From Prof. E. B. Wilson’s The Cell inDevelopment and Inheritance, by permission of the author and of the Macmillan Co., New York.a, Permanent spireme-nuclei in cells from the intestinal epithelium of a dipterous larva,Ptychoptera. (After van Gehuchten.)From Korschelt and Heider,Lehrbuch der verg. Entwicklungsgeschichte der wirbellosen Tiere, by permission of Gustav Fischer.b, Branched nucleus of the “nutritive” cell, from a portion of an ovarial tube ofForficula auricularia.
From Prof. E. B. Wilson’s The Cell inDevelopment and Inheritance, by permission of the author and of the Macmillan Co., New York.
a, Permanent spireme-nuclei in cells from the intestinal epithelium of a dipterous larva,Ptychoptera. (After van Gehuchten.)
From Korschelt and Heider,Lehrbuch der verg. Entwicklungsgeschichte der wirbellosen Tiere, by permission of Gustav Fischer.
b, Branched nucleus of the “nutritive” cell, from a portion of an ovarial tube ofForficula auricularia.
Striking modifications resulting from specialization in secretion are frequently presented by the nucleus. In many secreting cells this structure is extensively branched,e.g.many gland cells and ovarian nutritive cells of insects (fig. 4,b). In some cases the nucleus of the gland cell contains a persistent spireme thread (fig. 4,a); while almost all actively secreting cells are characterized by the possession of large or numerous nucleoli.
A and B, Ganglion cells from the cerebral cortex; in A the only slightly branched axon may extend the whole length of the spinal cord. (After Schäfer.)C, Body of a ganglion-cell showing “Nissl’s granules.”D, Sensory cells from olfactory epithelium. (After Schultze.)E, Diagrammatic representation of the sensory epithelium of retina (rod and cone layer). (After Schwalbe.)
A and B, Ganglion cells from the cerebral cortex; in A the only slightly branched axon may extend the whole length of the spinal cord. (After Schäfer.)
C, Body of a ganglion-cell showing “Nissl’s granules.”
D, Sensory cells from olfactory epithelium. (After Schultze.)
E, Diagrammatic representation of the sensory epithelium of retina (rod and cone layer). (After Schwalbe.)
(c)Specialization for the Reception and Conduction of Stimuli.—One of the most striking of the fundamental attributes of living protoplasm is its “irritability,” that is to say, its power of responding to external impressions, “stimuli,” by movement, which, both in kind and intensity, is wholly independent of the amount of energy expended by the stimulus. The stimulus conveyed by the nerve fibre to the muscle is out of all proportion to the amount of work it may cause the muscle to do. Although protoplasmic irritability is thus incapable of a simple mechanical explanation, science has rejected the assumption of a special “vital force,” and interprets protoplasmic response as being a long series of chemico-physical changes,13initiated, but only initiated, by the original stimulus; the latter thus standing in the same relation to the response it produces as the pull on the trigger to the propulsion of the rifle bullet. The function of receiving stimuli from the outer world, originally possessed to a greater or less extent by all cells, has, in the Metazoa, been relegated to one class of cells, the sensory cells14(fig. 5, D and E). Another class of cells—the “ganglion cells” or “neurones” (fig. 5, A and B), are concerned with the conduction of the stimuli so received. The contractile elements in the Metazoa are thus dependent for their stimuli on the nervous elements—the sensory cells and neurones.
Origin of Cells.—In the preceding sections we have considered the structure of the cell in relation to the fundamental attributes of cell-metabolism, irritability, and movement. We have nowto consider the cell in relation to yet another vital attribute, that of reproduction. Just as we now know that the phenomena of assimilation, respiration, excretion, response, movement and so forth, characteristic of living things, are but the co-ordinated expressions of the corresponding activities of the constituent cells, so we now know that the reproduction of the organism is, in its ultimate analysis, a cell-process. Our knowledge of the essential fact that cells only arise by the division of pre-existing cells, now a fundamental axiom of biology, and of the details of this process, have been acquired during recent years by the strenuous efforts of numerous workers.15Matthias Jakob Schleiden (1838) supposed that in plants the new cell arose from the parent cell by a sort of “crystallizing” process from the cell fluid or “cytoblastema”; the nucleolus appearing first, then the nucleus, and finally the cell-body. Theodor Schwann (1839) extended Schleiden’s theory to animal tissues, with this yet greater error, that new cells might arise, not only within the mother cell as Schleiden had supposed, but also in the intercellular substance so common in animal tissues (to which he also gave the term “cytoblastema”). By 1846, however, the botanists, thanks mainly to the efforts of Hugo von Mohl and Nägeli, recognized as a general law that cells only arise by the division of a pre-existing cell. But it was long before the universal application of this law was recognized by zoologists; the delay being largely due to pathological phenomena. The work of Kölliker (1844-1845), Karl Bogislaus Reichert (1841-1847), and Remak (1852-1855), however, finally enabled Virchow in 1858 to maintain the law of the genetic continuity of cells in the since famous aphorismomnis cellula e cellula. At this time, however, nothing was known of the details of cell-division,—one school (Reichert, L. Auerbach, and the majority of the botanists) maintaining that the nucleus disappeared prior to cell-division, the other school (von Baer, Remak, Leydig, Haeckel, &c.) maintaining that it took a leading part in the process. It is not until the appearance of Anton Schneider’s work in 1873, followed by those of Fol, Auerbach, Strasburger and many others, that we begin to gain an insight into the process. In 1882 W. Flemming was able to extend Virchow’s aphorism to the nucleus also:omnis nucleus e nucleo.
Outline of Cell-division.—There are two very distinct methods of cell-division. The more general and also more complicated method is accompanied by the formation of a complex fibrillar mechanism, and was on this account termed “mitosis” (μίτος, a thread) by W. Flemming (1882), and “karyokinesis” (κάρυον, nut, nucleus, andκίνησις, change, movement) by W. Schleicher (1878). The other method, “amitosis,” or direct division, is unaccompanied by any visible mechanism and is of relatively exceptional occurrence. In the more usual method of cell-division, or “mitosis,” we can distinguish two distinct but parallel processes, the one undergone by the chromatin and resulting in the “chromatic figure,” the other usually only concerning the cytoplasm and resulting in the “achromatic figure.”16
We will consider the chromatin changes first. The chromatin granules lose their scattered arrangement on the nuclear reticulum, and become instead arranged in a linear series to form a coiled and deeply staining “spireme thread”17(fig. 6,a). As the thread contracts, its granular origin becomes less evident, and at the same time the coils become fewer in number; the “close” spireme of earlier stages becomes the “loose” spireme of later stages. As the spireme thread contracts, it segments into a number of short, and usuallyU-shaped, segments—the “chromosomes” (Waldeyer, 1888). The number of these chromosomes is always constant for the cells of any given species of plant or animal, but varies greatly in number in different species. Thus in the parasitic wormAscaris megalocephala, var.univalens, there are only two. In the crustaceanArtemiaBauer found 168, while in the amphibianSalamandra maculata, as also in the lily, the number is 24. While these changes have been proceeding in the nucleus, changes in the cytoplasm have resulted in the formation of the achromatic figure. These cytoplasmic changes are initiated by the division into two of a minute body, the “centrosome,” originally discovered by P. J. van Beneden in 1883,18and usually lying not far from the nucleus (fig. 6,a). The daughter centrosomes separate from one another, travelling to opposite poles of the nucleus. At the same time radiations extend out into the cytoplasm from the centrosomes, and, as the nuclear membrane disappears, invade the nuclear area (fig. 7,a). Some of the fibrillae in the latter region become attached to the chromosomes and are termed “mantle fibres”; others become continuous from one centrosome to the other and constitute the “spindle fibres.” The remaining radiations at the two poles of the spindle are the “astral rays.” (The details of the formation of the achromatic figure vary considerably, some indication of this is given in the next section in connexion with the question of the origin of the mitotic mechanism.) The chromosomes now arrange themselves in the “equatorial plate” of the spindle and each splits longitudinally into two19(fig. 6,bandc). The sister chromosomes now pass to opposite poles of the spindle (fig. 6,d), and there, returning to the “resting” condition, constitute the daughter nuclei. Division of the cell follows, usually, in animals, by simple constriction. Both Theodor Boveri and van Beneden, in their papers of 1887, regarded the centrosome as initiating, not only the division of the cell-body but that of the chromatin also; Beneden even suggested that the pull of the mantle fibres caused the division of the chromatin in the equatorial plate. W. Pfitzner in 1882 was the first to show that the splitting of the chromosomes in the equatorial plate was only the reappearance of a split in the spireme thread and was due to a correspondingdivision into two of each of the chromatin granules. In the spermatogenic cells ofAscaris, A. Brauer has shown that the chromatin granules divide while still scattered over the nuclear reticulum and before either the formation of a spireme thread or the division of the centrosome. In many other cases the reverse of this condition occurs, the centrosome dividing long before there is any indication of division in the nucleus (e.g.salamander spermatogenic cells, Meves, &c.). We must therefore, with Boveri and Brauer, regard the division of the chromatin in mitosis as a distinct reproductive act on the part of the chromatin granules, the chromosomes being merely aggregates (temporary or permanent,vide infra) of these self-propagating units.
For convenience of description it is usual to recognize four periods in mitosis: (i.) Prophase, (ii.) Metaphase, (iii.) Anaphase, and (iv.) Telophase (Strasburger, 1884). The prophase covers all changes up to the completion of the mitotic figure. The metaphase is the parting of the sister chromosomes in the equatorial plate; their passage to opposite poles of the spindle constitutes the anaphase; and their reconstruction to form the resting daughter nuclei, the telophase.
The Achromatic Figure.—The mode of origin of the achromatic figure varies greatly. In some cases a distinct and continuous spindle, the “central spindle” of F. Hermann, is visible from the very first separation of the daughter centrosomes (e.g.salamander spermatogenic cell)20(fig. 7,b). In other cases the rays only invade the nuclear area and become continuous in the equatorial plane after the centrosomes have assumed their definitive positions at the two poles of the nucleus, and may even appear to indent the disappearing nuclear membrane as they invade the nuclear area.21In the salamander testis cell (fig. 7,b), and in many other cases, the whole of the achromatic figure is obviously of cytoplasmic origin. In many cases, however, it equally obviously arises within the nucleus,22while in yet other cases23the spindle fibres are of mixed origin. The question, therefore, of the cytoplasmic or nuclear origin of the achromatic figure, at one time regarded as of considerable importance, is wholly immaterial. Various elaborate theories have been propounded to explain the mechanism of the mitotic figure. H. Fol (1873) regarded the centrosomes as centres of attractive forces, and compared the mitotic figure to the lines of force in the magnetic field, a comparison made by numerous subsequent workers. E. Klein’s hypotheses of two opposing systems of contractile fibrillae, elaborated by van Beneden (1883, 1887) and accepted by Boveri (1888), was still further extended by R. Heidenhain in relation to the leucocytes of the salamander, in which there is a permanent centrosome and astral rays to which the contractile movements of the cell appear to be due24(fig. 7, a). Hermann on the other hand confined the contractility to the astral and mantle fibres; while L. Druner regarded the spindle as exerting a pushing force, for not only do the interzonal spindle fibres elongate during the anaphase, but they were often at this period contorted, while on the other hand astral rays may be entirely absent (e.g.Infusoria), and in some cases the spindle pole may be caused to project at the surface of the cell. The futility of these attempted mechanical explanations of mitosis is sufficiently clearly shown, not only by the contradictory nature of the explanations themselves, but by the fact that, in amitosis, nuclear and cytoplasmic division occur without any fibrillar mechanism whatever.
Centrosome.25—This minute body was first detected at the spindle poles by Flemming in 1875, and independently by P. J. van Beneden in 1876. The important part played by the centrosome in fertilization,26first described by van Beneden and Theodor Boveri in their papers of 1887-1888, together with the behaviour of this structure in mitosis, led these authors to regard the centrosome not only as the dynamic centre of the cell but as a permanent cell-organ, which, like the nucleus, passed by division from one cell-generation to the next. This conclusion appeared to receive considerable support from the recognition of the centrosome in various kinds of resting cells,27and especially from the relation this structure frequently shows to the locomotor apparatus of the cell (e.g.its position in the centre of the radiating fibrillae in the contractile lymph and pigment cells, and its relation to the vibratile flagellum in spermatozoa and some protozoa,e.g.Trypanosoma).28In almost all cases the centrosome of the resting cell, when this can be detected, lies in the cytoplasm, and is often already divided in preparation for the next mitotic division (e.g.spermatogenic cells of the salamander; Meves). In some cases, however, it resides in, or arises from, the nucleus (Brauer; spermatogenesis ofAscaris, var.univalens). This indifferent nuclear or cytoplasmic position for the centrosome is paralleled by the attraction sphere or homologue of the centrosome in many Protozoa. Thus in many forms,e.g.Euglena(Keuten), it lies within the nucleus, while in other forms,e.g.Noctiluca(Ishikawa, 1894, 1898; Calkins, 1898) andParamoeba(F. Schaudinn, 1896), it lies in the cytoplasm, while inTetramitusit coexists with a “distributed” nucleus. In the Heliozoa conditions are exceptionally interesting; not only is the centrosome—here resembling in appearance that of the higher forms—permanently visible and extranuclear, lying at the centre of the radiations characteristic of these forms, but there is the strongest possible evidence for its formationde novo. For Schaudinn has shown inAcanthocystisthat, in the formation of the swarm spores, the nucleus divides amitotically, the centrosome remaining visible and unchanged at the centre of the radiating processes. Yet a centrosome appears later in the nucleus of the swarm spores and migrates into the cytoplasm. The experiments of T. H. Morgan and E. B. Wilson, in which numerous centrosomes and asters (“cytasters”) are caused to appear in unfertilized sea-urchin eggs by a brief immersion in a 13% solution of magnesiumchloride in sea-water,29as also the possibility in many cases that even in normal fertilization the cleavage centrosomes may arisede novo,30make it no longer possible to regard the centrosome as a permanent cell-structure.
Significance of Mitosis.—Whatever may be the nature of the chemico-physical changes occurring during cell-division, of which the achromatic spindle and astral rays are the visible expression, it is certain that the whole of this complicated process has for its function, not the division of the chromatin, for that has already occurred on the spireme thread or even earlier, but the distribution of the divided chromatin granules to the two daughter nuclei. It is indeed usually assumed that the mitotic mechanism is not merely for the distribution, but for theequaldistribution, of the sister granules to the two daughter nuclei. The conspicuous part the chromatin is seen to play in the whole mechanism of heredity—in maturation, fertilization and development—indicating as it does that the chromatin is the chief, if not the only, bearer of the specific qualities of the organism, sufficiently clearly emphasizes the importance of the equal distribution of this substance between the daughter cells at successive cell-divisions. There are, however, serious objections to the interpretation of mitosis as an adaptation to ensure this equal distribution of the chromatin. Not only does the occurrence of amitosis show that the mitotic mechanism is not essential for either nuclear or cytoplasmic division, but direct division may occur31in the life-history of the germ cells, the very point at which it should not occur had mitosis the significance usually attached to it. On the other hand, the most elaborate mitosis occurs in cell-tissues (e.g.skin of salamander larva) which can take no possible share in the reproduction of the species. Moreover, we have no reason for supposing that the division of the chromatin in amitosis is not as meristic, and its subsequent distribution as equal, as is so visibly the case in mitosis.32It is necessary, therefore, to seek for some other explanation of the elaborate mechanism of mitosis than that which assumes it necessary for the equal distribution of the divided chromatin granules. The present writer believes the true explanation to be found in that great economic law of nature, “division of labour.” The same economy which, working under the control of natural selection, has produced the complexly differentiated tissues of the higher metazoa, which has led to the sexual differentiation between the conjugating gametes and thus to the sexual differentiation of the parents, has resulted in the production of mitosis. Only here the economy finds expression in division of labour, not in space, but in time. The work of the self-propagating chromatin granules is so ordered that periods of undisturbed metabolic activity alternate with periods of reproductive activity. The brief space of time occupied by the latter process has necessitated a more elaborate specialization of the forces—whatever their nature—controlling cell-division; a specialization which has resulted, just as a similar specialization in so many other cases has resulted, in a visible differentiation of the cell-protoplasm. This explanation is in harmony with the occurrence of typical mitosis in active tissue cells on the one hand, and of amitosis in the relatively quiescent primary germ cells on the other.
Individuality of the Chromosomes.—The most striking feature in the behaviour of the chromatin in mitosis is its resolution, at each division, into a—for any particular species—constant number of chromosomes. This constant recurrence of the specific number of chromosomes at every cell-division is capable of explanation in two radically different ways. One explanation assumes for the organism a specific peculiarity determining the segmentation of the spireme thread into a definite number of segments (Delage, 1899 and 1901).33The other regards chromosomes as independent units of the cell, retaining their identity between successive cell-divisions. The latter “Individualitäts Hypothese” was originally put forward by Theodor Boveri in 1887 as a result of C. Rabl’s observation (1885) that in epidermal cells of the salamander larva the chromosomes reappear in the mitosis of the daughter cells with the same arrangement as they possessed in the prophase of the mother cell—the angles of theU-shaped chromosomes being all directed towards one pole (Rabl’s “Poleseite”) of the nucleus. In the formation of the “resting” nucleus, the chromatin, becoming metabolically active, flows out on to the linin reticulum, all trace of the chromosomes being for the time lost. InAscaris, Boveri (1888) obtained similar but still more striking results. The thickened ends of the four elongated chromosomes cause projections on the nuclear surface throughout the resting period, and the ends of the reappearing chromosomes always coincided with these protuberances; cf. also Sutton (1902) on locust spermatagonia. Moreover, the arrangement of the chromosomes must follow one of three well-marked groupings, and this is determined for each individual in the cleavage spindle of the egg and maintained throughout later development (fig. 8).
In the same worm (var.univalens) Boveri (1888 and 1899) found that occasional abnormalities in maturation resulted in the suppression of the first polar body and the inclusion of its chromosomes in the second maturation spindle; the egg-nucleus at the time of fertilization thus having two chromosomes instead of one, while the spermatozoon nucleus has only one. Three chromosomes instead of two reappear in subsequent divisions. Boveri’s “Individualitäts Hypothese” received striking support from the work of Herla (1893), L. R. Zoja (1895) and O. zur Strassen (1898). Herla and Zoja showed that if the egg ofAscaris megalocephala(var.bivalens), which possesses two chromosomes, be fertilized with the spermatozoon of var.univalens, in which the germ cell has only one chromosome and that smaller than either of the two in the other variety, three chromosomes reappear, two large and one small, in the cleavage divisions of the resulting hybrid embryo. Zur Strassen’s observations on the giant embryos ofAscarisalso support Boveri’s theory. These embryos arise by the fusion of eggs, either before or after fertilization. The number of chromosomes in the subsequent cleavage-figures is proportional to the number of nuclei that have fused together. Similar results are given by Boveri’s (1893-1895) and T. H. Morgan’s (1895) experiments on the fertilization of enucleated sea-urchin egg-fragments; all the nuclei of the resulting embryo having only half the number of chromosomes characteristic of the species (e.g.inEchinus9 instead of 18). All the above facts point to the conclusion that, as Boveri expressed it in hisGrundgesetz der Zahlenkonstanz(1888), “the number of chromosomes arising from a resting nucleus is solely dependent on the number which originally entered into its composition.”34
Boveri’s Law of Proportional Nuclear Growth.—The chromatin in the nucleus is exactly halved at every cell-division. As the bulk of the chromatin remains constant from one cell-generation to another, it must double its bulk between successive divisions. That this proportional growth of the chromatin is dependent solely on the chromatin mass, and not on that of the cell, is very clearly indicated by cases where the normal chromatin mass has been artificially increased or reduced,35the chromatin in either case doubling its bulk between successive cell-divisions, and neither the mass of the chromatin nor the number of the chromosomes undergoing any readjustment. By double or partial fertilization, different regions in the same embryo may show nuclei of different sizes (Boveri). We must therefore distinguish in the cell between “young” and “adult” chromatin. In other words the chromatin must be regarded as being composed of individual units, each with a definite constant structure and maximum growth (Boveri, 1904). This conclusion is strongly suggested, not only by the evidence in favour of the individuality of the chromosomes considered above, but also by the independent reproductive activity of the chromatin granules in the prophase of mitosis.
Differentiation among the Chromosomes.—If we grant the assumption of a persistent individuality for the chromosomes, then it becomes possible to consider whether in one and the same nucleus these structures may not take varying parts in controlling the cell’s activity in development and in inheritance. Such a differentiation among the chromosomes would be due to independent ancestry rather than to the economy resulting from a division of labour; nevertheless a division of labour of a sort would be the result of this gradual divergence of the chromosomes from one another, and we might therefore expect that, in some cases at least, amorphologicalwould accompany thephysiologicaldifferentiation. Examples of such a morphological differentiation do indeed occur in the “accessory” chromosomes first described by H. Henking (1891) for the spermatogonia ofPyrrhocoris, and since described for numerous other insects, Arachnids and Myriapods. W. Sutton’s work on the spermatogenesis ofBrachystola magnais of especial interest in this connexion. Not only does the “accessory chromosome” in this insect form a resting nucleus independent, and obviously physiologically differentiated from that formed from the remaining chromosomes (fig. 9,a), but the latter are themselves differentiated by size, there being one pair of chromosomes of each size (fig. 9,b), a point of considerable interest when we remember that half the chromosomes in each cell are necessarily derived from each parent.36
Although this morphological differentiation among the chromosomes is undoubtedly to be regarded as indicating a corresponding physiological differentiation, it by no means follows that the latter need always, or even generally, be accompanied by the former. Since, however, the specific characters of the organism must be due to the combined activity ofallthe chromosomes, any physiological differentiation among the latter should result in abnormal development if the full complement of chromosomes be not present.37Boveri,38utilizing Herbst’s method39for separating echinoderm blastomeres, has interpreted in this manner the abnormal development which H. Driesch40found almost invariably to follow the double fertilization of the sea-urchin egg. In such eggs the first cleavage spindle is four-poled. The chromosomes are half again as numerous as in normally fertilized eggs (54 instead of 36), but each is only divided once, so that in the distribution of the resulting 108 chromosomes the four daughter nuclei receive each only 27 instead of 36 (assuming the distribution to be fairly equal, which is by no means usually the case in four-poled mitosis). Driesch had already (1900) shown that any one of the first four blastomeres of a normally fertilized egg will, if isolated, develop normally. Boveri found that in the case of the doubly fertilized egg the isolated “¼” blastomeres develop very variously, a variability only to be accounted for by their varying chromosome equipment. Occasionally a three-poled instead of a four-poled figure resulted from double fertilization. In such cases Driesch found, as we should expect from Boveri’s interpretation, that the percentage of approximately normal larvae was considerably greater; for not only would the chances of an equal distribution of the chromosomes be much greater, but the number received by each of the three daughter cells would approximate to, or even equal, the normal.
Reduction.—In all the Metazoa the prevailing, and in the higher forms the only, method of reproduction is by the union (conjugation) of two “sexually” differentiated germ-cells or “gametes”; a small motile “microgamete” or spermatozoon and a large yolk-laden “macrogamete” or ovum (see Reproduction). This differentiation between the germ-cells is another example of the advantages of division of labour; for while the onus of bringing about the union of the germ-cells is thrown entirely on the spermatozoon, the egg devotes itself to the accumulation of food-material (yolk) for the subsequent use of the developing embryo. Far more yolk is thus secreted than would be possible by the combined efforts of both the germ-cells had each of these at the same time to preserve its motility. The fundamental physiological difference which this division of labour has produced in the germ-cells is reflected on to the general metabolism of the parents and underlies the sexual differentiation of the latter.41Beyond this, however, sexual differentiation does not go. The two germ nuclei which enter into the formation of the first mitotic figure of the developing egg are not only physiologically equivalent, but, at the time of their union in the egg, are usually morphologically identical.42The essence of fertilization is, therefore, the union of two germ nuclei only differing from one another in that they are derived from separate individuals.43Since the number of chromosomes appearing in mitosis is solely dependent on the number whichoriginally entered into the composition of the nucleus (Boveri’s Law of Chromosome-Constancy), it follows that, in the mitotic figures of the developing embryo, the chromosomes will be half maternal, half paternal in origin;44the germ nuclei thus necessarily possessing only half the number of chromosomes characteristic of the ordinary tissue cells of species,i.e.the somatic number.45The manner in which this “reduction” in the number of chromosomes in the germ-cells is brought about, and the significance to be attached to the process, constitute the most hotly debated questions in cytology. In all the metazoa the phenomenon of reduction is associated with the two last and, usually, rapidly succeeding “maturation” divisions by which the definitive germ-cells—ova or spermatozoa—are produced.46
Assuming the persistent individuality of the chromosomes, then there are only three conceivable methods by which this numerical reduction can be brought about (Boveri, 1904, p. 60). (1) One-half the chromosomes degenerate. (2) The chromosomes are distributed entire, half to one daughter cell, half to the other (reducing division of Weismann, 1887). (3) The chromosomes fuse in pairs (Conjugation of the Chromosomes, Boveri, 1892). The first possibility—that of an actual degeneration of a part of the chromatin originally suggested by van Beneden and adopted by August Weismann, Boveri and others, has been long abandoned, and a steadily increasing bulk of evidence is tending to prove the general, if not universal, occurrence of the second method—the distribution between the daughter cells of undivided chromosomes. The occurrence of such a “reducing division” was postulated on theoretical grounds by Weismann (1887)47and by Boveri (1888); by the former as a result of his adoption of de Vries’s hypothesis of self-propagating and qualitatively varying units for the chromatin; by the latter in relation to his theory of chromosome individuality. The actual occurrence of this reducing division was first demonstrated by Henking (1891) forPyrrhocoris, and afterwards by Häcker, vom Rath and many others, but especially by Rückert (1894) forCyclops(fig. 10). In this latter type the chromatin of the oocyte, as this prepares for the first maturation division, resolves itself into 12 (instead of 24) longitudinally split chromosomes (fig. 10,a). As these continue to thicken and contract a transverse fission appears (fig. 10,c). This is to be regarded as a belated segmentation of the spireme thread, and shows that the reduction so far is only a “pseudo-reduction” (Rückert), the chromosomes being really all present but temporally united in pairs,i.e.“bivalent” (Häcker). A striking confirmation of this interpretation is provided by Korschelt’s description of reduction in the annelidOphryotrocha. In this type the full somatic number of split chromosomes (here only four) appears, and these secondarily associate end to end in pairs, thus forming split “diads” (i.e.tetrads), in every way similar to those described by Rückert forCyclops. In the latter type, at the first maturation division, the sister diads are separated from one another, an “equating” division thus taking place. At the second division the diads are resolved into their constituent parts, and the “univalent” chromosomes are distributed to the daughter cells (reducing division). A similar process has since been described for numerous other types (e.g.various arthropods, Häcker, 1895-1898; vom Rath, 1895; and by Sutton forBrachystola, 1902-1903). InOphryotrocha, as inPyrrhocoris(Henking),Anasa(Paulmeir),Peripatus(Montgomery), &c., reduction occurs at the first maturation division (“pre-reduction” of Korschelt and Heider, 1900), instead of at the second division (post-reduction) as in most Copepods and Orthoptera. In many cases the tetrads (i.e.split chromosomes associated in pairs) have the form of rings, the genesis of which was first clearly determined by vom Rath (1892) in the mole cricketGryllotalpa(fig. 11). In this form the sister diads remain united by their ends but widely separate in the middle (fig. 11,b). As inCyclops, the belated transverse segmentation appears as the condensation of the chromatin proceeds (fig. 11,d), but the symmetrical tetrads which this process here produces make it impossible to determine at which of the two divisions reduction is effected. An essentially similar ring formation occurs inEnchaetaandCalanus(vom Rath), and in the CopepodsHeterocopeand Diaptomus (Rückert), and in other types.48
All the above cases, in which the reduction is effected by the distribution of entire chromosomes at one or other of the maturation divisions, may be grouped together as “pseudomitotic” (Häcker, and Korschelt & Heider). In sharp contrast to the pseudomitotic method is the “Eumitotic” method, in which the chromosomes are longitudinally divided at both divisions. Such a method not only robs the process of any “reducing” value in Weismann’s sense, but is in serious conflict with the chromosome-individuality hypothesis. Nevertheless it is in this sense that Boveri (1881) and van Beneden (1883-1887) described the maturation of the egg, and at a later period Brauer (1893) that of the spermatozoon, inAscaris. In each case the tetrads are formed by the double longitudinal splitting of the chromosomes, the latter appearing in the prophase in the reduced number. Not only was the eumitotic method ofAscaristhe first method to be described, but the descriptions are fully equal in point of clearness to that of Hertwig for the pseudomitotic maturation ofCyclops.49A similar eumitotic maturation has been described for other types also,e.g.Sagittaand the Heteropods, but nowhere more frequently than in the Vertebrates among animals and the Phanerogams among plants. In these two latter groups the chromosomes of the reducing division only rarely have a ring form comparable to that seen inGryllotalpa, &c. When such rings do occur their genesis is very obscure, and at no time do they present the appearance of “tetrads.” It is the characteristic appearance these looped chromosomes give to the first maturation division in many Vertebrates, and especially in the Amphibia (fig. 12), that originally led Flemming (1887) to term this type of mitosis “heterotypical”; the second division, lacking this peculiar appearance, being distinguished as “homotypical.” Until quite recently these looped chromosomes of the heterotypical mitosis of Vertebrates (and plants) were described as arising by the opening out of longitudinally split chromosomes, exactly as this occurs in the early prophase of the maturation divisions in such types asGryllotalpa,Diaptomus, &c. In the heterotype mitosis, however, no transverse segmentation appears, and the halves of the rings, as they separate in the first division, show an obvious longitudinal split in preparation for the second division.50Both divisions were thus interpreted as equating divisions.51The more recent works of Farmer and Moore (1903-1905), Montgomery (1903, Amphibia), and (for plants) Strasburger (1903-1904) have shown, however, that even for the higher plants and animals, a reducing division in Weismann’s sense occurs in an essentially similar manner to that so convincingly described by Rückert, vom Rath and others, for Invertebrate types. For the chromosomes of the heterotype mitosis arise by the looping round, not opening out, of the bivalent chromosomes. The first division is thus a reducing division, while the split appearing in the anaphase of the heterotype and presumably reappearing in the prophase of the homotype is the original split of the spireme thread.
The widespread, if not universal, formation of tetrads,i.e.the temporary union in pairs of split chromosomes, in reduction, and the relation this latter process always bears totworapidly succeeding maturation divisions—those completing the gametogenic cycle in animals and terminating the sporophytic generation in plants,—has received a suggestive explanation at the hands of Boveri (1904). The growth of the chromatin is an indispensable prelude to its reproduction (Boveri’s Law of Proportional Growth). The chromatin is therefore incapable of undergoing reproductive fission in two successive mitotic divisions when these are not separated by a resting (i.e.growth) period. In addition to this, the “bipolar” condition of the adult chromosomes, which determines its mode of attachment to mantle fibres frombothpoles of the spindle, is not possessed by the unripe chromatin. The undivided,i.e.unripe, chromosomes are therefore incapable of utilizing the mitotic mechanism for such a transverse fission as Weismann originally postulated. The difficulty is, however, at once overcome if the unripe chromosomes are associated in pairs in the equatorial plate, for the bivalent chromosomes so produced are bipolar just as are the adult (i.e.split) chromosomes in the ordinary and homotype mitosis.52
Synopsis(συνάπτειν, to fuse together).—During the prophase of the reducing or heterotype divisions the whole of the chromatin becomes temporarily massed together at one pole of the nucleus (Moore, 1896, for Elasmobranchs). Montgomery (1901) has suggested that this is to facilitate the temporary union in pairs, or “conjugation” of homologous paternal and maternal chromosomes. InAscaris megalocephalavar.univalens, where the somatic number is only two, the association must necessarily be between homologous chromosomes. The assumption that this “selective pairing” of equivalent chromosomes is universal is supported by the behaviour of the “Heterochromosomes” (Montgomery) of the Hemiptera. These chromosomes, distinguished by their size, are paired before, and single after, the “pseudo-reduction” has taken place. Even more convincing is Sutton’s account of reduction inBrachystolaalready referred to.53Boveri (1904) has suggested that this temporary association of the chromosomes—presumably facilitated by the synapsis—has a much deeper meaning than to ensure their correct distribution between the daughter nuclei in the heterotype mitosis; the associated chromosomes exchanging material in a manner analogous to conjugation inParamoecium.54
Present Position of the Cell-theory.—Since the time of Schleiden and Schwann a wealth of evidence has accumulated in support of the “cell-theory”—the theory which regards the cell as the unit of organic structure. “The organism consistsmorphologically, of cells, and subsists, physiologically, by means of the ‘reciprocal action’ of the cells,”—this was the cell standpoint of Schleiden and Schwann, and it is no exaggeration to say that this same conception has dominated the cell-theory almost to the present day.55The frequently striking correlation between cell-division and cell-differentiation in development has caused this process to be regarded as dependent on cell-division, while a wholly exaggerated importance has been attached to the distinction between “unicellular” and “multicellular” organisms—between “intercellular” and “intracellular” organs. The influence of the “cells” upon one another, the subordination of the cell’s growth, division and differentiation, to the requirements of the whole organism—seen in normal growth, but nowhere more strikingly than in development and regeneration,—is, however, very difficult of explanation in terms of the cell-theory as this was, until quite recently, generally understood. The very elaborate regional differentiation of the protoplasm often seen in the Protozoa sufficiently indicate that multicellular structure is no essential condition for complex regional differentiation. That the regional differentiation of the protoplasm in the Metazoa should usually correspond with cell-limits is scarcely surprising. Nor is it to be wondered at that, with so convenient a mechanism for segregation to hand as cell-division, the progressive differentiation seen during development should often appear to go hand in hand with this process. In recent years, however, evidence has been steadily accumulating to show that this association between cell-division and regional differentiation of the protoplasm in development is a casual one—as casual, and as natural, as the correspondence between cell limits and regional differentiation in the formed tissues. The fact that the regional differentiation may be foreshadowed in the egg before cleavage begins,56—that as Driesch has shown, the mode of cleavage may be artificially altered without affecting the ultimate organization of the embryo,—and many other similar observations, tend to emphasize the importance of the “organism” standpoint (C. O. Whitman, 1903, p. 642) in contradistinction to the widely prevalent “cell” standpoint. The occurrence of syncytial organs and organisms, and the increasing frequency with which protoplasmic continuity is being demonstrated between all kinds of cells, are facts tending in the same direction. In the plant kingdom the growth of themasshas been recognized as the primary factor in development;57die Pflanze bildet Zellen, nicht die Zelle bildet Pflanzen(de Bary). For the animal kingdom this “Inadequacy of the Cell-Theory of Development” has been maintained amongst others by Whitman,58and by Adam Sedgwick.59The latter author, mainly as the result of work on the development ofPeripatusand of Elasmobranch embryos, regards the developing embryo as a continuous protoplasmic reticulum, for the nuclei of which the limiting epithelial layers constitute as it were a breeding ground. Differentiation is a regional specialization of this nucleated meshwork, and is not to be regarded as the result of the proliferation and subsequent specialization of cells predestined by cleavage for this end.
It is possible to suggest a mechanico-physical explanation of multicellular structure which will deprive the cell of much of its assumed significance as a unit of organization. The fact that surface area becomes relatively less extensive as bulk increases would alone set a limit to the size of “unicellular” organisms; for not only is there a constant reaction between nucleus and cytoplasm through the nuclear membrane, but the surface of the cell serves both for the intake of food and the elimination of waste material. In addition to the limit thus imposed upon the cytoplasmic area which can be effectually controlled by the nucleus, and the necessity for a minimum surface area to the protoplasmic mass, the advantages of the more or less complete subdivision of the living substance into—as far as their metabolism is concerned—semi-autonomous units, is indicated by the mechanical support derived from the specialized cell walls and turgescent cells of the plant, and the intercellular secretions of the animal tissues. It is more than possible that these two conditions—i.e.surface area for diffusion, and mechanical support—are alone responsible for theoriginof multicellular structure, and that the sharply defined character this now so generally possesses has been secondarily acquired as a result of the facilities it undoubtedly offers for regional specialization in the protoplasmic mass.