Experiments are described that examine the state of organisation of the presumptive mesoderm and ectoderm of the Xenopus embryo at stages up to the onset of gastrulation. It is shown that a process during blastula stages, establishing the normal proportions in which this cell population is partitioned to found the two outer ‘germ layers’, has a positive regulative property. An operation has been performed to excise the yolky endodermal core, at the beginning of gastrulation, leaving only the presumptive territories of mesoderm, neural tissue, epidermis and supra-blastoporal endoderm. This reveals that by this time a stable capacity exists within the induced tissue to express the craniocaudal sequence of the normal pattern, including the proper numbers of somite segments. The mediolateral organisation of such body patterns is however abnormal. The relevance of the observations to understanding mechanisms of axial pattern control is discussed.

In early Xenopus development, ooplasmic localisations cause the process of cleavage to generate two cell types. One, derived from the yolky vegetal region, develops in isolation as endoderm, while tissue from the remainder, in similar circumstances, forms a structurally abnormal but wholly epidermal tissue (Smith et al. 1985). The mesoderm, giving rise to skeletal, muscular, excretory and mesenchymal tissues as the middle one of three cell layers after gastrulation, originates in an equatorial portion of the ectodermal territory because of specific inductive signals initiated from the endodermal one. The system has been studied by co-culturing of the normal signalling and response-competent regions, and more recently by use of isolated soluble proteins which are candidates for identity or close relationship with the endogenous signals (Nieuwkoop, 1985; Smith et al. 1985; Gurdon et al. 1985; Warner & Gurdon, 1987; Slack et al. 1987; Kimelman & Kirschner, 1987; Smith et al. 1988; reviewed in Smith, 1989). Since molecular understanding of signal and response mechanisms may be at hand, it is important to complement this with knowledge of the ‘system properties’ of mesoderm formation, considered as an episode of pattern formation (Wolpert, 1971; Meinhardt, 1982; Cooke, 1985a) within the animal cap region.

Quantitative proportions between the territories of the larval mesodermal structure have been reported to be regulated against natural variation in egg size, and hence mesoderm size (Cooke & Webber, 1985a). In this paper, I initially present data supporting this, and also showing that a similar constancy of proportion is kept between cell populations of mesoderm and ectoderm as a whole, in naturally different-sized embryos. These constancies could all in principle occur because eggs come each equipped with appropriate sized sources of inducers. In an older terminology, eggs may rely upon a balanced set of animal (ectodermal) and vegetal (endodermal) development tendencies at the outset. Alternatively, since establishment of the mesodermal territory is ultimately the stabilisation of a frontier between induced and remaining tissue in the wall of the blastula, the process of response to induction may have actively regulatory properties. A test for such properties is to equip very early embryos with experimentally enlarged or reduced initial territories of ectodermally prespecified tissue in relation to their undisturbed (vegetal) sources of induction, and to compare these with normally constructed siblings as to the proportions of tissue ultimately allocated to mesodermal and to ectodermal structure. I report here that such an experiment indeed reveals regulation of the position of the mesoectodermal boundary in relation to total tissue size, although the severe operating procedures also happen to destabilise the mechanisms that then allocate relative territory sizes for components within the mesoderm pattern.

How stable is the information for the extent of mesodermal territory, and for future organisation within it, that exists within the tissue of the animal cap by onset of gastrulation at around stage 10 (Nieuwkoop & Faber, 1967)? Several recent studies have re-emphasised the largely cell-autonomous nature of the pattern of activities that generates force for shape-change during gastrulation (Keller et al. 1985; Keller & Danilchik, 1988; Cooke & Smith, 1989), suggesting that the sequence of such change is largely preorganised by the time of its onset. The sequence of cell movement is also thought to correlate in general with the craniocaudal sequence of axial pattern (e.g. Keller, 1976) and has been proposed to be a first expression of the positional variable in preinvoluted mesoderm that sets up at least the dorsal axial pattern (Cooke, 1985a,b; Cooke & Smith, 1989). But the extent of that autonomy of pattern, by the initiation of gastrulation, has never been tested directly. I describe here the major features of bodies that follow complete removal of yolky endoderm, probably including all initial sources of induction, at stages close to the start of gastrulation. They reveal great autonomy for craniocaudal aspects of pattern including segment number, despite great alteration in the geometry of the mesoderm at gastrulation. This has relevance to models for control of segmentation. By contrast, the results reveal that the full sequence of lateral and ventral differentiation in normal mesoderm requires conditions in the normal entire embryo which are not fulfilled in endodermless ones. ‘Correct’ allocation of cells to dorsal axial vs. lateroventral modes of mesodermal differentiation appears to require normal geometry during or after gastrulation.

Operations were performed in strength NAM (Slack, 1984), pH adjusted to 7·0 with dilute HO, on beds of 2% Noble Agar, using electrolytically sharpened tungsten needles and mounted loops of baby hair. For further details see Figs 1 and 3. After operations of both the types reported, ionic strength of the bathing solution was reduced to y NAM or less within one hour, and not returned to until the close of the gastrulation period.

Fig. 1.

The alteration of mesoectodermal extent. (A) The origin of mesoderm. A lateral view of a common pattern of Xenopus egg cleavage, 32-cell stage, is shown at left. A sectional view from camera lucida drawing of fixed specimens two cell cycles later is shown at right, with descendants of the two upper (animal pole) tiers from the 32-cell stage hatched. The limits of expiants that include those cells, with or without some immediate neighbours, are indicated. The replica halfsection has those regions of blastomeres stippled, from which cleavage products normally fated to become mesoderm are derived. These embrace material more vegetal than ‘tier 2’ derivatives but also include appreciable material from tier 1. (B) Specification at the stage of operation (shown in A). Expiants that include just the hatched blastomere material form only wrinkled epidermis when cultured in simple salines (left). Inclusion of a little more vegetal material often causes development of slightly elongated, mesenchyme-filled expiants (right) in which cement gland induction (arrows) occurs. (C) The operation to alter size of the territory competent to induction. Embryos are shown about one cleavage cycle after the reciprocal grafting of larger and smaller groups of blastomeres at 128-cell stage. The actual grafts are now shown hatched, but both are from within the hatched region of A. Embryo shape is maintained by snugly fitting agar wells as indicated. Individuals not healing to surface continuity within 40 minutes, or showing any clone of cleavage-retarded cells, were rejected. (D) Transverse sections through operated siblings with smaller (left) and larger (right) ectomesodermal populations, at late neural stage. The mass of endoderm is unchanged by operation, but there has usually been relatively more elongation, giving a smaller cross section in siblings with larger mesoderms presumably by the greater force of convergent extension movement. Note relative thinness and stretching of mesoderm and epidermis of embryo with size-reduced animal cap territory. Left, 32,000 mesoderm and ectoderm cells, 46% mesoderm. Right, 54,000 mesoderm and ectoderm cells, 42 % mesoderm, np = neural plate; nc = notochord; p = paraxial mesoderm; Ip = lateral plate mesoderm; ep = epidermis; end = endoderm. Scale bar = 1 mm approx.

Fig. 1.

The alteration of mesoectodermal extent. (A) The origin of mesoderm. A lateral view of a common pattern of Xenopus egg cleavage, 32-cell stage, is shown at left. A sectional view from camera lucida drawing of fixed specimens two cell cycles later is shown at right, with descendants of the two upper (animal pole) tiers from the 32-cell stage hatched. The limits of expiants that include those cells, with or without some immediate neighbours, are indicated. The replica halfsection has those regions of blastomeres stippled, from which cleavage products normally fated to become mesoderm are derived. These embrace material more vegetal than ‘tier 2’ derivatives but also include appreciable material from tier 1. (B) Specification at the stage of operation (shown in A). Expiants that include just the hatched blastomere material form only wrinkled epidermis when cultured in simple salines (left). Inclusion of a little more vegetal material often causes development of slightly elongated, mesenchyme-filled expiants (right) in which cement gland induction (arrows) occurs. (C) The operation to alter size of the territory competent to induction. Embryos are shown about one cleavage cycle after the reciprocal grafting of larger and smaller groups of blastomeres at 128-cell stage. The actual grafts are now shown hatched, but both are from within the hatched region of A. Embryo shape is maintained by snugly fitting agar wells as indicated. Individuals not healing to surface continuity within 40 minutes, or showing any clone of cleavage-retarded cells, were rejected. (D) Transverse sections through operated siblings with smaller (left) and larger (right) ectomesodermal populations, at late neural stage. The mass of endoderm is unchanged by operation, but there has usually been relatively more elongation, giving a smaller cross section in siblings with larger mesoderms presumably by the greater force of convergent extension movement. Note relative thinness and stretching of mesoderm and epidermis of embryo with size-reduced animal cap territory. Left, 32,000 mesoderm and ectoderm cells, 46% mesoderm. Right, 54,000 mesoderm and ectoderm cells, 42 % mesoderm, np = neural plate; nc = notochord; p = paraxial mesoderm; Ip = lateral plate mesoderm; ep = epidermis; end = endoderm. Scale bar = 1 mm approx.

Sectioning of embryos was always at stage 19 in the horizontal plane, except in the case of one endodermless set and its controls where the stage was 35 and the plane transverse. The complete set of 7 ftm sections was taken, after double-wax embedding, and stained with Feulgen/Light Green/Orange G (Cooke, 1979). Estimates for the cell numbers in the clearly visible tissue layers and structures (see Figs ID and 4A-J), and the proportions between them, were made by counting nuclei on regularly spaced subsets of sections and using appropriately derived correction factors in view of cell and nuclear size, section thickness and the proportion of sections sampled (Abercrombie, 1946; Cooke, 1979a, 1985a). Sampling density was a standard one in fifteen sections for the larval (st. 35) form, and one in ten for the nearly spherical neurula form (one in five for the 3 grazing sections at ends of the series). Notochord, representing a relatively small cell number, was sampled more densely to reduce error, and the proportion of its contribution to total mesoderm was computed appropriately.

Additional larval bodies after endoderm removals, and their normal siblings, were fix-dissected under the binocular microscope, a few minutes after onset of fixation with 2% acetic acid in a solution of 1 % K2Cr2O7 in water (Cooke et al. 1987; Cooke & Smith, 1989).

Early size regulation under natural and experimental conditions

Table 1 shows the relative sizes of mesoderm and neuroectoderm (epidermis and derivatives plus rudiment of the central nervous system) in two stage-matched (st. 19) sets of neurulae with members of widely contrasting natural sizes. Set I was of synchronously obtained material from two different females characterised by the ovulation of unusually small and large eggs, respectively, while set II was of siblings collected from an unusual female ovulating eggs of two contrasting size-classes. All embryos had developed side by side, undisturbed at 22°C until fixation. It can be seen that both the overall allocation of ectodermally prespecified material to the mesoderm as defined by the end of gastrulation (Keller, 1976; Cooke, 1979b), and relative allocation to the notochord within the mesodermal pattern, are significantly controlled against overall size of the tissue field. Naturally smaller embryos may tend to have, relatively, somewhat large mesoderms and notochords. But this tendency if it exists is by no means as great as expected on the null hypothesis that the embryos of each set are a population in which whole mesoderms, and notochords, each tend towards one absolute size that is the average of those seen (see note to Table 1). Notochord extent is a sensitive index of the overall balance of pattern in Xenopus -in the sense of over-presentation, normality or under-representation of dorsal and anterior body structure -after various agencies that disturb this (Cooke, 1985a,b, 1986). External body form, which in turn correlates well with the relative extent of notochord internally, is also normally balanced in the naturally smallest and largest embryos seen. The probability thus is that positive natural regulation of the relative extents of whole germ layers and of mesoderm parts occurs in embryos of this type.

Table 1.

Partitioning between ectoderm and mesoderm, and allocation of mesoderm to notochord, in naturally large and small embryos at stage 19

Partitioning between ectoderm and mesoderm, and allocation of mesoderm to notochord, in naturally large and small embryos at stage 19
Partitioning between ectoderm and mesoderm, and allocation of mesoderm to notochord, in naturally large and small embryos at stage 19

Fig. 1A shows how an extensive group of blastomeres can be defined, in the 128-cell blastula, as not yet having stably recorded any inductive signals even though many of its descendants will normally be recruited into the mesoderm (Cooke & Webber, 19856; Dale & Slack, 1987a). The descendants of the two animal octets from a typical pattern of 32-cell stage, when dissected at this later stage and cultured alone in saline, form no mesoderm but only ciliated epidermal tissue. If more vegetal material is included in such expiants, evidence of mesoderm formation tends to be seen in the result (Fig. 1B). Accordingly, smaller and larger groups of blastomeres from within the stated uninduced territory have been reciprocally exchanged between sibling embryos as shown in Fig. 1C. Control operations in each set of siblings were constructed by replacing a moderate sized group of the blastomeres with a similar group from a donor. The pigmentation pattern across the animal hemisphere is normally an indicator of presumptive dorsal/ventral polarity. In view of the possibility that dorsal/ventral differences may exist even at these early stages across ectodermal material, graft and surrounding pigment patterns were matched up as well as possible.

All cells from the graft and its surroundings will be allocated to ectoderm or to mesoderm, except for any isolated loose blastomeres which could become incorporated into the endoderm in a way most unlikely to affect the events of pattern formation within the animal cap. Normal cell contact appearance and cellular relationships, as apparent in fix-dissections (see Materials and Methods), are rapidly restored between newly joined regions within blastular animal caps. The vegetal zone of origin of inductive signals is undisturbed mechanically by the operations. The tissue proportions in which the allocation to the-two outer germ layers has been made is measured at stage 19, shown in transverse section in Fig. 1D although material for the cell counts was in horizontal section. This stage is chosen because essentially all mesoderm specified in the primary inductive process has then involuted and migrated as a distinctively behaving, middle cell layer in the body structure, with minimal distortion of the original proportions of germ layer allocations by specific mitotic schedules (Keller, 1976; Cooke, 1979a,b). Secondary complexities whereby mesoderm of the posterior axis may be recruited from the neural plate in the tailbud (Woodland & Jones, 1989) are thus avoided.

The derived estimates for absolute sizes of total ectomesodermal (epidermal+neural+mesodermal) and of mesodermal cell populations, for embryos of two operated sibling sets, are plotted in Fig. 2. The two sets contained 12 and 13 experimental embryos, and 8 control-operated embryos each. ‘Siblings’ are from eggs of similar pigmentation, cleavage pattern and size in one ovulation and in vitro fertilisation. The means and spreads in the two parameters plotted were indistinguishable in the two sets, which are therefore plotted together. The ectodermal-plus-mesodermal measurement represents, in each individual, the size of the field of tissue which the inductive mechanism had partitioned into a deep equatorial zone of mesoderm and a remaining, ectodermally specified part around the animal pole. The relative size of the mesodermal measurement then represents the final proportion according to which that partitioning had occurred. A highly significant trend is seen whereby mesoderm sizes are positively correlated with those of the fields of tissue in which they were induced. Some feedback mechanism, operating to adjust or limit the proportional allocation to mesoderm, is a dominant component in setting its size.

Fig. 2.

The relationship between mesoderm cell population and total population of mesoderm plus ectoderm (epidermis and neural tube) at the later neurula stage. Absolute size of the mesoderm allocation is strongly correlated (r = 0·76, P< 0·001) with size of the field of competent tissue among individuals after the latter has been altered in operations as illustrated in Fig. 1C. Within the experimentally accessible range of sizes, the calculated best-fit regression is shown as a solid line, while the expectation on the basis of a constant-sized mesoderm territory (i.e. determined by local mechanisms only) is shown as a horizontal dotted line. Axes give estimates of absolute cell numbers, derived from counting of nuclei and use of the Abercrombie (1946) correction factor for nuclear diameter and section thickness. Outline circles and diamonds represent experimental embryos from 2 sets, solid discs and diamonds, their control-operated siblings.

Fig. 2.

The relationship between mesoderm cell population and total population of mesoderm plus ectoderm (epidermis and neural tube) at the later neurula stage. Absolute size of the mesoderm allocation is strongly correlated (r = 0·76, P< 0·001) with size of the field of competent tissue among individuals after the latter has been altered in operations as illustrated in Fig. 1C. Within the experimentally accessible range of sizes, the calculated best-fit regression is shown as a solid line, while the expectation on the basis of a constant-sized mesoderm territory (i.e. determined by local mechanisms only) is shown as a horizontal dotted line. Axes give estimates of absolute cell numbers, derived from counting of nuclei and use of the Abercrombie (1946) correction factor for nuclear diameter and section thickness. Outline circles and diamonds represent experimental embryos from 2 sets, solid discs and diamonds, their control-operated siblings.

Fig. 3.

The autonomous state of organisation for axial pattern from the onset of gastrulation. (A) Sectional view of a beginning gastrula, showing how the entire endodermal core (End) and the small amount of already involuted mesoderm (Inv) are cleanly removed with tungsten needles, the remaining shell of tissue allowed to contract for 10 minutes and then the outer continuity restored with a plate of gastrula animal pole ectoderm (below). (B) Sketches of the normal (above) and operated (below) larval appearance. Greater axial extension has occurred, and the forebrain structure is absent or reduced in the operated case, but a high degree of structure is nevertheless apparent, eg = cement gland; oc = optic cup; ev = ear vesicle. (C) Section showing brain structures normally associated with mesodermal pattern at the anteriormost notochord level, but in an endodermless body as in B. These are now bent around the notochord tip. oc = brain region showing retinal pigmentation and associated hypophyseal anatomy adjoining notochord (nc). The other brain region is probably midbrain. (D) Horizontal section at mid-axial level of endodermless body showing massive notochord and perfect somites. (E) The tail, the last region to undergo morphogenesis, in another endodermless specimen. Notochord (nc), tailbud somite (s) and spinal cord are all formed grossly normally, although there is sometimes a loss of coherence in segmentation accompanied by failure of tail expansion in the absence of post-anal endoderm. Scale bar for C,D,E = 100 μm approx.

Fig. 3.

The autonomous state of organisation for axial pattern from the onset of gastrulation. (A) Sectional view of a beginning gastrula, showing how the entire endodermal core (End) and the small amount of already involuted mesoderm (Inv) are cleanly removed with tungsten needles, the remaining shell of tissue allowed to contract for 10 minutes and then the outer continuity restored with a plate of gastrula animal pole ectoderm (below). (B) Sketches of the normal (above) and operated (below) larval appearance. Greater axial extension has occurred, and the forebrain structure is absent or reduced in the operated case, but a high degree of structure is nevertheless apparent, eg = cement gland; oc = optic cup; ev = ear vesicle. (C) Section showing brain structures normally associated with mesodermal pattern at the anteriormost notochord level, but in an endodermless body as in B. These are now bent around the notochord tip. oc = brain region showing retinal pigmentation and associated hypophyseal anatomy adjoining notochord (nc). The other brain region is probably midbrain. (D) Horizontal section at mid-axial level of endodermless body showing massive notochord and perfect somites. (E) The tail, the last region to undergo morphogenesis, in another endodermless specimen. Notochord (nc), tailbud somite (s) and spinal cord are all formed grossly normally, although there is sometimes a loss of coherence in segmentation accompanied by failure of tail expansion in the absence of post-anal endoderm. Scale bar for C,D,E = 100 μm approx.

The control data points, clustered at a middle position on total size as would be expected, show the inherent variability in ectomesodermal partitioning in embryos after the operating procedures. This is little if at all greater than that in unoperated controls from the laboratory collection, as exemplified in Table 1, so we assume the operation per se to have little effect upon this particular aspect of the control of induction. Experimental data points span a twofold range of total mesoectodermal sizes, a variation greatly exceeding what would occur in any one batch of normal sibling embryos.

The situation is probably different as regards the balance of the differentiations that constitute pattern within mesoderm. The only measure of this in the present embryos is relative cell number allocated to the notochord, a small structure already distinctive and scorable in the neurula. As noted above, notochord proportion correlates well with overall balance of mesodermal pattern and is well regulated among naturally different-sized ectomesodermal fields in Xenopus. The operating procedure involved in constructing the present embryos appears to destabilise pattern formation as judged by this criterion (data not shown). Among both the large and small mesoderms, relatively normal and abnormally small notochords are found apparently at random, and only the largest of them constitute what would be normal allocations for their mesoderms (5·5−7%, see Table 1). In addition, many of the experimental mesoderms have less than normal of the convergent extension that precedes somite formation, and a small narrow neural plate. In the large mesoderm of Fig. 1D, for instance, which shows relatively normal cross-sectional appearance, a slender notochord and abundant ventrolateral tissue is nevertheless evident. I return in the discussion to the question of the ‘regulation’ of intramesodermal pattern after blastular microsurgery in Xenopus.

Development without endoderm from onset of gastrulation

Fig. 3 shows the operation whereby the entire yolky endodermal core is removed from embryos, at some time during a period extending 20 minutes either side of the first appearance of the stage 10 indented dorsal blastoporal lip. This operation is clearly easier to perform accurately once the lip appears, but in all cases results in a ‘shell’ of tissue comprising presumptive epidermis and nervous system (ectoderm), some large proportion of the presumptive mesoderm, and the superficial cell layer immediately above the lip position in the dorsal sector known as the suprablastoporal endoderm (presumptive archenteron roof). This latter structure, despite its endodermal nature, is almost certainly specified by induction, whereas the remaining yolky mass is descended from the egg region that is the initial source of induction. Preliminary observations showed that a very complete and normally ordered dorsal axial pattern is obtained, despite the abnormally complete convergence and thus elongation that is mechanically allowed by absence of the yolky endoderm (see Fig. 3B-E). A series of 35 such operations was therefore performed exactly at onset of stage 10 (3 separate egg batches). These were fixed for comparison with control siblings as regards epidermal, neural and mesoderm cell numbers, at stage 19 or as stage 35 larval bodies, or were explored by fix-dissection at the latter stage.

Fig. 4 shows a series of histological appearances in transverse section, and the appearance of the whole body stripped of skin in endodermless and control developments compared. A normal, if contracted, pharyngeal architecture usually exists, presumably derived from the suprablastoporal endoderm with appropriate anteroventral mesoderm. Heart is always absent, and throughout the abnormally long thin trunk region the pattern is completed ventrally by at most a slender tract of lateral plate type mesoderm. Some axes are essentially entirely of notochord and somite in section behind the head region. Tail is of rather normal gross morphology but may show incoherent segmentation. Pronephros is well-formed, and may almost complete the mesoderm mid-ventrally in the absence of heart, lateral plate and liver rudiment. The central nervous system appears massive in the more ventral parts of its cross-section, even though, perhaps as a mechanical effect of the difficulty of inrolling to a tube within the small body cross-section, its roof is often thin and stretched. A frequent abnormal feature of pattern is the absence of territories anterior to the eyes, and the fusion of their paired fields to give a large synophthalmic mass of retina.

Fig. 4.

The larval body after the operation of Fig. 3, in relation to control development. (A-D) Transverse sections through the normal larval body at (grazing) forebrain/eye; hindbrain/ear vesicle, heart and pronephric levels, respectively. (E-H) Sectional levels, equivalent to those shown above them, through a sibling larval body after endoderm excision at stage 10 onset. In relation to the normal pattern, note in E the fused retinal area wrapped around truncated forebrain field, in F the visibly expanded sections of notochord and hindbrain and loss of ventral structures, in G the continuation of previous features and thin expanded neural roof, and complete absence of heart, and in H the restricted strip of mesoderm more lateroventral than the expanded pronephros. (I, J) Midtrunk sectional levels of the same normal and endodermless siblings. Note notochord size and greatly diminished allocation to ventral mesoderm in J. (K, L) Normal and endodermless sibling whole larvae fixed in K2Cr2O7/Acetic acid and stripped of skin. Note absence of gut and truncated anterior neuraxis (with synophthalmia) in L, but the great normality of somite segmentation pattern (cp. Fig. 3D), ev = ear vesicle, fb = forebrain, hb = hindbrain, nc = notochord, s = somite, g = gill mesenchyme (N.B. neural crest and mesoderm-derived), h = heart, pn = pronephros, Ipl = lateroventral mesodermal types. Scale bar for A-J = 100 μm approx.

Fig. 4.

The larval body after the operation of Fig. 3, in relation to control development. (A-D) Transverse sections through the normal larval body at (grazing) forebrain/eye; hindbrain/ear vesicle, heart and pronephric levels, respectively. (E-H) Sectional levels, equivalent to those shown above them, through a sibling larval body after endoderm excision at stage 10 onset. In relation to the normal pattern, note in E the fused retinal area wrapped around truncated forebrain field, in F the visibly expanded sections of notochord and hindbrain and loss of ventral structures, in G the continuation of previous features and thin expanded neural roof, and complete absence of heart, and in H the restricted strip of mesoderm more lateroventral than the expanded pronephros. (I, J) Midtrunk sectional levels of the same normal and endodermless siblings. Note notochord size and greatly diminished allocation to ventral mesoderm in J. (K, L) Normal and endodermless sibling whole larvae fixed in K2Cr2O7/Acetic acid and stripped of skin. Note absence of gut and truncated anterior neuraxis (with synophthalmia) in L, but the great normality of somite segmentation pattern (cp. Fig. 3D), ev = ear vesicle, fb = forebrain, hb = hindbrain, nc = notochord, s = somite, g = gill mesenchyme (N.B. neural crest and mesoderm-derived), h = heart, pn = pronephros, Ipl = lateroventral mesodermal types. Scale bar for A-J = 100 μm approx.

Experimentls is from three sets of operations were fix-dissected, along with synchronous controls, when segmentation had reached the stage with around 20 segmented somites in controls (st. 26/27). Table 2 shows that in addition to the frequently clear organisation of somites (see Fig. 3D), the number segmented, per time, follows the same course in experimental as in control embryos despite the considerably more converged, elongated and apparently more massive paraxial tissue (see below). Final segment counts in endodermless specimens having well segmented tailbuds (Fig. 4L) are not distinguishable from their controls.

Table 2.

Number of somites segmented with time in endodermless embryos and synchronous control siblings

Number of somites segmented with time in endodermless embryos and synchronous control siblings
Number of somites segmented with time in endodermless embryos and synchronous control siblings

Table 3 shows cell number estimates for whole germ layers and their major identifiable territories, notochord, epidermis and neural plate/tube, in examples of endodermless development and their synchronous control siblings at stage 19. The data make it clear that the pattern deficits -as opposed to geometrical changes -in the experimentally altered development are finked with altered allocation of available material to particular territories. The differences between control and experimental allocations cannot be solely due to a variable inclusion of preinvoluted mesoderm at operation. Some of the isolated ectomesoderms possess total cell numbers in the normal range for these structures in whole siblings. Dorsal axial mesoderm pattern, and the nervous system that is thought to be induced by ‘dorsal’ mesoderm are allocated expanded amounts of tissue, at the expense of ventral territories and of epidermis, respectively. Overall allocation as between mesoderm and ectoderm is often remarkably near normal (cp. Table 1 and Fig. 2), confirming that essentially all presumptive mesoderm has been included and suggesting that its induction is in an advanced state of completion by the onset of activity in the first-involuting cells.

Table 3.

Partitioning between ectoderm and mesoderm, and allocation of ectoderm to nervous system and mesoderm to notochord, in endodermless embryos and normal siblings at stage 1

Partitioning between ectoderm and mesoderm, and allocation of ectoderm to nervous system and mesoderm to notochord, in endodermless embryos and normal siblings at stage 1
Partitioning between ectoderm and mesoderm, and allocation of ectoderm to nervous system and mesoderm to notochord, in endodermless embryos and normal siblings at stage 1

Early size regulation

There is little evidence that the process of mesoderm induction in itself alters the early mitotic schedule (Symes & Smith, 1988) or that regional specificity within the mesoderm alters cell cycles on a large scale before neurula stages (Cooke, 1979a,b) when the cycle slows progressively in dorsal axial mesoderm. The data on mesoderm/ectodermal partitioning of Table 1 and Fig. 2 must therefore represent a rather constant, but not large, under-estimate of the proportion of the original animal cap cell descendants respecified by the spread of induction. There is no reason to believe the estimate will itself be of altered accuracy when pattern is formed to a different scale in different individuals. It therefore seems that around 50% of the cleavage products of the region competent to respond to induction are in fact recruited into the primary mesoderm, in a way that is substantially adjusted to tissue size. This adjustment operates both in undisturbed embryos and in those where size of the field of competent tissue is manipulated in relation to a constant sized origin of inducers at the 128-cell stage. Because of this, and in view of in situ lineage-labelling data from 32- and 64-cell stages (Dale & Slack, 1987a; Cooke & Webber, 1985b), it is most unlikely that mesodermal lineages are founded only as a ring, one or two cells wide, directly contacting endodermal cells. Release of inducer appears possible from the first cleavages, and competence to respond, from the 64-cell stage (Jones & Woodland, 1987), but the in situ process appears to take place by diffusion or active propagation of signals across responsive tissue over much of the blastular stage. Cooke et al. (1987) have reported the transmission of induction from competent newly induced cells to further uninduced ones.

Gurdon (1989) has recently proposed that mesoderm extent, as monitored by formation of the muscle phenotype in cultured combinâtes, is largely a function of proximity to source of inducer and the limited spread of signal, i.e. of local mechanisms. The present results show that for mesoderm as a territory in the whole embryo, some more positive principle that sets proportion is in play. How might the spatial extent of induction be controlled in the required way? It should be noted that there may to some extent be an initial ‘pepper and salt’ distribution of mesodermised cells in the blastula wall, with the final boundary established or refined by a local cell sorting phenomenon or indeed a co-operative ‘community’ effect (Gurdon, 1988). The extent to which such a mechanism operates inXenopus, or in other vertebrates where the geometry of mesoderm formation is different, is for future investigation. But the proportioning mechanism probably operates fundamentally by position-dependent specification. There is much indirect evidence that a feedback inhibitor of the response to inducers, as well as inducers or ‘activators’ themselves, are present in the natural system (Smith, 1987; Cooke et al. 1987; Cooke et al. 1989), and that appreciable levels of inducing signal actually reach regions of competent tissue beyond those where a mesoderm response is normally seen. These are the arrangements expected for a reaction-diffusion based ‘morphogen’ gradient system, in which the action of rapidly transmitted inhibition allows regulation of positional information against overall size variation (Gierer & Meinhardt, 1972; Meinhardt, 1982). The control principle proposed is best illustrated by analogy. Suppose a transmissable disease first infects one edge of a lawn or forest (the competent field of ectodermal cells). Infected plants soon infect their immediate neighbours via a local transmission process, and so on. But, at the same time, diseased plants release an agent that can confer resistance, and which diffuses rapidly within the community as a whole (e.g. because air- or insect-borne). Further transitions to the diseased state will stop at a frontier, whose position could reflect the proportion of the population, rather than the absolute number, that are infected.

One view would be that the whole animal region is the field for positional information (Wolpert, 1971), organised from signal origins at the boundary with endoderm, that first delimits the mesoectodermal boundary and later determines the territories within mesoderm by a set of such regulatory mechanisms as that outlined above. Performance of the first step is evidently little affected after the opening up of the blastocoel to simple saline medium, necessary in the present operation, but the later steps controlling intra-mesodermal pattern proportions are caused to perform much more variably. It has been this author’s general observation that embryos in which the blastocoel has been opened and flushed with saline at all extensively at early stages tend to form ‘tail-heavy’ bodies with small or incomplete head parts and notochords, even without further surgery. The real meaning of an apparent ability to adjust notochord size to altered mesoderm size after operations that lay open the Xenopus blastocoel (Cooke, 1981) must be reassessed, since this does not normally occur when the embryos are altered in size by very early manipulations, but ones that should be less physiologically damaging (Cooke & Webber, 1985a,b reviewed Cooke, 1985a, b). Work to this end is in progress, but it already seems likely that for single patterns, the only circumstance ensuring proper proportioning of the body plan in this species is the arrival at a proper balance of extents between zones that will produce ‘dorsal axial’ vs. ‘ventrolateral’ types of inducing activity. The ability further to regulate this balance normally ends before the 2- to 4-cell stage (Cooke, 1985a,b; 1986). The present results (Fig. 2) show that the extent of mesoderm is controlled independently of the balance of pattern within it.

The state of organisation at onset of gastrulation

The observations of Figs 3 and 4 show that, while induction may be a prolonged process during pregastrular stages, an advanced and stable state has been reached at the very onset of gastrulation. This state contains the preconditions that control many aspects of mesoderm organisation, though it is not meaningful to say that the pattern is determined within the tissue at this stage. The sequence of neural induction, probably reflecting craniocaudal sequence of specificities in axial mesoderm, is normal. The absence of preoptic area and fusion of eye fields could reflect omission of a small but crucial group of the first-involuting, presumptively anterior mesoderm at the time of separating the tissues, since such omission cannot be compensated for by regulation (Cooke, 1985c). Alternatively the normal separation of the optic field into two areas may require the broadening of the neural plate, which is somehow dependent on the outer germ layers being stretched around a massive, relatively incompressible endodermal core.

The most striking aspect of endodermless development from stage 10 is the normal segmentation sequence and number, preserved despite the quite altered geometry of gastrulation, and changed shape, cell number and packing of the almost fused columns of paraxial mesoderm. This simple but singular fact argues strongly against model mechanisms for segmentation proposing that somite number is controlled as some indirect consequence of the local mechanical conditions and shape in the segmenting tissue. Instead, without speaking directly to the molecular nature of the machinery involved, it argues for the concept that some physiological variable having to do with rates of development, and possibly with periodicities, is already so organised within the pregastrular mesoderm cell population that the proper number of discrete anatomical units are allocated and express their segmentation behaviour in the correctly timed and arranged sequence. It must increase the belief that, however it is set up, segment number is a profound and not an incidental aspect of regionality in the vertebrate body (Cooke & Zeeman, 1976; Primmett et al. 1989).

The absence of heart, and shift of allocation away from iateroventral’ in favour of dorsal axial structure after stage 10 endoderm excision, reflect a lesser state of preorganisation with respect to mediolateral pattern as compared with dorsal craniocaudal sequence at onset of gastrulation in Xenopus. Jacobson & Sater (1988) have drawn attention to possible differences in timing, as between subclasses of amphibians, for requirements for endoderm in the induction of heart mesoderm. One possible meaning of the present results is that in Xenopus an inductive signal from endoderm, of a regionally specific or else a general ‘permissive’ kind, is required later than stage 10 for stabilisation of heart formation.

Another possible explanation, embracing both the absence of heart and the expanded notochord and paraxial mesodermal territories, is in terms of pattern formation by gradients of morphogen signals within the plane of the mesoderm itself during and after gastrulation. Such interactions have an ‘asymmetrical’ character whereby dorsal tissue emits signals which will dorsalise ventral tissue, but not vice versa (Slack & forman, 1980; Smith & Slack, 1983; Dale & Slack, 1987b). Following the endoderm excision operation, tissue presumably repacks as gastrulation proceeds, as the cross-sectional cellular distance between dorsal and ventral meridians becomes much less than would be allowed in the normal gastrula/neurula, and the tissue has in any case an inbuilt mechanism leading to dorsal convergence. In these circumstances a simple ‘diffusion’ mediated, mediolateral gradation in a signal from a dorsal origin might be expected to ‘flood out’, giving a partial, dorsally overbalanced sequence of signal intensities which will not permit the specification of fully ventral structure. The expanded allocation to nervous system, in the cross-section of the ectodermal layer, would follow from the wider tract of mesoderm now active in the next major inductive episode, neural induction. The results of Table 2, for notochord and for nervous system allocations taken together, make clear that the precise allocation between notochord and somite territories, and that between the whole dorsal axial and more lateral mesodermal domains, is not stabilised until well into gastrulation (see Dale & Slack, 1987b; Smith & Slack, 1983; Cooke, 1983). Heart is relatively ‘dorsally’ derived from the early cleavage fate map, so that its omission from these patterns may not seem consistent with the above hypothesis. It is nevertheless worth noting that XTC-MIF, an inducer of distinctively dorsal mesoderm, can directly or indirectly prevent heart specification in the embryo when injected into the blastocoel of blastulae (Cooke & Smith, 1989).

An operation drastically changing geometry and cell packing in earliest mesoderm can allow striking normality of expression of craniocaudal pattern while substantially disturbing mediolateral pattern. This should strengthen belief that the vertebrate mesodermal body plan is built up at later stages by interacting but different mechanisms of organisation that primarily relate to these two anatomical dimensions, even though many early manipulations on whole embryos produce only co-ordinated effects upon them.

I thank Dennis Summerbell for help with statistics, Wendy Hatton for expert histology, and Marilyn Brerman for typing the manuscript.

Abercrombie
,
M.
(
1946
).
Estimation of nuclear population from microtome sections
.
Anal. Rec
.
94
,
239
247
.
Cooke
,
J.
&
Zeeman
,
E. C.
(
1976
).
A clock and wavefront model for control of the numbers of repeated structures during animal morphogenesis
.
J. theoret. Biol
.
58
,
455
476
.
Cooke
,
J.
(
1979a
).
Cell number in relation to primary pattern formation in the embryo of Xenopus laevis. I. The cell cycle during new pattern formation in response to implanted organisers
.
J. Embryol. exp. Morph
.
51
,
165
182
.
Cooke
,
J.
(
1979b
).
Cell number in relation to primary pattern formation in the embryo of Xenopus laevis. II. Sequential cell recruitment, and control of the cell cycle, during mesoderm formation
.
J. Embryol. exp. Morph
.
53
,
269
289
.
Cooke
,
J.
(
1981
).
Scale of body pattern adjusts to available cell number in amphibian embryos
.
Nature, Lond
.
290
,
775
778
.
Cooke
,
J.
(
1983
).
Evidence for specific feedback signals underlying pattern control during vertebrate embryogenesis
.
J. Embryol. exp. Morph
.
76
,
95
114
.
Cooke
,
J.
(
1985a
).
The system specifying body position in the early development of Xenopus, and its response to perturbations
.
J. Embryol. exp. Morph
.
89
,
Suppl
.
69
87
.
Cooke
,
J.
(
1985b
).
Early specification for body position in mes-endodermal regions of an amphibian embryo
.
Cell Diff
.
17
,
1
12
.
Cooke
,
J.
(
1985c
).
Dynamics of the control of body pattern in the development of Xenopus laevis. III. Timing and pattern after UV-irradiation of the egg and after excision of presumptive head endo-mesoderm
.
J. Embryol. exp. Morph
.
88
,
135
150
.
Cooke
,
J.
(
1986
).
Permanent distortion of positional system of Xenopus embryo by brief early perturbation in gravity
.
Nature, Lond
.
319
,
60
63
.
Cooke
,
J.
&
Smith
,
J. C.
(
1989
).
Gastrulation and larval pattern in Xenopus after blastocoelic injection of a Aenopus-derived inducing factor: experiments testing models for the normal organisation of mesoderm
.
Devi Biol
.
131
,
383
400
.
Cooke
,
J.
,
Smith
,
J. C.
,
Smith
,
Emma J.
&
Yaqoob
,
M.
(
1987
).
The organisation of mesodermal pattern in Xenopus laevis’. experiments using a Xenopus mesoderm-inducing factor
.
Development
101
,
893
908
.
Cooke
,
J.
,
Symes
,
K.
&
Smith
,
Emma J.
(
1989
).
Potentiation by the lithium ion of morphogenetic responses to a Xenopus inducing factor
.
Development
105
,
549
558
.
Cooke
,
J.
&
Webber
,
J. A.
(
1985a
).
Dynamics of the control of body pattern in the development of Xenopus laevis. II. Timing and pattern in the development of single blastomeres (presumptive lateral halves) isolated at the 2-cell stage
.
J. Embryol. exp. Morph
.
88
,
113
134
.
Cooke
,
J.
&
Webber
,
J. A.
(
1985b
).
Dynamics of the control of mesoderm induction in Xenopus laevis’. I. Timing and pattern in the development of dorso-anterior and of posterior blastomere pairs, isolated at the 4-cell stage
.
J. Embryol. exp. Morph
.
88
,
85
112
.
Dale
,
L.
&
Slack
,
J. M. W.
(
1987a
).
Fate map for the 32-cell stage of Xenopus laevis
.
Development
99
,
197
210
.
Dale
,
L.
&
Slack
,
J. M. W.
(
1987b
).
Regional specification within the mesoderm of early embryos of Xenopus laevis
.
Development
100
,
279
295
.
Gierer
,
A.
&
Meinhardt
,
H.
(
1972
).
A theory of biological pattern formation
.
Kybernetik
12
,
30
39
.
Gimlich
,
R. L.
&
Gerhart
,
J.
(
1984
).
Early cellular interactions promote embryonic axis formation in Xenopus laevis
.
Devi Biol
.
104
,
117
130
.
Gurdon
,
J. B.
(
1988
).
A community effect in animal development
.
Nature, Lond
.
336
,
772
774
.
Gurdon
,
J. B.
(
1989
).
The localisation of an inductive response
.
Development
105
,
27
33
.
Gurdon
,
J. B.
,
Fairman
,
S.
,
Mohun
,
T. J.
&
Brennan
,
S.
(
1985
).
Activation of muscle-specific actin genes in Xenopus development by an induction between animal and vegetal cells of a blastula
.
Cell
41
,
913
922
.
Jacobson
,
A.
&
Sater
,
A. K.
(
1988
).
Features of embryonic induction
.
Development
104
,
341
361
.
Jones
,
E. A.
&
Woodland
,
H.
(
1987
).
The development of animal cap cells in Xenopus’. a measure of the start of animal cap competence to form mesoderm
.
Development
101
,
557
564
.
Keller
,
R. E.
(
1976
).
Vital dye mapping of the gastrula and neurula of Xenopus laevis. II. Prospective areas and morphogenetic movements of the deep region
.
Devi Biol
.
51
,
118
137
.
Keller
,
R. E.
&
Danilchik
,
M.
(
1988
).
Regional expression, pattern and timing of convergence and extension during gastrulation of Xenopus laevis
.
Devi Biol
.
103
,
183
210
.
Keller
,
R. E.
,
Danilchik
,
M.
,
Gimlich
,
R. L.
&
Shih
,
J.
(
1985
).
The function and mechanism of convergent extension during gastrulation of Xenopus laevis
.
J. Embryol. exp. Morph
.
89
(
suppl
.),
185
209
.
Kimelman
,
D.
&
Kirschner
,
M.
(
1987
).
Synergistic induction of mesoderm by FGF and TGF/3 and the identification of an mRNA coding for FGF in the early Xenopus embryo
.
Cell
51
,
869
877
.
Meinhardt
,
H.
(
1982
).
Models of Biological Pattern Formation
.
London
.
Academic Press
.
Nieuwkoop
,
P. D.
(
1985
).
Inductive interactions in early amphibian development and their general nature
.
J. Embryol. exp. Morph
.
89
(
Suppl
.),
333
347
.
Nieuwkoop
,
P. D.
&
Faber
,
J.
(
1967
).
Normal Table of Xenopus laevis (Daudin
. 2nd edition.
Amsterdam
:
North Holland Publ. Co
.
Primmett
,
D. R. N.
,
Norris
,
W. E.
,
Carlson
,
G. J.
,
Keynes
,
R. J.
&
Stern
,
C. D.
(
1989
).
Periodic segmental anomalies induced by heat shock in the chick embryo are associated with the cell cycle
.
Development
105
,
119
130
.
Slack
,
J. M. W.
(
1984
).
Regional biosynthetic markers in the early amphibian embryo
.
J. Embryol. exp. Morph
.
80
,
289
319
.
Slack
,
J. M. W.
,
Darlington
,
B. G.
,
Heath
,
J. K.
&
Godsave
,
S. F
: (
1987
).
Mesoderm induction in early Xenopus embryos by heparin binding growth factors
.
Nature, Lond
.
326
,
197
200
.
Slack
,
J. M. W.
&
Forman
,
D.
(
1980
).
An interaction between dorsal and ventral regions of the marginal zone in amphibian embryos
.
J. Embryol. exp. Morph
.
56
,
283
299
.
Smith
,
J. C.
(
1987
).
A mesoderm-inducing factor is produced by a Xenopus cell line
.
Development
99
,
3
14
.
Smith
,
J. C.
(
1989
).
Mesoderm induction and mesoderm inducing factors in early amphibian development
.
Development
105
,
665
677
.
Smith
,
J. C.
,
Dale
,
L.
&
Slack
,
J. M. W.
(
1985
).
Cell lineage labels and region-specific markers in the analysis of inductive interactions
.
J. Embryol. exp. Morph
.
89
Suppl
. p.
317
.
Smith
,
J. C.
&
Slack
,
J. M. W.
(
1983
).
Dorsalization and neural induction, properties of the organiser in Xenopus laevis
.
J. Embryol. exp. Morph
.
78
,
299
317
.
Smith
,
J. C.
,
Yaqoob
,
M.
&
Symes
,
K.
(
1988
).
Purification, partial characterisation and biological effects of the XTC mesoderminducing factor
.
Development
103
,
591
600
.
Symes
,
K.
&
Smith
,
J. C.
(
1988
).
Gastrulation movements provide an early marker of mesoderm induction in Xenopus laevis
.
Devlopment
101
,
339
350
.
Warner
,
A
&
Gurdon
,
J. B.
(
1987
).
Functional gap junctions are not required for muscle gene activation by induction in Xenopus laevis
.
J. Cell Biol
.
104
,
557
564
.
Wolpert
,
L.
(
1971
).
Positional information and pattern formation
.
Current Topics in devl Biol
.
6
,
183
223
.
Woodland
,
H. R.
&
Jones
,
E. A.
(
1989
).
Mesoderm induction in the future tail region of Xenopus
.
Roux Arch, devl Biol
.
197
,
441
446
.