The resting nuclei of Cucurbita pepo and Cucurbita maxima contain a single, spherical nucleolus that is relatively very large. Part, at least, of this nucleolus persists during nearly the entire process of mitosis. During the metaphase it lies in the equatorial plate and becomes elongated in the direction of the long axis of the spindle, to form first a cylinder and then, with further elongation, a dumb-bell-shaped structure, which finally separates into two fragments that migrate to the opposite poles of the spindle. This entire movement occurs prior to any anaphase migration of the chromosomes. The nucleolar fragments apparently are not included in the daughter nuclei during the telophasic reconstruction, but degenerate in the cytoplasm.
Examination of the literature suggests that a similar process of nucleolar division probably occurs in a wide variety of plant-cells.
The bearing of this type of nucleolar division on theories of the dynamics of anaphase migration of the chromosomes is briefly discussed.
The process of mitosis is one of the most impressive and intriguing of biological phenomena. Since its discovery it has been responsible for an increasing amount of speculation as to its nature and exact significance. With the development of the chromosomal theory of heredity the intent of so elaborate a procedure has become clear, but we are still in almost complete ignorance of the dynamic factors involved in most of its phases. Theories as to the exact nature of these forces are very numerous, but no one of them seems adequate to explain more than a limited number of the facts.
It has become clearly recognized from a mass of observation and experiment that the apparent continuity of the whole process of mitosis is the result of a remarkable synchronization of a number of different phases, each one of which must probably be explained independently. One of these phases that is largely independent of the others is the anaphasic movement of the chromosomes. This paper is chiefly concerned with observations 1 that seem to have some significance in relation to theories of the factors involved in bringing about the poleward movement of chromosomes.
Material and Methods
The material for this study comprised members of the genus Cucurbit a. 2 Roottips of the Hubbard squash variety of Cucurbita maxima, and of the Connecticut Field, Winter Luxury, and English Vegetable Marrow varieties of the pumpkin, Cucurbita pepo, were prepared for study. The seeds were sprouted on wet filter-paper in moist chambers, and the root-tips were cut off with a razor and fixed, when 2-3 mm. in length.
The following fixatives were used: Mottier, Bouin, Hermann, strong and weak Flemming, vom Rath, and a saturated aqueous solution of corrosive sublimate. Fixation in Mottier’s fluid was followed by the regular Benda method of mordanting and staining. Auerbach’s acid fuchsin-methyl green combination was used on sections of root-tips that had been fixed in the sublimate solution. After all the other fixing fluids either Fe-haematoxylin or safranin was used as a basic dye. This was sometimes followved by counterstaining with light green or eosin. Both longitudinal and cross-sections were cut at thicknesses of 4-6 micra.
A. The Resting Nucleus
The cells of the meristematic region of the root-tip of both species of Cucurbita studied have relatively large nuclei, the greater part of the contents of which consists of a relatively huge,’spherical nucleolus 3 (fig. 1, Pl. 11). In the sectioned and stained material this nucleolus is commonly surrounded by a clear, unstained area, which probably is a shrinkage artifact. The nucleolus stains heavily with haematoxylin and safranin. With Auerbach’s acid fuchsin-methyl green combination it is stained red; with Benda it is a brick-red colour. Outside the nucleolus there is a small region containing scattered chromatin granules. The identification of this material as chromatin is based particularly upon the fact that it is stained green by the Auerbach combination. The granules are very small, and it was not possible to determine their exact relation to a general nuclear reticulum. Except for these granules the region between the nucleolus and the nuclear membrane appears rather homogeneous.
B. The Prophase
In Cucurbita maxima, prior to the breakdown of the nuclear membrane in the late prophase, there is formed a coarse spireme of fairly uniform diameter (fig. 2, Pl. 11). The relation of this spireme to the scattered chromatin granules seen in the resting nucleus is obscure, the small size of the cells making it difficult to distinguish early stages in the spireme development. In Cucurbita pepo we have been unable to find a similar spireme. The formation of the spireme in Cucurbita maxima appears to be accompanied by no marked change in the nucleolus beyond a possible slight decrease in size. Either contemporary with, or prior to, the disappearance of the nuclear membrane the spireme apparently becomes broken up into a number of small, spherical or cylindrical chromosomes, clustered on the periphery of the nucleolus, which is now irregular in shape (fig. 3, Pl. 11). By the time the spindle has developed and the chromosomes have moved to their equatorial position upon it the nucleolar material has in many cases entirely disappeared, as, indeed, is customary in mitosis in both animal and plant cells. In some cases, 4 however, the nucleolus behaves as though its material was relatively too abundant for dissolution to be completed during the prophase period. 5 It then persists throughout the entire prophase period and is caught in the centre of the spindle-area, becoming elongated to form a cylinder with its long axis parallel to that of the developing spindle (fig. 4, Pl. 11). This alteration in shape is significant in view of the metaphase phenomena to be described below.
C. The Metaphase
Nucleolar material thus frequently persists up to the beginning of the metaphase. In such cases, in the early metaphase the nucleolus, in the form roughly of a cylinder, is always found in the centre of the equatorial plate of chromosomes, with its long axis parallel to that of the spindle (figs. 5, 10, and 11, Pl. 11). Polar views (figs. 10 and 11, Pl. 11) emphasize the fact that the nucleolus is actually within the spindle region and surrounded by a broad ring of chromosomes. Perhaps a better description would be that the nucleolus ‘perforates ‘the plate of chromosomes more or less centrally, being often much contracted at the point of perforation (fig. 10, Pl. 11), and enlarged at both extremities. This orientation is not always symmetrical with respect to the equatorial plate, since it frequently happens that the chromosomes surround the nucleolus in a plane much closer to one end of the cylindrical nucleolus than the other.
The nucleolus next becomes drawn out to a shape usually resembling a dumb-bell (figs. 6, 7, and 8, Pl. 11). Occasionally nearly all of the cylinder lies in one-half of the spindle, and in such a case the nucleolus becomes drawn out to a shape shown in fig. 9, Pl. 11. In either case, the two ends of the nucleolus continue their movement towards opposite spindle-poles with a consequent stretching of the part in the equatorial region, until finally the strand connecting the two polar portions is ruptured (figs. 13 and 14, Pl. 11). The two masses of nucleolar material thus formed continue their poleward migration and eventually reach the poles of the spindle (figs. 12, 15, and 16, Pl. 11). As would be expected from the variation in orientation of the nucleolus before and during its division, the two spheres may be nearly equal in size (fig. 12, Pl. 11), or markedly unequal as in fig. 15, Pl. 11. Sometimes the size differences are even more pronounced. When the nucleolar fragments are approximately equal, as in fig. 16, Pl. 11, their simulation of the centrioles in some animal-cells is rather striking. During all this time the nucleolus appears to be undergoing a progressive shrinkage in size, although positive determination of this fact is difficult because of initial variations in nucleolar size at the close of the prophase. However, we infer that there is a shrinkage from the fact that the sum of the volumes of the largest pair of spheres observed in the polar position never appears to be as great as that of the larger nucleoli before metaphase. During the metaphase phenomena the nucleolus is commonly seen in fixed and stained preparations to be surrounded by a narrow clear zone. One is tempted to assume that this represents a region where dissolution of the nucleolus, resulting in a reduction of its size, is taking place, but possibly it is an artifact resulting from shrinkage, like that seen in the resting nucleus.
D. The Anaphase
As is usually the case, the anaphase migration of the chromosomes in Cucurbita appears to be a very rapid process, and intermediate stages of it are very rare. Apparently no significant change takes place in the nucleolar masses, and, being slightly pushed away from their exact polar locations, they are seen at the end of the anaphase lying adjacent to the chromosome plates and appearing very much as at the close of their metaphase migration (fig. 17, Pl. 11).
E. The Telophase
During the telophase reconstruction of the nucleus the nucleolar masses continue to remain clearly apart from the chromosome groups, although they may be in their near vicinity (figs. 18 and 19, Pl. 11). The shrinkage process apparently continues (fig. 19, Pl. 11). During the nuclear reconstruction a new mass of nucleolar material appears in the midst of each group of chromosomes. It has no apparent connexion with the old nucleolar mass although the occasional close juxtaposition of the chromosome plates and these old nucleolar fragments sometimes makes this point difficult to determine. But in cases where the polar movement has proceeded to such an extent as in fig. 17, Pl. 11, it would seem practically certain that an intimate topographical relationship could never occur without a further movement of either the chromosomes or the nucleolar mass, and for such movements there is no evidence whatsoever. All the facts indicate that during the reorganization of the daughter nuclei the nucleolar fragments of the parent cell remain entirely extranuclear, and ultimately undergo complete disintegration in the polar cytoplasm. In sister cells in which the nuclei have been completely reconstructed it has not been possible to identify any nucleolar material in the cytoplasm. Small granules are occasionally seen, but they cannot certainly be distinguished from cytoplasmic granules that never had any relation to the nucleolus. Probably by this stage the dissolution of the nucleolar remnants has proceeded to such an extent that they have either disappeared entirely or shrunken to insignificant and unrecognizably small masses. We wish to be very emphatic upon one point, namely, that, in spite of the simulation of chromosome behaviour shown by the equatorial orientation and subsequent division of the nucleolus, the evidence practically demonstrates that this is merely a necessary result of the more or less accidental catching of the nucleolus in the spindle-region. The nucleolus is strikingly different from the chromosomes in that there is no direct continuity of its substance from one cell-generation to another.
In this study our attention has been chiefly directed toward the nucleolar phenomena, and observations of other structures in the cell have been entirely incidental. This discussion will be largely confined to the same topic, namely, the behaviour of the nucleolus during mitotic division.
A resting nucleus of the same general character as that in Cucurbita has been described in Azolla by de Litardière (1921). The similarity centres largely in the presence of numerous small granules of chromatin of fairly uniform size. In the case of Azolla the evidence strongly suggests that these granules are chromosomes which are persistent throughout the interkinetic phase. These chromosomes appear to go on the metaphase plate without ever forming a definite spireme. Our observations indicate the possibility of similar phenomena in Cucurbit a pepo, but positive determination of this fact awaits more careful investigation with this particular object in mind.
The nucleolus is usually described as approximately spherical in resting plant-nuclei, and the clear space around it is commonly shown in published figures. Our tentative conclusion that this clear area is largely an artifact—the result of shrinkage—is based in part on the statement of Lundegårdh (1912) that in living cells such a space is very seldom observed, even in those forms which show it most clearly in the resting nuclei of fixed preparations.
The irregularity in shape of the prophase nucleolus has also been frequently described. Nucleoli have been observed to undergo amoeboid changes of form during the prophase in living cells of Chara (Zacharias, 1902) and Vici a (Lundegårdh, 1912). We have no clue to any broad significance which may attach to this phenomenon and it probably merely indicates some change in the physical character of the nucleolus or nuclear sap that is incidental to the prophase condition in the nucleus as a whole. The spherical form of the resting nucleolus is assumed by Lundegårdh to indicate a very fluid consistency. If this be true, the prophasic amoeboid form might be construed as an indication either of an increase in viscosity or of a decrease in surface tension.
During the prophase of cell division the nucleolus commonly undergoes a progressive shrinkage, which has often been emphasized because of its supposed indication of the transfer of material from the nucleolus to the growing chromosomes. This shrinkage commonly results in the disappearance of the nucleolar material in the earlier prophase, or at the latest before the breakdown of the nuclear membrane, as in Allium, &c. In other cases, however, the nucleolar substance may persist into the late prophase and thus be cast out into the cytoplasm at the breakdown of the nuclear membrane, as in Pustularia (Bagchee, 1925), where the final stages in nucleolar dissolution take place in the cytoplasm adjacent to the equatorial region of the spindle. Finally, in some cases, the persistent nucleolus may be caught in the spindle and involved in the actual division processes, as in Cucurbita. The fact that the decrease of nucleolar mass progresses throughout the anaphase and telophase stages in the last-mentioned cases seems to us significant in view of the hypothesis occasionally put forward that the prophase shrinkage of the nucleolus indicates some direct contribution from the nucleolus to the concomitant increase in the chromatin. Certainly chromosome growth usually ceases in the late prophase, yet the nucleolar shrinkage continues, while the nucleolar remnants become widely separated from the chromosomes (fig. 14, Pl. 11). This whole series of events seems to indicate that we have to do here with two coincident but otherwise unrelated processes, namely, increase of chromatin and decrease of nucleolar substance. The nucleolus disintegrates, not because it contributes to chromatin development, but more probably as a result of the new physical or chemical conditions in the nuclear sap—conditions that are probably incidental to the mitotic process as a whole.
The division of what superficially at least resembles nucleolar material, 6 and the polar migration of the division products coincident with or subsequent to the actual separation of the chromosomes has been described in a variety of forms. An example of such a process in the Protista is the case ofEuglena (Keuten, 1895; Tschenzoff, 1916, and others). It has been described among the filamentous green Algae by Berghs (1906) in Spirogyra, by Němec (1910a) in Cladophora simplicior, and by de T’Serclaes (1922) in Cladophora glomerata. It has been noted also in the myxomycete Spongiospora by Osborn (1911), and in the Plasmodio-phorales by Cook (1928). One example has been reported in the pteridophytes—the case of Mars ilia (Berghs, 1907) where the nucleolus is caught in the spindle, but dissolves during the metaphase. Many cases of this sort have also been more or less completely reported in the higher plants. It has been unmistakably described in Phaseolus by Posen (1896), in Roripa by Němec (1897), in Solanum by Němec (1899), in Alnus and Hibiscus by Němec (1901 a and b), in Phaseolus by Wager (1904) and Martins Mano (1905), in Ricinus and Cucurbita maxima by Němec (19105), in Cucurbita pepo by Lundegårdh (1912), in Lupinus by de Smet (1914), in Helianthus by Tahara (1915), in Olivia by Van Camp (1924), and in Canna and Lupinus by Schaede (1928). In all these cases in the higher plants it appears that nucleolar material sometimes persists until the metaphase, in which case it is caught in the spindle, and is probably divided essentially as we have described in Cucurbita. In the lower plants the division of nucleolar material usually accompanies the separation of the chromosomes instead of preceding it, and not infrequently the nucleolar derivatives are ultimately incorporated in the daughter nuclei as the definitive nucleolus. In most of the cases mentioned above among angiosperms the original descriptions are limited to a few sentences and but two or three figures. The work of Van Camp (1924) is distinguished by his more positive identification of the nucleolar material by means of staining with the Auerbach combination. Both Lundegårdh’s and Nemec’s descriptions of Cucurbita are limited to a few figures and casual references in the text, since the main purpose of their studies lay in another direction. Van Camp’s paper is fairly complete and includes some review of earlier work.
Apparently Němec at least clearly appreciated the significance of these observations in relation to general theories of the factors responsible for the anaphase movement of the chromosomes. At the time of his paper (1901) the most widely accepted hypothesis was that of Van Beneden and others that the chromosomes were pulled to the poles by the contraction of the attached spindle fibres. In a rather extensive discussion of this view Nemec strongly emphasized the point that the division and polar migration of a nucleolus, to which there were certainly no spindle fibres attached, was a practically insurmountable difficulty in the way of explaining anaphase chromosome movement as due to fibrillar contractility. This theory, Němec pointed out, could only be maintained if it assumed that the chromosomes and nucleolar fragments were moved to the poles by different forces.
This same objection to theories of fibrillar contractility was later indicated by Bonnet (1912). This author was more emphatic in his declaration that the division of the nucleolus definitely ruled out the theory of Van Beneden, and he concluded that all that is known of the process of anaphase migration is that there exist in the cell at mitosis certain factors which ordinarily cause the movement of certain cellular constituents towards the poles. He believed that these forces were not necessarily confined to the spindle region, since amyloplasts in the surrounding cytoplasm occasionally participated in the poleward movement. Bonnet was completely sceptical about the role of a spindle in the process, and believed that it had merely a remarkable chronological relation to the other phenomena—an extreme view to which we can hardly subscribe when we consider the universal presence of a spindle in cells undergoing mitosis.
In an instructive discussion of the mechanism of mitosis Tischler (1922) reiterates this argument against the fibrillar contractility theory. This author also considers its relation to the so-called ‘Stemmtheorie ‘originated by Drüner (1895) and recently revived by Bělař (1927). This theory holds that the chromosomes move to the poles because they are pushed apart by the growth of the interzonal fibres as a whole (Stemm-körper) in the region between the daughter chromosomes. As Tischler points out, this theory is quite adequate to explain the polar migration of all the chromosomes simultaneously, but completely fails to allow for a precocious division of nucleolar material. It might be added that the theory also fails to explain the precocious anaphase migration of some sex chromosomes, and even of some autosomes. These difficulties have recently been considered by Bělař (1928), who suggests means of circumventing them.
It is evident that our observations on Cucurbita offer no immediate solution of the problems presented by the mitotic spindle and the anaphase movements of the chromosomes. They do, however, clearly emphasize two features of importance. Firstly, that the anaphase migration of the daughter chromosomes is apparently not directly due to the so-called spindle-fibres with which they are in connexion, since the nucleolar portions behave in exactly the same way without any spindle-fibre attachments whatever; and secondly, that the products of the division of the nucleolus move to the spindle poles while the chromosomes, though already divided, remain unmoved in the equatorial region of the spindle.
With regard to the first of these points, our observations would seem to suggest that the spindle area represents a region in which are localized those forces of whatever kind which are responsible for anaphasic movements. This localization is most strikingly demonstrated in cases like that of Pustularia, where the nucleolus is in most instances left outside the spindle at metaphase and disintegrates in the equatorial cytoplasm, but on rare occasions divides as in Cucurbita simply because it is by chance caught in the spindle area. It is clear from other work that movements are also on foot in the general cytoplasm looking toward the bipolar orientation of chondriosomes and archiplasts during mitosis, but the identity of these with the factors at work within the spindle area seems to us by no means so clear as Bonnet, for example, assumed.
If the spindle then represents a specialized area within which very definite, directed movements take place, it seems probable that whatever the force at work, it would be equally potent regardless of the nature of the bodies which found themselves in the spindle region—whether chromosomes or nucleoli. Thus, in the present case, the circumstances strongly suggest that the presence of the nucleolus within the spindle region is a pure matter of chance, dependent primarily on the fact that its disintegration during the prophase has proceeded too slowly. The nucleolus may thus be thought of in terms of some foreign body inserted into the spindle area, and its resulting division and movements depend entirely on just where the nucleolus happens to lie with respect to the mid-region of the spindle. Presumably any other small mass of proper consistency inserted into the spindle would behave in the same way. All these things indicate that the ‘something ‘which moves chromosomes towards the spindle poles is not especially associated with the chromosomes as such. Neither does this ‘something ‘reside in the chromosomes themselves. Rather do the chromosomes move in response to the same forces which would move anything placed in the same position.
It is, then, abundantly clear that the movements of the chromosomes cannot depend on the ‘contraction ‘of spindlefibres, or any other such special apparatus. Some hypothesis involving protoplasmic streaming could doubtless be suggested, and, indeed, the recent work of Chambers (1917) on protoplasmic currents in the asters of the cleaving egg, and the studies of Spek (1918) on artificial simulacra of cleavage processes, recall the old suggestion of Bütschli (1876) that chromosome migration may be effected by streaming movements in the spindle; also the view later held by Berthold (1886) that the chromosomes might be pushed apart by the streaming of cytoplasmic material into the mid-plane of the spindle. Unfortunately we have absolutely no definite information concerning the dynamic con-ditions which prevail in the spindle, and the presence of anything comparable to ‘streaming movement ‘has not been recognized. It seems to us, therefore, at present of doubtful usefulness to attempt to trace the behaviour of bodies in the spindle area to mere protoplasmic currents. The important point is that whatever the nature of the force at work in the spindle area, it is nothing which has to do specifically with the chromosomes. It is rather part and parcel of the whole achromatic division figure.
The difficulty remains that the chromosomes do not move until a definite moment in the mitotic cycle has been reached, in spite of the clear evidence that the factors responsible for their migration are operative relatively early in spindle formation. The reasons for this stability of the equatorial chromosome complex seem to be bound up with the well-known fact that the spindle area is, in part at least, a region of higher viscosity than the surrounding cytoplasm. This is true not only of the metaphase, but particularly of the late anaphase, as Bělař’s (1927) recent study of living cells has so clearly demonstrated. But concerning any details of the morphological structure of this viscous spindle area, again we must confess an almost complete ignorance. It is, however, difficult to see how the viscosity of the spindle can be directly invoked as an ex. planation for the retention of the chromosomes in an equatorial position, since the nucleolar fragments are meanwhile moving poleward presumably propelled by the same mechanism which finally moves the chromosomes as well. In this connexion it is perhaps of interest to recall the demonstration recently given by Nassonov (1918) that in plant mitosis the ‘fibres ‘attached to the chromosomes are definitely demonstrable by osmic-acid impregnation, while the so-called central spindle remains unblackened and apparently structureless. If the ‘fibres ‘thus demonstrated represent regions of unusually high protoplasmic viscosity, the metaphase retention of the chromosomes in the equator would receive an obvious explanation. The ultimate movement of the chromosomes could be brought about by the progressive liquefaction of the fibres at one or both ends. The interzonal fibres, or ‘Stimmkörper developed between the diverging chromosomes, represent presumably the formation of a new area of very high viscosity which leaves the spindle in a remarkably firm condition and perhaps in its development assists in the extreme pushing apart of the daughter chromosome groups as a whole—a phenomenon which is indeed well known from the work of several observers, most recently Bělař (1927).
These suggestions are here put forward by way of further approach to a problem which has hitherto proved extra ordinarily baffling. The ultimate solution of the structure and dynamics of the ‘spindle ‘clearly demands much further investigation.
EXPLANATION OF PLATE 11
All of the figures have been outlined as far as possible with the camera lucida at an initial enlargement of approximately 1,675 diameters, and subsequently completed free hand. In
The major facts described in this paper were first tentatively made out during the course of an extended study of plant material with another purpose in view. A subsequent prolonged search through the literature of plant cytology has brought to light a considerable number of apparently similar cases among angiosperms, in none of which, however, have the phenomena been very completely described.
We have from time to time also noted stages indicative of phenomena similar to those here to be described in Cucurbita in several other angiosperms (and, with some modifications of detail, in Equisetum).
In root-tip cells of many plants the occurrence of an unusually large amount of nucleolar material is a striking characteristic. This may be concentrated in one very large nucleolus, as in Cucurbita, or distributed in several distinct masses, as in Equisetum, &c.
Particularly in the outer layers of the plerome.
See also Tischler (1922).
Although the general opinion seems to be that in lower forms this material is not actually equivalent to the true nucleolus of the higher plant-cell.
There are also cases in which the nucleus characteristically contains more than one nucleolus, the nucleoli being transported intact to the opposite spindle poles during the prophase-metaphase transition—for example, Karsten (1893) in Psilotum, and Bargagli-Petrucci (1905) in Equisetum. See also Zimmermann (1893) and Lenoir (1926).