On June 10 I still had four tadpoles which had never had thyroid, but only pieces of meat, earthworm, or fish. These were very much larger than any of the others, were active and vigorous, and the largest one showed small rudiments of hind-legs, the others none at all.

CHAPTER VII

Metamorphosis And Recapitulation

As one of the most remarkable examples of metamorphosis and recapitulation in connexion with adaptation we will consider once more the case of the Flat-fishes which I have already mentioned in an earlier chapter. These fishes offer perhaps the best example of the difference between gametogenic mutations and adaptive modifications. In several species specimens occur occasionally in which the asymmetry is not fully developed. [Footnote: See 'Coloration of Skins of Fishes, especially of Pleuronectidae,' _Phil. Trans. Royal Soc_., 1894.] These abnormalities are most frequent in the Turbot, Brill, Flounder, and Plaice. The chief abnormal features are pigmentation of the lower side as well as of the upper, the eye of the lower side, left or right according to the species, on the edge of the head instead of the upper side, and the dorsal fin with its attachment ceasing behind this eye, the end of the fin projecting freely forwards over the eye in the form of a hook. Such specimens have been called ambicolorate, but it is an important fact that they are also ambiarmate--that is to say, the scales or tubercles which in the normal Flat-fish are considerably reduced or absent on the lower side, in these abnormal specimens are developed on the lower side almost as much as on the tipper. Minor degrees of the abnormality occur: in Turbot with the hook-like projection of the dorsal fin the lower side of the head is often without pigment, while the rest of the lower side is pigmented. Less degrees of pigmentation of the lower side occur without structural abnormality of the eye and dorsal fin.

There is no evidence that these abnormalities are due to abnormal conditions of life. One specimen of Plaice of this type was kept alive in the aquarium, and it lay on its side, buried itself in the sand, and when disturbed swam horizontally, like a normal specimen. The abnormalities are undoubtedly mutations of gametic origin. The development of one of these abnormal specimens from the egg has not to my knowledge been traced, but there is no reason to suppose that the fish develops first into the normal asymmetrical condition and then changes gradually to the abnormal condition described. On the contrary, everything points to the conclusion that the abnormality is an arrest or incomplete occurrence of the normal process of development, _i.e._ of the normal metamorphosis. T. H. Morgan, in a volume published some years ago, [Footnote: _Evolution and Adaptation_.] put forward the extraordinary view that the Pleuronectidae arose from symmetrical fishes by a mutation which was entirely gametogenetic and entirely independent of habits or external conditions, and then finding itself with two eyes on one side of its head, and no air-bladder, adopted the new mode of life, the new habit of lying on the ground on one side in order to make better use of its asymmetrically placed eyes. According to this view habits have been adapted to structure, not structure to habits.

We are thus to believe that Amphibia came out of the water and breathed air because by an accidental mutation they possessed lungs and a pulmonary circulation capable of atmospheric respiration. Such is the result of applying conclusions derived from phenomena of one kind to phenomena of a totally different kind. One of the chief differences between structural features and correlations which are adaptive from those which are not is the process of metamorphosis, where we see the structure changing in the individual life history as the mode of life changes. The egg of the Flat-fish develops into a symmetrical pelagic larva similar to that of many other marine fishes. The larva has an eye on each side of its head and swims with its plane of symmetry in a vertical position: it has also colour on both sides equally. When the skeleton begins to develop the transformation takes place: the eye of one side, left in some species, right in others, moves gradually to the edge of the head and then on to the other side. The dorsal fin extends forward, preserving its original direction, and so pa.s.ses between the eye that has changed its position and the lower side of the fish, on which that eye was originally situated. In some cases this extension of the fin takes place earlier and the eye pa.s.ses beneath the base of the fin to reach the other side. Any one who takes the trouble to make himself acquainted with the facts will see that the three chief features of the Pleuronectid--namely, the position of the eyes, the extension of the dorsal and ventral fins, and the absence of pigment from the lower side--are not structurally correlated with one another at all as changes in different parts of the organism in a mutation are said to be, but are all closely related to their functions in the new position of the body. A mutation consisting in general asymmetry would be comprehensible, but the head of the Pleuronectid is not asymmetrical in a general sense, but only so far as to allow of the changed position of the eyes. The posterior end of the skull is as symmetrical as in any other fish, and in some cases the mouth and jaws are also symmetrical, entirely unaffected by the change in the position of the eyes. In other cases the jaws are asymmetrical in a direction opposite to that of the eyes, there is no change of position but a much greater development of the lower half of the jaws, reduction, with absence of teeth, of the upper half. In the latter case the fish feeds on worms and molluscs living on the ground and seized with the lower half of the jaws, in the former the food consists of small fish swimming above the Flat-fish and seized with the whole of the jaws (Turbot, Halibut, etc.).

I contend, then, that the mode in which the normal Flat-fish develops is quite different from that in which mutations arise. T. H. Morgan [Footnote: _A Critique of the Theory of Evolution_ (1916), p. 18.] states that a variation arising in the germ-plasm, no matter what its cause, may affect any stage in the development of the next individuals that arise from it. In certain cases this is true, that is to say, when there are very distinct stages already. For example, a green caterpillar becomes a white b.u.t.terfly with black spots. A mutation might affect the black spots, an individual might be produced which had two spots on each wing instead of one, and no sign of this mutation would be evident in the caterpillar.

But my contention is that when this mutation occurred, the original condition of one spot would not be first developed and then gradually split into two. Morgan proceeds to state clearly what I wish to insist upon concerning mutations. He writes that in recent times the idea that variations are discontinuous has become current. Actual experience, he tells us, shows that new characters do not add themselves to the line of existing characters, but if they affect the adult characters, they change them without as it were pa.s.sing through and beyond them.

Now in the case of the ancestors of the Flat-fish the adult and the larva must have had the same symmetry with regard to eyes and colour and the dorsal fin terminated behind the level of the eyes. Thus the variations which gave rise to the Flat-fish were not discontinuous but continuous. In each individual development now, not merely hypothetically in the ancestor, the condition of the adult arises by an absolutely continuous change of the eyes, fins, and colour. Such a continuous change cannot be explained by a discontinuous variation, _i.e._ a mutation. The abnormalities above mentioned on the other hand, although they doubtless arise from the same kind of symmetrical larva as the normal Flat-fish, and develop by a gradual and continuous process, do not presumably pa.s.s through the condition of the normal adult Flat-fish and then change gradually into the condition we find in them. As compared with the normal Flat-fish they arise by a discontinuous variation, they are mutations, whereas the normal Flat-fish as compared with its symmetrical ancestor arises by a continuous change.

In order to make my meaning clear I must point out that I have been using the word continuous in a different sense from that in which it is used by other biologists, Bateson for example. The word has been applied previously to variations which form a continuous series in a large number of individuals, each of which differs only slightly from those most similar to it. No two individuals are exactly alike, and thus such continuous variations are universal. According to the theory of natural selection the course of evolutionary change in any organ or character would form a similar continuous series, the mean of each generation differing only by a small difference from that of the preceding. According to the modern mutationists such small differences are to be called fluctuations, and have no effect on evolution at all, are not even hereditary, are not due to genetic factors in the gametes. Discontinuous variations, on the other hand, are as a rule differences in an individual from the normal type and from its parents of considerable degree, and are conspicuous: these are what are called mutations.

The mutationists and Mendelians have not shown how the essential characteristics of mutations are to be reconciled with the facts of metamorphosis, or with recapitulation in development which is so often a.s.sociated with metamorphosis. T. H. Morgan is the only mutationist, so far as my reading has gone, who has attempted to do this, and he seems to me to have failed to understand the difficulties or even the nature of the problem. He points out that the embryos of Birds and Mammals have gill slits representing the same structures as those of the adult Fish, but the young stage of the Fish also possessed gill slits, therefore it is 'more probable that the Mammal and Bird possess this stage in their development simply because it has never been lost.' He concludes therefore that the gill slits of the embryo Bird represent the gill slits of the embryo Fish, and not the adult gill slits of the Fish, which have been in some mysterious way pushed back into the embryo of the Bird.

Morgan evidently does not realise that the Birds and Reptiles must have been derived from Amphibia, and that the embryo Reptile or Bird with gill slits and gill arches is merely a tadpole enclosed in an egg sh.e.l.l. The Frog in its adult state differs much from a Fish, while the larva in its gill arches and gill slits resembles a Fish. Morgan contends that the new characters do not add themselves to the end of the line of already existing characters. But in the case of the Frog this is exactly what they have done. The existing characters were in this case the gill arches and slits. Those who believe in recapitulation do not suppose that the animal had to live a second life added on to the life of its ancestors and that the new characters appeared in the second life. They believe that in the ancestor a certain character or general structure of body when developed persisted without change throughout life like the gill arches and slits in a Fish. At some stage of life before maturity this character underwent a change, and in the descendants the development of the original character and the change were repeated by heredity. There is no 'mysterious pushing back of adult characters into the embryo,' although it is possible or even probable that in some cases the change gradually became earlier in the life history: it is the new character which is pushed back, not the adult character of the ancestor.

It is perfectly true, as Morgan says, that new characters which arise as discontinuous variations--in other words, those kinds of variation which are called mutations--do not add themselves to the line of already existing characters, but 'change the adult characters without as it were pa.s.sing through and beyond them.' The mutations which Morgan describes in his own experiments on _Drosophila_ ill.u.s.trate this in every case. In no case is the original organ or character, _e.g._ wings, of the normal Fly first developed and then changed by a gradual continuous process into the new character. It might perhaps be said that this took place in the pupa, but that seems impossible, for the complete wing is not fully developed in the pupa. The same truth is equally apparent in the mutations described in _OEnothera_. It follows, therefore, that none of the evolutionary changes which have produced what are called recapitulations can have been due to changes of that kind which is known as mutation.

The abnormalities in Pleuronectidae to which I have referred are of the kind usually regarded as due to arrested development. But closer consideration gives rise to doubt concerning the validity of this explanation. It might be supposed that the attached base of the dorsal fin is unable to extend forward because the eye on the edge of the head is in the way, but if the metamorphosis is arrested, why should the fin grow forward in a free projection? I have described a very abnormal specimen of Turbot in a paper communicated to the Zoological Society of London, [Footnote: _Proc. Zool Soc._, 1907.] and in that paper have discussed other possible explanations of these mutations. In the specimen to which I refer the pigmentation instead of being present on both sides was reversed: the lower side was pigmented from the posterior end to the edge of the operculum (Plate II, fig. 2), while the upper side was unpigmented excepting a scattering of minute black specks and a little pigment on the head (Plate II., fig. 1).

[Ill.u.s.tration: PLATE II, Fig. 1 and Fig. 2, Abnormal Specimen Of Turbot]

I have suggested that the explanation here is that in the zygote the primordia of a normal body and a reversed head have been united together.

We may suppose that different parts of the body are represented in the gametes by different determinants or factors, and therefore it is possible that these factors may be separated. In the specimen we are considering the body is normal or nearly so, with the pigmentation on the left side, which is normal for the Turbot, while the head has both eyes with some pigment on the right side and the left side unpigmented. Reversed specimens occasionally occur in many species of Pleuronectidae, and if the determinants for a reversed head and a normal body were united in one zygote, the curious abnormality observed might be the result. It is just a possibility that if this fish which was only 4.4 cm. long had lived to adult size, the upper side would have become pigmented under the influence of light, while the strong hereditary influence would have prevented the disappearance of the pigment from the lower side. In that case the adult condition would have been similar to that of ordinary ambicolorate specimens, but reversed, with eyes on the right side instead of the left.

Other explanations of the more frequent ambicolorate mutation are possible: the body may consist of two left sides instead of a left and right, joined on to a normal head. But the first suggestion seems the more probable, as two rights or two lefts would not be symmetrical. Supposing the head and body not properly to belong to each other, one being reversed and one normal, we can in a way understand why the dorsal fin does not form the usual connexion with the edge of the head, because the determinants would not be in the normal intimate relation to each other.

In thus writing of reversed and normal it must be understood that the former word does not mean merely turned over, for in that case right side of the body would be joined to the left side of the head, and the dorsal fin would be next to the ventral side of the head, which is not the case.

What is meant is that a left side of the body which is normally pigmented is joined to a left side of the head which instead of having both eyes has neither, the two eyes being on the right side of the head which is joined to the right side of the body, and this is normal and unpigmented. The dorsal fin belonging to the normal sinistral body would therefore have a congenital tendency in the metamorphosis to unite with the head on the outer side of the original lower or right eye after it has moved to the left side. Actually, however, in this abnormal specimen it finds itself on the outer side of the left eye which has pa.s.sed to the right side, and it has no tendency to unite with this part of the head. At the same time it has no tendency to bend over at an angle to reach the outer side of the right eye, and therefore it grows directly forward without attachment to the head at all.

It will be seen, therefore, that what is changed in relative position in these mutations is not the actual parts of the body, but merely the _characters_ of those parts. In a sinistral Flat-fish, whether it is normally sinistral like the Turbot or abnormally like a 'reversed'

Flounder, the viscera are in the same position as in a dextral specimen: the liver is on the left side, the coils of the intestine on the right.

Thus in a reversed or sinistral Flounder, which is normally dextral, the left side which is uppermost is still the left side, but it has colour and two eyes, whereas in the normal specimen the right side has these characters and not the left. Thus we are forced to conceive of the determinants in the chromosomes of the fertilised ovum which correspond to the two sides of the body, as entirely distinct from the determinants which cause the condition or 'characters' of the two sides, unless indeed we suppose that determinants of right side with eyes and colour occur in some gametes and of right side without eyes and colour in others, and vice versa, and that h.o.m.ozygous and heterozygous combinations occur in fertilisation. On this last hypothesis the mutation here considered might be a heterozygous specimen, with the dextral condition dominant in the head and the sinistral in the body. Or it might be somehow due to what Morgan and his colleagues have called crossing over in the segregation of heterozygous chromosomes, so that a part corresponding to a sinistral body is united with a part corresponding to a dextral head.

My conclusion from the evidence is that any process of congenital development may in particular zygotes exhibit a mutation, a departure from the normal. We need not use the term heredity at all, or if we do, must remember that in the present argument it does not refer to any transmission from the parent. The factors in the gametes of the normal Flat-fish egg cause the normal metamorphosis to take place after the larval symmetry has lasted a certain time. In occasional individuals the factors whatever they are, portions of the chromosomes or arrangement of the chromosomes or anything else, are different from those of the normal egg, and in consequence the abnormalities above described are developed.

But the chief fact which I cannot too strongly emphasise is that the development of the abnormality from the symmetrical larva is direct, whether it is merely an arrest of development or an abnormal combination of reversed and normal parts. The abnormal development is not due to a change occurring _after_ the normal asymmetry has been developed. These abnormalities are true mutations.

The evolution of the normal Flat-fish, on the other hand, was obviously due to a change of a different kind. Here we are dealing with the change from a symmetrical fish to the asymmetrical. Judging from what takes place in other mutations, it was quite possible for asymmetry to have developed directly from the egg, in consequence of some difference in the chromosomes of the nucleus. It has been shown that placing a fish egg for a short time in MgCl[2] [Footnote: Stockard, _Arch. Eut. Mech._, xxiii.

(1907).] causes a cyclopean monstrosity to be developed in which the two eyes are united into one: but the two eyes do not develop separately first and then gradually approach each other and unite, the development of the optic cups is different from the first. In the normal Flat-fish the evolution that has occurred is the original development of the symmetrical fish, and the subsequent _continuous gradual_ change in eyes, fin, and colour to the adult Flat-fish as we see it. All the evidence acc.u.mulated by the experiments and observations of mutationists and Mendelians goes to prove that this change is of an entirely different kind from those variations which are described as mutations, or as loss or addition of genetic factors.

This being the case, we have to inquire what is the explanation of the evolution of the normal metamorphosis.

The important fact is that the original symmetrical structure of the larva and the asymmetrical structure of the adult Flat-fish correspond to the different positions of the body of the fish in relation to the vertical, the horizontal ground at the bottom of the water, and incidence of light.

The larva swims with its plane of symmetry vertical like most other fishes; its locomotion requires symmetrical development of muscles and fins; the two sides being equally exposed to light, it requires an eye on each side, and the pigment on each side is also related to the equal exposure to light. The adult lying with one side on the ground has its original plane of symmetry horizontal and parallel to the ground, and only the other side exposed to light, and on this side only eyes and colour, _i.e._ pigment. The change of structure corresponds with the change of habit. It consists in the change of position of the lower eye, the extension of the dorsal fin forwards, and the disappearance of pigment from the lower side. In the actual metamorphosis these changes take place as the skeleton develops, before the hard bones are fully formed, while the fish is still small, but the young Turbot reaches a much larger size before metamorphosis is complete, namely, about one inch in length, than the young Plaice or Flounder. It is of little importance to consider whether at the beginning of the evolution the change of position occurred late or early in life. It may have become earlier in the course of the evolution. The important matter is to consider the evidence in support of the conclusion that the relation to external conditions has been the cause of the evolutionary change. We have already seen that the nature of the change and the relation of the change of structure to the change of conditions necessarily tend to the inference that the latter is the cause of the former. But we have to consider the particular changes in detail.

To take first the loss of pigmentation from the lower side. I have shown experimentally that exposure of the lower sides of Flounders to light reflected upwards from below causes development of pigment on the lower side. At the same time the experiments proved that the loss of pigment in the fish in the natural state and the development of it under exposure to light were not merely direct results of the presence or absence of light in the individual, for in some cases the young fish were placed in the apparatus before the pigment had entirely disappeared from the lower side, and the metamorphosis went on, the lower side becoming quite white, and the pigment only developed gradually after long exposure to the light. In the princ.i.p.al experiment four specimens were placed in the apparatus on September 17, 1890, when about six months old and 7 to 9 cm. in length.

One of these died on July 1, 1891, and had no pigment on the lower side.

The other three all developed pigment on that side. In one it was first noticed in April 1891, and in the following November the fish was 22 cm.

long and had pigmentation over the greater part of the lower side (Plate III.). Microscopically examined, the pigmentation was found to consist of black and orange chromatoph.o.r.es exactly similar to those of the upper side. Some hundreds of young Flounders were reared at the same time under ordinary conditions and none of them developed pigment.

It is clear, therefore, that exposure of the lower side to light and reduction of the amount of light falling on the upper side (for the tops of the aquaria used were covered with opaque material) does not cause the two sides to behave in the same way in respect of pigment, as they would if the normal condition of the fish was merely due to the difference in the exposure to light of the two sides in the individual life. There is a very strong congenital or hereditary tendency to the disappearance of pigment from the lower side, and this is only overcome after long exposure to the light. On the other hand, if the disappearance of the pigment were due to a mutation, were gametogenic and entirely independent of external conditions, there would be no development of pigment after the longest exposure. To prove that an inherited character is an acquired character is quite as good evidence as to show that an acquired character is inherited.

The latter kind of evidence is very difficult to get, for the effect of conditions in a single lifetime is but slight, and is not likely to show a perceptible inherited effect. The theory that adaptations are due to the heredity of the effects of stimulation a.s.sumes that the same stimulus has been acting for many generations.

[Ill.u.s.tration: PLATE III - Flounder, Showing Pigmentation Of Lower Side After Exposure To Light]

It is necessary, however, to consider how far the conclusions drawn from these experiments are contradicted by the mutations occurring in nature, some of which have already been mentioned. We will consider first ambicolorate specimens. If the absence of pigment from the lower side in normal Flat-fishes is due to the absence of light, how is it that the pigmentation persists on the lower side of ambicolorate specimens, which is no more exposed to light than in normal specimens? The answer is that in the mutants the determinants for pigmentation are united with the determinants for the lower side of the fish. My view is that the differentiation of these determinants for the two sides was due in the course of evolution to the different exposure to light, was of somatic origin, but once the congenital factors or determinants were in existence they were liable to mutation, and thus in the ambicolorate specimens there is a congenital tendency to pigmentation on the lower side, which would only be overcome by exclusion of light for another series of generations.

Mutations also occur in which part or whole of the upper side is white and unpigmented. Several such specimens are mentioned in the memoir by myself and Dr. MacMunn in the _Phil. Trans._ already cited, one being a Sole which was entirely white on the lower side, and also on the upper, which was pigmented only over the head region from the free edge of the operculum forwards. Since the upper sides in these specimens are fully exposed to light in the natural state and yet remain unpigmented, it would appear impossible to believe that the action of light was the cause of the development of pigment on the lower sides of normal specimens in my experiments. To some it may be so, but in my own opinion the one fact is as certain as the other. I believe the two facts can be reconciled. I had one specimen of Plaice in the living condition which had the middle third of its upper surface white, and the whole of the lower side white as usual. This specimen was kept for 4-1/2 months with its _lower_ surface exposed to light and the upper side shaded. At the end of that period there were numerous small patches of pigment scattered over the lower side princ.i.p.ally in the regions of the interspinous bones, above and below the lateral line. In the area of the upper side, which was originally unpigmented, there were also numerous small pigment spots. I believe, therefore, that in this case there were determinants for absence of pigment not only on the lower side but on part of the upper side also, and that so long as light was excluded from the lower side the patch on the upper side remained unpigmented in sympathy. When the congenital tendency of the determinants on the lower side was overcome by the action of light, the white patch on the upper side also began to develop pigment.

Lastly, I may refer again to the specially abnormal Turbot mentioned above. In this case the lower side was over the greater part pigmented and the upper side white, and this would appear to contradict the conclusion just drawn concerning the piebald Plaice. But this Turbot was only 4.4 cm.

long, and is the only case known to me where so much of the lower side was pigmented with the upper side almost entirely white. The theory of sympathy or correlation might apply here since the lower side of the head was unpigmented, but from the small size of the specimen and the amount of pigment on the lower side, it seems to me most probable that if the specimen had lived to be adult the upper side would have developed pigment under the action of light and the specimen would have become ambicolorate.

When we compare the results reached by the mutationists with those obtained by the Mendelians we find that they tend to two different conceptions of the relation between the gametes and the organism developed from them. The effect of a change in the determinants of the gametes according to the mutationists is evident in every part of the plant. A factor in Mendelian experiments usually affects only one organ or one part of the organism. The factor for double hallux in fowls, for instance, may coexist with single comb or rose comb. The general impression produced on the mind by study of Mendelian phenomena is that the organism is a mosaic of which every element corresponds to a separate element in the chromosomes. Thus we know that what we call a single factor may cause the whole plumage of a fowl to have the detached barbs, which const.i.tutes the Silky character, but we also know that an animal may be piebald, strongly pigmented in one part and white or unpigmented in another. So we find in these Flat-fish mutations mosaic-like forms which evidently result from mosaic-like factors in the gametes, or in the chromosomes of the gametes.

Experimental evidence concerning the movement of the lower eye to the upper side and of the forward extension of the dorsal fin has not been obtained, though years ago I made some attempts, at the suggestion of Mr.

G. J. Romanes, to obtain such evidence with regard to the eye by keeping young Flounders, already partially metamorphosed, in a reversed position.

I did not succeed in devising apparatus which would keep the young fish alive in the reversed position for a sufficiently long time. We can only consider, therefore, whether those other changes can reasonably be attributed to the conditions of life. Anatomical investigation shows that the bony interorbital septum composed princ.i.p.ally of the frontal bones, which in symmetrical fish pa.s.ses between the eyes, is still between the eyes in the Flat-fish, but has been bent round through an angle of 90 degrees on the upper side, while in the lower side a new bony connexion has been formed on the outer side of the eye which has moved from the lower side. This connexion is due to a growth from the prefrontal backwards to join a process of the frontal, and is entirely absent in symmetrical fishes. It is along this bony bridge that the dorsal fin extends. The origin of the eye muscles and of the optic nerves is morphologically the same as in symmetrical fishes. On the theory of modification by external stimuli we must naturally attribute the dislocation of the eye of the lower side to the muscular effort of the fish to direct this eye to the dorsal edge, but something may also be due to the pressure of the flat ground on the eye-ball. There is little difficulty in attributing the bending of the interorbitl septum to pressure of the lower eye-ball against it, pressure which is probably due partly if not chiefly to the action of the eye muscles. The formation of the bony bridge outside the dislocated eye is more difficult to explain, as I have never had the opportunity to study the relation of this bridge to the muscles. It is worth mentioning that in the actual development of Turbot and Brill the metamorphosis takes place to a considerable degree while the young fish is pelagic, before the habit of lying on the ground is a.s.sumed, but of course this is no evidence that the change was not originally caused by the habit of lying on the ground.

With regard to the extension of the dorsal fin there is no difficulty in discovering a stimulus which would account for it. Symmetrical fishes propel themselves chiefly by the tail; in shuffling over the ground or swimming a little above it. Flat-fishes move by means of undulations of the dorsal and ventral fins. Increased movement produces hypertrophy, and according to the theory here maintained, not merely enlargement of parts existing, but phylogenetic increase in the number of such parts, here fin rays and their muscles. In Flat-fishes the dorsal and ventral fins extend along the whole length of the dorsal and ventral edges: the dorsal from the head, in some cases from a point anterior to the eyes, to the base of the tail, the ventral from the a.n.u.s, which is pushed very far forward, to the base of the tail, and in some species of Solidae these fins are confluent with the caudal fin.

Formerly it was dogmatically maintained that the effect of an external stimulus on somatic organs or tissues could have no influence on the determinants in the chromosomes of the gametes to which the hereditary characters of the organism were due. As we have tried to show, this dogma is no longer credible in face of the discoveries concerning hormones. The hormone theory supposes that the somatic modifications due to external stimuli--in the case of the Flat-fish the disappearance of pigment from the lower side, the torsion of the orbital region of the skull, and the extension of the dorsal fin--modify the hormones given off by these parts, increasing some and decreasing others, and that these changes in the hormones affect the determinants, whatever they are, in the gametocytes within the body.

Here arises an interesting question--namely, how does the hormone theory explain the phenomenon of metamorphosis any better than the mutation theory? It might be agreed that if the determinants are stimulated or deprived of stimulation, the effect of the change should logically show itself from the beginning of development, and that therefore the process of metamorphosis or indirect development does not support the hormone theory any more than the theory of gametogenic mutations. This objection may be answered in the following way. The reason why the determinants give rise to the original structure first and then change it into the new structure is probably the same as that which causes secondary s.e.xual characters to develop only at the stage of p.u.b.erty. By the hypothesis the new habits and new stimuli begin to act at some stage after the complete development of the original structure of the body. The differences in the original hormones of the modified parts are therefore acting simultaneously with the hormones, that is, the chemical substances derived from all other parts of the body in its fully developed condition. It is very probable that in the early stages of development the metabolism of the body would be considerably different from that of the adult stage, and the same combination of hormones would not be present. We may suppose, therefore, that the determinants of the zygote have acquired a tendency to produce the increases and decreases of tissue which const.i.tute a certain modification, _e.g._ the change in the position of the eyes in a Flat-fish, but the stimulus which caused this tendency has always acted when the adult combination of hormones was present. In consequence of this the developed tissues do not undergo the inherited modification until the adult combination is again present. In this way we can form a definite conception of the reason why an adaptive modification is inherited at the same stage in which it was produced, just as the antlers of a stag are only developed when the hormone of the mature testis is present. At the same time it is probable that the age at which the inherited development takes place tends to become earlier in later generations, to occur in fact as soon as the necessary hormone medium is present.

The diagnostic characters, of some of the species of Pleuronectidae have been mentioned in an earlier part of this volume, in order to point out that they have no relation to differences of habit or external conditions.

Here it is to be pointed out that there is no evidence that they arise by metamorphosis. The scales, for example, afford distinct and constant diagnostic characters both of species and genera, but their peculiarities have not been found to arise by modification of a primitive form. The rough tubercles of the Flounder, and the scattered thornlike tubercles of the Turbot, develop directly, not by the continuous modification of imbricated scales. There is, however, one scale-character among the Pleuronectidae which appears to stand in direct contradiction to the conclusions drawn by me concerning scales in general. It not only develops by a gradual change, but it is a secondary s.e.xual character developing in the males only at maturity. The character was described by E. W. L. Holt in specimens of the Baltic variety of the Plaice, _Pleuronectes platessa_, [Footnote: _Journ. Mar. Biol. a.s.sn._, vol iii. (Plymouth, 1893-95.)] and consists in the spinulation of the posterior edges of the scales, especially on the upper side, in mature males. The same condition, but to a much slighter degree, was afterwards shown by myself to occur constantly in Plaice from the English Channel and North Sea. [Footnote: _Ibid._, vol.

iv. p. 323.] It occurs also in _P. glacialis_, the representative of the Plaice in more northern seas. I have shown that the spinules develop in the mature males not as a modification of the scale, but as separate calcareous deposits the bases of which afterwards become united to the scale. It would seem that the development of this character is dependent on the hormone from the mature testis, and in order to conform with the arguments used by me in other cases, the spinulation should have some definite function in relation to the habits of the s.e.xes, and this function should involve some kind of external stimulation restricted to the mature male. So far, however, no evidence whatever of such function or such stimulation has been discovered. It is possible that the case differs from other secondary s.e.xual characters as the antlers of stags in one respect, namely, that the Dab (_P. limanda_), the Sole, and other species of _Solea._ and several other Pleuronectidae have what are called etenoid scales--that is, scales furnished with spines on the posterior edge--and since the ordinary scales of the Plaice are reduced, the spinulation of scales in the mature male Plaice is not a new character but the retention of a primitive character. Then the question would remain why the scales in the mature female and immature male have degenerated, or rather why the primitive character develops only in the mature stage of the male.

There is one point in which this s.e.xual dimorphism in the Plaice appears to differ from typical cases, and which suggests that the greater spinulation of scales in the males has no function at all in the relations of the s.e.xes, and is therefore not subject to and external stimulation.

This point is the remarkable way in which the degree of development of spiny armature differs in different regions and in local races, and seems to correspond to different climatic conditions. Both Plaice and Flounders in the Baltic are much more spiny than in the North Sea, although in the Flounder no s.e.xual difference in this respect has been noted. On the east coast of North America occurs _P. glacialis_, in which the scales of the male are strongly spinulate and those of the female smooth. On the coast of Alaska females of this species seem to be more spinulate than elsewhere. The Flounder does not occur in the Arctic, but on the west coast of North America occurs a local form called _P. stellatus_, scarcely distinct as a species, which has a strong development of spiny tubercles all over the upper side. The Flounders of the Mediterranean are much less spinous than those of the North Sea or Channel. The Dab (_P.

limanda_) occurs on the American coast in a local form called _Limanda ferruginea_, and in the North Pacific there is a rougher form called _L.

aspera_. In these three species therefore, apart from mutations, the northern forms all show a greater development of spines on the scales.

Whether this is an effect of colder temperature it is difficult to say. It is possible that the difference is due to external conditions, of which lower temperature of the water is the most obvious, and it may be that these conditions have a greater effect on the male than on the female in the Plaice.

s.e.xual differences in scales, which have a function in the relations of the s.e.xes, occur in a few other fishes, and these can be attributed with good reason to mechanical stimulation. For example, in the Rajidae among Elasmobranchs the males possess on each 'wing' or pectoral two series of large, recurved, hooked spines. It has been stated, [Footnote: Darwin, _Descent of Man_ (2nd edit., 1885), p. 331.] apparently by Yarrell, that these spines are developed only in the breeding season. It is doubtful if there is any marked breeding season in these fishes, but it is probable that the spines are absent in the immature male, as it is known that in _Raia clavata_ the adult male has sharp pointed teeth, while the young male and the female at all ages have broad flat teeth. It is supposed that the spines and perhaps the sharp teeth are used for holding the female, but it seems equally probable that these structures are really used by the males in fighting with each other. The habits of these marine fish have not been much observed, but there is little reason to doubt that these differences in scales and teeth correspond with differences of mechanical stimulation. This does not at all imply that the scales and teeth themselves have been produced by mechanical stimulation, or that the difference between the dermal denticles of Elasmobranchs and the scales of Teleosteans correspond to differences of stimulation. But the degree of development of a structure whose presence is due to gametic factors may very probably be modified by external stimulation, and the modification may become hereditary. If the views here advocated are true, the two processes mutation and modification must be always acting together and affecting the development not only of the individual but of any organ or structure. Thus the peculiarities of antlers in stags, it seems to me, prove that the mechanical stimulation due to fighting was the cause of the evolution of antlers, that without the habit of fighting in the males antlers would not exist. At the same time each species of the _Cervidae_ has its special characters in the antlers, in shape and branching, and it would be impossible to attribute these to differences in mode of fighting: they are due to mutation.

In connexion with the metamorphosis of Amphibia the case of the Axolotl has always been of very great interest. In the few small lakes near the city of Mexico where it occurs it has never been known to undergo metamorphosis but is aquatic throughout its life and breeds in that condition. Yet in captivity by reducing the quant.i.ty of water in which it is placed the young Axolotl can be forced to breathe air, and then it undergoes complete metamorphosis to the abranchiate condition. The same species in other parts of North America normally goes through the metamorphosis, like other species of the Urodela. It is evident, therefore, that the Mexican Axolotls, although they have been perennibranchiate for a great number of generations, have not lost the hereditary tendency to the metamorphosis which changes the larvae of _Amblystoma_ elsewhere into an air-breathing terrestrial animal. This may be regarded as evidence that the conditions of life which prevent the metamorphosis in the Mexican Axolotl have produced no hereditary effect.

The fact, however, that Axolotls require special treatment to induce metamorphosis seems to show that they have distinctly less congenital tendency to metamorphosis than larvae of the same species, _Amblystoma tigrinum_, in other parts of North America, and this difference must be attributed to the inherited effect of the conditions. The most important of these conditions seems to be abundance of oxygen in solution in the water, and the next in importance abundance of food in the water. Recently it has been shown that the metamorphosis may be induced by feeding Axolotls on thyroid gland. But there is no reason to suppose that a congenital defect of thyroid arising as a mutation was the original cause of the neoteny, _i.e._ the peisistence of the larval or aquatic, branchiate condition. Such a supposition would imply that the a.s.sociation between Axolotls and the peculiar Mexican lakes, supplied with oxygenated water by springs at the bottom, was purely accidental. Moreover, there is no evidence that there is any deficiency of thyroid in the Axolotl. The secretion of the thyroid gland is necessary for the normal growth and development of all Vertebrates, and we are only beginning to understand the effects of defect or excess of this secretion. There is nothing very surprising in the fact that excess in the case of the Axolotl causes the occurrence of the metamorphosis which had already in numerous experiments been produced by forcing the animals to breathe air.

Metamorphosis, as in the development of gill arches and gill slits in the embryos of Birds, Reptiles, and Mammals, exhibits a recapitulation of the stages of evolution of certain organs. But in the case of other organs the absence of recapitulation is remarkable by contrast. If, as I believe, the development of lungs and disappearance of gills was directly due to the necessity of breathing air, it is difficult to avoid the conclusion that the terrestrial legs were originally evolved from some type of fishes'

fins by the use of the fins for terrestrial locomotion. Yet neither the amphibian larva nor the embryo of higher Vertebrates develops anything closely similar to a fin. There is no gradual change of a fin-like limb into a leg, but the leg develops directly from a simple bud of tissue. The larva of the Urodela is probably more primitive than the tadpole of the Frogs and Toads, and in the former the legs develop while the external gills are still large, long before the animal leaves the water.