ABSTRACT
Patterns of individuation occurring in the primary embryonic axis of Xenopus following excision of the organizer region of the early gastrula are described. In some 70 % of cases the information for induction of the complete head is qualitatively restored by the time of cell determination, giving rise to an essentially normal embryo. In some 40 % of cases a second posterior axis of bilaterality is formed, causing development of a secondary anus, tail-fin and spinal cord, and often somites. The probabilities of twinning in the tailfield and of failure to complete apical regulation (= head formation) are largely independent.
After such excision of the head organizer region, a delay of some 3 h in the schedule of visible differentiation in the neurula/tail-bud embryo is commonly incurred, whether or not apical regulation is successful.
When the apex is excised from a host embryo which has already contained for some hours a second apex (= head organizer) as described in an earlier paper, that grafted apex then captures a considerably increased territory in the host material, as seen from the size of the individuation field finally caused by it. Such a shift across host cells, of the boundary between fields of positional information due to two organizers, is not seen under any conditions where these are left intact, or where host excision is carried out soon after implanting the donor organizer.
In discussing the results and reconciling them with earlier observations, it is shown that they strongly suggest the presence of local polar (i.e. vectorial) properties in the presumptive mesoderm, due to signals from restricted regions which have achieved a special apical state. Repolarization of cells by a new organizer is not very rapid, and may spread decrementally from the source. Data on further delays in development, caused by the presence of the second organizer during regulation in the host apex, suggest that one organizer may act directly on cells elsewhere to delay or prevent the restoration of the apical state there.
INTRODUCTION
Hydroids have great merits for the study of pattern-forming mechanisms in development, with their spatial simplicity as compared to the amphibian embryo (Wolpert, Hicklin & Hornbruch, 1971). One disadvantage is that, generally, only the status attained at boundaries of the field (hypostome and peduncle) is visible in the results of experiments (see Wolpert, 1969). In the Xenopus embryo each intermediate zone of positional value is made visible by a corresponding histodifferentiation. In the present paper, advantage is taken of this fact to explore further the properties of pairs of dominant (organizer) regions existing in single embryos, as begun in previous papers of this series (Cooke, 1972a, b).
In those papers an operation was described, closely similar to those of Spemann and Mangold (see Spemann, 1938), which originally established the organizing role of the dorsal lip in amphibian development. The present operation has shown that in Xenopus a plug-shaped mass of cells, incorporating the earliest dorsal lip and some presumptive head endoderm/mesoderm, when implanted into an early gastrula host at a maximal distance from its own organizer in the marginal zone, can cause a complete second axial individuation field to arise amongst host cells.
In the donor embryo following excision of the organizer, restoration of apical (i.e. head structure and forebrain-inducing) values for the field will often occur in the surrounding cells as the gastrulation movements continue after healing. The final result is then a normal embryo, making the whole phenomenon a paradigm case of regulation. In this paper other patterns that may result from organizer removal are described. They suggest that the presumptively posterior parts of the primary embryonic field are in a fairly labile condition at the time of the operation, and that the normal configuration of positional information there depends upon some fairly rapidly communicated influence along the extent of the presumptive embryonic axis.
This tentative conclusion is reinforced by the results next described. Here, the patterns of positional information finally attained in the whole mesoderm are compared, as between control embryos having two organizers, and experimental embryos where the host apical cells, homologous with the grafted plug, are removed either synchronously with the initial implantation operation or else some later. The time delays occurring in the normal schedule of differentiation following some of the compound operations (though not following simple implantations) are also described.
In discussing possible mechanisms involved in the build-up and maintenance of positional information in the primary organization field of Xenopus, these observations as well as those from previous papers in the series must be incorporated. The theory that has been evolved will be discussed only in outline in this paper, in order to show how it differs qualitatively from some others that have seemed appropriate to other systems (Wolpert et al. 1971).
MATERIALS AND METHODS
Details of solutions used, handling of blastulae and gastrulae during and after operations, and of operations themselves have been presented in full in a previous paper (Cooke, 1972a). The apical excision operation, always made at stage 10, the earliest gastrula, is identical with that performed on the donor in an organizer transfer, many of the subjects of these experiments being the actual donors for operations of the second type described in this paper. After excision and implantation operations, gastrulae were left in half- or third-strength Holtfreter for about 35 min, for healing, before transfer to 1/10 strength solution on glass for storage. The incidence of exogastrulation was low, and such embryos were discarded.
Careful pretreatment with EDTA in the manner described elsewhere (Cooke, 1972 a) enabled the host stage for early implantation of an organizer to be lowered to 8−, in which case almost 4 h sometimes elapsed, at 21 °C, before onset of host stage 10 (the visible onset of dorsal lip - see Nieuwkoop & Faber, 1956).
Embryos for histology were fixed overnight in Bouin, washed for 24 h in several changes of 70 % ethanol, dehydrated and wax-embedded, and finally sectioned at 12 μ m. Staining was with haematoxylin/eosin.
RESULTS
The head organizer excision operation causes, at most, a delay in completion of the gastrulation movements as judged by time of reaching stage 12. By stage 28 in initially synchronous controls, however, there is usually delay of between 2 and 5 h in the schedule of differentiation as judged by tail-fin and somite development, and by headparts when present.
The actual pattern of primary embryonic axes, resulting from the operation, presumably represents the configuration of positional information obtaining in the mesendoderm by the time of initial cell commitment, whereupon mosaic rather than field properties would supervene (see Cooke, 1972b). The spectrum of results is variable between experiments, due principally to the varying properties of the eggs laid by individual females. However, all successful operations over a series of some ten experiments gave rise to one of the nine categories of result shown in Table 1. These categories are themselves an expression of two variables: the degree of restoration of apical properties around the excised region, and thus of induction of head structure ; and the degree of destabilization of bilaterality in the trunk/tail inducing part of the field, resulting in a more or less profound accessory tail individuation. To a first approximation, these two features of the results vary independently. However, some egg-batches give much higher incidence of successful apical restitution, or regulation, than others, and a low incidence at least of profound doubling in the tailfield is slightly correlated with this. The overall percentages of successful apical regulation and of all degrees of tail doubling are, respectively, 73 % and 40 %, taken from the pooled data of Table 1.
The relation between bifurcation in the tailfield and degree of apical restoration ( = head induction) following head organizer excision from stage-10 gastrulae

Fig. 1 shows a selection of typical individual results together with a representation of the excision operation. Fig. 2 is a photograph of a particular configuration of the result which is prominent in making up the bottom right-hand category of Table 1. Here two almost equally profound sets of somites, trunk nervous systems and tail-fin structures run ventro-ventrally opposed, coalescing over the anterior end where information has not been restored to a more apical level than to give rise to a neural vesicle underlain by notochord. In one of these examples, histology revealed continuity of both sets of somites on either side of the vesicle.
The organizer excision operation, and examples of the morphogenesis resulting by stage 28 approx., ventrolateral view. (A)The operation. The boundary of the region removed at stage 10, including presumptive head endoderm and mesoderm, is shown dotted. (B)Complete apical regulation, though with the small head structures frequently seen. A small, isolated supplementary tail structure is also present. (C)Presence of a small, isolated apical region (= cement-gland induction) only, due either to partial regulation or to incomplete excision of original material. (D)No apical regulation evident; absence of structure from anterior to approx, ear vesicle level. Profound doubling in tailfield has given rise to two complete sets of tail/trunk structures at a wide angle, coalescing anteriorly.
The organizer excision operation, and examples of the morphogenesis resulting by stage 28 approx., ventrolateral view. (A)The operation. The boundary of the region removed at stage 10, including presumptive head endoderm and mesoderm, is shown dotted. (B)Complete apical regulation, though with the small head structures frequently seen. A small, isolated supplementary tail structure is also present. (C)Presence of a small, isolated apical region (= cement-gland induction) only, due either to partial regulation or to incomplete excision of original material. (D)No apical regulation evident; absence of structure from anterior to approx, ear vesicle level. Profound doubling in tailfield has given rise to two complete sets of tail/trunk structures at a wide angle, coalescing anteriorly.
Two examples of an extreme type of result, from organizer excision. In each case a second posterior axis, at about 180° to the original one and having its own anus, tail, small somites and neural tube, meets the original directly in an anterior neural vesicle with complete absence of head structure. Development was here delayed some 5 h by operation, as judged by tail-fin morphogenesis, .
Two examples of an extreme type of result, from organizer excision. In each case a second posterior axis, at about 180° to the original one and having its own anus, tail, small somites and neural tube, meets the original directly in an anterior neural vesicle with complete absence of head structure. Development was here delayed some 5 h by operation, as judged by tail-fin morphogenesis, .
In cases where regulation has succeeded qualitatively the complete head is often, though by no means always, significantly smaller than that of controls. Qualitative failure in regulation is expressed in two forms. In the first, induction of a small cement gland, often isolated on a proboscis-like termination of the axis, is accompanied by no other head structure or by a small neural vesicle representing the position of the forebrain field above the gland, and perhaps a small monophthalmic but recognizable eye. In the other form the axis is completely truncated apically, in which case the pattern ends at or just behind ear vesicle level.
A series of excisions was made in which special care was taken to cause no disruption of the actual presumptive trunk/tailfield material or of the neurectoderm lying behind the site of excision across the exposed blastocoel, and to leave no loose cells in the embryo. The results confirm the initially surprising conclusion that the formation of accessory tail and trunk structures is genuinely due to the necessity for apical regulation, and not to mechanical splitting of the presumptive material concerned. The fate map of Xenopus is not known in great detail so far as the author is aware, but the probability of accessory tail formations did not correlate with the cleanness or otherwise of operations. Also, it seems unlikely that under conditions where presumptive areas immediately around the wound regularly heal to give unitary, even if otherwise deficient patterns, those more distant should be mechanically split. Configurations typical of double organizer embryos (see below, and Cooke, 1972b) were never seen following excisions.
The smaller, more isolated secondary tail individuations are often revealed by histology to consist only of a small-lumened neural tube, underlying a fin-like specialization of the ectoderm with a local anus-like ingression into the yolky endodermal mass at some distance from the main axis of the embryo. There is then little evidence of shape-changes in the vicinity of the second anus and the field is isolated, anteriorly, from the main neural axis. At the other extreme, however, it is difficult to ascribe priority to one or the other of two equally pro-found, histologically complete sets of trunk structures, each with anus, tail-fin, notochord and somites forming an independent axis of elongation. In such cases the co-presence of a complete head is relatively rare and this is always of the abnormally small variety. Fig. 3 shows outlines of transverse sections demonstrating each type of secondary trunk field.
Outline figures of two transverse sections of stage-28 embryos, showing additional tailfield structures following organizer excision. (A)Small, isolated secondary axis showing at a posterior level, with neural tube and fin but no proper somites or notochord in underlying mesoderm. Corresponds to Fig. 1B. (B)Point of anterior coalescence of two well-developed trunk level fields, where the anterior-most level of neural development is simply a wide lumened vesicle. Two well-developed notochords and somite-rows (outer only) are visible here. Corresponds to Fig. 1 D. .
Outline figures of two transverse sections of stage-28 embryos, showing additional tailfield structures following organizer excision. (A)Small, isolated secondary axis showing at a posterior level, with neural tube and fin but no proper somites or notochord in underlying mesoderm. Corresponds to Fig. 1B. (B)Point of anterior coalescence of two well-developed trunk level fields, where the anterior-most level of neural development is simply a wide lumened vesicle. Two well-developed notochords and somite-rows (outer only) are visible here. Corresponds to Fig. 1 D. .
Inspection during the process of induction following excision operations reveals that the accessory anus-like site, which is the most constant feature of extra tailfields, arises neogenously in the ectoderm at a point discrete from the closed blastopore, and thus by a process unlike that in normal morphogenesis. Any subsequent neural induction in the area between it and the main neuraxis is delayed relative to the main neural plate, although the folds may coalesce before closure.
Many of the classes shown in Table 1 are too small for significance-testing despite the overall sample of 120 operations. However, it is apparent that only the extreme class of the bottom right, incorporating profound doubling and the least successful apical regulation, is greater than would be expected from a chance association of the two features that define it. Thus we may tentatively conclude that within individuals following organizer excision, there is no strong association between conditions leading to extra axes of bilaterality in subapical areas of the field and those decreasing the chance of successful apical restoration of positional information. Such conditions in their more extreme forms may be correlated, however. Inspection of the means and spread of developmental delays, entered with the totals for each class of tail and head configuration, suggests also that conditions associated with profound tailfield doubling and lack of apical regulation may both also be associated with the greater delays. Examination of the data for individual egg-batches (i.e. individual females) shows that in this case, lower probability of apical regulation is found in types of eggs where greatest developmental delays are observed. But these delays exhibit great variability always, and the numbers of operations that can be performed in one experiment preclude a statistical demonstration of this relationship.
A preliminary conclusion from these observations is that conditions obtaining at or near the apex of a field (apex, because the area excised and it alone can potentially organize a whole secondary field) are necessary in some rather immediate way for the stability of the positional information in regions far from the apex. It thus became of great interest to study the effect, upon the final pattern obtained in host material, of excising the host’s organizer after implanting a second one.
Table 2 gives the final results of two series of operations, where the effects of excising the host stage-10 organizer were compared undertwo sets of conditions. In the first series the second organizer had been in position, with proper cellcontact, for over 3 h at time of host excision, having been implanted at stage 8∼. In the other series, healing in of the second organizer coincided closely in time with removal of the host’s own. A small control series confirmed that without host excision, essentially similar configurations were seen after each timing, consisting (see Cooke, 1972 b) of dual anterior axial structures, merging together into a unitary axis at some anteroposterior level dependent upon interorganizer angle. The clear effect of host organizer excision, but one seen only when the implant has been in position for some hours beforehand, is a considerable encroachment of the territory of the secondary organized field into that of the host, at the boundary joining them. These are the classes 3 and 4 in Table 2. The increases in size of secondary axes in host material are absolute in these cases, not just the small, apparent increases that might be expected from the simple removal of presumptive head material from the host. The intermediate interorganizer angles used in these operations normally cause joining of axes just to the rear of the head region, and it should be realized that there is no evidence that host excision leads to change in this level, i.e. the value for positional information at the boundary between fields. The proportion of the total host material devoted to the secondary field is simply increased.
The morphogenesis resulting from combined organizer implantation and host organizer excision

As mentioned but not stressed in a previous paper (Cooke, 1972b), anterior axial structures due to the influence of implanted organizers tend to be smaller than those of the host, although complete back to the level of fusion. Also they are usually eccentrically placed and subsidiary, i.e. it is evident to external inspection (and checked by vital staining) which, axis is primary in the host and which due to the newly arisen field. There is variability in this respect among material from different females. Table 2 is composed only of results from egg-batches where even in the control unexcised operations the implants caused secondary axial fields of fair size in host material. Such secondary individuation does not always reach fully apical values, but this is considered not to prejudice the important feature of the results. Fig. 4 shows representative examples of results using such eggs.
Results of operations combining organizer implantation with host organizer excision. Drawings made at stage 28, approximately, in initially synchronous un-operated controls. (A) Control operation. Organizer implanted into host stage 9+, with no sub-sequent excision. Results for stage 8− implantations are similar. (B) Organizer implanted into stage 9+, followed within 20 min by host excision. The implanted organizer has caused apical individuation of a very subsidiary axis. Host regulation happens to be of the incomplete type, see Fig. 1 C. (C) Organizer implanted into stage 8−, followed only after 3 £ hours by host excision. In this result, of the first type discussed, the two axes (both apical) are well individuated and of almost equal size. (D) Operation as in (C), but result of the second type, with the developmental delay very great, and neither of the two massive axes well-organized or of apical development, .
Results of operations combining organizer implantation with host organizer excision. Drawings made at stage 28, approximately, in initially synchronous un-operated controls. (A) Control operation. Organizer implanted into host stage 9+, with no sub-sequent excision. Results for stage 8− implantations are similar. (B) Organizer implanted into stage 9+, followed within 20 min by host excision. The implanted organizer has caused apical individuation of a very subsidiary axis. Host regulation happens to be of the incomplete type, see Fig. 1 C. (C) Organizer implanted into stage 8−, followed only after 3 £ hours by host excision. In this result, of the first type discussed, the two axes (both apical) are well individuated and of almost equal size. (D) Operation as in (C), but result of the second type, with the developmental delay very great, and neither of the two massive axes well-organized or of apical development, .
Fig. 5 shows examples from a particular experiment where the effect of host organizer excision after early implantation is particularly clearly seen. This is due to the rather homogeneous dynamics of the eggs, where in the control and the simultaneous implantation/excision situations, fields due to implants were almost too small to give recognizable individuations.
Drawings of (A) control stage 8 implantation without host excision, (B) late stage 9+ and (C) early, stage 8 implantations combined with subsequent host organizer excision, ca. stage 32. The dynamics of the particular egg-batch giving these results was such that for three parallel series, of six operations each, no secondary axis was more individuated or massive than that shown in (A) within either series A or B. In each embryo of series C, however, paired anterior axes of approximately equal size, as shown, were developed. Examples shown were matched for an original inter-organizer angle of ca. 110° in the marginal zone, .
Drawings of (A) control stage 8 implantation without host excision, (B) late stage 9+ and (C) early, stage 8 implantations combined with subsequent host organizer excision, ca. stage 32. The dynamics of the particular egg-batch giving these results was such that for three parallel series, of six operations each, no secondary axis was more individuated or massive than that shown in (A) within either series A or B. In each embryo of series C, however, paired anterior axes of approximately equal size, as shown, were developed. Examples shown were matched for an original inter-organizer angle of ca. 110° in the marginal zone, .
The means and spreads of delay time given in Table 2 are measured against controls, having excisions only, performed synchronously with excisions in the experimental hosts reaching stage 10 at the same time. Thus we have the situation that :
delay averaging some 4 h in differentiation by tail-bud stages is incurred by the requirement for apical regulation begun at stage 10,
a second organizer, provided that it is present for a few hours before host excision, causes a pronounced additional delay averaging some 6 – 7 h, and
the presence of a second organizer for up to 4 h before onset of host stage 10 causes no delay (or advance) in the schedule of subsequent differentiation in the absence of host organizer excision.
Included among the results of class 4 in Table 2, and represented in Fig. 4 D, are embryos whose morphogenetic delay cannot be estimated, being altogether greater than that for any other type of result. These embryos characteristically show two massive but ill-defined axes, with little appearance either of anteropos-terior or of dorso-ventral differentiation within the material between them, and with greatly delayed neural folds. Neither axis shows apical individuation as evidenced by a cement-gland induction, even if the neural fold closure finally reaches a stage where this would normally be visible. Disintegration sets in before muscle-cell differentiation or fin formation and tail-bud extension occur. These results following the early implantation/excision operation are relatively distinct from all others, including those where either the graft or the host has failed to achieve fully apical development but where the two axes are well defined, and the schedule of differentiation is not much more than 6 h delayed by stage 28 in simple excised controls. When they are excluded from class 4 it appears that, within the limits of the small numbers involved, presence of a second organizer for several hours beforehand does not greatly alter the chance of complete apical regulation in the host axis following excision, which is found from the protocols to be 27/39 (= 69 %).
DISCUSSION
1. Apical regulation
Variability in the results of organizer excision may be largely due to inevitable variation in the amount and values of the presumptive material removed. It can, however, be stated that organizer plugs, whose removal has caused each of the degrees of regulation observed, are seen to cause apical individuations in hosts following implantation. Also there does appear to be a slight correlation between failure in anterior regulation, profound doubling in the tail/trunk field, and the greater delays in schedule of differentiation after gastrulation, particularly when the properties of individual female’s eggs are compared in protocols. Earlier work (Cooke, 1972 b) provides evidence that the mesendoderm causing cement-gland induction is at the apex of the primary embryonic field, in positional information terms, followed by forebrain-inducing material.
Doubling in the tail/trunk field is most generally interpretable as a destabili-zation of the bilaterality component of positional information there, somehow caused by removal of the apex of the field and occurring before any restoration of the special conditions there that constitute its dominance and organizing power. It would appear that maintenance of the labile gradient in the mesendo-derm at this stage depends rather immediately upon the presence of the material normally at its apex. Thus following the temporary removal of this, the bilaterality component of the gradient might become so shallow as to be below some threshold for reliable perception by the cells. There might then be a probabilistic process whereby, due to temporary loss of cells’ memory as to their ‘polarity potential’ (see Wolpert, 1969), the gradient field becomes reorganized around two axes of bilaterality, following restitution of the organizing power of the apex. Each of these would then become the basis for its own tail and trunk individuation and induction. It is much easier to understand the possibility of this doubling of a distant, bilaterally symmetrical region of a gradient landscape on the supposition that cells create substance gradients by actively assuming locally vectorial properties (e.g. polarized uphill transport of the substance), rather than existing in a gradient caused only by passive diffusion from a local source at the head region.
The delay of a small number of hours caused by the necessity for apical regulation, parallels that observed during an equivalent situation in the slime mould Dictyostelium (Robertson, 1972), then the anterior tip of a newly formed, developmentally labile slug is removed. There too, such a tip is capable of organizing a new slug, whilst the remainder of the old slug experiences a delay in its differentiation after regulation. In the present system, delay is found both when regulation finally succeeds and in cases where it fails qualitatively to restore all zones of anterior information, although the range of delays is greater for the extreme examples of the latter type. Thus we have the following combination of conclusions:
The final time of onset of differentiation, as visible in morphology and histology, depends upon the prior operation of a condition which is lost on apical excision and restored some 3 or 4 h afterwards. It is not set by some developmental clock within cells which is independent of the state of the mor-phogenetic field, with the latter merely determining the pattern of differentiation produced.
Nevertheless, the restoration of that condition in the field which leads in due course to differentiation does not by itself guarantee restoration of the whole pattern of positional information (e.g. the height of a gradient) apically.
Particularly in cases more delayed than average as judged by the setback in schedule of differentiation, such regulation may not have been achieved by the time cell commitment supervenes and development hence becomes mosaic.
2. The effect of a second apex during regulation
Reference to Fig. 6A defines the bilateral positional information profile normally caused by the presence of two head organizers during development. The horizontal axis defines actual extent of territory (essentially, numbers of cells) in the host material, the level of the curve the anteroposterior and/or dorsolateral levels of the positional gradient (whatever its nature) as registered in the final differentiation pattern. The level of the valley floor, representing level of fusion of fields due to the two organizers, is roughly a constant for organizers initially separated a given distance in the marginal zone of gastrulae. Levels of both graft and host apices must be set similarly, as both may lead to full head individuation. The steepness of the valley sides, however, differs, preserving approximate proportionality within the patterns of the axes despite their occupation of differently sized territories among the host cells. The asymmetry of this first profile, found when neither organizer is excised, is an equilibrium position in some sense, stable over time and under varying conditions of temperature and relative ages of graft and host organizers as discussed in a previous paper (Cooke, 1972b). This fact by itself, i.e. the different sizes of axes due to secondary and host original organizers, suggests that any substance gradient involved is a result of active properties of the host cells, which either try to maintain a ‘remembered’ position as a scalar quantity locally, or else maintain a direction of active transport of substance. Theories utilizing only passive diffusion gradients, with local sources and cells being non-vectorial sinks, would all predict an equilibrium in the present instance consisting essentially of an equal division of the host territory, in terms of area, back to the point of field fusion. The phase-shift model of Goodwin & Cohen (1969) in its original form, which does not utilize large-scale transport of substance, would also predict equal partition of the territory by two organizers, and has already been found inadequate for the present system on grounds given elsewhere (Cooke, 1972c).
Hypothetical representation of the effects of operations combining organizer implantations and excision. (A) Control operation with no excision, leading to a small secondary field in host tissue. Implantation and simultaneous excision lead to a similar final result. (B) Operation with excision following some 4 h after early (stage 8−) implantation. Two well-individuated but almost equally large axes. (C) Operation as in (B), but axes ill-individuated, neither being apical. Developmental delay considerably greater than in cases of other operations. Dashed line, profile of positional information landscape (including that following host apical regulation in (B)). Hatched, implanted head organizer. Stippled, homologous area, removed from host. Arrows, presumed directions of induced morphogen transport (= local polarity) within host territory.
Hypothetical representation of the effects of operations combining organizer implantations and excision. (A) Control operation with no excision, leading to a small secondary field in host tissue. Implantation and simultaneous excision lead to a similar final result. (B) Operation with excision following some 4 h after early (stage 8−) implantation. Two well-individuated but almost equally large axes. (C) Operation as in (B), but axes ill-individuated, neither being apical. Developmental delay considerably greater than in cases of other operations. Dashed line, profile of positional information landscape (including that following host apical regulation in (B)). Hatched, implanted head organizer. Stippled, homologous area, removed from host. Arrows, presumed directions of induced morphogen transport (= local polarity) within host territory.
An implanted organizer takes at least to achieve full, normal cell contact with its host, so that for one of the two present series of operations, introduction of the new apex and removal of that of the host are essentially contemporaneous. Under these conditions the final profile (see again Fig. 6 A) is not affected. That is, restoration of dominance properties at the host apex has occurred, restoring stability in non-apical parts of the host’s original field, before the new organizer can influence and thus capture as part of its field more than the normal equilibrium number of cells.
In the other series, where we should expect the normal equilibrium to have been attained by the time of host apex excision, the new organizer is able to capture, before restoration of organizer properties at the host field apex, an increased territory such that the two axes finally individuated are about equally massive (Fig. 6B). A significant additional time delay is also incurred by this last situation, and is in some cases very great, and associated with imperfect pattern formation in both the two equal axes. Experiments discussed in another paper (Cooke, 1973) involving mitotic inhibition, show that, as has been suspected by embryologists, differential cell division can play little or no part in regulating the overall size of primary embryonic axes.
Gradient models utilizing diffusion as the only type of global interaction within fields, but proposing that cells act as local homeostats for remembered concentrations without locally polar properties, seem also to be inadequate to explain these results. On such a model the region of the lower interface between fields due to rival upper boundaries should be the least subject to disturbance when one of these boundaries is temporarily removed. Also, if such cells are subject to readjustment between remembered values and those actually obtaining due to diffusion (as they should be to allow smooth regulation), then the normal equilibrium would again be expected to tend towards equal field division.
It is perhaps significant that the only atypical results from the double series of 96 implantation/excision operations were seen in egg-batches with somewhat atypical properties. Thus in six cases only, early implantation and later excision caused secondary axes of normally small size, i.e. class 1 results of Table 2. There was independent evidence of rapid and reliable apical regulative ability, as judged by results of the first type given in this paper, in the egg-batch concerned. In the two cases of class 4 results, i.e. enhanced size of secondary axes following simultaneous implantation/excision, deriving from separate experiments, other embryos in the same egg-batches similarly gave evidence of little apical regulative ability, with greater time delays caused by excision operations per se.
A model suggested by these data proposes, in outline, that regions having achieved organizer status, through prior chemodetermination or by subsequent regulation, become apical in a substance gradient by the emission of signals, which polarize the transport of the substance within the material around them. The substance may itself be made at a uniform rate in all cells. Alteration of transport polarity towards an implanted organizer, within regions already controlled by the host’s organizer, is difficult unless the new one is much nearer to them, or unless the influence of the original one is removed. Such repolarization and thus redirection of transport may then spread rather slowly outward from the new organizer until such time as the dominance properties are restored around the host excision site by local regulation.
The data require that the repolarization process be rather slow, and spreading, since even where the second organizer remains in situ for some hours before removal of the host influence, the boundary between fields of graft and host only advances to be roughly equidistant between them. Results given in Paper III (Cooke, 1972c) also demonstrate the retention, for some 2 h, of original polarity in small subapical squares from early gastrulae, when implanted with reversed orientation some distance from a host organizer.
Any complete theory must also explain quantitatively the delays seen under all these circumstances, in the final schedule of local differentiation. Delays can be caused even when apical positional information is finally restored completely (see Tables 1 and 2), whilst on the other hand, severe delays are often associated with failure of such restoration before cell commitment. This logically implies that two processes or aspects of the field can be distinguished. One, temporarily abolished by organizer removal but later reinstated, is required for cell differentiation and ultimately determines the time of its onset. The second, the gradient of positional information itself, starts to be degraded in the remaining host material at the time of organizer excision, and to be built up again when the organizer process is reinstated after regulation. Thus, the longer the delay before organizer reinstatement, the less the chance of recovery of a complete gradient before the onset of differentiation itself, some unit time after this reinstatement.
The delay in recovery of organizer properties and thus rebuilding of the gradient may be very variable if direct competition between the graft and the region around the host excision occurs via some signalling system, which may also be the one finally polarizing the cells and thus building the gradient. Fig. 6C is a hypothetical representation in the terms of the previous two diagrams, of the ill-differentiated, highly delayed type of result from an early implantation/ excision operation. Such a result might follow an instance of extra prolonged competition between graft and host (regulating) organizer regions, and thus delay in setting the polarity of the cells between them.
Detailed treatment of such a model, inappropriate to an experimental paper, will be presented elsewhere, and an alternative class of explanation for the phenomena reported here will now briefly be considered. This second type of explanation, essentially mechanical in nature, has been suggested by Professor Lewis Wolpert during discussion of these results. It supposes that cell stretching within the mesodermal mantle plays a dominant role in establishing the extents of fields for final axial individuation, and that such stretching, i.e. the assumption of elongated shapes by cells, is initially caused by pulling forces exerted due to cell-locomotory activity at the beginning organizer region only. Most of the tightly cohering cells of the sheet are then stretched passively, and in early phases of gastrulation their shape, and various properties depend upon it, is itself assumed to be continuously dependent upon the special status and activities of the immigrating cells near the apex of the field. Such an interpretation emphasizes the homology between amphibian gastrulation and that of the sea urchin, whose mechanics are relatively easier to study (Gustafson & Wolpert, 1967).
During the period before regulative restoration of apical status to cells, following excision of the original apex in early gastrulae, opportunity would arise for mechanical reorientation of the axis of passive stretching in cells, due to the established presence elsewhere of a second implanted group of cells with apical properties. Such an implanted organizer might require to be resident in a cell-sheet for some hours before achieving a state of graft/host contact, or graft cell activity, which allows the re-alignment and thus capture of host cells following host organizer excision.
On such a theory, the final boundary between fields due to two head organizers, in positional information terms, would result from the divide between zones of mechanical cell orientation due to each organizer, rather than from any ‘valley’ in an initial gradient landscape. This could remain true even if positional information within each field of orientated cells were subsequently to be established as a gradient with upper and lower boundaries.
In the normal equilibrium situation, without any excisions, fields due to implanted organizers are normally small in extent relative to the host’s original field. Thus the feature must remain, in this alternative mechanical theory, that cells once oriented by one organizer are relatively difficult to realign even by a new organizer situated closer to them.
Further close observation of gastrulae containing two organizers, and following organizer excision, may help in assessessing the two alternative conceptions of axial field initiation outlined here. The use of larger, slower-developing urodele species of amphibian may facilitate such work.
ACKNOWLEDGEMENTS
I am grateful for frequent discussions, during the work described in this and the previous paper, with Professor J. Maynard Smith and Drs Brian Goodwin and Stuart Newman, all of the University of Sussex. I thank Ann Blanshard and Rosi Tucker for skilled technical work. The research was supported by the Science Research Council.