to perceive that the normal effect of cross-breeding is a progressive synthetic evolution and not a stationary average, though we are having some fine examples of the lengths to which the specialist will sometimes go to escape facts too simple and obvious for his appreciation.

Individual 'Hybrids.'—Perhaps the loosest use of the word hybrid is for the offspring of crosses between so-called 'horticultural varieties' of domesticated plants propagated by cuttings or grafts. Everybody knows, though some forget, that the Baldwin apple, the Bartlett pear, the Niagara grape, and a great multitude of analogous sorts, are descending from single seedling trees or vines, and are thus for evolutionary purposes single individuals. The distinction between such individuals and those of wild species in nature is largely psychological; we have learned to regard differences between individual apple trees, but have not attained such close acquaintance with oaks and elms. If crosses between the normally diverse individuals of a species are to be termed hybrids then the word covers all sexually differentiated organisms and is utterly useless as a means of drawing biological distinctions. Mendel deliberately disregarded the question as to which of his pea hybrids were between different species, and which between varieties merely, and for the purposes of his inquiry this was a matter of little importance. But for his followers to draw general conclusions, while ignoring all distinction between the evolutionary conditions of the organisms which they study, is a reversion to the same general woolliness of evolutionary thinking to which Mendel constituted so brilliant an exception.

The millions of species with which nature has been experimenting for millions of years seem to make it very plain that individual diversity with free interbreeding is the optimum condition for evolutionary progress, since this is what we find everywhere among natural species. It is true that the diversity masks the slow and gradual motion of the species from perception by our momentary observations, and also that the interbreeding hinders the segregation of species; but we may take the results as evidence that evolutionary progress is not impeded by wide individual variation, nor by opportunities for the progressive accumulation of new characters. Nor need we turn our backs on this interpretation of the history of organic nature because Mendel and others have given new demonstrations of the old fact that there are degrees of evolutionary divergence in which the combination of parental characters is no longer possible.

Mendel's laws are of much practical importance because they make plain to breeders of economic plants and animals that they can not do what has been attempted so frequently, make improved breeds by combining the divergent characters of close-bred varieties. The Mendelian facts are of general evolutionary interest, not because they explain descent, but because they are incompatible with the commonly accepted static theories of development which hold that evolutionary progress is due to external or environmental influences and overlook the independent self-caused motion of species. More detailed presentation of the latter view can not be undertaken here;[9] it must suffice for the present to have pointed out that cytology has not proved the universality of Mendel's laws as 'principles of inheritance,' nor do the laws prove that the chromosomes are the long-sought 'hereditary mechanisms.'

Heredity should be thought of as a general property of organisms, and not as the function of a special organ of the cell or of the embryo. As a phenomenon it should be associated with crystallization, on the one side, and with memory, on the other. There may be simpler properties of matter which render crystallization, heredity and memory possible, but such properties are not yet recognized in physics and chemistry, so that the terms and theories of these sciences are of little use in the discussion of evolution.

Viewed as the basis of an independent generalization the Mendelian experiments ran counter to multitudes of the most obvious and best established data of biology, and it may have been on this account that they were so long disregarded. The apparent conflict is here explained as due to erroneous theories of evolution; the recognition of spontaneous organic motion enables Mendel's facts to find a place in the evolutionary series, and renders the general inferences of de Vries, Bateson and Wilson unnecessary. Nor need the present view be thought to depreciate the importance of Mendel's laws, since such discoveries are of much greater practical value after they have found their true place among related facts than while as novelties they are permitted to obscure all the adjoining fields of investigation.

  1. 'Origin and Metamorphoses of Insects.'
  2. Science, N. S., 16: 506, September 26, 1902.
  3. Science, N. S., 16:991, December 19, 1902.
  4. 'Mendel's Law,' Popular Science Monthly, 62: 269.
  5. The theory of Bateson that the germ-cells are pure with respect to characters seems to have been misunderstood both by Professor Wilson and by Mr. Cannon in his 'Cytological Basis for the Mendelian Laws' (Bulletin Torrey Botanical Club, 29:657, 1902).
  6. Chromosome differences utterly disproportional to the differences of the adult organisms have recently been described by Monkhouse in hybrid fish eggs.
  7. The question has been raised as to whether Mendel's discoveries should be called 'laws.' The present view would deny to them universal application as 'laws' or 'principles' of heredity, though it admits as probable their general truth for a certain evolutionary condition or stage.

    Laws of gases are not called laws of matter, and do not apply until matter reaches the gaseous state. Similarly, there can be no objection to such expressions as 'Mendel's laws of the disjunction of characters in hybrids,' or 'Mendel's laws of reciprocal hybrids.'

  8. * ' * * * hybrids show every possible gradation between the characters of the two parents.'
  9. See 'A Kinetic Theory of Evolution,' Science, N. S., 13:969, June 21, 1901.