Popular Science Monthly/Volume 85/September 1914/The Cellular Basis of Heredity and Development IV
| THE CELLULAR BASIS OF HEREDITY AND DEVELOPMENT. II |
By Professor EDWIN GRANT CONKLIN
PRINCETON UNIVERSITY
C. The Mechanism of Heredity
The mechanism of heredity, as contrasted with the mechanism of development, consists in the formation of particular kinds of germ cells and in the union of certain of these cells in fertilization. We have briefly traced the origin, maturation and union of male and female sex cells in a number of animals, and in these phenomena we have the mechanism of the hereditary continuity between successive generations. But in addition to these specific facts there are certain general considerations which need to be emphasized.
I. The Specificity of Germ Cells
The conclusion is inevitable that the germ cells of different species and even those of different individuals are not all alike. Every individual difference between organisms must be due to one or more differentiating causes or factors. Specific results come only from specific causes. These causes may be found in the organization of the germ cells or in environmental stimuli, i. e., they may be intrinsic or extrinsic, but as a matter of fact experience has shown that they are generally intrinsic in the germ. In the same environment one egg becomes a chicken and another a duck; one becomes a frog, and another a fish, and another a snail; one becomes a black guinea-pig and another a white one; one becomes a male and another a female; one gives rise to a tall man and another to a short man, etc. Since these differences may occur in the same environment they must be due to differences in the germ cells concerned.
On the other hand, different environmental conditions may he associated with similar developmental results. Loeb and others have found that artificial parthenogenesis may be induced by a great variety of environmental stimuli, viz., by salt solutions, by acids and alkalies, by fatty acids and fat solvents, by alkaloids and cyanides, by blood serum and sperm extract, by heat and cold, by agitation and electric current. There is certainly nothing specific in these different stimuli. Similarly, Stockard has discovered that cyclopia, or one-eyed monsters, may be produced by magnesium salts, alcohol, chloretone, chloroform, and ether. In all such cases it is evident that the specific results of such treatment are due to a specific organization of the germ rather than to specific stimuli.
Why does one egg give rise to a chicken and another to a duck, or a fish, or a frog? Why does one egg give rise to a black guinea-pig and another to a white one, though both may be produced by the same parents? Why does one child differ from another in the same family? Why does one cell give rise to a gland and another to a nerve, one to an egg and another to a sperm? If these differences are not due to environmental causes, and the evidence shows that they are not, they must be due to differences in the structures and functions of the cells concerned.
Many differences in the material substances of cells are visible, and many more are invisible though still demonstrable. These differences may not be detectable by chemical or physical tests, and yet they may be demonstrated physiologically and developmentally. The most delicate of all tests are physiological, as is shown by the Widal test in typhoid fever, the Wassermann reaction in syphilis, the reactions of immunized animals to different toxines, etc. Lillie has recently shown that egg cells give off a substance which he calls fertilizin, which can be detected only by the way in which spermatozoa react to it. No chemical or physical test can distinguish between the different eggs or spermatozoa produced by the same individual, but the reactions of these cells in development prove that they are different. Undoubtedly chemical and physical differences are here present, but no chemical methods at present available are sufficiently delicate to detect them. The developmental test proves that there must be as many kinds of germs as there are different kinds of individuals which come from germs. It is one of the marvelous facts of biology that every individual which has been produced sexually is unique, the first and last of its identical kind, and although some of these individual differences are due to varying environment, others are evidently due to germinal differences, so that we must conclude that every fertilized egg cell differs in some respects from every other one.
But are there molecules and atoms enough in a tiny germ cell, such as a spermatozoon, to allow for all these differences? Miescher has shown that a molecule of albumin with 40 carbon atoms may have as many as one billion stereoisomers, and in protoplasm there are many kinds of albumin and proteins, some with probably more than 700 carbon atoms. In such a complex substance as protoplasm the possible variations in molecular constitution must be well-nigh infinite, and it can not be objected on this ground that it is chemically and physically impossible to have as many varieties of germ cells as there are different kinds of individuals in the world.
Even with regard to morphological elements which may be seen with the microscope it can be shown that an enormous number of permutations is possible. It seems probable, as Boveri has shown, that different chromosomes of the fertilized egg differ in hereditary potencies, and where the number of chromosomes is fairly large the number of possible combinations of these chromosomes in the germ cells becomes very great. In woman, where there are probably 48 chromosomes, and, after synapsis, 24 pairs of maternal and paternal ones, the possible number of permutations in the distribution of these chromosomes to the different egg cells would be , or 16,777,036, and the possible number of different types of fertilized eggs or oosperms which could be produced by a single pair of parents would be (16,777,036), or approximately three hundred thousand billions. But probably other things than chromosomes differ in different germ cells, and it is by no means certain that individual chromosomes are always composed of the same chromomeres, or units of the next smaller order, and in view of these possibilities it may well be that every human germ cell differs morphologically and physiologically from every other one, in short that every oosperm and every individual which develops from it is absolutely unique.
Indeed, the production of unique individuals seems to be the chief purpose and result of sexual reproduction. In a sexual reproduction the individual variations which occur are chiefly, if not entirely, due to environment, but in sexual reproduction they are also due to new combinations of hereditary elements. The particular germinal organization transmitted from one generation to the next depends upon, (a) The ancestral organization, (b) The particular character of the cell divisions by which the germ cells are formed, (c) The particular kinds of egg and sperm cells which combine in fertilization. The ancestral organization determines all the general characteristics of race, species, genus, order, phylum. It determines the possibilities and limitations of individual variations. Given a certain ancestral organization, the individual peculiarities of the germ cells are determined by the particular character of cell division by which the germ cells are formed, and the peculiarities of the individuals or persons which develop from these cells are determined in large part by the particular kinds of germ cells which unite in fertilization.
The behavior of chromosomes in maturation and fertilization is like the shuffle and deal of cards in a game, and apparently with the same object, namely, never to deal the same hand twice. To make this comparison more complete suppose that kings be discarded from the pack, leaving 48 cards of two colors, red and black, which we will compare to the 48 maternal and paternal chromosomes in the human oocyte; suppose that in the shuffling of these cards corresponding cards of the red and the black suits are temporarily stuck together so that the ace of diamonds is united with the ace of clubs, the queen of hearts with the queen of spades, etc., thus forming 24 red-black pairs of the same denominations. If these cards are then dealt into two hands, one card of each pair going to one hand and the other to the other hand, we will have two cards of each denomination in each hand, but if the cards are dealt indiscriminately some of them will be red and some black. This description parallels what takes place in the maturation of the human ovum, except that there is no evidence that there are more than two suits of chromosomes, one maternal and the other paternal.
To carry out this comparison in the case of the maturation of the human sperm where there are only 47 chromosomes it is necessary to take another pack and discard an additional card, say the queen of clubs; then in the union of corresponding red and black cards into pairs the queen of hearts unites with the queen of spades, but the queen of diamonds remains alone, and when the cards are dealt into two hands as before one hand will contain 24 cards and the other 23.
If now we complete this comparison by extending it to what takes place in fertilization we must take one hand from each of these deals and put them together into one pack; though this pack would contain cards of every denomination there would be varying numbers of red and black cards and a mixture of cards from two distinct packs. In no game of cards do corresponding cards from different packs have slightly different values nor are half of the cards taken from one pack and half from another (at every game), but this is just what happens in the shuffle and deal of the chromosomes. Because of the mixture of chromosomes from distinct individuals in every generation, each of which has its own peculiar value, the game of heredity becomes vastly more complex than any game of cards.
This illustration may serve to make plain the fact that the purpose of maturation and fertilization is, in part, this shuffle and deal of the chromosomes, and its result is that every oosperm and every individual which develops from it is different from every other one.
This conception of the specificity of every germ cell, as well as of every developed individual, sets the whole problem of heredity and development in a clear light. The visible peculiarities of an adult become invisible as development is traced back to the germ, but they do not wholly cease to exist. Similarly, the multidinous complexities of an adult fade out of view as development is traced to its earliest stages, but it is probable that they are not wholly lost. In short, the specificity of the germ applies not merely to those things in which it differs from other germs, but also to characters in which it resembles others—in short, to hereditary resemblances no less than to hereditary differences.
The mistake of preformation was in supposing that germinal parts were of the same kind as adult parts; the mistake of epigenesis was in maintaining the lack of specific parts in the germ. The development of every animal and plant consists in the transformation of the specific characters of the germ into those of the adult, but not in the formation of structures or characters de novo. From beginning to end development in a series of morphological and physiological changes, but not of new formations or creations. It is only the incompleteness of our knowledge of development which allows us to say that the eye or ear or brain begin to form in this or that stage. They become visible at certain stages, but their real beginnings are indefinitely remote.
II. Correlations Between Germinal and Somatic Organization
All the world knows that the organization of the germ is not the same as that of the developed animal which comes from it, and yet the specificity of the germ indicates that there must be some correlation between the germinal and the developed organization—in short, there is not identity of organization, but correlation of organization between the germ and the adult. What correlations are known to exist between the oosperm and the developed animal?
1. Nuclear Correlations
Many biologists maintain that the nucleus and more particularly the chromosomes are the exclusive seat of the "inheritance material" and that all the "determiners" of adult characters are located in them.
There are certain general and a priori reasons for assuming that the chromosomes are important factors in heredity and differentiation; (1) they come in approximately equal numbers from the father and the mother, (2) one half of each of the maternal and paternal chromosomes is distributed to each cell of the developing organism, (3) in the formation of the egg and sperm cells the normal number of chromosomes is reduced by one half, and (4) in fertilization the normal number is restored by the union of the chromosomes of the egg and sperm. It is a remarkable fact that the determiners or factors of certain inherited characters come in equal numbers from both parents and that in spite of their ultimate association in an individual they may be separated or "segregated" in the formation of that individual's germ cells. Such inheritance is known as Mendelian and will be treated at length in the next lecture, but it may be said here that the association, distribution and segregation of Mendelian factors and of maternal and paternal chromosomes is exactly parallel. This is strong evidence that these factors are associated with the chromosomes.
There are also certain special reasons for considering that the chromosomes are important factors in heredity and development. (5) Boveri has studied the abnormal distribution of chromosomes to different cleavage cells in doubly fertilized sea-urchin eggs, and has found evidence that the hereditary value of different chromosomes is different.
(6) McClung, Stevens and Wilson have discovered that the determination of sex is associated with the presence or absence of a particular chromosome, the X-chromosome, in the spermatozoon which fertilizes the egg. If an egg is fertilized by a sperm which lacks the X-chromosome a male is produced; if fertilized by the other type a female results. (7) Finally, Morgan has found that there is a linkage of certain somatic characters with sex in the fruit fly, Drosophila, which can be readily explained by assuming that the determiners for these characters are in some way associated with the sex chromosome.
We have in these facts a remarkable correlation between the distribution of the chromosomes and the occurrence of certain characters of the adult animal. The association of maternal and paternal chromosomes in fertilization and their segregation in the maturation of the germ cells is parallel to the association of Mendelian characters in the zygote and their segregation in the gametes; if the distribution of chromosomes in cleavage is abnormal the larva shows abnormal characters (Boveri); sex determination is associated with the distribution of a particular chromosome to one half of the spermatozoa, and the fertilization of the egg by one type or the other of spermatozoa (Wilson); the linkage of certain characters with sex finds a ready explanation by assuming that these characters are associated with the sex chromosome (Morgan).
2. Cytoplasmic Correspondences
On the other hand, the most direct and the earliest recognized correlations between the oosperm and the developed animal are found in the polarity and symmetry of the fertilized egg and of the animal to which it gives rise.
(a) Polarity
In all eggs there is a polar differentiation, one pole, at which the maturation divisions take place, being known as the animal pole, and the opposite one being known as the vegetative pole. The substance of the egg in the vicinity of the animal pole usually gives rise to the ectoderm, or outer cell layer of the embryo; the portion of the egg surrounding the vegetative pole usually becomies the endoderm or inner cell layer. The axis which connects these poles, the chief axis of the egg, becomes the gastrular axis of the embryo and in every great group of animals bears a constant relationship to the chief axis of the adult animal. The polarity of the developed animal is thus directly connected with the polarity of the egg from which it came (Figs. 23, 26, 29, 30, 40, 41).
(b) Symmetry
In many cases the symmetry of the developed animal is foreshadowed in the symmetry of the egg. The eggs of cephalopods (Fig. 40) and of insects (Fig. 41) are bilaterally symmetrical, while they are still in the ovary; in other cases, such as ascidians, amphioxus and the frog, bilateral symmetry appears immediately after fertilization (Fig. 29, 1, 2), though in some of these cases there is reason to believe that the eggs are bilateral even before fertilization; in still other cases bilaterality does not become visible until later in development and we do not now know whether it is present in earlier stages or not; but wherever it can be recognized in the earlier stages it is certain that the bilateral
| Fig. 40. | Fig. 41. |
| Fig. 40. Outlines of the Unfertilized Egg of a Squid, Loligo, showing the polarity and symmetry of the egg with reference to the axes of the developed animal; d, dorsal; v, ventral; l, left; r, right; a, anterior; p, posterior. (After Watasé.) | |
| Fig. 41. Median Section through Egg of a Fly, Musca, just after fertilization, showing the relations of the polarity and symmetry of the egg to the axes of the developed animal; the long axis of the egg corresponds to the antero-posterior axis of the animal; d, dorsal; v, ventral; m, micropyle through which sperm enters the egg; g, glutinous cap over the micropyle; r, polar bodies; p, egg and sperm nuclei; do, yolk; k, peripheral layer of protoplasm; dh, vitteline membrane of egg; ch, chorion. (After Korschelt and Heider.) | |
symmetry of the egg bcomes the bilateral symmetry of the developed animal.
(c) Inverse Symmetry
In most animals bilateral symmetry is not perfect, certain organs being found on one side of the mid line and not on the other, or being larger or differently located on one side as compared with the other; among all such animals variations occasionally occur which show a complete reversal of these asymmetrical organs, i. e., in man the heart and arch of the aorta may occur on the right side instead of the left, the pyloris and chief portion of the liver on the left instead of the right, etc. Among certain snails this inversion of symmetry may occur regularly in certain species and not in others, the inverse form being known as sinistral and the ordinary form as dextral (Fig. 44). In these sinistral snails, and probably in all animals showing inverse symmetry, the embryo is inversely symmetrical and every cleavage of the egg from the first to the last is the inverse of that which occurs in dextral snails (Figs. 42, 43). There is good reason to believe that in such cases the unsegmented egg is also inversely symmetrical as compared with the more usual type (Kg. 42). In all of these cases there is a direct correspondence between the polarity and symmetry of the oosperm and the polarity and symmetry of the developed animal (Fig. 40-44).
Fig 42. Inverse Symmetry in the Unsegmented Egg and in the First and Second Cleavages. Figs. 42, 43, 44. The cause of inverse symmetry in snails. In each case the right-hand column represents dextral forms, the left-hand column sinistral ones.
(d) Localization Pattern
In many animals the ectoderm, endoderm and mesoderm may be traced back to areas of peculiar protoplasm in the oosperm, but in addition to this one can recognize in the ascidian egg areas of peculiar protoplasm which will give rise to mesenchyme, muscles, nervous system and notochord, and these substances are present in the oosperm in the approximate positions and proportions which they will have in the embryo and larva (Figs. 28-31).
Indeed, there are types of localization of these cytoplasmic materials in the egg which are characteristic of certain phyla; thus there are the ctenophore, the flat-worm, the echinoderm, the annelid, mollusk and the chordate types of cytoplasmic localization (Fig. 45). The polarity, symmetry and pattern of a jelly fish, star-fish, worm, mollusk, insect or vertebrate are forshadowed by the characteristic polarity, symmetry and
Fig. 43. Inverse Symmetry of the 3d, 4th. 5th and 6th cleavages. The cells 1a-1d, 2a-2d and 3a-3d give rise to all the ectoderm; 4d or M gives rise to mesoderm; A, B, C, D to endoderm.
pattern of the cytoplasm of the egg either before or immediately after fertilization. In all of these phyla eggs may develop without fertilization, either by natural or by artificial parthenogenesis, and in such cases the characteristic polarity, symmetry and pattern of the adult are found in the cytoplasm of the egg just as if the latter had been fertilized. The conclusion seems to be justified that these earliest and most fundamental differentiations which distinguished the eggs of various phyla are not dependent upon the sperm.
All of these correspondences between the polarity, symmetry and pattern of the egg and of the developed animal are found in the cytoplasm. It is possible that the polarity may be carried over from generation to generation through the egg cell, but the symmetry and localization
Fig. 44. Inverse Symmetry in Late Embryos and Adult Stages. In 1, cross hatched area is blastopore; cells shaded by lines mesoderm, other cells endoderm; the spiral twist of the snail begins in opposite directions in the two embryos. In 2 the adult organization is shown with all organs inversely symmetrical; os, olfactory organ; a, anus; L, lung; V, ventricle; K, kidney. In 3 sinistral and dextral shells of adult snails are shown.
pattern develop in the ovum before or just after maturation. In this differentiation and localization of the egg cytoplasm it is probable that certain influences have come from the nucleus of the egg, and perhaps from the egg chromosomes. There is no doubt that most of the differentiations of the egg cytoplasm have arisen during the ovarium history of the egg, and as a result of the interaction of nucleus and cytoplasm;
Fig. 45. Types of Egg Organization in Different Phyla; cross-hatched area, mesoderm or mesenchyme (mes); horizontal lines, endoderm (end); clear area, ectoderm (ect). In the first four figures the pattern of localization Is that which is found at the close of the first cleavage; in the annelid egg the localization of later stages is projected upon the egg.
but the fact remains that at the time of fertilization the hereditary potencies of the two germ, cells are not equal, all the early stages of development, including the polarity, symmetry, type of cleavage, and the pattern, or relative positions and proportions of future organs, being foreshadowed in the cytoplasm of the egg cell, while only the differentiations of later development are influenced by the sperm. In short, the egg cytoplasm fixes the general type of development and the sperm and egg nuclei supply only the details.
We are vertebrates because our mothers were vertebrates and duced eggs of the vertebrate pattern; but the color of our skin and hair and eyes, our sex, stature, and mental peculiarities were determined by the sperm as well as by the egg from which we came. There is evidence that the chromosomes of the egg and sperm are the seat of the differential factors or determiners for Mendelian characters, while the general polarity, symmetry and pattern of the embryo are determined by the cytoplasm of the egg.
It will be observed that the correlation between chromosomes and adult characters is different in kind from that between the cytoplasm of the egg and the adult characters; in the latter case polarity, symmetry and pattern are of the same kind in the egg and in the adult, and the correspondence is comparatively close; in the latter there is no correspondence in kind between the chromosomal peculiarities and the peculiarities of the adult. This fact might suggest that the chromosomal organization may be more fundamental than that of the cytoplasm. There are reasons for believing that many substances of the cell are formed by the interaction of nucleus and cytoplasm, and most probably the chromosomes are an important factor in this process. But in no case is the cytoplasm a negligible factor—in no case does it serve merely as food for the chromosomes. The entire cell, nucleus and cytoplasm, is concerned in heredity and differentiation.
D. The Mechanism of Development
Development consists in the transformation of the oosperm into the adult. What is the mechanism by which this transformation is effected? There is progressive differentiation of the germ into the developed organism, but by what process is this differentiation accomplished?
Many different processes are concerned in embryonic differentiation. From the standpoint of the cell the most important of these are (1) the formation of different kinds of substances in cells, (2) the localization and isolation of these substances, (3) the transformation of these substances into the various structures which are characteristic of the different kinds of tissue cells. We shall here describe only the first and second of these processes which are of more general interest than the last.
1. The Formation of Different Substances in Cells
Differentiation consists primarily in the formation of different kinds of protoplasm from the proptoplasm of the germ cells. It is plain that different kinds of protoplasm are present in the two germ cells before they unite in fertilization, but in the course of development the number of these substances and the degree of difference between them greatly increase.
Actual observation shows that by the interaction with one another of substances or parts originally present and by their reactions to external stimuli new substances and parts appear which had no previous existence, just as new substances result from chemical reactions. This is "creative synthesis" in philosophy, epigenesis in development. Differentiations appear chiefly in the cytoplasm, but only as the result of interaction between cytoplasm and nucleus. Similarly, it may be argued, smaller units of organization, such as chromosomes or chromosomeres, do not in themselves give rise to any adult part, but only as they interact upon one another are new parts formed.
In many cases the first formation of such new substances appears in the immediate vicinity of the nucleus and, like assimilation itself, is evidently brought about by the interaction of nucleus and cytoplasm. In certain cases it can be seen that the achromatin and oxychromatin which escape from the nucleus during division take part in the formation of new substances in the cell body, and since the oxychromatin is derived from the chromosomes of the previous cell division, it is probable that the chromosomes are a factor in this process.
Weismann maintained that the chromosomes and the inheritance units contained in them undergo differentiation by a process of disintegration and that these disintegrated units escape into the cell body and there produce different kinds of cytoplasm in different cells. A somewhat similar view was advanced by de Vries in his theory of intracellular pangenesis. However, as we have seen already, there is good evidence that the chromosomes do not undergo progressive differentiation in the course of development; they always divide with exact equality, and even in highly differentiated tissue cells their number and form remain as in embryonic cells.
On the other hand, the cytoplasm undergoes progressive differentiation, and when by pressure or centrifugal force such differentiated cytoplasm is brought into relations with strange nuclei the differentiations of the cytoplasm are not altered thereby, thus showing that the different nuclei are essentially alike and that differentiations are mainly limited to the cytoplasm. Thus the differentiations of cells are not due to the differentiations of their nuclei, but rather the reverse is true,—such differentiations of nuclei as occur are due to differentiations of cytoplasm in which they lie. Nevertheless, differentiations do not take place in the absence of nuclear material, and it seems probable that the interaction of nucleus and cytoplasm are necessary to the formation of the new cytoplasmic substances which appear in the course of development.
2. Segregation and Isolation of Different Substances in Cells
But differentiation consists not only in the formation of different kinds of substances in cells, but also in the separation of these substances from one another. This separation is brought about to a great extent by flowing movements within cells which are associated especially with differential cell division.
In all these processes of heredity and development cell division plays a particularly important part. If cell divisions were always exactly alike there could be no initial difference between the daughter cells, and unless acted upon by different stimuli all cells would remain exactly alike. But there is much evidence that daughter cells are often unlike from the time of their formation, and that different stimuli act upon them to still further increase this initial difference.
(a) Differential and Non-differential Cell Division
When each half of any dividing unit is like the other half the division is non-differential. So far as we now know the divisions of all the smallest elements of the cell are of this sort; there is no good evidence that the plastosomes, the chromomeres, or the chromosomes ever divide into unlike halves, though in the maturation divisions the separation of whole chromosomes leads to the appearance of a differential division of the chromosomes. But while all of the cell elements may be supposed to grow and divide into equivalent halves, there may be an unequal distribution of these halves in cell division, so that the two daughter cells are unlike. This is what is known as differential cell division and it plays a most important part in differentiation. While the chromosomes are equally distributed to the daughter cells, except in the case of the maturation divisions, the achromatin and the oxychromatin of the nucleus are not always distributed equally and this is probably a most important factor in development. The divisions of the cytoplasm of the egg are frequently differential and such divisions are known to play a great part in the embryonic differentiation.
In the differential divisions of the cytoplasm unlike substances become localized in certain parts of the cell body, chiefly by means of definite flowing movements of the cytoplasm, and when cell division occurs these substances become permanently separated by partition walls. In this way irreversible differentiations are formed. If the formation of partition walls is prevented, the different substances within the cell body may freely commingle, especially during nuclear division when the cytoplasmic movements are especially active; in such cases differentiation is arrested even though nuclear division continues. In the developing eggs of most animals partition walls between daughter cells are necessary to prevent the commingling of different kinds of substances, which are sorted by the movements within the cell and are isolated by the partition walls. In some cases, as for example, in certain protozoa, the commingling of different kinds of protoplasm within a cell may be prevented by the viscosity of portions of the protoplasm, or by the formation of intracellular membranes, or by a reduction to a minimum of the mitotic movements within the cell by the persistence of the nuclear membrane during division. In general the degree of differentiation may be measured by the degree of unlikeness btween different cells, and by the completeness with which the protoplasm of different cells is prevented from intermingling.
All the phenomena of life, including heredity and development, are cellular phenomena in that they include only the activities of cells or of cell aggregates. The cell is the ultimate independent unit of organic structure and function. The only living bond between one generation and the next is found in the sex cells and all inheritance must take place through these cells. Inherited traits are not transmitted from parents to offspring but the germinal factors or causes are transmitted, and under proper conditions of environment these give rise to developed characters. Every oosperm as well as every developed organism differs more or less from every other one and this remarkable condition is brought about by extremely numerous permutations in the distribution of certain parts of the sex cells in maturation and fertilization. Sex is an inherited character dependent upon an alternative distribution of certain chromosomes of the nucleus. There is much evidence that the factors for all sorts of alternative characters are associated with the chromosomes. The differentiation of the oosperm into the developed organism is accomplished in part by the associations and dissociations of germinal units which lead to the formation of new materials and by the segregation and localization of these materials in definite cells.
Germ cells and probably all other kinds of cells are almost incredibly complex. We know that former students of the cell greatly underestimated this complexity and there is no reason to suppose that we have fully comprehended it. What Darwin said of the entire organism may now be said of every cell.