In this article we will discuss about the concept of gradients in the development of embryo.

The Dorsoventral and Anteroposterior Gradients :

The phenomenon of re­gional specificity raises the question of whether the gray crescent does not contain, after all, two or even more different specific substances, responsible for the development of the head and the trunk regions (or the archencephalic, deuterencephalic, and spinocaudal regions), with their separate inductors or organizers. Certain results achieved with abnormal inductors have been adduced in favor of this supposition.

When different tissues of various adult animals are tested as inductors, it is found that they do not all act exactly alike; some of them act preponderantly as head inductors, and others as trunk inductors. Some induce only ectodermal parts (neural structures, sense organs); others also induce mesodermal organs and tissues.

For instance, the liver of the guinea pig, treated with alcohol, acts as an archencephalic inductor; that is, it induces large brain vesicles, bearing a resemblance to the telencephalon, diencephalon, and mesencephalon, sometimes with eyes. It also induces noses and balancers. The kidney of the adder is a rare case of a predominantly deuterencephalic inductor; it induces brain parts resembling the hindbrain and also ear vesicles.


The kidney of the guinea pig treated with alcohol is a spinocaudal inductor; it induces mainly spinal cord, notochord, bands of muscle segments arranged as they are normally in the trunk region of the embryo, and often also complete tails and fin folds around their edges. Lastly, alcohol-treated bone marrow of the guinea pig induces almost exclusively mesodermal parts of the trunk and tail – notochord, rows of somites, nephric tubules, and limb-buds.

The inducing properties of tissues may change depending on the type of treatment. In particular, it was found that heating (or boiling) the tissues reduces their ability to induce mesodermal and spinocaudal structures, whereas archencephalic inductions are still easily obtained. This and similar results have led to the conclusion that the regional nature of inductions produced by adult tissues is the result of the interaction of two factors contained in various proportions in the different tissues.

One factor is the “neuralizing agent” of Toivonen and Saxen, called “dorsalizing agent” by Yamada and “activating agent” by Nieuwkoop et al. When it is present alone it causes archencephalic inductions. It is the active principle of the alcohol-treated liver. The second principle is Toivonen and Saxen’s “mesodermalizing agent,” which is the same as the “caudalizing agent” of Yamada and the “transforming agent” of Nieuwkoop.

If present alone, this second factor induces only mesodermal part –  notochord, muscle, kidney, and limb-bud. This is the active principle of the alcohol-treated guinea pig bone marrow. A deuterencephalic induction is the result of the presence of a small amount of the mesodermalizing factor in addition to the neuralizing factor. A large amount of the mesodermalizing factor added to the neuralizing factor produces a spinocaudal induc­tion.


That deuterencephalic structures are a result of a certain balance between neuraliz­ing and mesodermalizing inducing substances has been shown by the following ingeni­ous experiment. Two pellets, one prepared from guinea pig liver and another prepared from guinea pig bone marrow, were implanted simulta­neously into the blastocoele of a young gastrula.

The liver preparation alone would have induced exclusively archencephalic structures; the bone marrow alone would have induced trunk mesoderm. Together, however, they induced organs belong­ing to all levels, including deuterencephalic and spinocaudal structures –  medulla, ear vesicles, and spinal cord.

A counterpart of the experiment of the Finnish embryologists was performed by the Tiedemann School. By a modification of the method of extraction and purification of inducing substances derived from chick embryos, the investigators obtained a preparation which acted as a deuterencephalic inductor, induc­ing mainly hindbrain and ear vesicles.

The preparation was then separated by chromatography on diethylaminoethyl cellulose into several fractions. Two active frac­tions were obtained, one which produced archencephalic inductions, and another which produced spinocaudal inductions. Deuterencephalic inductions disappeared almost completely.


The two factors are actually substances with different chemical properties. The neuralizing substance is thermostable and soluble in organic solvents. The mesodermalizing substance is insoluble in organic solvents and highly thermolabile.

The result of this latter property is that spinocaudal and deuterencephalic inductors upon heat treatment induce archencephalic structures; in the fresh state they contain both sub­stances, but the mesodermalizing substance is destroyed by heat, whereas the neuraliz­ing substance remains unchanged (at least after a short heat treatment—prolonged heat treatment inactivates the neuralizing substance as well). By graded heat treatment of a tissue its inducing properties may be changed gradually, as the ratio between the mesodermalizing and the neuralizing substances changes in favor of the latter.

It has been a matter of controversy for some time whether the neuralizing agent and the mesodermalizing agent are two distinct chemical substances, two aspects of the action of one substance, or two states of one substance, a more labile state (the mesodermalizing agent) and a more stable state (the neuralizing agent). Experiments have now resolved this problem.

It has been found that the neuralizing substance of the guinea pig liver can be isolated in the form of a ribonucleoprotein, either by sedimentation with streptomycin sulfate or by ultracentrifugation, as a result of which the microsomal fraction, containing the cytoplasmic nucleoprotein, shows the strongest archencephalic inductive action. The mesodermalizing substance of guinea pig bone marrow, on the other hand, is not sedimented by streptomycin sulfate and is not contained in the microsome fraction after ultracentrifugation.

The mesodermalizing substance is therefore a protein which does not tend to be coupled with nucleic acid. The two agents are thus distinctly different and separate substances, one possibly bound to the microsomes, the other not. There is no contradic­tion between the statement that the neuralizing substance can be obtained in the form of a ribonucleoprotein, and our previous finding that the active principle of the inducing substance is a protein and not a nucleic acid –  if the active nucleoprotein prepared from liver is treated with ribonuclease, the inducing action is retained; if it is treated with proteolytic enzymes, the preparation loses its inducing ability.

The normal cellular differentiation of isolated gastrula ectoderm reveals an important feature of the induction phenomenon. The signal coming from the inducing part affects a cellular system endowed with two or more possibilities for further differ­entiation. The induction determines which of the possibilities is to be realized.

In the case of gastrula ectoderm the sequence of events may be presented in the following way:

Gastrula ectoderm + neuralizing factor = neural system

Gastrula ectoderm + mesodermalizing factor = mesoderm


Gastrula ectoderm + no inductor = epidermis

The results obtained with abnormal inductors have been used to interpret normal development. It has been assumed that the same two substances are active as inducing substances of the archenteron roof, and that the regional differentiations in the normal embryo are controlled by a balance between the two substances distributed in the form of gradients.

The gradient of the neuralizing substance is highest mid-dorsally and declines toward the lateral and ventral parts of the embryo. The mesodermalizing substance is most highly concentrated at the posterior dorsal end of the embryo and forms a declining gradient both in the anterior direction and in a lateral direction. This distribution of substances as originally contained in the archenteron roof is transmitted to the overlying ectoderm.

The anterior part of the archenteron roof (including prechdal plate) emits only the neuralizing substance and induces the archencephalon. No notochord and no somites are developed at this level. At a slightly posterior level, the notochord and somites are already present. There is some admixture of the mesodermalizing substance with the neuralizing substance, and as a result deuterencephalic stuctures (medulla and ear vesicles) are induced in the ectoderm.

Still farther back, large amounts of the mesodermalizing substance are available, mesoderm develops into notochord and large masses of muscle, and the neural structure induced is a spinal cord. Each level of the embryo has a different concentration of the neuralizing and mesoder­malizing agents.

The concept of two mutually permeating gradients reminds us at once of the animal and vegetal gradients of the sea urchin egg, though the relationships of the gradients to the main axes of the embryo (and egg) are different.

Gradients, in the form of grading or fading out of certain properties, are obvious in the early development of amphibia and other vertebrates. The dorsoventral gradient (the gradient of the neuralizing agent) can be traced both in the ectodermal and in the mesodermal layers. In the ectoderm the highest level of the gradient is associated with the development of the neural plate and neural tube.

The next highest level of the gradient causes the differentiation of the neural crest. There is ample evidence that the development of the neural crest is the result of a weak induction of the same nature as the induction of neural tissue. When neural plates are induced, the cells on the periphery of the region exposed to the inductor, where the inductive stimulus is fading away, develop as neural crest.

Sometimes, if the inductor is too weak to induce a neural plate, neural crest cells can still be induced. Usually the presence of neural crest cells on the site of induction can be easily noticed, because some of them differentiate as pigment cells. Possibly a still lower level of the gradient is responsible for the differentiation of the various placodes, including nose and ear rudiments, the rudiments of cranial ganglia and of the lateral line sense organ system. 

Where the gradient of the neuralizing agent fades away, ectoderm differentiates as skin epidermis. In the chordomesodermal mantle, the dorsoventral gradient may be held responsible for the segregation of the notochord and the mesoderm in the gastrula of an amphibian. Those parts of the marginal zone in which the specific cytoplasmic substance is most highly concentrated become notochord, and those parts in which the concentration of the substance is lower become mesoderm (the somites).

Lastly, those parts in which the concentration of the cytoplasmic substance is minimal become un-segmented lateral-plate mesoderm. The gradient concept does not contradict our previous statement that the action of the primary organizer may include the induction of somites and other mesodermal parts.

This induction may be due to the presumptive notochord, as the highest level in the gradient system, establishing a new gradient in the surroundings into which it has been transplanted. This may involve a diffusion of the specific substance from higher levels of concentration into the surroundings where the concentration is lower.

Besides the gradient which has its highest level in the dorsomedian strip of tissues diminishing toward both sides, there is a second gradient, with a center of highest concentration at the anterior end of the embryo, which lessens in a posterior direction.

This anteroposterior gradient may be considered to be another manifestation of the caudocranial mesodermalizing gradient; its reverse side, as it was. This gradient is responsible for the differentiation of parts of the nervous system, the most anterior part of the head being the part where the gradient is at its highest.

Control of Development by Influencing the Gradient System:

The gradient unites the parts of a developing embryo into one whole, into one morphogenetic system. Any factor that affects the gradient will therefore affect the whole morphoge­netic system, no matter how simple or even elementary the factor itself may be. It is possible, for instance, to affect a morphogenetic system by depressing the level of the gradient at its highest point.

One way of achieving this is to expose the embryo to certain chemical substances, such as magnesium chloride or lithium chloride. Another way of depressing the high level of the gradient is to remove part of the archenteron roof underlying the anterior end of the nervous system.

The result in both cases is about the same – the most anterior parts of the nervous system fail to develop normally. The defects can best be traced in the structure of the eyes. The injured embryos develop a defect known as cyclopia, that is, the appearance of one median eye instead of two lateral ones.

The median eye is really the result of fusion in the midline of the two eye-forming areas, and every intermediate state may be found in embryos in which the injury has not been very severe. If large parts of the underlying chordomesoderm have been removed, or if the action of the chemical agent has been very strong, the defects in the structure of the eyes are still greater, the single eye being reduced in size and failing to develop altogether in the extreme cases.

The eyes are not the only organs affected in cyclopic embryos; other parts of the head are changed and reduced more or less in correspondence with the defects of the eyes. The parts which may be involved are the brain, the nose, the ear, the mouth, and the gill clefts. The defects spread from the front end backward as the injury is more and more pronounced.

The nose rudiments and the forebrain are the first to be affected, so that the olfactory sacs become unpaired and finally disappear. Then the eyes and the mouth follow, the mouth becoming narrower and eventually disappearing. After the eye and mouth follow the posterior parts of the brain, the gill clefts, and the ear vesicles. The gill clefts fuse in the mid-ventral line, before disappearing one by one. In extreme cases practically the whole head is lacking.

It must be understood that the complete presumptive ectoderm remains in place during the experiments just described; none of it is removed, and neither do any of its parts become necrotic. All cells are there, but their development is changed as a result of an interference with the gradient system.

The experiment with the lithium chloride or magnesium chloride reveals a further important phenomenon – although the whole embryo may be exposed to the chemical substance, it is only the topmost levels of the gradient system that are affected. This is a very common result, namely, that the highest point of a gradient is more easily damaged than the other parts.

In fact, a gradient can often be discovered by exposing the embryo to any mild injurious factor, such as weak poisons, abnormal temperature, or ultraviolet radiation. The effect will be observed first and sometimes only, at the highest point of the gradient if the intensity of the injurious factor has been chosen correctly. In vertebrates in stages following gastrulation, such a sensitive region is invariably the anterior part of the head.

Lithium can also affect the gradient in the chordomesodermal system. Treatment of the embryo during gastrulation with a weak solution of lithium chloride (LiCl) suppresses the development of the notochord, which has been postulated as representing the center of highest activity of the mesodermal gradient If the development of the notochord is suppressed, the presumptive notochordal cells differentiate according to the next high­est level of the gradient and develop into somite tissue. The right and left rows of somites are then continuous with each other across the midline, underneath the neural tube.

It has been reported that the opposite effect may be produced by treating frog embryos in the blastula stage with a solution of sodium thiocyanate (NaSCN). The result is the development of embryos in which the notochord is much larger and thicker than in control animals. This means that a greater than normal area has shown a differentiation characteristic of the highest level of the gradient. This action may be called raising the level of the gradient.

Raising the dorsoventral gradient in the ectoderm should lead to an increase in the size of the neural system. In fact, embryos with excessively broad and massive brains have been observed after sodium thiocyanate treatment in fishes and frogs.

The opposite effect of lithium and thiocyanate ions on amphibian regional differentiation has been confirmed in some experiments performed on explants in vitro. It was found that two-to four-hour exposure of early gastrula ectoderm of Ambystoma to either of these ions causes the ectoderm to differentiate (in the absence of the normal inductor—the chordomesoderm).

This is the same sort of abnormal induction –  induction with methylene blue or with urea. The point of interest in the present context is that the action of NaSCN leads to the development of archencephalic structures, and the influence of LiCl causes the differentiation of notochord, mesoderm, and endoderm. The analogy between the gradient system of amphibians and that of sea urchins is thus greatly enhanced.

Some further experiments serve to show that the gray crescent area of the amphib­ian egg is in fact the center of the gradient responsible for the development of the anterior end of the body. Irradiation of the amphibian egg with ultraviolet light soon after fertilization causes deficiencies in the development of the embryo of the same nature as those caused by lithium—microcephaly, anencephaly—while allowing the embryo to proceed with gastrulation with only a slight retardation.

In badly damaged embryos only the tail remains as the sole representative of the axial organs. It has been found, by careful control of the irradiation, that it is the gray crescent area that is particularly sensitive to the ultraviolet light.

The opposite effect, increase of the anterior end, of the head (macrocephaly), and of the suckers (in frog embryos), may be achieved by injecting into the blastocoele, in the blastula stage, material from the germinal vesicle (the nucleus) of the oocyte (or of an extract of the same).

At the maturation of the oocyte, substances are released from the nucleus into the cytoplasm and contribute to the composition of the cytoplasm of the fertilized egg. It would appear that certain substances from the oocyte nucleus serve to reinforce the particular prop­erties of the gray crescent, as the future “organization center” of the embryo.

The same material from the germinal vesicle injected into the blastocoele of embryos irradiated with ultraviolet light has been reported to counteract the action of the irradiation and partially restore the embryos to normal development.

A similar action, that is, counteracting the damage caused by ultraviolet irradiation, may be achieved by injection into the blastocoele of cytoplasm of untreated fertilized eggs, especially of cytoplasm taken from the gray crescent area. This cytoplasm would presumably contain some of the substances enhancing the organizer action, which had been released from the ruptured germinal vesicle.

Cyclopia is occasionally observed in natural conditions both in man and in domes­tic animals. It is therefore significant that the same effect can be produced experimentally. In this way we get an inkling of what goes on “behind the scenes” when a cyclopic monster is born. There must have been some injury to the gradient system concerned with the formation of the head in the embryo.

The injurious effect may be produced by both hereditary and environmental factors. Cyclopia or similar defects may be the result of some toxicosis of the mother occurring at an early stage of pregnancy. This is known to be the case if a woman contracts the illness known as German measles (rubella) during the early weeks of pregnancy.

The illness is not a serious one for the mother, but the toxin poisons the embryo, producing defects of the anterior part of the head. On the other hand, a gene has been found in guinea pigs which produces, in varying degrees, defects of a cyclopic nature, from the underdevelopment of the nose and mouth to almost complete absence of the head.

The gradient theory has also been applied to the development of insects. As an example of the application of the gradient concept to insect development we will quote some experiments carried out on the embryos of the chironomid midge, Smittia.

Most insect eggs are elongated, and the body of the embryo is formed along the length of the egg on one side, with the ventral surface of the embryo facing outwards. It has been assumed that the subdivision of the embryo rudiment into the various regions of the insect’s body depends on two gradients of “morphogens,” one of the morphogens being responsible for the differentiations of the anterior end of the insect’s body, and the other morphogen—for the differentiation of the posterior regions.

From the anterior end the sequence of the regions is –  A—prosencephalon, the part developing the brain and the sense organs, B—gnathencephalon, the part developing the mouth appendages and the suboesophageal ganglion, C—thorax, with the legs and wings where present, and D and E—the anterior and posterior regions of the abdomen, the posterior region bearing certain distinctive appendages, such as cerci, claspers in males, and other parts involved in copulation, and also, in Dipteran larvae, caudal spiracles.

The boundaries between the regions are supposed to be established depending on certain critical levels in the concentrations of the two morphogens.

As in echinoderm and amphibian embryos, the existence and the morphogenetic activity of the gradients are proved by suppressing a gradient. In the embryos of the midge Smittia, the anterior gradient may be suppressed by subjecting the anterior end of the embryo in the blastoderm stage to ultraviolet radiation.

The result is that the anterior parts of the embryo do not develop at all. Instead, the part of the blastoderm which should have formed the head and thorax now develops into an abdomen, in an inverted orientation, that is with the posterior tip of the abdomen formed where the prosencephalon should have been.

The heterotopic abdomen is thus symmetrical to the abdomen developing from the posterior half of the embryo. The second abdomen is complete except for one part –  it does not possess the genital rudiment. This is as should be expected, as the development of the sex cells is de­pendent on the “sex plasm,” which is present only at the posterior end of the egg.

It is not so easy to explain why the abdomen developing at the anterior end of the embryo is in an inverted position, in mirror image orientation to the abdomen developing from the posterior half of the embryo. This result is contrary to the effect of suppressing the “animal gradient” in a sea urchin embryo, where the result is that the vegetal gradient, and the morphogenetic processes controlled by the gradient, simply extend into the animal hemisphere, leading to the formation of an increased, but single gut.

Assuming that in Smittia the ultraviolet radiation has destroyed the “morphogen” responsible for the anterior gradient, we must assume that by this a posterior gradient of inverse orientation, present but dormant in the anterior end of the embryo, is unmasked. There may perhaps be other explanations.

The inverse orienta­tion of the secondary abdomen is not due to some sort of interaction with the original abdomen, as the embryo may be ligated in the middle, so that no influence from the posterior half can reach the anterior half, and yet, if UV irradiated, the anterior half produces an abdomen in reversed orientation. The value of the case lies in the possibility of getting a further characterization of the nature of the “anterior gradient,” and the “morphogen” which is assumed to be responsible for its existence.

The basic experiment leading to the formation of the “double abdomen” embryos consists of irradiating the anterior end of the embryo with ultraviolet rays. Ultraviolet rays are known to cause damage to nucleic acids, but also to proteins. The damage to nucleic acids is due to formation of linkages between adjacent pyrimidine groups in the nucleic acid chain (pyrimidine dimers).

This is a serious damage of the nucleic acid structure, which prevents its normal functioning. The formation of the pyrimidine dimers is, however, known to be reversible partially by visible light. Embryos of Smittia exposed to UV irradiation, and then treated with visible light produce a lower per­centage of double abdomens, provided that the “healing” visible light irradiation is done before the formation of the body of the embryo.

This is an indication that the action leading to the double abdomen formation goes via damage to a nucleic acid. Furthermore, double abdomens may be produced by injecting quantities of RNase into the anterior end of the embryo. It is thus likely that the “morphogen” responsible for the anterior gradient in the embryo of Smittia is a ribonucleic acid, perhaps some form of mRNA.

This is further confirmed by experiments of centrifugation of early embryos of Smittia. After centrifugation, the contents of the early embryo (before the formation of the blastoderm) are stratified, and each layer may be subjected to a narrow beam of UV light. Double abdomens were produced by irradiating the microsomal fraction in the anterior part of the embryo, but not the other fractions, in particular not the fraction containing nuclei and mitochondria.

The experiments on Smittia present so far the closest approach to the chemical characterization of a “morphogen” causing a gradient in an embryo. Whether all morphogenetic gradients are of the same nature is not evident. The weak point in the experiments described is that the morphogen has not been isolated and purified, and that it has not been possible to introduce (inject) the morphogen into an abnormal position, and thus test its activity.