Results are presented which offer strong evidence that extensive alteration of the fates of embryonic Xenopus cells occurs independently of the schedule of cell division, after operations which lead to a doubling of the axial pattern of mesodermal differentiation in the gastrula. The experimental strategy was to make estimates of total mesodermal cell numbers and mitotic index in closely matched sets, each of three synchronous sibling embryos, fixed during the ten hours following the close of gastrulation. Within each set two embryos, an unoperated control and a sham-operated embryo whose own dorsal-lip (organizer) cells had been replaced with an equivalent graft, were developing normally. The third, experimental embryo had received an organizer implant to replace an equivalent number of cells from its ventral marginal zone, and was thus developing two axial mesodermal patterns of differentiation in relation to two dorsal midlines, the extra pattern embracing much host tissue. Mitotic index was also determined, in specific regions and throughout the mesoderm, in similar sets of embryos but at mid-gastrula stages.

The conclusions are justified by the results of a control investigation which show that there is normally no difference in cell cycle time along the presumptive dorso-ventral mesodermal, dimension, during the interval between time of operations and the determination of patterm The lack of any enhancement of mesodermal cell number in late embryos with dual axia patterns, or intervening enhancement of mitotic index in younger operated embryos, thus suggests that new patterns may be determined in the Xenopus gastrula without generation of extra cells.

The results are discussed in relation to recent ideas about pattern formation, and the concepts of morphallaxis and epimorphosis.

In the study of the control of pattern during animal development, a distinction has come to be drawn in principle between two modes of regulative response to disturbance, known as Morphallaxis and Epimorphosis (Morgan, 1901; Wolpert, 1971). In Morphallaxis, restoration to wholeness of the final pattern of differentiation or the generation of new such patterns, after removal or grafting of embryonic cells, occurs without any requirement for special production of new cells or for special cell migration. The behaviour is more reminiscent of the redistribution of some spatially graded information, by an averaging or perhaps a diffusion process amongst the cells in the system, and the term positional information was recently introduced by Wolpert (1969, 1971) to discuss the mechanism of such regulation. In Epimorphosis on the other hand, pattern restoration or extra formation occurs by specific recruitment of a local cell population at the site of disharmony, or stimulation of the cell cycle there, to produce the new pattern in tissue that would not have been made but for the disturbance (e.g. Nothiger, 1972; Bryant 1977). Cells in such systems might therefore be said to have positional values which, sometimes in conjunction with the disparate values of cells that they find themselves contacting, determine the values of the daughter cells produced. But there is no evidence for overall intercellular positional signalling.

Although we can describe clearly the formal distinction between morphallaxis and epimorphosis and the variations on the basic rules that distinguish the latter (see discussions by Wolpert, 1971; Cooke, 1975a), it is currently unclear how many developing patterns show real behaviour which partakes of both characters, or is of one character at early stages and the other later on (Faber, 1971; Summerbell, Lewis & Wolpert, 1973; Kieny & Pautou, 1976). In a recent paper proposing a universal formalism to describe behaviour of epimorphically controlled patterns, French, Bryant & Bryant (1976) suggest that primary differentiation fields of embryos, i.e. those controlling the whole body pattern at early stages, will be found to be morphallactic (independent of growth), while secondary fields that control specific patterns such as limbs or appendages at later times will be found to be epimorphic, using growth both for normal pattern formation and for regeneration where this is possible. The present paper reports the results of experiments designed to investigate this question for the primary embryonic field of the vertebrate embryo. This field exists, during gastrula stages, largely within the layer of cells known as the mesoderm.

Regulation to restore wholeness of pattern after tissue removal, and production of pattern duplication in response to presence of an extra field boundary or organizer, can occur extensively at pre-gastrula stages in amphibians, birds and almost certainly mammals (Spemann & Mangold, 1924; Spemann, 1938; Waddington, 1941; Waddington & Yeo, 1950; Waddington, 1952; Eyal-Giladi & Spratt, 1965; Cooke, 1972a, b, 1973a). Strikingly, there has been no definitive information as to whether such regulative responses involve alteration of the schedule of cell division at their early stages. Amphibian embryos are highly suitable for addressing this question, allowing performance of an operation that causes extensive pattern duplication within a tissue consisting of a manageable number of cells. The geometry of normal mesoderm formation in anuran amphibians, such as Xenopus, has been investigated using various methods (Nieuwkoop & Florschutz, 1950; Nakatsuji, 1974, 1975; Cooke, 1975b, Keller, 1976), and grafting operations, producing a duplication of the mesodermal pattern of differentiation within host tissue, have been described (Cooke, 1972ac). During this latter work however, and in three further papers (Cooke, 1973a, b, 1975b) on analysis of pattern regulation, the unjustified assumption was made that the cellular responses involved were exclusively morphallactic.

The previous observation of appropriately small and few-celled structures, arising in tailbud larvae after early size-reduction operations on blastulae (Cooke, 1975c), justifies the assertion that completion of spatial patterns does not depend rigidly on production of particular cell numbers. Many complete secondary axes, after early organizer grafts, are also obviously short and slender compared with the host axis. The criterion for pure morphallaxis, however, is more stringent than this; the production of new or replacement pattern parts should take place without any effect upon the cell-cycle in the participating tissues.

On theoretical and experimental grounds (Dettlaff, 1964; Graham & Morgan, 1966; Flickinger, Freedman & Stambrook, 1967; Chulitskaia, 1970; see also Snow, 1977) it is very unlikely that real differences of cell cycle time, within one ‘germ layer’ of embryos at particular stages, could be concealed by a homogeneous mitotic index. Variation in the cell cycle occurs principally through variation in phases other than M; thus the incidence of M is a sensitive index of the relative length of the cycle within regions of an early cell layer. The determinations of mitotic index in this study therefore serve to indicate, in the event of a significant enhancement of ultimate cell number being observed, whether such enhancement is caused by relatively rapid division among a small population of cells near the graft-host boundary, or by only slightly elevated division rate spread through much of the mesodermal field. In the latter event, the small real differences in cycle time required could be expected to go unobserved, because of the low control mitotic index and the limitation upon the numbers of cells that can be scanned.

The results offer strong evidence that the responses to disturbance in the Xenopus primary embryonic field are truly morphallactic; that an endogenous schedule of division frequency in mesodermal cells, nearly homogeneous throughout the future body axis up until the time of determination, is unaffected by artificially provoked cell interactions that alter the pattern-forming or positional values of large numbers of those cells.

The results are discussed in relation to the concept of ‘quantal mitoses’ (Holtzer, 1964, 1970), to previous work showing independence of pattern development from mitosis (Cooke, 1973b), and to possible secondary effects of pattern size upon cell division (rather than, as here, the lack of any special role of cell division in pattern creation).

Embryos were obtained by artificially induced spawning (Chorionic Gonadotrophin ‘Pregnyl’-Organon, Ltd, 350 i.u. for females and 150 i.u. for males) in Xenopus laevis kept on a diet of raw beef heart and liver. Pairs were not used more frequently than every 5 weeks. Washing, handling and demembranation of pre-gastrula stage embryos was as previously described (Cooke, 1972 a, 1975b).

Operations were performed as previously described (Cooke, 1972a, 1973a), and as discussed in the results section, but using a solution of 48 % Holtfreter saline, 16 % Niu Twitty saline and 36 % glass-distilled water, brought to pH 7·2 with 1 N-HC1. Following operations implanting stage-10 (Nieuwkoop & Faber, 1956) organizers into hosts during the 2 hours of their development preceding stage 10, the healed embryos were stored on glass in 10% amphibian saline at temperatures between 17 and 21 °C, from the onset of their gastrulation (stage 10). Unoperated, demembranated control embryos were passed through exactly the same schedule as their synchronous operated siblings.

Sets of embryos were fixed at various stages, for 24 h, in a solution of 10% 40 (w/v) formaldehyde, 2 % glacial acetic acid, 50 % alcohol and 38 % amphibian saline, followed by washing in 2% formol saline for some days. They were wax embedded and sectioned at 7 μm in precise transverse section, care being taken that the plane of section was very similar in embryos within one set for comparison. Staining was by an adaptation of the Feulgen method, followed by counterstaining with 0-2% alcoholic light green, and then orange G in 2% phosphotungstic acid. By this means, while nuclei and mitotic figures showed up clearly, the different germ layers and the cellular structure within them were visualized well enough for camera lucida drawings and, at later stages, identification of patterns of cell behaviour and differentiation.

Mitotic index was scored within particular regions of the mesodermal mantle of embryos by counting metaphase and anaphase figures, every fourth section, for 2000 cells. Where both future nuclei in an anaphase figure were seen, one mitosis was scored.

Standard estimates of cell numbers were made by counting nuclei visible in every sixth section of the mesodermal mantle throughout embryos, including the thickening around the closed blastopore in neurula and later stages (see Keller, 1976). The total count was multiplied by a constant which was the appropriate Abercrombie correction factor (Abercrombie, 1946) multiplied by 6 to give a number which is referred to as ‘the number of cells in the mesoderm ‘, but which must be an estimate, variably biased at different developmental stages. In fact neither nuclear diameter nor mean internuclear distance, used to compute the correction factor, change consistently between stages 13 and 17, but after this time elongation and flattening of mesodermal cells in the transverse plane will systematically bias such counts as absolute estimates of cell number. Their use as sensitive indices of relative cell number is justified by comparison only of equivalently shaped, synchronous sibling embryos, sectioned in identical planes.

Table 1 shows the data from an overall control investigation, on mitotic index in dorsal and ventral mesoderm within sibling sets of unoperated embryos during gastrula stages. These stages, about halfway between the times of operations and of earliest scoring of cell numbers in the experimental series, are of the order of one cell-cycle time or less after such operations, and probably precede the time at which the configuration of positional values within the mesodermal mantle is made permanent by the onset of determinations for the basic pattern (Holtfreter & Hamburger, 1955). There is no evidence for overall dorso-ventral difference in mesodermal cell cycle times during these stages (although a very localized region of higher mitotic index in the presumptive prechordal mesoderm and anteriormost endoderm, seen at all later stages examined, is already present). Thus, replacement of ventral with dorsal mesoderm, per se, would have no effect on future total cell number that might be mistaken for a stimulation of the cell cycle caused by positional interaction. There is no cell death intrinsic to early amphibian development, so that total cell number, at stages some two average cycle-times subsequent to the experimental rearrangement of cells, will be a sensitive detector of any effect upon the division cycle during the intervening period of pattern determination.

Table 1.

Dorsal, lateral and ventral mitotic indices in unoperated gastrulae of two egg-batches (18 °C)

Dorsal, lateral and ventral mitotic indices in unoperated gastrulae of two egg-batches (18 °C)
Dorsal, lateral and ventral mitotic indices in unoperated gastrulae of two egg-batches (18 °C)

Beginning anteriorly at stage 13, the notochord is rapidly segregated physically from the remaining mesoderm. Some h later, and spreading more slowly posteriorly than notochord segregation, cells destined as somite tissue segregate by dorsal convergence from all the remainder of the mesoderm (presumptive nephros and lateral plate). As will be discussed in a later paper, mitosis within these two most dorsal parts of the axial pattern effectively ceases for a long time, the possibility of mitosis seeming to decrease sharply a set time before the visible change in cell behaviour that segregates the tissues off, at each successive level of the axis. Thus apart from the prechordal centre, there is hardly any mitosis in all but the most posterior dorsal midline by stage , and almost none in anterior presumptive somite by stage 14. However, this morphogenesis, and progressive loss of mitosis in the dorsal pattern areas, itself takes place rapidly compared with the average cell cycle (estimated in excess of 10 h, see Discussion). It is therefore considered that such a relatively late effect of pattern upon cell division, following a longer period of homogenous mitosis, cannot by itself have a significant effect upon the total cell count when comparing patterns in neurulae which are developing two dorsal midlines with those which are not.

Each set of observations derives from three carefully matched, synchronous sibling embryos, one experimental, one sham-operated and one unoperated control. First, estimates have been made of total mesodermal cell number in embryo sets fixed during the ten hours following the close of gastrulation, at which time the basic spatial pattern of mesodermal determinations is considered to have been fixed, and the first external signs of the existence of the secondary pattern are to be seen. Determinations of mitotic index, in mesodermal pattern parts where cell division is still to be expected (see Results and Discussion), have also been made on this material. Secondly, mitotic index determinations have also been made in various mesodermal regions of similar sets of embryos, but at much shorter times after the operations; that is, while the secondary pattern formation is actually occurring.

Figure 1a shows the standard operation and its sham control version. The graft, cut from a very early gastrula donor, is a group of a few hundred cells at most, extending from mid-dorsal marginal surface to the blastocoele. It embraces the bottle cells of Holtfreter (1943), a small sector of the presumptive anterior dorsal mesoderm internally, and of the endodermal surface (future archenteron of the head region) externally (Nieuwkoop & Florschutz, 1950). It is implanted in harmonious orientation (see Cooke, 1972a, c) into the ventral marginal zone of a host blastula within of onset of the latter’s gastrulation, after excision of an equivalent group of marginal cells to accommodate it. In the sham operation, where a blastula is simply given a new organizer to replace its own presumptive one, development is to a normal, single pattern. After healing in, a ventral graft sinks inward while maintaining surface contact, to produce an invagination which is usually met smoothly by the advancing lines of external gastrulation activity that extend ventrally from the host’s organizer to complete a blastoporal ring. ]n the best cases, this blastopore when closed thus overlies two archenterons, but this is not necessary for substantial second mesodermal patterns to be formed internally, correlated with second neural plates. At the end of gastrulation the internal mesodermal mantle, essentially a cylindrical sleeve of cells, has advanced further than usual ventrally, having there a second axis of bilateral symmetry around which an extra pattern of differentiations develops in opposition to those due to the host’s field.

Fig. 1.

(a) The operation (left) and its sham-control version (right). The core-shaped stage-10 organizer graft is shown in surface aspect. The two recipient sites left by excision of groups of cells from the ventral and dorsal (presumptive organizer) marginal zones of host blastulae are shown in vegetal view. In these relatively late hosts, stage 9+, and about to commence their own gastrulation, the internal and external cell-size gradient is indicated, and the incipient dorsal lip activity in the experimental host. The more substantial secondary axes tend to be formed (see Fig. 2 c) when rather earlier hosts are used. The necessary re-orientation of the graft relative to the experimental host is indicated. (6) A set of three siblings at stage 13, in surface view from the rear, showing the outlines and flattenings of the early neural plates, and the slanting blastoporal groove and smaller appearance (due to the double ‘gatherings’ of the neural folds) in the experimental embryo. In this case, host and secondary neural plates are of synchronous development, (c) A set of siblings at stage 17, from the front, showing deeply indented cephalic neural folds and the incipient induction of the ectodermal cement gland beneath them. In this case the experimental secondary neural induction, while of considerable size, is delayed in its morphogenesis. In such cases, the underlying mesodermal pattern also tends to be of primitive development (as in Fig. 2d).

Fig. 1.

(a) The operation (left) and its sham-control version (right). The core-shaped stage-10 organizer graft is shown in surface aspect. The two recipient sites left by excision of groups of cells from the ventral and dorsal (presumptive organizer) marginal zones of host blastulae are shown in vegetal view. In these relatively late hosts, stage 9+, and about to commence their own gastrulation, the internal and external cell-size gradient is indicated, and the incipient dorsal lip activity in the experimental host. The more substantial secondary axes tend to be formed (see Fig. 2 c) when rather earlier hosts are used. The necessary re-orientation of the graft relative to the experimental host is indicated. (6) A set of three siblings at stage 13, in surface view from the rear, showing the outlines and flattenings of the early neural plates, and the slanting blastoporal groove and smaller appearance (due to the double ‘gatherings’ of the neural folds) in the experimental embryo. In this case, host and secondary neural plates are of synchronous development, (c) A set of siblings at stage 17, from the front, showing deeply indented cephalic neural folds and the incipient induction of the ectodermal cement gland beneath them. In this case the experimental secondary neural induction, while of considerable size, is delayed in its morphogenesis. In such cases, the underlying mesodermal pattern also tends to be of primitive development (as in Fig. 2d).

In this work, only cases where the second axis consists of an appreciable part of the total mesodermal material are considered, the size of the second field being revealed only at early neurula stages by alterations in shape, because of dorsal elongation and convergence due to each axis, and by the flattening defining each area of neuralized ectoderm (Fig. 1 b). The secondarily induced ectodermal derivatives finally seen (e.g. Fig. 1 c) are relatively passive reflections of the patterns of mesodermal determination beneath them (Nieuwkoop, 1970; Cooke, 1972 b) that the present work is concerned with. Previouswork has shown that donor cells essentially form some anterior dorsal mesoderm, and anteriormost endodermal structures only. Much of the new pattern is within host tissue.

The data on cell numbers come from ten sets, each of three precisely synchronously developing sibling embryos, matched as closely as possible for shape, and plane of sectioning, and each involving one successful example of secondary pattern formation from stage 13 onwards, one sham operation developing normally, and one unoperated control. Mitotic indices were also scored for dorsal, lateral and ventral mesoderm in transverse sections of further such sets of synchronous sibling embryos, at gastrula stages midway between the time of operation and that of earliest cell counting. Figure 1 b and c show typical external appearance of sets of three, at neurula stages 13 and 17. Figure 2a–d show the mid-axis transverse sectional appearance of, respectively, a stage-unoperated gastrula, a stage-15 sham-operated neurula developing normally, and its two siblings with doubled axial patterns. In Figure 2a the boundaries are marked according to which mesoderm was scored as dorsal, lateral or ventral (by reference to the archenteron) for mitotic index purposes.

Fig. 2.

marked in, of transverse sections of control and experimental embryos at mid-axial levels. (a) Stage- -unoperated, ‘anterior’ to the blastoporal ring, to show the limits relative to the dorsal midline for scoring dorsal, lateral and ventral M.I. (b) A sham-operated, normally developing stage 15. At this axial level, although the notochord is well developed, the future somitic cells are only partially aligned and gathered dorsally to leave the lateral plate (and presumptive nephros). (c) An experimental stage 15, where the secondary axis is synchronous with the host’s pattern in differentiation, and associated with its own archenteron cavity in the endoderm, (d) A sibling of (b) and (c), experimental, but where the obvious secondary axial pattern in the mesoderm and its overlying ectoderm is of relatively delayed development. There is as yet no notochord differentiation or somitic alignment, and no archenteric cavity. On shape criteria, (d) was not paired with (6) for cell counting, whereas (c) was.

Fig. 2.

marked in, of transverse sections of control and experimental embryos at mid-axial levels. (a) Stage- -unoperated, ‘anterior’ to the blastoporal ring, to show the limits relative to the dorsal midline for scoring dorsal, lateral and ventral M.I. (b) A sham-operated, normally developing stage 15. At this axial level, although the notochord is well developed, the future somitic cells are only partially aligned and gathered dorsally to leave the lateral plate (and presumptive nephros). (c) An experimental stage 15, where the secondary axis is synchronous with the host’s pattern in differentiation, and associated with its own archenteron cavity in the endoderm, (d) A sibling of (b) and (c), experimental, but where the obvious secondary axial pattern in the mesoderm and its overlying ectoderm is of relatively delayed development. There is as yet no notochord differentiation or somitic alignment, and no archenteric cavity. On shape criteria, (d) was not paired with (6) for cell counting, whereas (c) was.

Table 2 shows the cell number estimates, and mitotic indices for appropriate mesodermal areas, within the sets of matched synchronous siblings, together with information about the particular operations, ambient temperatures of development and stage of fixation of the embryos. Variation in mitotic index and cell number between experiments and ovulations is much greater than that within them. Indeed, only the remarkable consistency of cell number between normal siblings makes the overall strategy a sensitive one to detect cell interactions with respect to the division cycle. The variation between experiments could reflect properties of eggs from particular toads, but there is evidence (reviewed in Dettlaff, 1964; Chulitskaya, 1970) for a differential effect of temperature on the cell cycle and on early morphogenesis during amphibian development, which could produce just the sort of variability seen.

Table 2.

Total estimated mesodermal cells, and examples of mitotic indices in still dividing regions, within matched sets of three embryos after operations

Total estimated mesodermal cells, and examples of mitotic indices in still dividing regions, within matched sets of three embryos after operations
Total estimated mesodermal cells, and examples of mitotic indices in still dividing regions, within matched sets of three embryos after operations

It can be seen that comparisons within sets give no indications of enhancement of cell number in mesoderms within which two patterns are developing, up to earliest tail-bud (20 s) stages, and no trend or significant differences in mitotic indices. Furthermore, operations as such seem to be without effect upon the cell cycle at any of these stages, even in the immediate region of the expected graft-host boundary. Particularly interesting is the lack of any difference from control values in mitotic index of the still dividing lateral mesoderm between closely situated dorsal midlines, in the stage-21 embryo shown in Figure 3.

Fig. 3.

Camera lucida outline drawing, with mesodermal nuclei marked in, of the stage-21 experimental embryo of set 10 (Table 2) at mid-axial level. Note the small cell number involved in pattern formation between the two dorsal midlines on the left of the section, as contrasted with the right. Lateral plate mitotic indices in the two areas were nevertheless indistinguishable, and of control level.

Fig. 3.

Camera lucida outline drawing, with mesodermal nuclei marked in, of the stage-21 experimental embryo of set 10 (Table 2) at mid-axial level. Note the small cell number involved in pattern formation between the two dorsal midlines on the left of the section, as contrasted with the right. Lateral plate mitotic indices in the two areas were nevertheless indistinguishable, and of control level.

Table 3 shows mitotic index data, at two stages a few hours after operations, in operated and control gastrulae. Again, no evidence for enhancement or depression of mitotic index is seen due to any operation.

Table 3.

Mitotic indices in dorsal and ventral mesoderm (graft-host border) of gastrulae after organizer implantations, in comparison with sham operations and unoperated siblings

Mitotic indices in dorsal and ventral mesoderm (graft-host border) of gastrulae after organizer implantations, in comparison with sham operations and unoperated siblings
Mitotic indices in dorsal and ventral mesoderm (graft-host border) of gastrulae after organizer implantations, in comparison with sham operations and unoperated siblings

Taken overall, these data make it very unlikely that any enhanced cell division is involved in creation of new patterns through cell interaction in the Xenopus primary embryonic field. Many of the secondary axial patterns in these experiments embrace a substantial proportion (estimated as about ) of the total mesodermal cells at neurula stages. By this developmental stage, at least in vivo, the fundamental fates of all mesodermal cells as to axial structures are almost certainly determined (Holtfreter & Hamburger, 1955). Thus the assumption, that the spatial pattern of differentiations in amphibian embryos until these stages is controlled by morphallactic interactions, is given strong support. Further circumstantial evidence for this type of interaction is seen in the transverse sections of several experimental embryos where the initial operation has been particularly asymmetrical, so that fewer cells have been situated initially between host and graft dorsal midlines, tracing one way around the embryo as compared with the other (Fig. 3). In these cases, evidence from cellsize and yolk platelet density in cells, and the positions of notochord rudiments which may fuse posteriorly (see also Cooke, 1972b), strongly suggest that the fates of cells have been so altered by the new interactions that the dorsal midlines of pattern (apexes of the fields in positional information terms) have not necessarily developed from the cells whose presumptive fates were to form them in the original graft and host fields. This phenomenon will be addressed fully in a later paper, where the use of 1 μm Epon sections will clarify such shifting of the fates of cells. Such lability of cell fate is characteristic of morphallactic interactions, as earlier studied for instance in sea urchin (Horstadius, 1939,1973) or Hydra morphogenesis (Wilby & Webster, 1970; Wolpert, Hicklin & Hornbruch, 1971), whereas in epimorphosis only a minority of the cells, those induced to special growth, produce descendents with altered fates. Figure depicts, in a highly schematic way, the interpretation of positional interactions in gastrula mesoderm required by the present findings, and contrasts it with extreme epimorphosis (Fig. 4b).

Fig. 4.

Diagrams to show extremes of a continuum of behaviour possible for cell positional interaction in pattern formation, (a) Pure morphallaxis, suggested to be in operation in the system described in this paper. (b) Extreme epimorphosis. Intermediate phenomenaare imaginable, where an essentially morphallactic redistribution of cell position values after operations, as in (a), is followed by enhanced cell division throughout the new pattern, stimulated in some way by the steeper new gradient of position values. The present results offer no evidence that this occurs by the Xenopus stages studied. Time T\ is that of operations, where donor cells (stippled territory) of mid-dorsal (A) pattern forming value are introduced among host presumptive mid-ventral (F) tissue. A–F represent presumptive pattern-forming values in the dorsal-ventral dimension, i.e. a labile positional gradient of cell state in model (a), or a fixed distribution of cell-labels in model (b). Time T2,several hours after T1 is that at which cells pattern-forming values become irrevocable commitments that they or their descendents will differentiate as particular pattern parts. Generation of a secondary set of such pattern values, between T1 and T2, is by cellular interactions provoking changes independent of division in (a), or by specific production of new cells with intercalated values in (b).

Fig. 4.

Diagrams to show extremes of a continuum of behaviour possible for cell positional interaction in pattern formation, (a) Pure morphallaxis, suggested to be in operation in the system described in this paper. (b) Extreme epimorphosis. Intermediate phenomenaare imaginable, where an essentially morphallactic redistribution of cell position values after operations, as in (a), is followed by enhanced cell division throughout the new pattern, stimulated in some way by the steeper new gradient of position values. The present results offer no evidence that this occurs by the Xenopus stages studied. Time T\ is that of operations, where donor cells (stippled territory) of mid-dorsal (A) pattern forming value are introduced among host presumptive mid-ventral (F) tissue. A–F represent presumptive pattern-forming values in the dorsal-ventral dimension, i.e. a labile positional gradient of cell state in model (a), or a fixed distribution of cell-labels in model (b). Time T2,several hours after T1 is that at which cells pattern-forming values become irrevocable commitments that they or their descendents will differentiate as particular pattern parts. Generation of a secondary set of such pattern values, between T1 and T2, is by cellular interactions provoking changes independent of division in (a), or by specific production of new cells with intercalated values in (b).

In view of these results, it seems unlikely that introduction of a competing second organizer into the primary field at an earlier stage than was accomplished in this work would lead to the type of cell cycle stimulation which characterizes epimorphosis. The latter may be restricted to later, secondary fields as suggested by French et al. (1976). It is known that still earlier organizer operations in Xenopus, or a graft into an equivalent host to that used here, but in slower developing urodele species where there is more time for interactions, can lead to essentially equal partition of the mesodermal cell population into the two axial patterns. If total mesodermal cell number were found to be increased in such newly determined twin patterns, it is predicted that this would be by abnormally extended, or radially symmetrical, cell recruitment from the animal neurectodermal zone to create the initial mesoderm, and not by any enhancement of mitosis. Such cell recruitment, or determination as mesoderm, is a yet earlier pattern-forming process whose normal course has been documented by Nieuwkoop (1969), and Weyer, Nieuwkoop & Lindermeyer (1977), who call it ‘induction’ of the mesoderm within the animal cap by the endoderm.

The question of the independence of pattern regulation from the generation of new cells is quite separate from that of the ultimate control of pattern size and proportions (in this case, the size and proportions of the whole body) by feedback control of growth. Adjustment of newly formed patterns to normal size by cell division, over a period of time, has been documented in insects (Bohn, 1970, 1971). There is also evidence that this occurs during later developmental stages, when whole vertebrate embryos are initially formed from abnormally few cells, as seen in operatively produced Xenopus (Cooke, unpublished observations) and in the widespread occurrence of normally sized identical twins, human and other. A cumulative increment of, say, two-fold in the number of cells in each part of a pattern, would be achieved over weeks of larval life by an enhancement of mitotic index undetectable at any one time. The normal M.I. in the latest of the embryos reported here, even in a region where pattern has been determined across many fewer cells than normal between host and graft centres (see Fig. 3, and set 10 of Table 2), therefore indicates only that there is no tendency even at this stage to complete the ventral parts of disturbed patterns by any local epimorphic growth process. We are not yet in a position to say at what point in development feedback interactions between growth and pattern size set in, or to speculate on their mechanism (i.e. whether by systemic, humoral controls, or by local intercellular interactions of the same type that initially regulate pattern itself).

These data are not relevant to the question of whether cells must pass through some particular phase of the cell cycle in order to register and react to a change of positional information, or to pass from a pluripotential to a more highly determined compartment in development (e.g. Holtzer, 1964, 1970, - the idea of a ‘quantal mitosis’). Both such strictures could apply during the development of these supernumerary patterns after operations, since the estimated cell-cycle time in the mesoderm is some 10 h, and a similar period elapses between operations that rearrange the cells and the subsequent determination of the patterns of differentiation among them by the close of gastrulation. Previous work (Cooke, 1973 b, c) established that if such a critical phase of the cell cycle does exist, it is not cytokinesis or mitosis itself. Xenopus embryos developed from pluripotential blastula stages up to highly determined tailbud stages under complete inhibition of the latter two aspects of the cell cycle. Subsequent work has shown that mesoderm cells in such embryos nevertheless incorporate thymidine into nuclear DNA at appreciable rates over this period of morphogenesis, thus presumably becoming endopolyploid (B. C. Goodwin, unpublished data). Thus, neither those data nor the present observations challenge the idea that some nuclear reprogramming or cellular transition through a part of the replication cycle is necessary during development (e.g. Gurdon, 1969). There is, however, no positive evidence for this, known sometimes as the ‘clean gene’ hypothesis, as regards the cellular transition between pluripotency and restriction in vertebrate pattern formation, but only for that between stem cell status and overt differentiation (Holtzer, 1964, 1970).

Of considerable interest is the appearance of typical gastrular levels of mitosis in the obvious presumptive somite-notochord region of several newly determined axial patterns, at host stages when the original dorsal pattern parts have for some hours been entered on their long, non-dividing phase of morphogenesis. In fact, in such cases no cells as advanced in histogenesis as those of the host’s own dorsal midline could be found, even though any foreign cells in them (from the donor) were even older developmentally than the host cells. An observation such as this calls into question the idea, often implicit in discussions of development, that the schedule of intracellular (genetic) events leading to progressive restrictions of developmental potential proceeds independently of any interactions modifying the positional information among cells, the latter interactions simply determining which restriction compartment the cells will enterat particular ages or choice-points.

In the present material for instance, by virtue of the new cellular positional interactions occurring, donor mid-dorsal cells have been delayed in their schedule of maturation, while host presumptive ventral cells are entering the dorsal differentiation compartments but again with a delay in schedule relative to the normal dorsal precursor cells. There is every evidence, from cell behaviour and the series of embryos as a whole, that in such cases of delayed morphogenesis a quite typical example of primary pattern is nevertheless forming. Each pattern compartment is determined after the lapse of a set ‘physiological time’ only in normal, undisturbed development.

Data in this laboratory (Steedman and Cooke, unpublished) suggest that penetration times for thymidine and for colcemid are variable in gastrulating amphibians after injections, probably because of differential access to mesodermal cells during morphogenetic movements. For this reason, it has not been considered practicable to attempt the already complex task of determining the pattern of absolute times of cell cycle phases in the mesoderm in vivo over the period of pattern determination and axis formation. The best estimate from the literature (Deuchar, 1958; Graham & Morgan, 1966; Flickinger et al. 1967), by comparing present mesodermal M.I.S with endodermal, and using the assumption of other authors that mitosis itself has a relatively constant duration within early embryos, is that Xenopus mesodermal cells from early gastrula onwards until differentiation divide once in about 10–15 h at 18–20 °C. The apparent variation in cycle time between different experiments in this paper might be explained by temperature variation (Dettlaff, 1964; Chulitskaia, 1970) as well as by individual female toad variability.

There appears to be a significant discrepancy within the data as reported so far. Over the 9 or 10 h of development between stages and 21, as seen in Table 2, there is something like a doubling of total mesoderm cell population. And yet, precisely over this period, the great majority of the cells are being removed from the cycling pool in conjunction with differentiation as notochord of somite, having only contributed on average a doubling every 10 or more hours, in an asynchronous manner, whilst in the cycle. The pre-chordal area of continuing and slightly elevated mitosis is far too small to account for this, and the tissue expansion is in any case posterior to the spreading zone of axial differentiation rather than in front of it. A subsequent paper, in reporting exactly how the cell complement of the mesoderm continues to be built up after the ‘official close’ of gastrulation, will resolve this apparent discrepancy.

I thank June Colville for expert technical work and my colleagues Susan Udin, Dennis Summerbell, Michael Gaze and Malcolm Maden for their critical discussion during preparation of the manuscript. The work is supported by the M.R.C.

Abercrombie
,
M.
(
1946
).
Estimation of nuclear population from microtome sections
.
Anat. Res
.
94
,
239
247
.
Bohn
,
H.
(
1970
).
Interkalare Regeneration und Segmentale Gradienten bei ein Extremitaten von Leucophaea Larven
.
Wilhelm Roux Arch. EntwMech. Org
.
165
,
303
341
.
Bohn
,
H.
(
1971
).
Interkalare Regeneration und Segmentale Gradienten bei ein Extremitaten von Leucophaea Larven
.
Wilhelm Roux Arch. EntwMech. Org
.
167
,
209
221
.
Bryant
,
S. V.
(
1977
).
Pattern regulation in amphibian limbs
.
In Symp. Soc. Devi Biol. 3. Vertebrate Limb and Somite Morphogenesis
(ed.
Ede
Hinchliffe
&
Balls
), pp.
313
327
.
Cambridge University Press
.
Chulitskaia
,
E. V.
(
1970
).
Desynchronization of cell divisions in the course of egg cleavage and an attempt at experimental shift of its onset
.
J. Embryol. exp. Morph
.
23
,
359
374
.
Cooke
,
J.
(
1972a
).
Properties of the primary organization field in the embryo of Xenopus laevis. I. Autonomy of cell behaviour at the site of initial organizer formation
.
J. Embryol. exp. Morph
.
28
,
13
26
.
Cooke
,
J.
(
1972b
).
Properties of the primary organization field in the embryo of Xenopus laevis. II. Positional information for axial organization in embryos with two head organizers
.
J. Embryol. exp. Morph
.
28
,
27
46
.
Cooke
,
J.
(
1972c
).
Properties of the primary organization field in the embryo of Xenopus laevis. Ill. Retention of polarity in cell groups excised from the region of the early organizer
.
J. Embryol. exp. Morph
.
28
,
47
56
.
Cooke
,
J.
(
1973a
).
Properties of the primary organization field in the embryo of Xenopus laevis. V. Regulation after removal of the head organizer, in normal early gastrulae and in those already possessing a second implanted organizer
.
J. Embryol. exp. Morph
.
30
,
283
300
.
Cooke
,
J.
(
1973b
).
Properties of the primary organization field in the embryo of Xenopus laevis. IV. Pattern formation and regulation following early inhibition of mitosis
.
J. Embryol. exp. Morph
.
30
,
49
62
.
Cooke
,
J.
(
1975a
).
The emergence and regulation of spatial organisation in early animal development
.
In Ann. Rev. Biophys. Bioeng
.
4
(ed.
Mullins et al.
)
, pp.
185
217
.
Annual Reviews Inc
.,
Palo Alto, California
.
Cooke
,
J.
(
1975b
).
Local autonomy of gastrulation movements after dorsal lip removal in two anuran amphibians
.
J. Embryol. exp. Morph
.
33
,
147
157
.
Cooke
,
J.
(
1975c
).
Control of somite number during development of a vertebrate, Xenopus laevis
.
Nature, Land
.
254
,
196
199
.
Cooke
,
J.
(
1977
).
The control of somite number during amphibian development: models and experiments
.
In Symp. Soc. Devi Biol. 3, Vertebrate Limb and Somite Morphogenesis
(ed.
Ede
Hinchliffe
&
Balls
), pp.
434
448
.
Cambridge University Press
.
Dettlaff
,
T. A.
(
1964
).
Cell divisions, duration of interkinetic states and differentiation in early stages of embryonic development
.
In Adv. in Morphogenesis
3
, pp.
323
362
.
New York and London
:
Academic Press
.
Deuchar
,
E. M.
(
1958
).
Regional differences in catheptic activity in Xenopus laevis embryos
.
J. Embryol. exp. Morph
.
6
,
223
237
.
Eyal-Giladi
,
H.
&
Spratt
,
N. T.
(
1965
).
The embryo-forming potencies of the young chick blastoderm
.
J. Embryol. exp. Morph
.
13
,
267
273
.
Faber
,
J.
(
1971
).
Vertebrate limb ontogeny and limb regeneration: morphogenetic parallels
.
Advan. Morphogen
.
9
,
127
148
.
Flickinger
,
R. A.
,
Freedman
,
A L
&
Stambrook
,
P. J.
(
1967
).
Generation times and DNA replication patterns of cells of developing frog embryos
.
Devi Biol
.
16
,
457
473
.
French
,
V.
,
Bryant
,
P. J.
&
Bryant
,
S. V.
(
1976
).
Pattern regulation in epimorphic fields
.
Science, N.Y
.
193
,
969
891
.
Graham
,
C. F.
&
Morgan
,
R. W.
(
1966
).
Changes in the cell cycle during early amphibian development
.
Devi Biol
.
14
,
439
460
.
Gurdon
,
J. B.
(
1969
).
Nucleocytoplasmic interactions during cell differentiation
.
Proc. XII Int. Congr. Genetics
3
,
191
203
.
Holtfreter
,
J.
(
1943
).
A study of the mechanics of gastrulation
.
J. exp. Zool
.
94
,
261
318
.
Holtfreter
,
J.
&
Hamburger
,
V.
(
1955
).
Embryogenesis: progressive determination: amphibians
.
In Analysis of Development
(ed.
Willier
Weiss
&
Hamburger
), p.
230
297
.
Philadelphia
:
Saunders
.
Holtzer
,
G.
(
1964
).
Control of chondrogenesis in the embryo
.
Biophys. J
.
4
,
239
250
.
Holtzer
,
H.
(
1970
).
Proliferative and quantal cell cycles and the differentiation of muscle, cartilage and blood cells
.
In Symp. Int. Soc. Cell Biol
.
9
, pp.
69
88
.
New York and London
:
Academic Press
.
Horstadius
,
S.
(
1939
).
The mechanisms of sea urchin development, studied by operative methods
.
Biol. Rev
.
14
,
132
179
.
Horstadius
,
S.
(
1973
).
Experimental embryology of echinoderms
.
Oxford
:
Clarendon Press
.
Keller
,
R. E.
(
1976
).
Dye mapping of the gastrula and neurula of Xenopus laevis. IT. Prospective areas and morphogenetic movements of the deep layer
.
Devi Biol
.
51
,
118
137
.
Kieny
,
M.
&
Pautou
,
M. P.
(
1976
).
Régulation des excédents dans le développements du bourgeon de membre de l’embryon d’oiseau. Analyse experimentale de combinaisons xenoplustiques caill/poulet
.
Wilhelm Roux Arch. Devl Biol
.
179
,
327
338
.
Morgan
,
T. H.
(
1901
).
Regeneration
.
London
:
Macmillan
.
Nakatsuji
,
N.
(
1974
).
Studies on the gastrulation of amphibian embryos: pseudopodia in the gastrula of Bufo japónicas and their significance to gastrulation
.
J. Embryol. exp. Morph
.
32
,
785
804
.
Nakatsuji
,
N.
(
1975
).
Studies on the gastrulation of amphibian embryos: cell movement during gastrulation in Xenopus laevis embryos
.
Wilhelm Roux Arch. Devi Biol
.
178
,
1
14
.
Nieuwkoop
,
P. D.
(
1969
).
The formation of the mesoderm in urodelean amphibians. I. Induction by the endoderm
.
Wilhelm Roux Arch. EntwMech. Org
.
162
,
341
373
.
Nieuwkoop
,
P. D.
(
1970
).
The organisation centre of the amphibian embryo; its origin, spatial organisation and morphogenetic action
.
Advan. Morphgen
.
10
,
1
39
.
Nieuwkoop
,
P. D.
&
Faber
,
J.
(
1956
).
Normal Table of Xenopus laevis (Daudiri)
, 2nd ed.,
Amsterdam
:
North-Holland
.
Nieuwkoop
,
P. D.
&
Florschutz
,
P. A.
(
1950
).
Quelques characteres spéciaux de la gastrulation et de la neurulation de l’oeuf de Xenopus laevis (Daudin) et de quelques autres Anoures. I. Etude descriptive
.
Archs Biol., Liege
61
,
113
150
.
Nothiger
,
R.
(
1972
).
In The Biology of Imaginai Discs
, p.
32
.
Springer-Verlag
.
Sñow
,
M. H. L.
(
1977
).
Gastrulation in the mouse: growth and regionalization of the epiblast
.
J. Embryol. exp. Morph
.
42
,
293
303
.
Spemann
,
H.
(
1938
).
Embryonic Development and Induction. Yale University Press, reprinted 1967
,
New York
:
Hofner
.
Spemann
,
H.
&
Mangold
,
H.
(
1924
).
Uber Induktioren von Embryonhanlagen durch Implantation von artfremder Organisatoren
.
Archiv. fur Mileroscopische Anatomie und Entwicklungsmechanic
100
,
599
638
.
Summberbell
,
D.
,
Lewis
,
J. H.
&
Wolpert
,
L.
(
1973
).
Positional information in chick limb morphogenesis
.
Nature, Lond
.
244
,
492
496
.
Waddington
,
C. H.
(
1941
).
Translocations of the organiser in the gastrula of Discoglosus
.
Proc. Zool. Soc. Lond. A
111
,
189
198
.
Waddington
,
C. H.
(
1952
).
The Epigenetics of Birds
.
London and New York
:
C.U.P
.
Waddington
,
C. H.
&
Yao
,
T.
(
1950
).
Studies on regional specificity within the organization centre of Urodeles
.
J. exp. Biol
.
27
,
126
144
.
Weyer
,
C. J.
,
Nieuwkoop
,
P. D.
&
Lindermeyer
,
A.
(
1977
).
A diffusion model for mesoderm induction in amphibian embryos
.
Acta Biotheoret
.
26
,
164
180
.
Wilby
,
O. K.
&
Webster
,
G.
(
1970
).
Experimental studies on axial polarity in Hydra
.
J. Embryol. exp. Morph
.
IA
,
595
613
.
Wolpert
,
L.
(
1969
).
Positional information and the spatial pattern of cellular differentiation
.
J. Theoret. Biol
.
25
,
1
48
.
Wolpert
,
L.
(
1971
).
Positional information and pattern formation
.
Curr. Top. Devi Biol
.
6
,
183
223
.
Wolpert
,
L.
,
Hicklin
,
J.
&
Hornbruch
,
A.
(
1971
).
Positional information and pattern regulation in regeneration of Hydra
.
Symp. Soc. Exp. Biol
.
25
,
391
415
.