CHAPTER III

THE THEORY OF EVOLUTION (Continued)

The Evidence from Embryology

THE RECAPITULATION THEORY

At the close of the eighteenth, and more definitely at the beginning of the nineteenth, century a number of naturalists called attention to the remarkable resemblance between the embryos of higher animals and the adult forms of lower animals. This idea was destined to play an important rôle as one of the most convincing proofs of the theory of evolution, and it is interesting to examine, in the first place, the evidence that suggested to these earlier writers the theory that the embryos of the higher forms pass through the adult stages of the lower animals.

The first definite reference[5] to the recapitulation view that I have been able to find is that of Kielmeyer in 1793, which was inspired, he says, by the resemblance of the tadpole of the frog to an adult fish.[6] This suggested that the embryo of higher forms corresponds to the adult stages of lower ones. He adds that man and birds are in their first stages plantlike.

Footnote 5:

The earlier references of a few embryologists are too vague to have any bearing on the subject.

Footnote 6:

Autenrieth in 1797 makes the briefest possible reference to some such principle in speaking of the way in which the nose of the embryo closes.

Oken in 1805 gave the following fantastic account of this relation: “Each animal ‘metamorphoses itself’ through all animal forms. The frog appears first under the form of a mollusk in order to pass from this stage to a higher one. The tadpole stage is a true snail; it has gills which hang free at the sides of the body as is the case in _Unio pictorum_. It has even a byssus, as in Mytilus, in order to cling to the grass. The tail is nothing else than the foot of the snail. The metamorphosis of an insect is a repetition of the whole class, scolopendra, oniscus, julus, spider, crab.”

Walther, in 1808, said: “The human fœtus passes through its metamorphosis in the cavity of the uterus in such a way that it repeats all classes of animals, but, remaining permanently in none, develops more and more into the innate human form. First the embryo has the form of a worm. It reaches the insect stage just before its metamorphosis. The origin of the liver, the appearance of the different secretions, etc., show clearly an advance from the class of the worm into that of the mollusk.”

Meckel first in 1808, again in 1811, and more fully in 1821 made much more definite comparisons between the embryos of higher forms and the adult stages of lower groups. He held that the embryo of higher forms, before reaching its complete development, passes through many stages that correspond to those at which the lower animals appear to be checked through their whole life. In fact the embryos of higher animals, the mammals, and especially man, correspond in the form of their organs, in their number, position, and proportionate size to those of the animals standing below them. The skin is at first, and for a considerable period of embryonic life, soft, smooth, hairless, as in the zoophytes, medusæ, many worms, mollusks, fishes, and even in the lower amphibians. Then comes a period in which it becomes thicker and hairy, when it corresponds to the skin of the higher animals. It should be especially noted here, that the fœtus of the negro is more hairy than that of the European.

The muscular system of the embryo, owing to its lack of union in the ventral wall, corresponds to the muscles of the shelled, headless mollusks, whose mantle is open in the same region. Meckel compares the bones of the higher vertebrates with the simpler bones of the lower forms, and even with the cartilages of the cephalopod. He points out that in the early human embryo the nerve cord extends the whole length of the spinal canal. He compares the simple heart of the embryo with that of worms, and a later stage, when two chambers are present, with that of the gasteropod mollusk. The circulation of the blood in the placenta recalls, he says, the circulation in the skin of the lower animals. The lobulated form of the kidney in the human embryo is compared with the adult condition in the fishes and amphibians. The internal position of the reproductive organs in the higher mammals recalls the permanent position of these organs in the lower animals. The posterior end of the body of the human embryo extends backwards as a tail which later disappears.

Some of these comparisons of Meckel sound very absurd to us nowadays, especially his comparison between the embryos of the higher vertebrates, and the adults of worms, crustaceans, spiders, snails, bivalve mollusks, cephalopods, etc. On the other hand, many of these comparisons are the same as those that are to be found in modern text-books on embryology; and we may do well to ask ourselves whether these may not sound equally absurd a hundred years hence. Why do some of Meckel’s comparisons seem so naïve, while others have a distinctly modern flavor? In a word, can we justify the present belief of some embryologists that the embryos of higher forms repeat the adult stages of lower members of the same group? It is important to observe that up to this time the comparison had always been made between the embryo of the higher form and the adult forms of existing lower animals. The theory of evolution had, so far, had no influence on the interpretation that was later given to this resemblance.

Von Baer opposed the theory of recapitulation that had become current when he wrote in 1828. According to Von Baer, the more nearly related two animals are, or rather the more nearly similar two forms are (since Von Baer did not accept the idea of evolution), the more nearly alike is their development, and so much longer in their development do they follow in the same path. For example two similar species of pigeons will follow the same method of development up to almost the last stage of their formation. The embryos of these two forms will be practically identical until each produces the special characters of its own species. On the other hand two animals belonging to different families of the same phylum will have only the earlier stages in common. Thus, a bird and a mammal will have the first stages similar, or identical, and then diverge, the mammal adding the higher characters of its group. The resemblance is between corresponding embryonic stages and not between the embryo of the mammal and the adult form of a lower group.

Von Baer was also careful to compare embryos of the same phylum with each other, and states explicitly that there are no grounds for comparison between embryos of different groups.[7]

Footnote 7:

In one place Von Baer raises the question whether the egg may not be a form common to all the phyla.

We shall return again to Von Baer’s interpretation and then discuss its value from our present point of view.

Despite the different interpretation that Von Baer gave to this doctrine of resemblance the older view of recapitulation continued to dominate the thoughts of embryologists throughout the whole of the nineteenth century.

Louis Agassiz, in the Lowell Lectures of 1848, proposed for the first time the theory that the embryo of higher forms resembled not so much lower adult animals living at the present time, as those that lived in past times. Since Agassiz himself did not accept the theory of evolution, the interpretation that he gave to the recapitulation theory did not have the importance that it was destined to have when the animals that lived in the past came to be looked upon as the ancestors of existing animals.[8] But with the acceptation of the theory of evolution, which was largely the outcome of the publication of Darwin’s “Origin of Species” in 1859, this new interpretation immediately blossomed forth. In fact, it became almost a part of the new theory to believe that the embryo of higher forms recapitulated the series of ancestral adult forms through which the species had passed. The one addition of any importance to the theory that was added by the Darwinian school was that the history of the past, as exemplified by the embryonic development, is often falsified.

Footnote 8:

Carl Vogt in 1842 suggested that fossil species, in their historical succession, pass through changes similar to those which the embryos of living forms undergo.

Let us return once more to the facts and see which of them are regarded at present as demanding an explanation. These facts are not very numerous and yet sufficiently apparent to attract attention at once when known.

The most interesting case, and the one that has most often attracted attention, is the occurrence of gill-clefts in the embryos of reptiles, birds, and mammals. These appear on each side of the neck in the very early embryo. Each is formed by a vertical pouch, that grows out from the wall of the pharynx until it meets the skin, and, fusing with the latter, the walls of the pouch separate, and a cleft is formed. This vertical cleft, placing the cavity of the pharynx in communication with the outside, is the gill-slit. Similar openings in adult fishes put the pharynx in communication with the exterior, so that water taken through the mouth passes out at the sides of the neck between the gill filaments that border the gill-slits. In this way the blood is aerated. The number of gill-slits that are found in the embryos of different groups of higher vertebrates, and the number that open to the exterior are variable; but the number of gill-openings that are present in the adults of lower vertebrates is also variable. No one who has studied the method of development of the gill-slits in the lower and higher vertebrates will doubt for a moment that some kind of relation must subsist between these structures.

In the lowest adult form of the vertebrates, amphioxus, the gill-system is used largely as a sieve for procuring food, partly also, perhaps, for respiration. In the sharks, bony fishes, and lower amphibians, water is taken in through the mouth, and passes through the gill-slits to the exterior. As it goes through the slits it passes over the gills, that stand like fringes on the sides of the slits. The blood that passes in large quantities through the gills is aerated in this way. In the embryos of the higher vertebrates the gill-slits may appear even before the mouth has opened, but in no case is there a passage of water through the gill-slits, nor is the blood aerated in the gill-region, although it passes through this part on its way from the heart to the dorsal side of the digestive tract. It is quite certain that the gill-system of the embryo performs no respiratory function.[9]

Footnote 9:

This statement is not intended to prejudice the question as to whether the presence of the gill-slits and arches may be essential to the formation of other organs.

In the higher amphibians, the frogs for example, we find an interesting transition. The young embryo, when it emerges from the egg-membranes, bears three pairs of external gills that project from the gill-arches into the surrounding water. Later these are absorbed, and a new system of internal gills, like those of fishes, develops on the gill-arches. These are used throughout the tadpole stage for respiratory purposes. When the tadpole is about to leave the water to become a frog, the internal gills are also absorbed and the gill-clefts close. Lungs then develop which become the permanent organs of respiration.

There are two points to be noticed in this connection. First, the external gills, which are the first to develop, do not seem to correspond to any permanent adult stage of a lower group. Second, the transition from the tadpole to the frog can only be used by way of analogy of what is supposed to have taken place ancestrally in the reptiles, birds, and mammals, since no one will maintain that the frogs represent a group transitional between the amphibians and the higher forms. However, since the salamanders also have gills and gill-slits in the young stages, and lose them when they leave the water to become adult land forms, this group will better serve to illustrate how the gill-system has been lost in the higher forms. Not that in this case either, need we suppose that the forms living to-day represent ancestral, transitional forms, but only that they indicate how such a remarkable change from a gill-breathing form, living in the water, might become transformed into a lung-breathing land form. Such a change is supposed to have taken place when the ancestors of the reptiles and the mammals left the water to take up their abode on the land.

The point to which I wish to draw especial attention in this connection is that in the higher forms the gill-slits appear at a very early stage; in fact, as early in the mammal as in the salamander or the fish, so that if we suppose their appearance in the mammal is a repetition of the adult amphibian stage, then, since this stage appears as early in the development of the mammal as in the amphibians themselves, the conclusion is somewhat paradoxical.

The history of the notochord in the vertebrate series gives an interesting parallel. In amphioxus it is a tough and firm cord that extends from end to end of the body. On each side of it lie the plates of muscles. It appears at a very early stage of development as a fold of the upper wall of the digestive tract. In the cartilaginous fishes the notochord also appears at a very early stage, and also from the dorsal wall of the digestive tract. In later embryonic stages it becomes surrounded by a cartilaginous sheath, or tube, which then segments into blocks, the vertebræ. The notochord becomes partially obliterated as the centra of the vertebræ are formed, but traces of it are present even in adult stages. In the lower amphibians the notochord arises also at an early stage over and perhaps, in part, from the dorsal wall of the digestive tract. It is later almost entirely obliterated by the development of the vertebræ. These vertebræ first appear as a membraneous tube which breaks up into cartilaginous blocks, and these are the structures around and in which the bone develops to form the permanent vertebræ.

In higher forms, reptiles, birds, and mammals, the notochord also appears at the very beginning of the development, but it is not certain that we can call the material out of which it forms the dorsal wall of the archenteron (the amphibians giving, perhaps, intermediate stages). It becomes surrounded by continuous tissue which breaks up into blocks, and these become the bases of the vertebræ. The notochord becomes so nearly obliterated in later stages that only the barest traces of it are left either in the spaces between, or in, the vertebræ.

In this series we see the higher forms passing through stages similar at first to those through which the lower forms pass; and it is especially worthy of note that the embryo mammal begins to produce its notochord at the very beginning of its development, at a stage, in fact, so far as comparison is possible, as early as that at which the notochord of amphioxus develops.

The development of the skull gives a somewhat similar case. The skulls of sharks and skates are entirely cartilaginous and imperfectly enclose the brain. The ganoids have added to the cartilaginous skull certain plates in the dermal layer of the skin. In the higher forms we find the skull composed of two sets of bones, one set developing from the cartilage of the first-formed cranium, and the other having a more superficial origin; the latter are called the membrane bones, and are supposed to correspond to the dermal plates of the ganoids.

In the development of the kidneys, or nephridia, we find, perhaps, another parallel, although, owing to recent discoveries, we must be very cautious in our interpretation. As yet, nothing corresponding to the nephridia of amphioxus has been discovered in the other vertebrates. Our comparison must begin, therefore, higher up in the series. In the sharks and bony fishes the nephridia lie at the anterior end of the body-cavity. In the amphibia there is present in the young tadpole a pair of nephridial organs, the head-kidneys, also in the anterior end of the body-cavity. Later these are replaced by another organ, the permanent mid-kidney, that develops behind the head-kidney. In reptiles, birds, and mammals a third nephridial organ, the hind-kidney, develops later than and posterior to the mid-kidney, and becomes the permanent organ of excretion. Thus in the development of the nephridial system in the higher forms we find the same sequence, more or less, that is found in the series of adult forms mentioned above. The anterior end of the kidney develops first, then the middle part, and then the most posterior. The anterior part disappears in the amphibians, the anterior and the middle parts in the birds and mammals, so that in the latter groups the permanent kidney is the hind-kidney alone.

The formation of the heart is supposed to offer certain parallels. Amphioxus is without a definite heart, but there is a ventral blood vessel beneath the pharynx, which sends blood to the gill-system. This blood vessel corresponds in position to the heart of other vertebrates. In sharks we find a thick-walled muscular tube below the pharynx; the blood enters at its posterior end, flows forward and out at the anterior end into a blood vessel that sends smaller vessels up through the gill-arches to the dorsal side.

In the amphibia the heart is a tube, so twisted on itself that the original posterior end is carried forward to the anterior end, and this part, the auricle, is divided lengthwise by a partition into a right and a left side. In the reptiles the ventricle is also partially separated into two chambers, completely so in the crocodiles. In birds and mammals the auricular and ventricular septa are complete in the adult, and the ventral aorta that carries the blood forward from the heart is completely divided into two vessels, one of which now carries blood to the lungs. When we examine the development of the heart of a mammal, or of a bird, we find something like a parallel series of stages, apparently resembling conditions found in the different groups just described. The heart is, at first, a straight tube, it then bends on itself, and a constriction separates the auricular part from the ventricular, and another the ventricular from the ventral aorta. Vertical longitudinal partitions then arise, one of which separates the auricle into two parts, and another the ventricle into two parts, and a third divides the primitive aorta into two parts. In the early stages all the blood passes from the single ventral aorta through the gill-arches to the dorsal side, and it is only after the appearance of the lung-system that the gill-system is largely obliterated.

We find here, then, a sort of parallel, provided we do not inquire too particularly into details. This comparison may be justified, at least so far that the circulation is at first through the arches and is later partially replaced by the double circulation, the systemic and the pulmonary.

A few other cases may also be added. The proverbial absence of teeth in birds applies only to the adult condition, for, as first shown by Geoffroy Saint-Hilaire, four thickenings, or ridges, develop in the mouth of the embryo; two in the upper, two in the lower, jaw. These ridges appear to correspond to those of reptiles and mammals, from which the teeth develop. It may be said, therefore, that the rudiments of teeth appear in the embryo of the bird. This might be interpreted to mean that the embryo repeats the ancestral reptilian stage, or, perhaps, the ancestral avian stage that had teeth in the beak; but since only the beginnings of teeth appear, and not the fully formed structures, this interpretation would clearly overshoot the mark.

The embryo of the baleen whale has teeth that do not break through the gums and are later absorbed. Since the ancestors of this whale probably had teeth, as have other whales at the present time, the appearance of teeth in the embryo has been interpreted as a repetition of the original condition. Some of the ant-eaters are also toothless, but teeth appear in the embryo and are lost later. In the ruminants that lack teeth in the front part of the upper jaw, _e.g._ the cow and the sheep, teeth develop in the embryo which are subsequently lost.

One interpretation of these facts is that the ancestral adult condition is repeated by the embryo, but as I have pointed out above in the cases of the teeth in whales, since the teeth do not reach the adult form, and do not even break through the gums in some forms, it is obviously stretching a point to claim that an adult condition is repeated. Moreover, in the case of the birds only the dental ridges appear, and it is manifestly absurd to claim in this case that the ancestral adult condition of the reptiles is repeated.

That a supposed ancestral stage may be entirely lost in the embryo of higher forms is beautifully shown in the development of some of the snakes. The snakes are probably derived from lizardlike ancestors, which had four legs, yet in the development the rudiments of legs do not appear, and this is the more surprising since a few snakes have small rudimentary legs. In these, of course, the rudiments of legs must appear in the embryo, but in the legless forms even the beginnings of the legs have been lost, or at any rate very nearly so.

Outside the group of vertebrates there are also many cases that have been interpreted as embryonic repetitions of ancestral stages, but a brief examination will suffice to show that many of these cases are doubtful, and others little less than fanciful. A few illustrations will serve our purpose. The most interesting case is that given by the history of the nauplius theory.

The free-living larva of the lower crustaceans—water-fleas, barnacles, copepods, ostracods—emerges from the egg as a small, flattened oval form with three pairs of appendages. This larva, known as the nauplius, occurs also in some of the higher crustaceans, not often, it is true, as a free form, as in penæus, but as an embryonic stage. The occurrence of this six-legged form throughout the group was interpreted by the propounders of the nauplius theory as evidence sufficient to establish the view that it represented the ancestor of the whole group of Crustacea, which ancestor is, therefore, repeated as an embryonic form. This hypothesis was accepted by a large number of eminent embryologists. The history of the collapse of the theory is instructive.

It had also been found in one of the groups of higher crustaceans, the decapods, containing the crayfish, lobster, and crabs, that another characteristic larval form was repeated in many cases. This larva is known as the zoëa. It has a body made up of a fused head and thorax carrying seven pairs of appendages and of a segmented abdomen of six segments. The same kind of evidence that justified the formulation of the nauplius theory would lead us to infer that the zoëa is the ancestor of the decapods. The later development of the zoëa shows, however, that it cannot be such an ancestral form, for, in order to reach the full number of segments characteristic of the decapods, new segments are intercalated between the cephalothorax and abdomen. In fact, in many zoëas this intercalated region is already in existence in a rudimentary condition, and small appendages may even be present. A study of the comparative anatomy of the crustaceans leaves no grounds for supposing that the decapods with their twenty-one segments have been evolved from a thirteen-segmented form like the zoëa by the intercalation of eight segments in the middle of the body. It follows, if this be admitted, and it is generally admitted now, that the zoëa does not represent an original ancestral form at all, but a highly modified new form, as new, perhaps, as the group of decapods itself. We are forced to conclude, then, that the presence of a larval form throughout an entire group cannot be accepted as evidence that it represents an ancestral stage. We can account for the presence of the zoëa, however, by making a single supposition, namely, that the ancestor from which the group of decapod has evolved had a larva like the zoëa, and that this larval form has been handed down to all of the descendants.

The fate of the zoëa theory cast a shadow over the nauplius theory, since the two rested on the same sort of evidence. The outcome was, in fact, that the nauplius theory was also abandoned, and this was seen to be the more necessary, since a study of the internal anatomy of the lowest group of crustaceans, the phyllopods, showed that they have probably come directly from many segmented, annelidian ancestors. The presence of the nauplius is now generally accounted for by supposing that it was a larval form of the ancestor from which the group of crustaceans arose.

The most extreme, and in many ways the most uncritical, application of the recapitulation theory was that made by Haeckel, more especially his attempt to reduce all the higher animals to an ancestral double-walled sac with an opening at one end,—the gastræa. He dignified the recapitulation theory with an appellation of his own, “The Biogenetic Law.” Haeckel’s fanciful and extreme application of the older recapitulation theory has probably done more to bring the theory into disrepute amongst embryologists than the criticisms of the opponents of the theory.

In one of the recognized masterpieces of embryological literature, His’s “Unsere Körperform,” we find the strongest protest that has yet been made against the Haeckelian pretension that the phylogenetic history is the “cause” of the ontogenetic series. His writes: “In the entire series of forms which a developing organism runs through, each form is the necessary antecedent step of the following. If the embryo is to reach the complicated end-forms, it must pass, step by step, through the simpler ones. Each step of the series is the physiological consequence of the preceding stage and the necessary condition for the following. Jumps, or short cuts, of the developmental process, are unknown in the physiological process of development. If embryonic forms are the inevitable precedents of the mature forms, because the more complicated forms must pass through the simpler ones, we can understand the fact that paleontological forms are so often like the embryonic forms of to-day. The paleontological forms are embryonal, because they have remained at the lower stage of development, and the present embryos must pass also through lower stages in order to reach the higher. But it is by no means necessary for the later, higher forms to pass through embryonal forms because their ancestors have once existed in this condition. To take a special case, suppose in the course of generations a species has increased its length of life gradually from one, two, three years to eighty years. The last animal would have had ancestors that lived for one year, two years, three years, etc., up to eighty years. But who would claim that because the final eighty-year species must pass necessarily through one, two, three years, etc., that it does so because its ancestors lived one year, two years, three years, etc.? The descent theory is correct so far as it maintains that older, simpler forms have been the forefathers of later complicated forms. In this case the resemblance of the older, simpler forms to the embryos of later forms is explained without assuming any law of inheritance whatsoever. The same resemblance between the older and simpler adult forms, and the present embryonic forms would even remain intelligible were there no relation at all between them.”

Interesting and important as is this idea of His, it will not, I think, be considered by most embryologists as giving an adequate explanation of many facts that we now possess. It expresses, no doubt, a part of the truth but not the whole truth.

We come now to a consideration of certain recently ascertained facts that put, as I shall try to show, the whole question of embryonic repetition in a new light.

A minute and accurate study of the early stages of division or cleavage of the egg of annelids has shown a remarkable agreement throughout the group. The work of E. B. Wilson on nereis, and on a number of other forms, as well as the subsequent work of Mead, Child, and Treadwell on other annelids, has shown resemblances in a large number of details, involving some very complicated processes.[10]

Footnote 10:

On the other hand it should not pass unnoticed that Eisigh as shown in one form (in which, however, the eggs are under special conditions being closely packed together) that the usual type of cleavage is altered.

Not only is the same method of cleavage found in most annelids, but the same identical form of division is also present in many of the mollusks, as shown especially by the work of Conklin, Lillie, and Holmes. This resemblance has been discussed at some length by those who have worked out these results in the two groups. The general conclusion reached by them is that the only possible interpretation of the phenomenon is that some sort of genetic connection must exist between the different forms; and while not explicitly stated, yet there is not much doubt that some at least of these authors have had in mind the view that the annelids and mollusks are descended from common ancestors whose eggs segmented as do those of most of the mollusks and annelids of the present day. This conclusion is, I believe, of more far-reaching importance than has been supposed, and may furnish the key that will unlock the whole question of the resemblance of embryos to supposed ancestral forms. It is a most fortunate circumstance that in the case of this cell lineage the facts are of such a kind as to preclude the possibility that the stages in common could ever have been ancestral adult stages. If this be granted then only two interpretations are possible: the results are due either to a coincidence, or to a common embryonic form that is repeated in the embryo of many of the descendants. That the similarity is not due to a coincidence is made probable from the number and the complexities of the cleavage stages.

I believe that we can extend this same interpretation to all other cases of embryonic resemblance. It will explain the occurrence of gill-slits in the embryo of the bird, and the presence of a notochord in the higher forms in exactly the same way as the cleavage stages are explained. But how, it may be asked, can we explain the apparent resemblance between the embryo of the higher form and the adult of lower groups. The answer is that this resemblance is deceptive, and in so far as there is a resemblance it depends on the resemblance of the adult of the lower form to its own embryonic stages with which we can really make a comparison. The gill-slits of the embryo of the chick are to be compared, not with those of the adult fish, but with those of the embryo of the fish. It is a significant fact, in this connection, that the gill-slits appear as early in the embryo of the fish as they do in the bird! The notochord of the embryo bird is comparable with that of the embryo of amphioxus, and not with the persistent notochord in the adult amphioxus. Here also it is of the first importance to find that the notochord appears both in the embryo bird and in amphioxus at the very beginning of the development. The embryo bird is not fishlike except in so far as there are certain organs in the embryo fish that are retained in the adult form. The embryo bird bears the same relation to the embryo fish that the early segmentation stages of the mollusk bear to the early segmentation stages of the annelid. There are certain obvious resemblances between this view and that of Von Baer, but there are also some fundamental differences between the two conceptions.

Von Baer thought that within each group the embryonic development is the same up to a certain point. He supposed that the characters of the group are the first to appear, then those of the order, class, family, genus, and, finally, of the species. He supposed that two similar species would follow the same method of development until the very last stage was reached, when each would then add the final touches that give the individual its specific character. We may call this the theory of embryonic parallelism. Here there is an important difference between my view and that of Von Baer, for I should not expect to find the two embryos of any two species identical at any stage of their development, but at most there might exist a close resemblance between them.

Von Baer’s statement appears to be erroneous from a modern point of view in the following respects. We know that in certain large groups some forms develop in a very different way from that followed by other members of the group, as shown by the cephalopods, for instance, in the group of mollusks. Again, it is entirely arbitrary to assume that the group-characters are the first to appear, and then successively those of the order, family, genus, species. Finally, as has been said above, we do not find the early embryos of a group identical; for with a sufficient knowledge of the development it is always possible to distinguish between the embryos of different species, as well as between the adults, only it is more difficult to do so, because the embryonic forms are simpler. The most fundamental difference between the view of Von Baer and modern views is due to our acceptation of the theory of evolution which seems to make it possible to get a deeper insight into the meaning of the repetition, that carries us far ahead of Von Baer’s position. For with the acceptance of this doctrine we have an interpretation of how it is possible for the embryonic stages of most members of a group to have the same form, although they are not identical. There has been a continuous, although divergent, stream of living material, carrying along with it the substance out of which the similar embryonic forms are made. As the stream of embryonic material divided into different paths it has also changed many of the details, sometimes even all; but nevertheless it has often retained the same general method of development that is associated with its particular composition. We find the likeness, in the sense of similarity of plan, accounted for by the inheritance of the same sort of substance; the differences in the development must be accounted for in some other way.

Among modern writers Hurst alone has advanced a view that is similar in several respects to that which I have here defended. It may be well to give his statement, since it brings out certain points of resemblance with, as well as certain differences from, my own view.[11] He says: “Direct observation has shown that, when an animal species _varies_ (_i.e._ _becomes_ unlike what it was before) in adult structure, those stages in the development which are nearest the adult undergo a similar, but usually smaller, change. This is shown in domestic species by the observations of Darwin, and the result is in exact harmony with the well-known law of Von Baer, which refers to natural species, both nearly related and widely dissimilar. Von Baer’s observations as well as Darwin’s, and as well as those of every student who has ever compared the embryos of two vertebrate species, may be summarized as follows:—

Footnote 11:

Hurst, C. H., “Biological Theories, III,” “The Recapitulation Theory,” _Natural Science_, Vol. ii., 1893.

“Animals which, though related, are very similar in the adult state, resemble each other more closely in early stages of development, often, indeed, so closely as to be indistinguishable in those early stages. As development proceeds _in such species_, the differences between the two embryos compared become more and more pronounced.” On this point, which is an essential one, I cannot agree with Hurst; for I do not think that the facts show that the early stages of two related forms are necessarily more and more alike the farther back we go. The resemblance that is sometimes so striking in the earlier stages is due to the fewer points there are for comparison, and to the less development of the parts then present. Hurst continues: “If similar comparisons could be instituted between the ancestral species and its much modified descendants, there is no reason for doubting that a similar result would be reached. This, indeed, has been done in the case of some breeds of pigeons, which we have excellent reasons for believing to be descended from _Columba livia_. True, _C. livia_ is not a very remote ancestor, but I do not think that will vitiate the argument. Let me quote Darwin verbatim: ‘As we have conclusive evidence that the breeds of the pigeon are descended from a single wild species, I have compared the young within twelve hours after being hatched; I have carefully measured the proportions (but will not here give the details) of the beak, width of mouth, length of nostril, and of eyelid, size of feet, and length of leg in the wild, parent species, in pouters, fantails, runts, barbs, dragons, carriers, and tumblers. Now some of these birds when mature differ in so extraordinary a manner in the length and form of the beak, and in other characters, that they would certainly have been ranked as distinct genera if found in a state of nature. But when the nestling birds of these several breeds were placed in a row, though most of them could just be distinguished, the proportional differences in the above specified points were incomparably less than in the full-grown birds. Some characteristic points of difference—for instance, that of the width of the mouth—could hardly be detected in the young. But there was one remarkable exception to this rule, for the young of the short-faced tumbler differed from the young of the wild-rock pigeon, and of the other breeds in almost exactly the same proportions as in the adult state.’”

Hurst concludes that: “The more the adult structure comes to be unlike the adult structure of the ancestors, the more do the late stages of development undergo a modification of the same kind. This is not mere dogma, but it is a simple paraphrase of Von Baer’s law. It is proved true not only by the observations of Von Baer and of Darwin, already referred to, but by the direct observation of every one who takes the trouble to compare the embryos of any two vertebrates, provided only he will be content to see what actually lies before him and not the phantasms which the recapitulation theory may have printed on his imagination.”

The growth of the antlers of stags is cited by Hurst in order to illustrate that what has been interpreted as a recapitulation may have a different interpretation. “Each stag develops a new pair of antlers in each successive year, and each pair of antlers is larger than the pair produced in the previous year. This yearly increase in the size of the antlers has been put forward as an example of an ontogenetic record of past evolution. I, however, deny that it is such a record.”

“The series of ancestors may have possessed larger antlers in each generation than in the generation before it. It is not an occasional accidental parallelism between the ontogeny and the phylogeny which I deny, but the causal relation between the two. Had the ancestors had larger antlers than the existing ones, there is no justification for the assumption that existing stags would acquire antlers of which each pair, in later years, would be smaller than those of the previous year.”

Hurst concludes: “There are many breeds of hornless sheep, but they do not bear large horns in early years and then shed them. If a rudiment ever appears in the embryo of such sheep, its growth is very early arrested.” The case of the appendix in man might have been cited here as a case in point. It is supposed to have been larger in the ancestors of man, but we do not find it appearing full size in the embryo and later becoming rudimentary. The preceding statements will show that, while Hurst’s view is similar in some respects to my own, yet it differs in one fundamental respect from it, and in this regard he approaches more nearly to the theory of Von Baer.

Hertwig has recently raised some new points of issue in regard to the recapitulation theory, and since he may appear to have penetrated farther than most other embryologists of the present time, it will be necessary to examine his view somewhat carefully. He speaks of the germ-cell (egg, or spermatozoön) as a species-cell, because it contains, in its finer organization, the essential features of the species to which it belongs. There are as many of these kinds of cells as there are different kinds of animals and plants. Since the bodies of the higher animals have developed from these species-cells, so the latter must have passed in their phylogeny through a corresponding development from a simple to a more and more complex cell-structure. “Our doctrine is, that the species-cell, even as the adult, many-celled representative of the species, has passed through a progressive, and, indeed, in general a corresponding development in the course of phylogeny. This view appears to stand in contradiction to the biogenetic law. According to the formula that Haeckel has maintained, the germ development is an epitome of the genealogy; or the ontogeny is a recapitulation of the phylogeny; or, more fully, the series of forms through which the individual organism passes during its development from the egg-cell to the finished condition is a short, compressed repetition of the longer series of forms which the forefathers of the same organism, or the stem-form of the species, has passed through, from the earliest appearance of organisms to the present time.” “Haeckel admits that the parallel may be obliterated, since much may be absent in the ontogeny that formerly existed in the phylogeny. If the ontogeny were complete, we could trace the whole ancestry.” Hertwig states further, that “The theory of biogenesis[12] makes it necessary to change Haeckel’s expression of the biogenetic law, so that a contradiction contained in it may be removed. We must drop the expression ‘repetition of the form of extinct forefathers,’ and put in its place the repetition of forms which are necessary for organic development, and lead from the simple to the complex. This conception may be illustrated by the egg-cell.”

Footnote 12:

This term, by which Hertwig designates a particular view of his own, has been already preoccupied in a much wider sense by Huxley to mean that all life comes from preëxisting life. Hertwig means by the theory of biogenesis that as the egg develops there is a constant interchange between itself and its surroundings.

Since each organism begins its life as an egg we must not suppose that the primitive conditions of the time, when only single-celled amœbas existed on our planet, are repeated. The egg-cell of a living mammal is not, according to Hertwig’s hypothesis, an indifferent structure without much specialization like an amœba, but is an extraordinarily complex end-product of a long historical process, which the organized substance has passed through. If the egg of a mammal is different from that of a reptile, or of an amphibian, because in its organization it contains the basis of a mammal, just so much more must it be different from the hypothetical one-celled amœba, which has no other characteristics than those that go to make up an amœba. Expressed more generally, the developmental process in the many-celled organisms begins, not where it began in primitive times, but as the representation of the highest point which the organization has at present reached. The development commences with the egg, because it is the elemental and fundamental form in which organic life is represented in connection with the reproductive process, and also because it contains in itself the properties of the species in its primordia.

“The egg-cell of the present time, and its one-celled predecessor in the phylogenetic history, the amœba, are only comparable in so far as they fall under the common definition of the cell, but beyond this they are extraordinarily different from each other.”

“The phyletic series must be divided into two different kinds of processes:—First. The evolution of the species-cell, which is a steady advance from a simple to a complex organization. Second. The periodically repeated development of the many-celled individual out of the single cell, representative of the species (or the individual ontogeny), which in general follows the same rules as the preceding ontogeny, but is each time somewhat modified according to the amount to which the species-cell has itself been changed in the phylogeny. Similar restricting and explanatory additions to the biogenetic law, like those stated here for the one-celled stage, must be made in other directions. Undoubtedly there exists in a certain sense a parallel between the phylogenetic, and the ontogenetic, development.

“On the basis of the general developmental hypothesis on which we stand, all forms which in the chain of ancestors were end-products of the individual development are now passed through by their descendants as embryonic stages, and so in a certain degree are recapitulated. We also admit that the embryonic forms of higher animals have many points of comparison with the mature forms of related groups standing lower in the system.

“Nevertheless, a deeper insight into the conditions relating to these resemblances shows that there are very important differences that should not be overlooked. Three points need to be mentioned: 1. The cell-material which in the ancestral chain gives the basis for each ontogenetic process is each time a different material as far as concerns its finer organization and primordia. Indeed, the differences become greater the farther apart the links of the original chain become. This thought may be formulated in another way: The same ontogenetic stages that repeat themselves periodically in the course of the phylogeny always contain at bottom a somewhat different cell-material. From this the second rule follows as a consequence. 2. Between the mature end-form of an ancestor and the corresponding embryonic form of a widely remote descendant (let us say between the phylogenetic gastræa and the embryonic gastrula stage of a living mammal, according to the terminology of Haeckel) there exists an important difference, namely, that the latter is supplied with numerous primordia which are absent in the other, and which force it to proceed to the realization of its developmental process. The gastrula, therefore, as the bearer of important latent forces, is an entirely different thing from the gastræa, which has already reached the goal of its development. 3. In the third place, at each stage of the ontogeny outer and inner factors are at work, in fact even more intensely than in the fully formed organism. Each smallest change that acts anew in this way at the beginning of the ontogeny can start an impulse leading to more extensive changes in later stages. Thus the presence of yolk and its method of distribution in the egg alone suffice to bring about important changes in the cleavage, and in the formation of the germ-layers, the blastula, and gastrula stages,” etc. “Moreover, the embryo may adapt itself to special conditions of embryonic life, and produce organs of an ephemeral nature like the amnion, chorion, and placenta.”

“A comparison of ontogenetic with antecedent phylogenetic stages must always keep in view the fact that the action of external and internal factors has brought about considerable changes in the ontogenetic system, and, indeed, in a generally advancing direction, so that in reality a later condition can never correspond to a preceding one.”

Hertwig sums up his conclusion in the statement that ontogenetic stages give us, therefore, a greatly changed picture of the phylogenetic series of adult ancestors. “The two correspond not according to their actual contents but only as to their form.” Hertwig also repeats His’s idea, that the reason that certain kinds of form repeat themselves in the development of animals with a great constancy depends principally on this, that they supply the necessary conditions under which alone the following higher stage of the ontogeny can be formed. The development, for instance, begins with the division of the egg, because this is the only way that a one-celled condition can give rise to a many-celled form. Again, the organs can be formed only when groups of cells have made a closer union with one another. Thus the gastrula must begin with the antecedent blastula, etc. Definite forms are, despite all modifying influences, held to firmly, because by their presence the complicated end-stages can be reached in the simplest and most suitable way.

Thus Hertwig adopts here a little from one doctrine and there a little from another, and between his attempt to reinstate the old biogenetic law of Haeckel, and to adopt a more modern point of view, he brings together a rather curious collection of statements which are not any too well coördinated. Take, for example, his description of the relation between Haeckel’s gastræa and the embryonic gastrula stage. The latter he maintains is a repetition of the other, but only in form, not in actual contents. And in another connection we are told that the cause of this repetition is that the gastrula is the simplest way in which the later stages can be reached, and, therefore, it has been retained. It seems to me that Hertwig has undertaken an unnecessary and impossible task when he attempts to adjust the old recapitulation theory to more modern standards. His statement that the egg is entirely different from its amœba prototype is, of course, only the view generally held by all embryologists. His mystical statement that the embryonic form _repeats the ancestral adult stage in its form, but not in its contents_, will scarcely recommend itself as a model of clear thinking. Can we be asked to believe for instance that a young chick repeats the ancestral adult fish form but not the contents of the fish?

In conclusion, then, it seems to me that _the idea that adult ancestral stages have been pushed back into the embryo, and that the embryo recapitulates in part these ancestral adult stages is in principle false_. The resemblance between the embryos of higher forms and the adults of lower forms is due, as I have tried to show, to the presence in the embryos of the lower groups of certain organs that remain in the adult forms of this group. It is only the embryonic stages of the two groups that we are justified in comparing; and their resemblances are explained on the assumption that there has been an ancestral adult form having these embryonic stages in its development and these stages have been handed down to the divergent lines of its descendants.

Since we have come to associate with the name of the recapitulation theory the idea of the recurrence of an ancestral adult form, it may be better to find a substitute for this term. I suggest, therefore, for the view, that the embryos of the higher group repeat the modified form of the embryos of the lower groups, the term, the theory of embryonic repetition, or, more briefly, the repetition theory.

Conclusions

In the light of the preceding discussion concerning the evidence in favor of the transmutation theory, we may now proceed to sum up our general conclusions, and at the same time discuss some further possibilities in regard to the descent theory.

The most widely accepted view in regard to the theory of organic evolution is that which looks upon the resemblances between the members of a group as due to their common descent from one original species that has broken up, as it were, into a number of new forms. Strictly applied, this means that all the vertebrates have come from one original species, all the mollusks from another, the echinoderms from a third, etc. Even farther back there may have been a common ancestral species for any two of the large groups, as, for example, the annelids and the mollusks; and if the relationship of all the many-celled forms be looked upon as probable, then they too have originated from one ancestral species.

Many zoologists appear to hesitate to apply strictly this fundamental idea contained in the transmutation theory, because, perhaps, they feel that it does not fit in with their general experience of living forms. Yet there can be no doubt that it is the primary conception of the transmutation theory. This is, however, not the whole question, for we must further consider the number of individuals of a species that are involved.

In some species there are smaller groups of individuals that are more like one another than like other individuals of the same species. Such groups are called varieties, and are often associated with certain localities, or with a special environment. In the latter case they are called local varieties. Some of these appear to breed true, not only when kept under the same conditions, but even when transferred to a new environment. Others change with the environment. It is not improbable that the varieties are of a different kind in these two cases, as shown by their different behavior when put under new and different surroundings. The variety that owes its peculiarities, not to the immediate environment, but to some internal condition independent of the surroundings, is recognized by some biologists as a smaller species. Such species appear to be commoner in plants than in animals, although it is possible that this only means that more cases have been found by the botanists, owing to the greater ease with which plants can be handled. These smaller species, in contradistinction to the ordinary Linnæan species, differ from the latter in the smaller amount of differences between the groups, and probably also in that they freely interbreed, and leave fertile descendants; but whether this is only on account of the smaller differences between them than between larger species, or because of some more fundamental difference in the kind of variation that gives rise to these two kinds of groups, we do not know.

These smaller species, or constant varieties, as we may call them, may be looked upon as incipient Linnæan species, which, by further variations of the same, or of other sorts, may end by giving rise to true species. A genus composed of several species might be formed in this way, and then, if each species again broke up into a number of new groups, each such group would now be recognized as a genus, and the group of genera would form a family, etc. The process continuing, a whole class, or order, or even phylum, might be the result of this process that began in a single species.

But we must look still farther, and inquire whether the start was made from a single individual, that began to vary, or from a number of individuals, or even from all the individuals, of a species. If we suppose the result to depend on some external cause that affects all the individuals of a species alike, then it might appear that the species, or at least as many individuals of a species as are affected, will give the starting-point for the new group. But if the new variation arises not directly as a response to some change in the surroundings, then it might appear in one or in a few individuals at a time. Let us consider what the results might be under these two heads.

If amongst the descendants of a single individual a new form or a number of new forms were to arise, then, if they represented only a variety, they would cross with the other forms like the parent species; and, under these conditions, it is generally assumed that the new variety would be swamped. If, however, the new forms have the value of new species, then, _ex hypothese_, they are no longer fertile with the original forms, and might perpetuate themselves by self-fertilization, as would be possible in some of the higher plants, and in those animals that are bisexual. But as a rule even bisexual forms are not self-fertilized, and, therefore, unless a number of offspring arose from the same form the chance of propagation would be small.

If, however, a number of new forms appeared at the same time and left a number of descendants, then the probability that the new group might perpetuate itself is greater, and the chance that such a group would arise is in proportion to the number of individuals that varied in the same direction simultaneously. In this case the new species has not come from a single individual or even from a pair of individuals, but from a number of individuals that have varied more or less in the same direction.

This point of view puts the descent theory in a somewhat unforeseen light, for we cannot assume in such a case that the similarities of the members of even the same species are due to direct descent from an original ancestor, because there are supposed to have been a number of ancestors that have all changed in the same direction. The question is further complicated by the fact that the new individuals begin to interbreed, so that their descendants come to have, after a time, the common blood, so to speak, of all the new forms. If with each union there is a blending of the substances of the individuals, there will result in the end a common substance representing the commingled racial germ-plasm.

A new starting-point is then reached, and new species may continue to be formed out of this homogeneous material. Thus, in a sense, we have reached a position which, although it appears at first quite different from the ordinary view, yet, after all, gives us the same standpoint as that assumed by the transmutation theory; for, while the latter assumes that the resemblances of the members of a group are due to descent from the same original form, and often by implication from a single individual, we have here reached the conclusion that it is only a common, commingled germ-plasm that is the common inheritance.

When we examine almost any group of living animals or plants, whether they are low or high in organization, we find that it is composed of a great many different species, and so far as geology gives any answer, we find that this must have been true in the past also. Why, then, do we suppose that all the members of the higher groups have come from a single original species or variety? Why may not all, or many, of the similar species of the lower group have changed into the species of the higher group,—species for species? If this happened, the resemblance of the new species of the group could be accounted for on the supposition that their ancestors were also like one another. The likeness would not be due, then, to a common descent, and it would be false to attempt to explain their likeness as due to a common inheritance. But before going farther, it may be well to inquire to what the resemblances of the individuals of the original species were due; for, if they have come from an older group that has given rise to divergent lines of descent, then we are only removing the explanation one step farther back. If this original group has come from numerous species of a still older group, and this, in turn, from an older one still, then we must go back to the first forms of life that appeared on the globe, and suppose that the individuals of these primitive forms are the originals of the species that we find living to-day. For instance, it is thinkable that each species of vertebrate arose from a single group of the earliest forms of life that appeared on the surface of the earth. If this were the case, there must have been as many different kinds of species of the original group as there are species alive at the present time, and throughout all the past. This view finds no support from our knowledge of fossil remains, and, although it may be admitted that this knowledge is very incomplete, yet, if the process of evolution had taken place as sketched out above, we should expect, at least, to have found some traces of it amongst fossil forms. Since this question is an historical one, we can, at best, only expect to decide which of all the possible suggestions is the more probable.

We conclude, then, that it is more probable that the vertebrates, the mollusks, the insects, the crustaceans, the annelids, the cœlenterates, and the sponges, etc., have come each from a single original species. Their resemblances are due to a common inheritance from a common ancestral species. Even if it be probable that at the time when the group of vertebrates arose from a single species, there were in existence other closely related species, yet we must suppose, if we adhere to our point of view, that these other related species have had nothing to do with the group of vertebrates, but that they have died out. Moreover, we must suppose that each order, each class of vertebrate, has come from a single original species; each family has had a similar origin, as well as each genus, but, of course, at different periods of time. Let us not shrink from carrying this principle to its most extreme point, for, unless the principle is absolutely true, then our much boasted explanation of the resemblances of forms in the same group will be thrown into hopeless confusion.

Let us ask another question in this connection. If a single species gave rise to a group of new species that represented the first vertebrates, they would have formed the first genus; and if the descendants of these diverged again so that new genera were formed, then a group which we should call a family would have been formed.

As the divergence went on, an _order_ would be developed, and then a _class_, and then a _phylum_. The common characters possessed by the members of this phylum would have been present in the original species that began to diverge. Hence, we find the definition of the phylum containing only those points that are the features possessed by all of the descendants, and in the same way we should try to construct the definition of each of the subordinate groups. This is the ideal of the principle of classification based on the theory of descent with divergence. If we admit the possibility of the other view that I have mentioned above, or of any other of the numerous possibilities that will readily suggest themselves, then we must be prepared to give up some of the most attractive features of the explanation of resemblance as due to descent.

That all biologists believe strictly in divergent descent, to the exclusion of any other processes, is not the case. And, as I have said before, since we are dealing with an historical question, it would be very unwise, in our present ignorance on many points, to pretend that we have any direct proof of the explanation that we find generally given to account for the resemblances of the species of a group to each other. At most we can claim that it is the simplest point of view, and that most biologists believe it to be also the most probable. It has been suggested that, in some cases, the new forms that arise from two or more species run a parallel course. If the original forms from which they came were very much alike, it would soon be impossible to say what the parentage of a particular form was; that is, to which of the two original forms it belonged. It has also been suggested that even a convergence has at times taken place, so that the descendants of different species have become more alike than the original forms, _at least in some one or more respects_. This last limitation is the saving clause, for species differ in so many points that, even when they converge in a few, it is unlikely that they will do so in all, and, therefore, the deception may be discovered by the acute observer. One famous paleontologist has gone so far even as to suppose that a species may change its generic characters, so that it goes over bodily into a new genus without losing its specific characters. If such things do occur, then our classifications may well be the laughing-stock of Nature.